Each Mini-Simposium consists of a minimum of one session of two hours (6 presentations of 20 minutes each). The number of sessions for a MS are determined by the number of accepted presentations. A MS cannot be split in parallel sessions.
A session with Keynote Talk is allowed every two full sessions, with a 40 min. Keynote Talk plus four presentations of 20 min. Each mini-symposium has a Corresponding Organizer as the liaison with the Conference Secretariat. The mini-symposium organisers can speak as part of the mini-symposium.
List of Accepted Mini-Symposia
Isogeometric Analysis (IGA) has been originally introduced and developed by T.J.R. Hughes, J.A. Cottrell, and Y. Bazilevs, in 2005, to generalize and improve finite element analysis in the area of geometry modeling and representation. However, in the course of IGA development, it was found that isogeometric methods not only improve the geometry modeling within analysis, but also appear to be preferable to standard finite elements in many applications on the basis of per-degree-of-freedom accuracy. Non-Uniform Rational B-Splines (NURBS) were used as a first basis function technology within IGA. Nowadays, a well-established mathematical theory and successful applications to solid, fluid, and multiphysics problems render NURBS functions a genuine analysis technology, paving the way for the application of IGA to solve a number of problems of academic and industrial interest. Further fundamental topics of research within IGA include the analysis of trimmed NURBS, as well as the development, analysis, and testing of flexible local refinement technologies based, e.g., on T-Splines, hierarchical B-Splines, or locally-refined splines, in the framework of unstructured multipatch parameterizations. Moreover, an important issue regards the development of efficient strategies able to reduce matrix assembly and solver costs, in particular when higher-order approximations are employed. Aiming at reducing the computational cost still taking advantage of IGA geometrical flexibility and accuracy, isogeometric collocation schemes have attracted a good deal of attention and appear to be a viable alternative to standard Galerkin-based IGA. Another more than promising topic, deserving a special attention in the IGA context, is finally represented by structure-preserving discretizations. Along (and/or beyond) these research lines, the purpose of this symposium is to gather experts in Computational Mechanics with interest in the field of IGA with the aim of contributing to further advance its state of the art.
Keywords: Computational Mechanics, Numerical methods, Partial Differential Equations
In fluid dynamics, designing accurate numerical schemes or solvers is a challenging issue. The Navier-Stokes equations are non-linear partial differential equations that describe the motion of fluids. In the case of Newtonian flows, the non-linear term vanishes and one can solve the Stokes problem. On the one hand, the Stokes problem can accurately model creeping flows. On the other hand, the study of Stokes problem helps to build an appropriate approximation of the Navier-Stokes equations. Indeed, it is well-known that poor mass conservation in the numerical scheme leads to instabilities, which are amplified when solving the Navier-Stokes equations. Thus, the search for a robust discretization of the Stokes problem is a preliminary step to solving the Navier-Stokes equations. We propose to study these problems from three complementary angles: theoretical, technical and modelling. In particular, the following issues will be addressed for Stokes problem: Dirichlet control on non-smooth boundaries, new variational formulations via T-coercivity, nonconforming Crouzeix-Raviart finite elements method with enriched discrete pressure space (applied to an industrial CFD solver). Concerning Navier-Stokes equations, we will focus on the modelling of the fluid dynamics of fuel cell channels and on model reductions of fluid flows in biological media.
Five prospective European speakers:
Pr. Th. Apel, Universität der Bundeswehr München, Germany.
Pr. P. Ciarlet, Jr., ENSTA Paris, Institut Polytechnique de Paris, France.
Dr. E. Jamelot, Université Paris-Saclay, CEA, France.
Dr. P. Mollo, Eindhoven University of Technology, Netherlands.
Dr. R. Tittarelli, FEMTO-ST Institute, Université de Bourgogne Franche-Comté, France.
Keywords: Finite element method, Fluid Dynamics, Modeling and Simulation
Over the last two decades, a large variety of non-standard discretization schemes, aiming at improving the interoperability between CAD designs and finite element discretizations, have emerged in the field of the numerical approximation of partial differential equations. Amongst them, some are based on the isogeometric concept and/or fictitious domain approaches. More than facilitating the geometry modeling within analysis, these methods offer the opportunity to drastically reduce the numerical resources required for simulation by exploiting the smooth, higher-order nature of the underlying CAD-based approximation spaces [1,2,3,4].
This mini-symposium aims to bring together multidisciplinary viewpoints that facilitate fruitful discussions and exchange of ideas on how to develop efficient CAD-based discretization techniques, whether in terms of operator assembly or system solutions. We welcome contributions from both method developments and applications. Typical topics are expected to be, but not restricted to: spline-based discretizations, fictitious domain approaches, collocation methods, reduced-order modelling techniques, tensor methods, efficient numerical integration rules and fast formation of discrete operators, as well as linear systems’ solvers such as domain decomposition and multigrid techniques.
[1] Calabro, F., Sangalli, G., & Tani, M. (2017). Fast formation of isogeometric Galerkin matrices by weighted quadrature. CMAME, 316, 606-622.
[2] Mantzaflaris, A., Jüttler, B., Khoromskij, B. N., & Langer, U. (2017). Low rank tensor methods in Galerkin-based isogeometric analysis. CMAME, 316, 1062-1085.
[3] Hofreither, C., & Takacs, S. (2017). Robust multigrid for isogeometric analysis based on stable splittings of spline spaces. SINUM, 55(4), 2004-2024.
[4] Hirschler, T., Antolin, P., & Buffa, A. (2022). Fast and multiscale formation of isogeometric matrices of microstructured geometric models. Comput. Mech., 1-28.
Keywords: Domain decomposition, Fast assembly, Immersed boundary methods, Multigrid methods , IsoGeometric Analysis
In recent years there has been a growing interest regarding the numerical modeling of fracture in mechanical problems. For that purpose, several innovative approaches, using element or nodal enrichment strategies, were proposed. The aim of this minisymposium is to address all these enrichment techniques, both from a theoretical and a practical perspective. The topics to be covered include, but are not limited to:
- embedded and generalised/extended finite element formulations for cracks or heterogeneities;
- modeling of material interfaces and/or microstructure of a material;
- computational efficiency, convergence and stability of enriched elements;
- new techniques to overcome convergence problems in the modeling of fracture in brittle and granular materials;
- enriched elements for coupled processes, such as corrosion, thermomechanical problems etc.;
- 3D and large scale problems.
Keywords: 3D, convergence techniques, couple processes, heterogeneities, interfaces, large scale problem, Enriched finite-elements
Porous media applications often face the challenge of the presence of multiple scales in space and time as well complex geometries. These have a direct impact on discretization and solution strategies, in particular for coupled problems introducing a range of different and potentially dynamically changing problem characteristics. With applications in mind ranging from geological carbon storage to perfusion of biological tissue, both accurate, robust and efficient numerical methods are highly desirable. The focus of this session lies on advanced discretizations and efficient solvers for coupled problems in porous media targeting these and related mathematical and computational issues.
Keywords: Finite element method, numerical analysis, porous media
Meshfree, particle, and peridynamic methods offer a new class of numerical methods that play an increasingly significant role in the study of challenging engineering and scientific problems. New and exciting developments of these methods often go beyond the classical theories, incorporate more profound physical mechanisms, and become the exclusive numerical tools in addressing the computational challenges that were once difficult or impossible to solve by conventional methods.
The goal of this minisymposium is to bring together experts working on these methods, share research results, and identify the emergent needs towards more rapid progress in advancing the important fields of meshfree, particle, and peridynamic methods. Topics of interest for this minisymposium include, but are not limited to the following:
• Recent advances in meshfree, particle, and peridynamic methods, and their coupling with other computational methods, such as IGA, material point method and finite element method
• Immersed approaches for non-body-fitted discretizations
• Enrichment of basis functions for non-smooth approximations
• Integration of physics-based and data-enabled approaches
• Enhancement of meshfree, particle, and peridynamic methods by machine learning algorithms
• Strong form collocation methods
• Stabilization for under-integrated Galerkin methods
• Methods for coupling multiple physics and/or multiple scales
• High performance computing, solvers, and large-scale simulations
• Multi-phase interactions under extreme conditions with the material point, meshfree, and peridynamic methods for multi-scale simulations
• Recent advances for challenging industrial applications: modeling extreme loading events, additive manufacturing, and mitigating disasters
• Methods enabling a rapid design-to-analysis workflow
Keywords: Meshfree, Particle, Peridynamics
Injury to head and neck systems represents a profoundly underestimated global public health issue. The sudden rise of micro-mobility such as e-scooters has been causing an alarming number of injuries and hospitalizations. At the same time, sports, cars, military and home accidents are sources of many different types of injuries. In fact, a more comprehensive understanding of injury-causing factors, encompassing diverse populations, human activities, and situations where excessive loads lead to harm, necessitates further elucidation. By delving into the biomechanics of injury and disability, we can aspire to achieve better protective measures against injury.
Injury causation entails examining the biomechanical responses of the human body, as well as the function and structure of cells and tissues, in response to dynamic loading. This analysis also investigates the mechanisms and tolerances of various body regions to injury. Research in this domain encompasses macroscopic motion analysis of human volunteers or surrogates, alongside microscopic measurements of cell and tissue function and structure. Computational and translational models are utilized to extend experimental findings to encompass a wide range of real-world environments.
This symposium welcomes computational studies and models - validated from experiments - that aim to identify and define injury mechanisms, quantify biomechanical responses, determine impact tolerance levels, and develop and employ injury assessment devices and techniques for evaluating injury prevention systems. Researchers are encouraged to submit contributions covering any aspect related to the biomechanics of head and neck injuries.
Keywords: Biomechanics, Brain, Computational Mechanics, Constitutive modeling, Finite element method, Injury, Neck
Cancers are highly heterogeneous diseases that involve diverse biological mechanisms, interacting and evolving at various spatial and temporal scales. Multiple experimental, histopathological, clinical, and imaging methods provide a means to characterize the heterogeneous and multiscale nature of these diseases by providing a wealth of temporally and spatially resolved data on their development and response to therapies (e.g., cancer architecture, mechanics, and vascularity; cancer cell mobility and proliferation; drug transport and effects). These multimodal, multiscale datasets can be exploited to constrain biophysical models of cancer growth and treatment response both in preclinical and clinical settings. These models can then be leveraged to test hypotheses, produce individualized cancer forecasts to guide clinical decision-making, and, ultimately, to design optimized therapies. The overall goal of this minisymposium is to provide a forum to present and discuss recent developments in data-informed computational models and methods for predicting cancer growth and treatment response, with special focus on the following research areas: (i) biology-based mechanistic models of cancer growth and treatment in vitro and in vivo; (ii) computational methods for model initialization, parameterization, and patient-specific simulation; (iii) model-oriented, personalized optimization of treatment regimens; (iv) uncertainty quantification and model selection methods; (v) hybrid strategies combining machine learning and mechanistic modelling; and (vi) digital twins in clinical oncology.
REFERENCES
[1] G. Lorenzo, D.A: Hormuth II, A.M. Jarrett, et al., “Quantitative In Vivo Imaging to Enable Tumour Forecasting and Treatment Optimizationâ€. In: Balaz, I., Adamatzky, A. (eds), Cancer, Complexity, Computation, 46, 55-97, Springer Cham, (2022).
[2] M. Alber, A. Buganza Tepole, W.R. Cannon, et al., “Integrating machine learning and multiscale modelling - perspectives, challenges, and opportunities in the biological, biomedical, and behavioral sciencesâ€, npj Digit. Med., 2, 115, (2019).
Keywords: cancer, cancer forecasting, computational oncology, digital twins, inverse problems, machine learning, mechanistic modeling, model selection, optimal control theory, uncertainty quantification
Personalized simulations based on patient-specific data are a relevant task in orthopaedic trauma surgery. The starting point for all individualized models are the material properties of bone given from experiments as well as from medical imaging. For the understanding of the behaviour of bone-implant-systems under realistic loading conditions, monitoring of patients is essential. In order to establish simulation-based workflow concepts into the clinical routine, besides the mechanical modelling also imaging techniques such as segmentation, material assignment and CAD integration are relevant for success. In addition, patient-specific simulation strategies are playing a key role for a personalization of osteosynthesis implants. With realistic material parameters and individual loading scenarios, an individualization of implants as well as rehabilitation processes are possible.
The main objectives of this mini-symposium are centred on bringing together engineers, mathematicians, physicist, biologists, computer scientists, experimentalists and (end)users in orthopaedic trauma surgery to discuss concepts and strategies from experiment and simulation to clinical applications. Topics of interest include, but are not restricted to, the following:
• simulation
• novel concepts in computational biomechanics
• novel aspects in (soft) tissue modelling
• novel strategies for personalized trauma implants
• applications with potential relevance for orthopaedic trauma surgery
• personalized simulations with a focus on clinical applications
Keywords: clinical biomechanics, fractured long bones, patient monitoring, patient-specific simulations, personalized implants, Individualized osteosynthesis
ABSTRACT
Cardiovascular and cerebrovascular diseases are a major cause of mortality and morbidity in developed countries [1]. Atherosclerotic disease is a degenerative process that is characterized by the development of atheromatous plaques on the wall of arteries. The achievement of a supporting tool to help in clinical reasoning may help cardiologists, neurologists and surgeons to better manage the atherosclerotic disease. This tool can serve to obtain the real and detailed hemodynamics such as pressure, velocity and WSS fields along the patient artery [2,3].
However, modelling hemodynamics with real physiological conditions of each patient using principles of engineering is still a challenge. A validated numerical tool is a non-invasive method that can bring many potential benefits: enhanced non-invasive evaluation of specific lesion with more judicious and selective referral, leading to reduced risk of complications and more cost-effective strategies in the diagnosis and management of patients with atherosclerosis.
Thus, this symposium is opened to a vast number of computational works used to model the hemodynamics and specific parameters. In addition, it can cover a range of topics of experimental and clinical data acquisition for further validation of the computational studies.
The proposed symposium is a joint venture between important research areas like biomedical engineering, computational biomechanics, mathematics, computational vision, medical imaging and cardiovascular-cerebrovascular medicine.
REFERENCES
[1] D. Mozaffarian et al, “Heart disease and stroke statistics – 2016 updateâ€, Circulation Vol. 131, e38-360, (2016).
[2] E. Boileau et al., “Estimating the accuracy of a reduced-order model for calculation of fractional flow reverse (FFR)â€, Int J Numer Meth Biomed Engng, Vol. 34, e2908, (2018).
[3] D.F. Bechsgaard et al., “Myocardial perfusion assessed with cardiac computed tomography in women without coronary heart diseaseâ€, Clinical physiology and functional imaging, Vol. 39, 65-77, (2019).
Keywords: Cardiovascular and Cerebrovascular Medicine, Computational Methods, Computational Vision, Fluid Dynamics, Medical Imaging
In recent years, the application of computational modeling to biomechanics has increased significantly. This growth has reached a point where in silico simulation of complex biological processes and surgical procedures is becoming an integral part of the regulatory approval process for new medical devices. Advances in computational methods recognize the close relationship between structure and functionality in living systems. The mechanical properties of tissue building blocks including extracellular matrix, e.g., collagen fibers and cells, as well as their ability to respond to external stimuli or internal changes, are unique features of living biomechanical systems. Furthermore, mechanobiological cues at the cell level are considered fundamental to the healthy functioning of organs and provide insights into the development of pathologies.
As a result, research in this field is progressing towards the integration of biomechanics, biochemistry, and mechanobiology in a multiscale and multiphysics computational environment. This mini-symposium aims to collect contributions on innovative computational methods applied to various biomechanical systems including lung, bone, cartilage, brain, cardiovascular system, and biocompatible materials used in biomedical applications. Relevant topics include but are not limited to:
• Multiscale modeling of fibrous tissues
• Growth, remodeling, and healing of tissues
• Damage and fracture
• Poroelasticity and multiphase models
• Multi-physics models (e.g., chemo-mechano-biological models)
• Mechanical models of living cells
• Structure-mechanics relationship
• Tissue engineering applications
Keywords: Biomechanics, Computational Methods, Microstructure, multi-scale
The functions of human organs are often regulated by the presence of muscle tissue. There are various types of muscle tissues that are typical of different organs: smooth muscle (arteries, intestines, hollow organs), striated muscle (skeletal muscles), cardiac muscle tissue (heart). The mechanics of muscles is due to the combination of active and passive behaviours. Under the action of a potential induced by the transfer of ionic charges through cell walls, muscles contract (active action), giving rise to a mechanical reaction (passive action). In this MS, we address the theoretical conception and the numerical implementation of constitutive models for active biological tissues and their application in boundary value problems concerning patient-specific human organs. Beside traditional phenomenological models, we welcome constitutive models of higher complexity, including multiscale, multi-phase (porous materials), time-dependent, evolutionary (growth and degeneration) behaviours, micro-mechanical models, and innovative models instructed and surrogated by Machine Learning algorithms. The MS will focus also on methods developed for the numerical coupling between the different physics and data assimilation techniques (functional and geometries from medical images such as CT scans) for the customization of the integrated model, in order to carry out in-silico scenario studies and analyses. The MS promotes a strong interdisciplinary and synergistic approach between mathematical and numerical modelling, micro-mechanics of biological tissues, clinical practice and computational medicine.
Keywords: active materials, Biomechanics, Constitutive modeling, geometrically nonlinear analysis
The numerical modelling of the vascular system and its pathologies becomes increasingly relevant in clinical practice due to great advances in the field of computational biomechanics and related disciplines. Increasing reliability and accuracy of the developed mathematical tools grants access to otherwise unavailable data to foster the understanding of the vascular system and its intricacies. In this regard, promising fields of application are, e.g., digital twinning, studies on virtual cohorts, patient-specific modelling of surgical procedures or detailed analysis of prototypical configurations [1, 2]. The insights gained through mathematical modelling can guide decision making, allow for comparison of various treatment options or can inspire medical device design.
This mini-symposium focuses on emerging topics in the context of the modelling of the vascular system by means of computational fluid mechanics, solid mechanics, fluid-structure interaction, contact mechanics and more. Hence, potential contributions might contain, but are not limited to topics such as blood flow, tissue modelling, drug transport, electrophysiology, thrombus formation, tissue growth and remodelling, virtual treatment, hemodynamic indicators, patient-specific models, stent deployment, heart valve dynamics, or numerical modelling of pathologies such as aneurysma, dissection, in-stent restenosis and more. The mini-symposium’s scope further encompasses reduced order modelling, machine learning techniques, sensitivity analysis, uncertainty quantification and other statistical methods applied to the vascular system. The proposed mini-symposium brings together experts from the involved fields of applied mathematics and biomedical engineering to exchange the latest results and discuss future challenges to advance personalized computational medicine.
[1] Z. Su and T. Chaichana, Int. J. Cardiol., Vol. 210, pp. 26-31, (2016).
[2] C.A. Taylor and D.A. Steinman, Ann. Biomed. Eng., Vol. 38(3), pp. 1188-203, (2010).
Keywords: Computational Medicine, Fluid-Structure Interaction, Hemodynamics, Pathology, Patient-specific Modelling, Sensitivity Analysis, Tissue Mechanics, Vascular System
Biological tissues are living systems able to actively respond and adapt to physical cues from their environment. Mechanical loads at the tissue length scale play a particular role. They can be associated, e.g., with body motion or hemodynamics but, potentially less evident, changes in the mechanical field quantities accompany a variety of physiological and pathological conditions. The understanding of how these mechanical states translate into cues perceived by cells, and how the cells' sensory apparatus registers and converts them into intracellular cues that activate the cell's signalling pathways and finally elicit the biological response, poses problems at the interface between the disciplines of mechanics and cell biology.
During the recent years, computational analysis has played an increasing role in addressing such problems, e.g., to estimate cellular forces, to probe new biomechanisms, or to support studies on mechano-regulated cell behaviour and its role in tissue engineering [1, 2].
This minisymposium focuses on computational methods and models that help unravelling the complex interplay between mechanical loads at tissue length scale, changes in the physical properties of the local cell environment and the cells’ mechanobiological response at short and long term. The topics addressed include but are not limited to models of the extracellular matrix, active cell forces and adhesion, nucleus mechanics, intra- and extracellular fibre networks, simulations of cells and cell layers and their interaction with substrates, analysis of signalling pathways, as well as experiments to inform and validate these approaches. Contributions from all related disciplines are welcome.
REFERENCES
[1] M. Bergert et al., "Confocal reference free traction force microscopy", Nat Commun, Vol. 7, 12814 (2016)
[2] M. Y. Emmert et al., "Computational modeling guides tissue-engineered heart valve design for long-term in vivo performance in a translational sheep model", Sci Transl Med, Vol. 10, 4587 (2018)
Keywords: Cell Mechanics, Extracellular Matrix, Mechanotransduction, Cellular Forces, Mechanobiology
Agent-based models are stochastic modelling procedures, whose most notable characteristic is to be rooted on the specification of local rules attached to the simulated objects –hence “agent basedâ€â€“ informing in turn the computed behaviours (i.e., movement, interaction, replication, death, transformation, etc.). Adoption of agent-based modelling (ABM) in biomedicine and life sciences has seen a rapid growth in the last decades, first and foremost, because AÎ’M is very appropriate to simulate systems about which researchers have accumulated parcelled local knowledge of datasets acquired by a variety of modalities, and that are interested in exploring how to relate to emerging behaviours. Furthermore, ABM is a rule-based approach that is well-suited to integrate multiple spatio-temporal scales (e.g., intracellular processes, microenvironmental remodelling, multicellular population dynamics, cell-to-tissue and tissue-to-cell interactions, angiogenesis, immuno-surveillance, etc.).
The purpose of this minisymposium is to act as a forum for investigators to present the state of the art in ABM in biology and life sciences, with a focus on pathophysiology of chronic (e.g., neuro-developmental and neurodegenerative disorders, cancer, diabetes, chronic inflammation) and communicable diseases (viral infections, parasitic disorders, etc.). Together, we aim to foster the exchange of knowledge and ideas across multiple disciplines: mathematics, physics, computer science, engineering, biology, medicine.
We welcome contributions addressing challenges related to mathematical and computational modelling using ABM with particular emphasis in:
• cross-scale (from organ to tissue, cell, protein and molecular level) ABM simulations;
• challenges in numerical techniques –including high-performance computing procedures– for single scale or/and multiscale agent-based models;
• benchmarking and calibration of ABM in problems related to systems biology;
• simulations that integrate in vitro and/or in vivo laboratory experiments for ABM initialization, parametrization and/or validation;
• prognostic ABM in drug delivery, nanomedicine, immunotherapy, and radiation treatment, with special focus on optimisation of trials and in vivo/ex vivo models substitution/ complementarity.
Keywords: agent-based model, cell biomechanics, disease progression model, high-performance computing, mechano-biology, Multiscale Fluid Dynamics
Computational Biomechanics is an exciting and challenging area of Computational Mechanics. Computational modeling and simulation of living tissues allow for new insights as well as quantitative analyses from the molecular up to the organ level. Reliable predictions support clinical diagnoses and treatment of diseases, as well as the design of new biomedical devices.
The mini-symposium on Computational Biomechanics and Applications has the aim of presenting and discussing recent developments in Biomechanical and Biomedical Engineering numerical simulation techniques, either in terms of mathematical modeling, computer simulation and validation. The mini-symposium includes, but it is not limited to, works related with several topics in biomechanics, such as:
- biomechanics of the musculoskeletal system
- joint biomechanics - hip/knee/ankle/shoulder/hand
- soft tissue - ligaments/tendons/cartilage/skin
- implants/orthotics/prosthetics
- orthopedic biomechanics
- cardiovascular and hemodynamic bio-fluids
- respiratory biomechanics
- oral-facial biomechanics
- tissue biomechanics
- tissue engineering
- sports biomechanics
Keywords: Computational Biomechanics, Mechanobiology, Biomechanical Applications, Biomechanics
Cardiovascular diseases remain a leading cause of death worldwide. Many different Mechanical Circulatory Supports (MCSs) technologies have been developed and implemented to mitigate the diseases. However, blood damage remains an important issue for such devices. The main blood damage aspects are hemolysis, the release of intracellular hemoglobin from red blood, and thrombosis, the formation of blood clot. The design of MCSs includes consideration of the hemolytic and thrombotic potential of a device using different means laboratory tests, animal models and numerical simulations. This mini-Symposium interests include experimental methodologies, animal protocols and data analysis and numerical models for assessing blood and tissue damage related to MCS. In particular, • hemolytic and thrombotic models, • non-linear biofluid and soft tissue models (non-newtonian fluids, hyperelastic, viscoelastic, plastic solids), • diffusion and transport models, • interface models, • rupture models and • growth models from the micro- to the macro-scale.
Keywords: biofluid models, blood damage, hemolysis, rupture multiscale models, soft tissue biomechanics models, thrombosis, transport, Mechanical circulatory devices
Vascular adaptation is a crucial process implicated in major cardiovascular diseases, including atherosclerosis, aneurysm formation, and responses to therapeutic interventions like endovascular or surgical procedures. Vascular adaptation processes are governed by multifactorial and multiscale networks of events, involving intricate feedback mechanisms, cause-and-effect relationships and mutual interactions of components across multiple spatio-temporal scales. Despite significant progress, a comprehensive understanding of these events remains elusive, posing a significant challenge to advancing our ability to manage cardiovascular diseases. It is clear that moving forward in the understanding and prediction of vascular adaptation processes requires a comprehensive capture of the complex, multiphysic and multiscale nature of the phenomenon. To address this challenge, multiscale mechanobiological computational models have recently emerged as promising tools to investigate arterial growth and remodelling in different vascular regions, such as aorta, coronary arteries and peripheral arteries. By providing a comprehensive framework to capture the complex interplay between biomechanical forces, cellular behaviour, and molecular pathways, these models hold great potential for advancing our understanding of vascular adaptation processes. Various modeling strategies, including continuum, discrete and hybrid (e.g., integrating continuum and discrete methods) have been proposed for investigating vascular adaptation processes, ranging from simulations in idealized vessel geometries to patient-specific ones. Moreover, the integration of multi-omics data in multiscale models of vascular adaptation, allowing the identification of patient-specific pathophysiological pathways, can provide a remarkable contribution to the understanding of cardiovascular diseases. Consequently, it can contribute to disease prevention, diagnosis and treatment, aligning with the emerging field of personalized medicine. However, as the complexity of the computational models increases, aspects such as verification, uncertainty quantification, calibration and validation of the model, as well as the computational time, become challenging priorities that need to be addressed for their clinical translation. In this context, this mini-symposium will cover recent advancements in the multiscale modeling of vascular growth and remodeling, which are expected to address some of the major curren challenges.
Keywords: Mechanobiology, Multiscale models, Vascular remodeling, Cardiovascular system
Heart failure (HF) is a pandemic currently affecting up to 15 million people in Europe. It is a complex clinical syndrome presenting with impaired heart function and is associated with poor quality of life for patients and high healthcare costs. There is a high clinical demand for novel computational modeling and artificial intelligence (AI) tools that will facilitate risk stratification, early diagnosis, and disease progression assessment in HF. STRATIFYHF [1] aims to develop, validate and implement the computational modeling and AI-based, decision support system (DSS) for risk stratification, early diagnosis, and disease progression assessment in HF to accommodate both primary and secondary care clinical needs. The DSS will integrate patient-specific demographic and clinical data using existing and novel technologies and establish. A coupled model which includes multiscale modelling of realistic sarcomeric system, genetics patient profile, electrophysiology, realistic directions of muscle fibers, solid-fluid interaction coupled to electrophysiology of the heart was implemented. Initial results give influence of left ventricle deformations on deformations of mitral valve, and on general blood flow in heart. Also drug distribution in the heart and effects of different drugs are tested for heart failure disease. STARTIFYHF [1] will change the way in which HF is diagnosed today and thereby improve the quality and length of patients’ lives and lead to efficient and sustainable healthcare systems by reducing the number of HF-related hospital admissions and unnecessary referrals from primary to secondary care in Europe and beyond.
Keywords: artificial intelligence, modeling, cardiovascular, heart failure
Geomaterials, such as soils, rocks, concretes and snows, are the porous geological media that show the complex mechanical responses and multiscale failure characateristics in the multiphysics geological environments. The better understanding of deformation and failure mechanisms in geomaterials plays an important roles in geophysics (i.e., fault weakening and instability, and earthquake rupture), geohazards (i.e., landslides and rock avalanches towards the extreme climate change), and geotechnical engineering (i.e., geothermal engineering, and CO2 storage reservoirs). Numerical analysis is crucial in the modern geomechanics and geotechnical engineering, which will be helpful to bridge the knowledge gaps between laboratory experiments and field-scale investigation. This mini-symposium is intended to provide a forum for presentation and discussion of recent advances in computational approaches to geomechanics. Topics within the scope of interests include, but are not limited to, the following aspects:
(1) Constitutive models, including development, implementation and validation;
(2) Mesh-based and grid-based computational methods for soils, rocks, concretes and snows
(3) Multiscale modeling techniques;
(4) Multiphysics coupling analysis in geomaterials;
(5) Data-driven and machine learning techniques in geomechanics
(6) Large-deformation modelling of geohazards and other geotechnical engineering
(7) Numerical simulations of damage, fracture and strain localization processes.
(8) Large-scale modeling and high-performance computing of geomaterials and geotechnical engineering
Keywords: Computational geomechanics, Geomaterials, Multiphysics environments, Multiscale modelling
Heterogeneous porous media are present in diverse natural and engineering systems, ranging from biological tissues and geological formations to ceramics and foams. These systems exhibit inherent multi-scale characteristics, spanning a wide range of spatial scales. Porous materials are also known for their distinctive properties such as low resistivity, thermal conductivity, and density just to name a few. In most porous systems, the functionality of the material is influenced and controlled by the movement of fluids, solutes, particles, electrical charges, and heat through their porous network. Understanding the physical, chemical, thermal, and biological processes within these porous systems, including fluid flow, diffusion, dissolution, degradation, shrinkage, fracturing, and electrical charges, is crucial for understanding their response to different loading conditions as well as optimizing their performance. However, the presence of diverse pore structures with different geometrical shapes, orientations, and configurations across various length scales presents challenges in numerically characterizing these systems. This session invites scientific and engineering contributions to this field by improving or developing computational methods, including but not limited to:
• modeling Thermo-Hydro-Mechanical-Chemical processes in porous media
• poro-mechanical coupling schemes
• multi-scale modeling method for fracture in porous media
• data-driven modeling and computation in porous media
• integration of experiments and modeling
We specifically invite participation from undergraduate and graduate students, postdocs, early careers, and individuals from minority backgrounds to share their research outcomes and contribute to a more inclusive session.
Keywords: Computational damage, Computational Material Science, multi-physics, Multi-scale modelling, Poromechanics, porous media, Scientific Machine Learning
Join our mini-symposium on Computational Applied Mathematics, featuring topics such as kinetic theory, machine learning models, particle-laden flows and complex fluids and energy systems.
The symposium will cover a wide range of subjects, including the behaviour of particles in kinetic systems and the development of numerical methods to study them. We will also explore detonation wave solutions, investigating their formation, propagation, and stability through computational simulations. Machine-learning models will be explored in the context of solving complex mathematical problems. Their applications in numerical analysis, optimization, and data-driven modelling will be discussed, emphasizing their ability to extract meaningful information and accelerate computations. The relationship between fluids and energy will be investigated, examining energy conversion processes, fluid-structure interactions, and the optimization of energy systems. Computational simulations and data-driven approaches will provide insights into sustainable and efficient solutions. Viscoelastic fluid flows will be examined, focusing on the behaviour of fluids with both viscous and elastic properties. Computational approaches will be explored to analyse non-Newtonian behaviour, elasticity-induced instabilities, and complex flow patterns
Join us as we bring together researchers and practitioners to explore the frontiers of Computational Applied Mathematics.
Keywords: Complex Fluids, Energy Systems, Kinetic Theory, Particle-laden Flows
In many processes in industrial applications and natural sciences, the evolution of interfaces is of paramount importance. Examples occur in a wide range of research areas including multi-phase flows, crack propagation, fluid-structure interaction, solidification, crystal growth and biomembranes. The phase-field methodology is a powerful mathematical modeling approach for systems with moving interfaces like these. In the phase-field method, moving boundary problems are reformulated as PDEs on fixed domains in which the interface evolution is governed by a PDE of a scalar order parameter (the phase field). Phase-field models are diffuse-interface models meaning that the interface is a smooth region described by the smooth phase field.
The phase field method has favorable properties, such as a rigorous thermodynamical structure and a physical interface description, but introduces new challenges for computations. Important challenges include the discretization of higher order spatial derivatives that typically occur in phase- field models, the design of thermodynamically stable numerical methods (both in space and time) and the treatment of a relatively sharp interface. This minisymposia is dedicated to modeling and computation with the phase-field method. We welcome talks on novel phase-field modeling approaches and numerical algorithms as well as applications in fluids, solids and biomechanics.
Keywords: Additive Manufacturing, Advanced Materials, coarse graining, Complex Fluids, Computational damage, computational fluid dynamics, Computational Mechanics, Constitutive modeling, Failure mechanics, Finite element method, Homogenization, Microstructure, Moving Boundaries, Multiphysics, Multiscale, Numerical methods, Partial Differential Equations
With the emergence of advanced numerical methods for solving partial differential equations using spline-based functions, such as Isogeometric Analysis, the construction of appropriate function spaces capable of representing complex geometries with high regularity or possessing local refinement/coarsening properties has become increasingly crucial. The primary objective of this symposium is to bring together experts who can share their theoretical findings, algorithms, and efficient implementations, as well as applications, related to the development of complex analysis-suitable geometries, potentially involving multiple patches, with high regularity. Additionally, topics of interest for the symposium include local refinement and coarsening, adaptivity, and other relevant areas. Numerical studies from any field of Computational Mechanics, ranging from simple model problems to complex multi-physics ones, both in the boundary-fitted or in the immersed framework, are highly encouraged.
Keywords: Geometric Modeling, IsoGeometric Analysis, numerical simulation, Partial Differential Equations, Spline
Turbulence is a nonlocal and multi-scale phenomenon that presents several computational challenges. Resolving all scales implies nonlocality is addressed implicitly, as it does not require a closure model, but it is computationally prohibitive. Inorder to keep the computational cost feasible, spatially or temporally averaged fields have been considered in the literature in conjunction with modeling the discarded scales explicitly, with an eddy-viscosity closure model. These methods had relatively good success, but non-negligible limitations. Thus, properly addressing and capturing turbulent behaviour still remains an open challenge.
In this mini-symposium we gather experts that have taken non-standard approaches to the simulation of turbulent behaviour, including nonlocal derivatives, implicit large eddy simulations (LES), and deep learning. As an example, paper [1] shows how the classical eddy-viscosity model can be generalized by introducing a nonlocal, fractional stress-strain relationship. In [2] it is shown that physical dissipation can be matched to numerical dissipation when the contributions of discarded scales are small; this approach is known as implicit LES. Finally, paper [3] is an example of a more popular deep learning approach to physics modeling; here a case of separated flow is solved without any turbulence model, via neural networks.
REFERENCES
[1] Mehta PP. Fractional and tempered fractional models for Reynolds-averaged Navier-Stokes equations. arXiv preprint arXiv:2305.00770. 2023 May 1.
[2] Girfoglio M, Quaini A, Rozza G. A POD-Galerkin reduced order model for a LES filtering approach. Journal of Computational Physics. 2021 Jul 1;436:110260.
[3] Eivazi H, Tahani M, Schlatter P and Vinuesa R. Physics-informed neural networks for solving Reynolds-averaged Navier–Stokes equations. Physics of Fluids. 2022 Jul 7;34(7):075117.
Keywords:
Scientific Machine Learning, Fluid Mechanics, Implicit Large Eddy Simulations, Reduced Order Modeling, Turbulence Modeling, CFD, Fractional Calculus, Turbulence
Reduced order methods (ROMs) are crucial for fast and accurate numerical predictions in engineering applications, especially when dealing with many-query scenarios in optimization, uncertainty quantification, and parameter estimation. Many classic model order reduction approaches - such as proper orthogonal decomposition or reduced basis methods - have a solid mathematical foundation that guarantees approximation accuracy and keeps the ROM interpretable to the governing physical laws. However, many practical applications are too complex (e.g., large Kolmogorov n-width) or inaccessible (e.g., private or legacy codes) for classic ROMs to approximate reliably. Therefore, in recent years, many data-driven and non-intrusive techniques have been introduced to enhance ROMs by exploiting additional data from various sources of data such as experiments or full-order computations. Under the umbrella of scientific machine learning, this combination of domain knowledge, physical principles, and artificial intelligence offers the advantages associated with machine learning techniques while remaining physically interpretable. However, many open problems still need to be solved to reliably merge these techniques and create stable ROMs, especially for practical applications with computationally infeasible high-fidelity simulations.
This mini-symposium aims to present recent computational strategies to improve the construction of ROMs with data. We also wish to foster discussions about potential future research directions in linear and nonlinear model order reduction, scientific machine learning, and data-driven methods.
REFERENCES
[1] Benner P, Grivet-Talocia S, Quarteroni A, Rozza G, Schilders WHA, Silveira LM., eds. Model Order Reduction Volumes 1-3. Berlin, Boston: De Gruyter. 2021
[2] Ghattas O, Willcox K. Learning physics-based models from data: perspectives from inverse problems and model reduction. Acta Numerica 30:445-554, 2021.
Keywords: Data-driven Models, Reduced Order Modeling, Scientific Machine Learning
The physical processes of interest to contemporary science and engineering are growing ever more complex. As a result, their governing equations are becoming high-dimensional or even partially unknown. For the efficient analysis, prediction, design, uncertainty quantification and control of these processes, reduced-order models capturing the core of the underlying physical phenomena are a must.
This minisymposium focuses on the recent developments in computational methods and tools for rigorous model reduction for nonlinear computational mechanics problems. In addition to the advances in classic projection-based reduced-order models, this symposium aims to highlight modern reduction methods that leverage dynamical systems theory of invariant manifolds for equation-driven reduced-order modelling on the one hand, and dynamics/mechanics-informed machine learning methods for data-driven model reduction on the other hand. The speakers at this minisymposium will show applications of data-driven as well as equation-driven model reduction methods in various fields of nonlinear computational mechanics such as structural dynamics, fluid mechanics, fluid-structure interaction, micro-electromechanical systems, soft robots among others.
Keywords: Data-driven model reduction, Invariant manifolds, Model order reduction, Nonlinear finite elements, Nonlinear model reduction, Nonlinear normal modes, Spectral submanifolds, Computational mechanics
Fusion based on magnetic confinement aims at producing power by using the energy liberated by deuterium and tritium nuclei reacting at extremely high temperatures (107 - 108 K), thus resulting in a plasma that is confined by magnetic fields in machines of the toroidal shape known as tokamaks. Numerous technological and scientific challenges remain that require a sustained research effort. Foremost among these challenges is the issue of power exhaust. The control of heat fluxes onto the tokamak walls in high energy confinement configurations and for both steady-state and transient regimes must be addressed to run future ITER experiments successfully. The proposal's primary goal is to review advanced numerical tools in finite volume and finite element frameworks, explicitly focusing on highly magnetized plasmas.
1) Consider the divergence-free property of the magnetic field, either by divergence cleaning or the use of the magnetic potential. When using magnetic potential, the C1-regularity of the numerical approximation is often desired.
2) Design well-balanced schemes that can preserve a given equilibrium for a very long time without initial perturbations.
3) Transport models for large scales MHD instabilities as disruptions.
4) Design of Efficient numerical strategies for MHD and reduced-MHD models.
Our interests are modeling and design of accurate and efficient numerical strategies. We review numerical difficulties and highlight some recent axes of improvements.
Keywords: magnetohydrodynamic, Multiphysics problems, Nonlinear finite elements, Numerical methods
Shallow flow models provide a far more practical, from the computational standpoint, engineering alternative to the full Euler or Navier-Stokes equations to model free surface flows. This model integrates vertically incompressible flow equations from the topography bed to the flow-free surface (depth integration). Derivations often assume 1) a relatively thin layer flow; 2) minor velocity fluctuations along the flow depth (weakly sheared flow); 3) hydrostatic pressure distribution; 4) a slight bed slope. Depth integration leads to the removal of the need to resolve the free surface explicitly, the reduction of space dimension, and, hopefully, the number of equations to be solved. Despite the reduction and associated simplifications, shallow flow models yield reasonable predictions of some natural process as debris flows, landslides, avalanches, river flows, and even more. However, the classical shallow water model fails in the context of strongly sheared geophysical flows on complex topography. In this context, we must go beyond some modeling assumptions on velocity fluctuation, slopes, and curvature. The usual modeling assumes that the horizontal velocity is weakly varying in the vertical coordinate, which implies that the vertical shear is negligible. The horizontal velocity is the depth average of the three-dimensional velocity field. Since the classical shallow water models assume negligible vertical shear, they cannot model large-scale eddies (rollers) that appear near the surface and behind the hydraulic jump. Under the assumption of the smallness of horizontal vortices, a more general model, the shear shallow water model (SSW), can be derived, including the second-order velocity fluctuation terms. However, the model is principally hyperbolic and nonconservative, posing difficulty in its numerical resolution.
We propose to gather active researchers focussing on this model to clarify the situation on different numerical strategies available: advantages and drawbacks. Then we will discuss future directions for investigations.
Keywords: geophysical fluid dynamics, Modelling and simulation, Numerical methods, turbulence modeling
Kinetic partial differential equations play a fundamental role in describing various phenomena that involve a large number of interacting particles. These models have been adopted effectively in several research fields, ranging from classical rarefied gas and plasma models to novel dynamics in socio-economical, life, and computing sciences. In this mini-symposium, we aim to bring together experts on modern applications of kinetic and mean-field equations in the context of uncertainty quantification and optimization (see [1-2] and the references therein).
The construction of numerical methods for kinetic equations with random inputs is indeed a problem attracting the attention of many researchers, due to the difficulties linked, e.g., to the high dimensional structure of the equations, the conservation of the structural physical properties, and the preservation of the equilibrium state. These challenges open new fascinating questions, both from an analytical and numerical viewpoint, that nowadays are of great interest.
In the context of computing science, many optimization methods or machine learning models can be modeled as stochastic dynamical systems in a high dimensional space. The high dimensionality may stem from the data, the model architecture, or the many agents involved. Due to their statistical nature, kinetic and mean-field equations have recently proved to be essential tools for a mathematical understanding of such systems. The mini-symposium intents to provide an overview of recent applications of kinetic theory in this field, with a focus on optimization and inverse problems.
Prof. Mattia Zanella, Prof. Liu Liu, Dr. Rafael Bailo, Dr. Yuhua Zhu, Dr. Urbain Vaes, Dr. Elisa Iacomini, Dr. Alessandro Scagliotti, Konstantin Riedl, and Anjali Nair have already shown interest in participating at the mini-symposium.
[1] J. A. Carrillo, Y.-P. Choi, C, Totzeck and O. Tse, An analytical framework for consensus-based global optimization method. Mathematical Models and Methods in Applied Sciences 28.06 (2018): 1037-1066.
[2] S. Jin and L. Pareschi, Uncertainty quantification for hyperbolic and kinetic equations, Cham, Switzerland: Springer International Publishing, 2017.
Keywords: plasma models, stochastic interacting particle systems, uncertainty quantification, Kinetic equations, optimization
Numerical simulation in electromagnetism has undergone a relevant development in recent years since it represents a key tool in the design of devices and processes, as well as in obtaining improvements in the operation of existing prototypes.
Many industrial applications involve electromagnetic problems which are often coupled with other physical models that require the solution of a coupled multiphysics problem. Therefore, it is essential to have suitable and efficient numerical schemes to approximate the electromagnetic submodel derived from Maxwell's equations. Furthermore, nowadays the numerical solution of electromagnetic goes frequently accompanied with numerical control and optimization; thus, the development of efficient algorithms is crucial to get an affordable optimization model from a computational point of view.
In the literature we can distinguish two great approaches in the research field of electromagnetism: development of new models and algorithms and mathematical and numerical analysis of the models to give a theoretical support to the numerical techniques. Both approaches have provided relevant results both in the applied mathematics and in the electric engineering fields. In this context, the aim of the proposed minisymposium is to offer a forum for the discussion of recent advances in computational electromagnetism with a special focus in industrial applications. The minisymposium will try to cover several topics: techniques in modelling and discretization, multiphysics approaches, coupling between distributed and lumped models, numerical optimization, material modelling, mathematical and numerical analysis, etc.
This minisymposium yields in the framework of ECMI's Special Interest Group MSOEE (Modeling, Simulation and Optimization in Electrical Engineering).
Keywords: computational electromagnetism, discretization techniques, multi-physics, optimization
In recent years, the production of huge amount of data in computational sciences has made attractive the capability to exploit such data to extract knowledge and enhance the prediction level. In aerodynamics, parametric studies, trade-off analyses and optimizations represent a precious information tank which could foster the usage of data-driven and data-fusion models in engineering practice. However, the maturity level of such models is quite low and the associated best practice is still in the preliminary stage: on one hand, machine learning techniques and neural networks are well-known and offer a wide range of choice for different purposes, from cluster analysis and dimensionality reduction to classification and regression; on the other hand, the type and preparation of aerodynamic/geometric data to be handled is not straightforward and may strongly depend on the real scope of the task, giving rise to widely different interpretations and forms of the data-driven application. Machine learning techniques commonly used in the area of Artificial Intelligence (AI) and Data Mining (DM) can represent a valuable support to reduce the computational cost required for aerodynamic analysis and uncertainty quantification.
This minisymposium aims at collecting and disseminating new ideas in application of machine learning and data-driven approaches for aerodynamic analysis and uncertainty quantification focusing on real world problems. This minisymposium also aims to disseminate the main activities and results of the GARTEUR action group AD/AG60 on this topic.
Keywords: aerodynamic analysis, machine learning
The advent of advanced manufacturing and materials technologies now provides the capabilities to architect microstructured materials such as 3D printed lattice structures, fiber-reinforced or multiphase composites, foams, electro- or magneto-active polymers, etc. The mechanical and multifunctional behaviors of these metamaterials can be tailored to their specific engineering applications and are often highly nonlinear, anisotropic, inelastic, and multiphysical. Thus, classical constitutive models are typically not flexible enough to model their effective material behavior in multiscale and multiphysics simulations, while concurrent multiscale approaches are inherently computationally expensive and slow. Thus, in recent years, the formulation of constitutive models using highly flexible machine learning and surrogate modeling methods such as artificial neural networks and deep learning, Gaussian processes, radial basis functions, clustering methods, etc. has gained momentum. Nevertheless, many challenges remain to be addressed for machine learning-based material models, such as their accuracy, reliability and physical soundness, their efficiency, the consideration of parametric dependencies or uncertainties, etc.
This minisymposium welcomes contributions on the state-of-the-art of machine learning methods for multiscale and multiphysics materials modeling. In particular, the areas of interest include, but are not limited to:
- Material models based on feed-forward, deep, recurrent, convolutional, graph and other types of neural networks, or Gaussian processes, radial basis functions, clustering methods, etc.
- Models for elastic, as well as dissipative, inelastic (elasto-plastic, visco-elastic, etc.), and multiphysically coupled (electro, magneto, thermo, chemo, mechanical, etc.) material behaviors
- Physics-enhanced/informed/augmented machine learning methods for thermodynamically consistent, physically, and mathematically sound material models
- Consideration of parametric dependencies, uncertainties, adaptivity, error estimates, etc. in machine learning methods for material modeling
- Efficient implementation and application of machine learning methods for multiscale and multiphysics simulations
Keywords: Computational Mechanics, machine learning, Multiphysics, Multiscale
MS014 Accelerating scientific discovery for dynamical systems with physics-informed machine learning
Machine learning (ML) algorithms continue to see increased prominence in the domain of scientific computing. Such algorithms are particularly promising in that they seek to extract useful models from large data sets while bypassing the classical challenges of the curse of dimensionality[1]. Physics-based machine learning provides added benefits in the matter of applying such algorithms to computational science problems, by virtue of preserving or imposing desirable qualities in an algorithm to avoid non-physical behavior, enhanced interpretability, stability, data-efficiency, etc.
This minisymposium focuses on the development and utilization of novel physics-based machine learning strategies to accelerate scientific discovery for, and enhanced understanding of complex dynamical systems. Examples of adjacent topics include the use of reduced-order modeling techniques to effectively capture the essential dynamics of high-dimensional systems, strategies that leverage data-driven approaches to enhance our understanding of complex dynamics, the fusion of physics-based knowledge with machine learning algorithms, and efficient optimization or control of dynamical systems using machine learning algorithms. We welcome investigations of novel techniques, such as state-of-the-art deep learning architectures, reinforcement learning, and generative models, to accelerate scientific discovery and extract hidden patterns in dynamical systems. We also welcome submissions from researchers utilizing such techniques for various applications and envision interdisciplinary collaboration through their presentations.
Keywords: Control, Dynamical systems, machine learning, Surrogate models
High-performance and relatively inexpensive hardware components are capable of supporting advanced Structural Health Monitoring (SHM) paradigms, producing large datasets growing in size and complexity. Machine learning (ML) allows dealing with such large amounts of data and ensures learning from them, even when no prior knowledge of the investigated phenomena is available. However, purely data-driven approaches can mostly learn and predict behavioural patterns that were somehow experienced or simply provide evidence of deviations from baseline conditions. Thus, they require training on data that are representative across a broad range of possible environmental and operational conditions, are prone to overfitting and poorly generalise to out-of-sample scenarios. Furthermore, they are less interpretable than physics-based models, which, despite their limitations, allow for a more transparent understanding of the relationships between variables and the reasoning behind the applied inference.
The integration of physics-based modelling, more intuitive for engineers, with ML techniques allows for the combination of the advantages of data-driven methods with the insights delivered by physics-based principles. This ensures an enhanced learning and predictive capability, particularly when training data are limited, by leveraging domain-specific knowledge. Various degrees of combination can be pursued within hybrid architectures, depending on the characteristics and objectives of the application.
The scope of this mini-symposium is to present advancements and emerging trends in physics-informed SHM, discussing open challenges and promising solutions to tackle them. It aims to constitute a platform for the exchange of ideas and experiences and foster more coordinated and interdisciplinary research on the wide spectrum of modelling strategies to enforce physical laws or constraints within the learning process.
We welcome contributions on the integration of physics-based with data-driven approaches focused on, but not limited to:
• Advanced Bayesian approaches for the fusion of SHM and non-destructive evaluation data.
• Empowering model-based SHM approaches with surrogate models.
• Leveraging digital twins to enhance SHM strategies.
• Network-level SHM using transfer learning techniques.
• Cutting-edge methods for computation-physical domain adaptation in SHM.
• Integrating physics and ML for enhanced SHM.
Keywords: Bayesian Inference, Damage Identification, Deep Learning, Digital Twins, Grey-box Modelling, Physics-Informed Machine Learning, Scientific Machine Learning, Structural Health Monitoring, Surrogate models, System Identification
Data-driven simulation methods are becoming extremely important as a tool to get insight in complex flows and multi-physics problems [1]. Firmly rooted in advances in data science and scientific machine learning, data-driven methods are having a tremendous impact in digital twins, flow control, forecasting, and many other fields. The purpose of this mini-symposium is to gather experts from the computational fluid mechanics community, as well as applied mathematicians and computer scientists to discuss the advancements in data-driven methods for simulation of flow and multi-physics problems. Contributions are welcome in the applications of data-driven methods in challenging problems, new methods and algorithms, computational aspects such as adaptive mesh refinement and coarsening, parallelism, data management and I/O, and libraries to support such developments.
Keywords: Data-driven methods, Fluid Flow, multi-physics, Scientific Machine Learning
Scientific machine learning (SciML) has emerged as a transformative field at the intersection of scientific computing and machine learning. This mini-symposium aims to explore the latest advancements on efficient SciML algorithms and their application to computational fluid dynamics (CFD) problems. This includes (but is not restricted to):
+ Neural network-based reduced order models for (high-dimensional) parametrized and inverse problems
+ Neural network-augmented numerical schemes for CFD problems
+ Multilevel and multiscale machine learning models
+ Scalable machine learning algorithms based on (space-time) domain decomposition + Data- and model-based parallelization of machine learning algorithms
CFD applications may range from model problems of simple linear flow models to highly nonlinear multiphase and multiscale flows as well as multiphysics problems which include a fluid component from real-world applications.
Keywords: Computational Fluid Dynamics, Domain Decomposition, Multilevel, Partitioning, Scalable Machine Learning, Multiscale models, Scientific Machine Learning
In 2019, "The European Green Deal" set an ambitious target to make Europe the first climate-neutral continent by 2050. A vital aspect of this transition is the development of sustainable and smart mobility systems, with a specific focus on railway transport [1]. This symposium aims to promote the use of advanced techniques in numerical analyses and experimental field tests to improve railway safety, efficiency, and resilience. By bringing together experts from academia and industry, the symposium will facilitate knowledge exchange, foster collaborations, and encourage the development of adaptive infrastructure systems to achieve the goals set forth by "The European Green Deal". The proposed symposium has the following objectives: i) explore innovative approaches and technologies for enhancing railway safety, efficiency, and resilience; ii) facilitate collaboration and partnership opportunities among researchers, practitioners and industry experts; iii) discuss strategies to address the challenges associated with increasing rail capacity high-speed rail lines and maintenance costs; iv) showcase successful case studies and best practices in railway infrastructure development and management; v) identify future research directions and potential policy interventions to support sustainable and resilient railway systems.
Keywords: bridge dynamics, computer vision and image analysis, data science, data-driven approach, machine learning, predictive maintenance, prognostic models, smart mobility, track-bridge interaction, vehicle-bridge interaction, wayside/drive-by condition monitoring, Wind turbines
The revolutionary integration of artificial intelligence (AI) and machine learning (ML) in earthquake engineering is driven by the expansion of seismic networks, resulting in a future abundant with extensive data encompassing building damage and earthquake records. This integration facilitates the acquisition of valuable insights from extensive data for seismic risk assessment, thus contributing to the development of disaster-resilient cities. Seismic risk assessment encompasses crucial components such as seismic hazard, vulnerability, and exposure. The integration of AI and ML algorithms has significantly enhanced the accuracy and reliability of these analyses, enabling seismic risk assessment on a larger scale and with increased accuracy.
This minisymposium proposal aims to provide a comprehensive overview of recent advancements, focusing on AI and ML approaches in deterministic and probabilistic seismic hazard analysis, including seismological and ground motion simulation techniques. Aligned with the importance of seismic hazard assessment, the minisymposium aims to explore a progressive approach that encompasses dynamic analysis of structures, vulnerability assessment, the impact of earthquakes on building responses, and exposure. Through a comprehensive exploration of these topics, we aim to encompass the latest studies utilising AI/ML techniques in the field of risk assessment, all with the ultimate goal of fostering the development of resilient cities. Participants are invited to share their research, methodologies, and case studies, fostering knowledge exchange and collaboration to advance seismic risk assessment and promote the development of resilient communities.
The minisymposium will cover various topics related to AI and ML techniques in seismic risk assessment, including:
• Big data analysis for signal processing and microzonation studies
• Seismic hazard assessment using deterministic and probabilistic frameworks
• Ground motion simulation and modelling techniques
• Dynamic analysis of structures under seismic loading
• Seismic vulnerability assessment
• Development of exposure models
• Development of seismic risk assessment framework
• Optimisation approaches in seismic risk mitigation studies
These topics, among others, will be explored to uncover new insights and approaches for effective seismic risk assessment and resilient city planning.
Keywords: Exposure modelling, Seismic hazard assessment, Seismic risk assessment , Vulnerability analysis, Artificial intelligence (AI)
ABSTRACT
Optimal control of dynamical systems has become a crucial task in the scientific and engineering community. Some examples are related to drag and turbulence reduction, efficient heat transfer, and shape optimization. Many of these systems are described by high-dimensional, nonlinear, and time-dependent partial differential equations (PDEs). The complexity of these systems hinders the use of traditional optimal control methods, such as, e.g., the Hamilton-Jacobi-Bellman equation or the adjoint method, due to poor scaling properties with respect to the problem dimensionality.
Deep learning has emerged as a powerful tool for tackling the challenges of reduce order modelling, and operator inference for high-dimensional, and complex PDEs. However, only in recent years, data-driven approaches have been applied to optimal control of PDEs. In particular, deep reinforcement learning (DRL) results to be a way to perform scalable optimal control from measurements. DRL is a mathematical framework for solving sequential decision-making problems from the interaction of the agent, i.e., the decision maker, and its environment. DRL has achieved outstanding success in different fields such as video games, robotics, finance, and even health applications. However, despite the success, DRL has only recently been employed to control complex PDEs leaving a vast gap in the literature [1].
This minisymposium will gather a broad spectrum of contributions in the very active and recent research field of DRL for optimal control strategies discovery, emphasizing theoretical aspects, computational performance, and practical examples in scientific applications.
REFERENCES
[1] Colin, V., Rabault, J. and Vinuesa, R. “Recent advances in applying deep reinforcement learning for flow control: Perspectives and future directions†Physics of Fluids 35.3 (2023).
[2] Botteghi, N. et al. “Low dimensional state representation learning with robotics priors in continuous action spaces†2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2021).
Keywords: Deep reinforcement learning, optimal control, PDEs
Machine learning (ML) in scientific applications including computational fluid dynamics (CFD) is a growing field of research. The broad range of ML techniques available and their ability to learn unknown and complex correlations enable a large spectrum of applications. However, in CFD, physical constraints are invaluable to obtain reasonable predictions. Another problem is that numerical simulations, especially high-order methods, are susceptible to instabilities due to inaccurate predictions, which the learning algorithm has to account for. Moreover, the definition of a suitable input space and loss function is a crucial and difficult task due to the highly nonlinear and mostly unknown mapping which has to be learned. Thus, the ML algorithm has to be consistent to the numerical discretization used, which a supervised learning method is never fully aware of.
With these shortcomings of supervised learning techniques in mind, recently, the focus of research has concentrated on reinforcement learning (RL) or physics-informed methods applied to CFD. While state-of-the-art RL approaches are widely utilized in the ML community, their applicability to, e.g., CFD, is an ongoing research topic. The relatively slow adaptation of state-of-the-art RL techniques to CFD can be attributed to the rapid development of new RL methods and the increasing complexity of problems in CFD compared to typical applications such as game theory. Common examples for RL in CFD are flow control, turbulence modeling and shock capturing. An additional and continuously growing field of research which alleviates the common problems of ML in CFD are physics-informed neural networks (PINNs). In general, classical PINNs are specifically suited for smooth problems and suffer from stability problems if discontinuities are present in the solution. This is the case in many real applications such as transonic flows. Recently, modified versions of classical PINNs have been proposed to push their limitations and to enable PINNs which are more tailored to CFD. With these considerations in mind, the objective of this minisymposium is to discuss the applicability, predictive performance and limitations of state-of-the-art ML methods in CFD. This can include data-enriched numerical methods based on RL such as closure models for turbulence or flow control as well as more advanced PINNs pushing the limits of their classical counterpart.
Keywords: Deep reinforcement learning, Numerical methods, Physics-Informed Machine Learning, Scientific Machine Learning
The development of high-fidelity efficient solvers is critical for accurate and reliable simulation of large-scale, multi-physics problems in computational science and engineering applications. However, many-query problems such as inverse design and parameter inference often place a computationally intractable burden on these solvers. Moreover, the characterization and propagation of systemic variability and uncertainties in input parameters are also of fundamental importance in model-based computations of various science and engineering problems. Recent advances in data-driven analysis of systems and machine learning (ML) approaches have revolutionized the modeling of engineering systems by presenting effective ways to tackle the computational bottleneck. Development of fast surrogate models, data-driven augmentation of existing high-fidelity solvers or identification of governing laws, the design of optimized hybrid methods that combine data-driven learning processes with physics-based models, and the integration of ML algorithms with classical solution techniques for inverse problems using tools like learning regularizers and deep generative priors, are some of the exciting new avenues that have helped to address many previously unattainable challenges in modeling and analysis of complex engineering systems and structures.
This minisymposium will be devoted to fundamental developments as well as challenging applications of data-driven methods for such forward and inverse problems in computational fluids and environmental flows, among others. Topics of interest include but are not limited to, advanced physics-informed ML strategies for engineering systems, multi-fidelity training, applications of transfer learning, neural operators, and deep generative modeling
Keywords: Computational Fluid Dynamics, Environmental Flows, Inverse Problem, Model Order Reduction, Operator Learning, Scientific Machine Learning
Reduced Order Models (ROMs) and Artificial Intelligence (AI) have gained significant attention in industrial applications thanks to their ability to reduce the computational cost and time of physics-based computational methods, while still maintaining a sufficient degree of accuracy, which enables the time-effective completion of decision-making processes. On the one hand, ROMs reduce the dimensionality of high-fidelity models while retaining essential features for faster and still accurate modelling of complex physical and engineering systems. On the other hand, AI extracts relevant features from available data to develop (1) effective Machine Learning (ML) surrogate models that temporarily replace expensive calculations and (2) generative models (GMs) that judiciously guide the exploration of physical parameter spaces for effective design and calibration of complex systems.
The goal of the minisymposium is to bring together recent scientific advances in ROMs, ML surrogate models and GMs, from both academia and industry, in the most diverse tasks regarding industrial production. Key applications and areas of interest where ROMs and AI are particularly relevant include i) manufacturing process optimization; ii) structural analysis and design, particularly in industries such as aerospace and automotive, for virtual testing and predictions on structural responses; iii) energy systems and grid optimization for managing power generation, transmission and distribution; iv) fluid dynamics and aerodynamics such as airfoil design, flow control and turbulence modelling in the areas of aircraft and turbines production; v) robotics, for the reduction of complex engineering processes and advanced control strategies; up to vi) smart predictive maintenance and supply chain optimization.
This minisymposium is thought as a mathematical interdisciplinary platform for the discussion on the state-of-the-art concerning a smarter industry with a lighter impact on computational costs and times. As far as the numerical mathematical modelling is concerned, we encourage the dissemination of works both methodologically and more application-oriented related - and not limited – to artificial intelligence, deep and reinforcement learning, data-driven (reduced order) models and computational sciences.
Keywords: Artificial intelligence (AI), Deep Learning, Dimensionality Reduction, Neural Networks, Reinforcement Learning
Presently, the literature offers some relevant works applying artificial intelligence technology in the field of computational mechanics [1]. Nevertheless, this is still a novel field that has not been properly shared and discussed among the computational mechanic's community [2]. Thus, the aim of this symposium is to present and discuss how to apply artificial intelligence methodologies, such as machine and deep learning, to computational mechanics. Thus, we welcome research works related to computational mechanics combined with artificial intelligence, aiming to deal with experimental data, to increase the simulation's computational efficiency, by increasing its accuracy and reducing its computational cost, enabling more comprehensive computational models and digital twins.
References:
[1] Ribeiro JP, Tavares SM, Parente M. Stress–strain evaluation of structural parts using artificial neural networks. Proceedings of the Institution of Mechanical Engineers, Part L: Journal of Materials: Design and Applications. 2021;235(6):1271-1286. doi:10.1177/1464420721992445.
[2]Jorge Belinha. (2021). Artificial Intelligence in Computational Mechanics and Biomechanics. Journal of Computation and Artificial Intelligence in Mechanics and Biomechanics, 1(1), 1–6. https://doi.org/10.5281/zenodo.4669522
Keywords: Artificial intelligence, Deep learning, Machine learning, Computational mechanics, Digital twins
Over the past few decades, sectors like aerospace and wind energy (among others) have experienced significant impacts from the integration of high-performance composites. However, a challenge in modelling and designing composites is the limited computational efficiency of precise high-fidelity models. Conventional optimization methods for design often lead to complex procedures due to the large dimensions of the design space and the computational burden associated with high-fidelity simulations. In recent times, machine learning approaches have emerged as promising techniques to enhance the efficiency and reliability of various approaches for modelling of composites [1]. These techniques offer a powerful approach to capture and understand the intricate behavior of composite materials by leveraging vast datasets and identifying underlying patterns. By training models on different sources of data, machine learning can accurately predict physical and mechanical properties, failure mechanisms, and durability, enabling efficient material design and reducing the need for extensive and costly experimental and computational testing programs [2]. Furthermore, machine learning can facilitate the optimization of composite processing parameters, aiding in the achievement of desired material properties and reducing manufacturing defects. The application of machine learning techniques in composite materials not only enhances our understanding of these advanced materials, but also enables their broader utilization across industries such as aerospace, automotive, and renewable energy, where lightweight and high-strength materials are in high demand. In this mini symposium, we will discuss the state-of-the-art developments in usage of machine learning techniques for modelling, design, and process optimization of composite materials. Contributions on a wide range of topics are encouraged, from traditional and physics-enhanced machine learning for surrogate modelling, to learning and exploiting latent representations of material behavior, generative learning for material optimization, efficient data assimilation and active learning, among others.
Keywords: Composite, Machine learning
Partial Differential Equations (PDEs) are central to the simulation and understanding of complex physical phenomena in Computational Mechanics. However, traditional numerical methods have difficulties coping with the computational cost of solving these PDEs due to the small spatiotemporal scales that need to be resolved. The recent advances in Artificial Intelligence and machine learning have offered promising avenues by introducing novel strategies. For instance, reduced order-models based on non-linear variational auto-encoder architectures or neural operators to directly solve forward and/or inverse problems have shown promising results. In these methods, the quantification of uncertainty plays a crucial role as incomplete models, the information loss due to the dimensionality reduction as well as uncertain input parameters have to be taken into account.
This mini symposium will be discussing how data- or knowledge-based strategies can be used for Computational Mechanics to increase both speed and accuracy of obtain a solution to a given forward or inverse problem. Given the aforementioned uncertainties in both model and data, a special focus is given to probabilistic modeling approaches as well as the algorithmic tools that enable those innovative approaches.
We are inviting researchers to share their research findings, challenges, and potential applications of machine learning and Uncertainty Quantification in solving high-dimensional PDEs. The symposium will foster an interdisciplinary environment where researchers from the fields of Computational Mechanics, Artificial Intelligence, and Statistics can exchange ideas, collaborate, and explore new directions in this evolving domain.
Keywords: Artificial Intelligence, Computational Mechanics, Data-driven Models, Scientific Machine Learning, Uncertainty quantification (UQ)
The understanding of processes and phenomena in science and engineering is radically transformed by machine learning. Many scientists and engineers embrace machine learning as an important tool. At the same time, obstacles and challenges are becoming apparent: most machine learning approaches require large amounts of data, but in many applications data is scarce. Furthermore, the performance and reliability of artificial neural networks − the dominant type of ‘learned machines’ − is usually difficult to interpret.
The ultimate goal of this mini-symposium is: to reveal how neural networks can be made more effective and efficient, and better understood, by incorporating mathematical and physical knowledge into their design. The direct goal of the mini-symposium is to contribute to the development of theory for mimetic neural networks for data-efficient and well-understood use in computational science and engineering.
To achieve the above, it is necessary to have contributions from multiple disciplines. The mini-symposium will have expert speakers from mathematics, computer science, machine learning, physics and astronomy.
From a mathematical point of view, there are many open questions. Overarching is the question: How can we optimally embed prior knowledge into neural networks? More specific questions are: How can we create interpretable machine-learning models? How can we further optimize training algorithms for neural networks? What are the nonlinear stability conditions and other requirements of neural networks?
Answers to these questions are very important for among others fluid mechanics and astro-mechanics, with essential opportunities for cross-pollination and mutual benefits for both. Work by astronomers on the N-body problem demonstrates that neural networks can be used to make N-body computations, with N»1, much more efficient. In fluid mechanics, machine learning methods for the analysis, modelling, and control of turbulent flows are currently developed to answer both fundamental and applied questions. The intrinsic chaotic behaviour of multi-body systems in astro-mechanics resembles the non-trivial statistical properties of turbulence in fluid mechanics.
The mathematically inclined contributions to the mini-symposium will concentrate on fundamental properties of neural networks. The fluid mechanics and astro-mechanics contributions concentrate on specific challenges which serve as test cases for potentially more general strategies.
Keywords: chaos, differential equations, machine learning, mimetic methods, N-body systems, stability, structure preservation, turbulence models, neural networks
This mini-symposium centers on the profound role of Machine Learning (ML) technologies in amplifying computational performance of simulation tools, a critical determinant in the advancement of Computer Aided Engineering (CAE) in the industrial sector. In the last years, it has been demonstrated through various Proof of Concepts (PoCs) that ML can significantly enhance the computational performance of solvers, in areas ranging from computational fluid dynamics to system simulations. However, the application of these technologies in real-world contexts and industrial workflows presents a unique set of challenges.
Hence, this symposium aims to delve into specific ML methods uniquely tailored for tangible applications, focusing on attributes such as error guarantees, minimal data requirements, potent extrapolation capabilities, and competitiveness compared to existing solvers. Contributions will be presented along real-world challenges and applications benchmarked along state-of-the-art industrial technologies.
We will conclude the mini symposium with a small panel discussion on required future research directions for ML solutions boosting industrial computer aided engineering.
Keywords: Numerical Solvers, Scientific Machine Learning, Computer Aided Engineering, Hybrid Methods, Industrial Applications
Scientific machine learning has gained large interest in the engineering community as a balance between classic, physics based approaches and emerging data based techniques [1]. On one hand, the robustness and reliability of the physical equations can be enhanced by the potentially fast machine learning methods, that further allow incorporation of experimental, real-life data. On the other hand, including physics information into the learning process enhances the training efficiency by e.g. imposing important physical properties such as conservation and symmetry laws.
Advances in scientific machine learning has also led to the development of hybrid computational methods where machine learning is used to aid classical solution methods such as finite elements, finite volumes, numerical time integration, etc. Amongst others, trained algorithms can serve as surrogate models that are fast to evaluate [2,3]. Furthermore, new or augmented models can be inferred, allowing to include or learn unknown physical behavior [4].
The aim of this mini-symposium is to bring together researchers in the field of scientific machine learning to enhance the modelling or simulation process in applications both from the engineering domain as well as natural sciences. Topics include (but are not limited to):
• Surrogate modelling with machine learning methods for uncertainty quantification
• Structure-preserving machine learning techniques
• System and parameter identification methods for differential equations
REFERENCES
[1] Strang, G. (2019). Linear algebra and learning from data. Wellesley-Cambridge Press.
[2] Lu, L., et al. (2021). Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators. Nature machine intelligence, 3(3), 218-229.
[3] Kovachki, N., et al. (2021). Neural operator: Learning maps between function spaces. arXiv preprint arXiv:2108.08481.
[4] Prume, E., et al. (2023). Model-free data-driven inference in computational mechanics. Computer Methods in Applied Mechanics and Engineering, 403, 115704.
Keywords: Computational Mechanics, differential equations, Scientific Machine Learning, Structure Preservation, Surrogate Modelling, System Identification
The field of scientific machine learning (SciML) deals with the combination of scientific computing and machine learning. A major impact of SciML on computational modeling has been made by the development of efficient surrogates that expand the capabilities of traditional numerical methods by using data-driven models. However, data-driven surrogates also encounter challenges, including the computationally intensive training and the lack of explicit error control and robustness, which limit their reliability in real-world applications.
In this mini-symposium, we will discuss recent developments on the enhancement of data-driven SciML surrogates via numerical methods, and vice versa, to combine the strengths and alleviate weaknesses. We invite submissions on various topics concerning the integration of deep learning and classical numerical methods , encompassing, but not limited to:
• The development of novel types of SciML surrogates, including approaches that incorporate physical and geometric constraints into the architectural design of neural networks.
• The integration of state-of-the-art numerical methods with SciML techniques to enhance the training process.
• Techniques that combine SciML surrogates with state-of-the-art numerical solution methods for increased efficiency and efficacy.
• Hybridization of standard numerical methods (Finite Element, Iso-geometric analysis, etc.) with deep learning algorithms.
Keywords: Hybrid Methods, numerical methods, Scientific Machine Learning
The mini-symposium “Explainable AI for Computational Mechanics†will bring together researchers from academia and industry working on the development of efficient techniques that are suitable for enhancing the transparency, accountability, and understanding of Artificial Intelligence (AI) methods when applied in Computational Mechanics.
Computational Mechanics plays a crucial role in the design, analysis, and optimization of complex mechanical systems, where precise predictions and decisions are essential for ensuring safety and efficiency in real-world situations. However, merely providing inference is no longer sufficient. Thus, Explainable AI (XAI) is getting an increasingly important role in Computational Mechanics. Before deploying AI methods, domain experts and engineers seek answers to questions such as how a machine learning model works, what parameters play a role in the prediction and what uncertainties and biases are incorporated in the data and the models.
Our mini-symposium welcomes contributions aimed at providing innovative perspectives on the usage of XAI in Computational Mechanics tasks related to the study of structures, materials, and fluids. We will consider both applied and methodological studies, offering valuable insights into (1) concrete results of using XAI to make better-informed decisions in engineering design problems, and (2) interesting XAI methods that show high application potential in real-world situations, including the field of Computational Mechanics.
Contributions may address new developments in the following subfields (non-exclusive list):
• Model prediction interpretation
• Feature Importance and Sensitivity Analysis
• Uncertainty Quantification
• Algorithm Design and Model Selection
• Human-centered Design
• Anomaly Detection
We look forward to receiving your contribution!
Keywords: Algorithm Design, Anomaly Detection , Computational Fluid Mechanics, Computational Solid Mechanics, Interpretable Machine Learning, Sensitivity Analysis, Uncertainty Quantification, Variable Screening, Explainable AI
Digital twins are virtual representations of physical objects, systems, or processes that allow for real-time monitoring, analysis, and simulation. Essentially, they are digital replicas used to model and optimize the behavior and performance of the physical asset. Regardless of challenges and limitations in their full deployment and implementation in real practice, there seems to be only one school of thought and a consensus toward the future adoption of Digital Twins for design, construction, management, operation, and decommissioning of assets among the scientific community and practitioners [1]. Current research gaps in the practical development and implementation of Digital Twins are mainly related to: (i) the lack of interoperability among the different proprietary and open-source software used along the Digital Twin model generation pipeline; (ii) performance improvement of currently available anomaly detection, location, description, and prognosis algorithms; and (iii) the direction for the creation of macro-digital twins that integrate Digital Twins of individual assets. Furthermore, future potential developments on this topic are related to the implementation of Industry 5.0 concepts and ideas within Digital Twin frameworks such as sustainability, human-centrism, and resilience [2].
Contributions aiming at tackling any of these issues by means of novel developments and applications of computational methods are welcome to participate in this mini symposium. Either basic methodologies, scientific developments and/or industrial applications are equally accepted.
Keywords: Anomaly Detection Algorithms, Computational Mechanics, Digital Twins, Industry 5.0
The Digital Twin is a cornerstone technology concept within the ongoing 4th industrial revolution. By connecting physical assets to their digital replicas, Digital Twins offer unprecedented opportunities for innovation in industrial design and operation, for example, with respect to design optimization, decision support, and online (real-time) monitoring, to name but a few relevant applications. To enable disruptive Digital Twin innovations, methods stemming from the fields of scientific computing and computational science and engineering are continuously developed and optimized to address the corresponding challenges. This minisymposium aims to address novel developments with respect to computational methods that are crucial for the realization of Digital Twins and their application for challenging industrial problems. Exemplary topics of interest include: (a) Executable Digital Twins by means of model order reduction, reduced order modelling, or surrogate modelling, such that fast albeit accurate model evaluations during system operation are possible. (b) Uncertainty-aware Digital Twins by means of probabilistic simulation, uncertainty propagation, and (Bayesian) model calibration. (c) Hybrid Digital Twins that fuse physic-based and data-driven modelling, simulation, and optimization methods, thus combining “the best of two worldsâ€. This list is by no means exhaustive, and several other topics of interest can also be considered, provided that they feature a connection to Digital Twin technologies and applications.
Keywords: Computational Methods, Digital Twins, Model Order Reduction, Modelling and simulation, Reduced Order Modeling, Scientific Machine Learning, uncertainty quantification
Digital twins are a powerful tool to design, optimize, monitor, operate and service real (physi-cal) objects by allowing for holistic and realistic simulations and predictions. The digital repre-sentation of a real object combines all relevant and available models, data and information about its real counterpart. Thereby, the coexisting digital and real twin are able to exchange information bi-directionally [1]. While the term digital twin arose in the context of manufac-turing (Industry 4.0), the concept is more and more explored in various other fields such as health care, education, meteorology and construction, too [2]. Using the great potential of digital twins for various and critical infrastructures, like road systems, bridges, water treat-ment facilities and energy networks, which are valuable and expensive goods of our society, is meaningful to increase among others safety, sustainability and operability. However, the de-velopment of digital twins at hand of sub-models, data and interfaces requires huge interdis-ciplinary knowledge and contributions on, e.g., coupling of models, domain knowledge, so-phisticated data science and machine learning.
The objective of our mini-symposium is to bring experts in the field of digital twins for infra-structures and their enabling technologies together to increase the interdisciplinary knowledge and to foster scientific exchange and collaboration.
Topics of interest include, but are not limited to:
• efficient physical and data-driven models,
• obtaining and processing of sensor data from real objects and experiments as data source for digital twins,
• scientific machine learning for digital twins, e.g. physics-informed neural networks,
• approaches to combine different models and data of a real object in one digital repre-sentation,
• twinning approaches to keep the real object and its digital representation consistent,
• architectures and use cases of digital twins,
• treatment of uncertainties within digital twins.
REFERENCES
[1] M. Asch, A Toolbox for Digital Twins: From Model-Based to Data-Driven, Philadelphia, PA: Society for Industrial and Applied Mathematics, 2022.
[2] A. Rasheed, O. San and T. Kvamsdal, “Digital Twin: Values, Challenges and Enablers From a Modeling Perspectiveâ€, IEEE Access, Vol. 8, pp. 21980ï€22012, (2020).
Keywords: Data-driven Models, Infrastructure, Physical Models, Digital Twins, Experimental Data Sources, Sensors as Data Source
Recent years have seen a humongous proliferation of personalized cardiac models of varying degree of complexity – from image-driven computational fluid dynamics (CFD) simulations, to electro-mechanical models, until recent fully-coupled multi-physics digital twins of the heart. The long-term objective of these developments is twofold: i) to replicate the status of a patient’s heart with access to information otherwise not obtainable by clinical imaging; ii) to predict hypothetical post-treatment scenarios and guide clinical decisions. Despite the enormous advances, the validation of these models against in-vivo clinical data of the same patient (e.g. CT, cine-MRI, 4D flow MRI, echo-PIV, etc.) remains elusive. Identifying the minimum set of modeling ingredients needed to replicate fundamental physiological features of heart function, as well as detailed flow, stress/strain or electric fields is a crucial point towards the incorporation of in-silico replicas into clinical practice. The impact of the uncertainties associated with the estimation of modeling parameters and with boundary conditions, as well as with the clinical data itself is also of utmost importance.
The proposed mini-symposium will bring together experts form the modeling, biomedical and clinical imaging communities to discuss the state-of-the-art in this field and explore future research directions. Topics of interest include (but are not limited to):
• medical imaging techniques: validation of algorithms for cardiac shape/motion reconstruction, physics-based and data-driven enhancement of imaging data;
• image-driven CFD simulations: comparisons against clinical imaging data (e.g. 4D flow MRI), impact of detailed valve modeling, incorporation of complex anatomical features, effect of boundary conditions;
• multi-physics cardiac modeling: calibration of model parameters; comparisons against stress/strain, flow and electric fields;
• role of machine-learning techniques in patient-specific cardiac simulations;
Studies focusing on the application of patient-specific cardiac models to real or virtual patient populations are also welcome.
Keywords: Cardiac modeling, Digital Twins, Medical Imaging, multi-physics modeling, patient-specific simulations
A digital twin is a personalized virtualization of a physical asset or process that evolves in time. It relies upon a set of computational models that dynamically update to persistently mirror its physical counterpart, enabling informed decisions that realize value. The models and parameters comprising the digital twin are continually updated through the assimilation of real-world data, or sensor recordings. This enables diagnostic and predictive capabilities not achievable with static digital models. Up-to-date and comprehensive models are suitable to be exploited within dynamic decision-making frameworks, informing actions tailored to the physical setting of interest.
Computational models, in the form of physics-based simulators and data-driven models, and a synergistic coupling between the physical asset and its virtualization, are critical enablers for effective digital twins. Enhanced computational efficiency is also required to handle the continuous assimilation of noisy and big data, as well as to accommodate the quantification and propagation of uncertainties related to, e.g., environmental conditions, modeling assumptions, and operational variabilities.
This session aims to collect contributions highlighting the impact of physics-based and data-driven methodologies for digital twins of engineering systems and processes. Contributors are invited to discuss topics ranging from, but not limited to, structural health monitoring and predictive maintenance of mechanical, aerospace, and civil engineering systems, simulation and reduced-order models for digital twins, process surveillance and decision-making through digital twins, data assimilation techniques for parameter and state estimation, digital twins for process optimization, multi-fidelity methods, and surrogate modeling strategies. Topics covering both methodology development and real-life applications in engineering and applied sciences are welcome.
Keywords: Digital Twins, Engineering systems, predictive maintenance, Structural Health Monitoring
This MS will have a special emphasis on enabling technologies for Digital Twins, where we adopt the following definition of a Digital Twin:
A digital twin is defined as a virtual representation of a physical asset, or a process enabled through data and simulators for real-time prediction, optimization, monitoring, control, and decision-making.
To enable predictive twins, one may utilize Hybrid Analysis and Modelling (HAM) that combines classical Physic-Based Methods (PBM) accelerated by means of Reduced Order Modelling (ROM) together with Data-Driven Methods (DDM) based on sensor measurement analysed by use of Machine Learning (ML). Pure Data-Driven Methods based on sensor measurement analysed by any means of AI is also welcome. In general, this MS welcome contributions on enabling technologies that can facilitate Predictive Digital Twins. Advanced applications of Predictive Digital Twins are also welcome.
Keywords: Hybrid Analysis and Modelling, Predictive Digital Twins, Reduced Order Modeling
Most flows present in nature and engineering are turbulent and characterized by a wide range of flow scales whose interaction is highly non-linear and difficult to model and simulate in practical flow configurations. These features make coarse graining formulations the only effective tool to predict complex flow problems, often comprising laminar, transitional, and turbulent flow, non-equilibrium phenomena, regions of incompressible or compressible flow, multi-physics, or turbulent kinetic energy generated by multiple mechanisms. This class of turbulence models comprises all formulations capable of resolving a fraction of the turbulent field - from detached-eddy simulations (DES), partially-averaged Navier-Stokes equations (PANS), large-eddy simulation (LES), implicit LES (ILES), to high-fidelity direct numerical simulation (DNS). This mini-symposium discusses the development and/or application of these models to complex practical flows, e.g., inertial confinement fusion, flows past vehicles, oceanography, offshore flows, materials mixing, or climate change. We invite contributions addressing any areas of fluid mechanics with these mathematical models. Contributions quantifying the simulations' accuracy are highly incentivized, as well as experimental studies that can help validate and improve coarse-graining methods. The MS will also present the work to be published in the book Coarse Graining Turbulence: Modeling and Data-Driven Approaches and their Applications [1].
[1] F.F. Grinstein, F.S. Pereira, M. Germano, “Coarse Graining Turbulence: Modelling and Data-Driven Approaches and Their Applicationsâ€, Cambridge University Press, (in
elaboration).
Keywords: coarse graining, computational fluid dynamics, prediction, turbulence modeling
The modeling and simulation of turbulent flows of practical interest is challenging due to the complex physics and wide range of scales. Direct numerical simulation (DNS) and large-eddy simulation (LES) are the ideal formulations to predict this class of flows as they resolve all or most turbulent flow scales. The high degree of resolution is responsible for the high-fidelity of these simulations, but also for their computational cost, which is excessive for engineering applications in the foreseeable future. This limitation made the Reynolds-averaging Navier-Stokes equations (RANS) the workhorse of engineering computational fluid dynamics. RANS model is expected to represent the physics of all turbulent scales through a closure model, reducing the simulations’ cost significantly. However, developing robust and accurate closures to model all turbulent scales is difficult and limits the accuracy of RANS in numerous flows of current practical interest. In 2003, Girimaji [1,2] proposed the partially averaged Navier Stokes equations (PANS) method to bridge the spectral gap between DNS/LES and RANS and combine their main advantages. Such an objective is accomplished by only resolving the flow scales not amenable to modeling, and representing the remaining through a turbulence closure. This feature enables the concept of accuracy on demand, responsible for PANS’ efficiency (cost vs accuracy) and success among CFD practitioners. After 21 years, PANS application in engineering problems is still growing and is becoming important to many engineering fields: automotive, combustion, offshore, or materials mixing. This mini- symposium celebrates the 21st anniversary of PANS. It aims to bring together PANS developers and users to promote discussions about the model’s current state, progress, challenges, and future directions. Thus, we invite contributions from experienced and new users using PANS in any areas of fluid mechanics.
[1] S.S. Girimaji, R. Srinivasan, E. Jeong, “PANS Turbulence Model for Seamless Transition Between RANS and LES: Fixed-Point Analysis and Preliminary Resultsâ€, In Fluids Engineering Division Summer Meeting, FEDSM2003-45336, pp. 1901-1909, (2003).
[2] S.S. Girimaji, “Partially-Averaged Navier-Stokes Model for Turbulence: A Reynolds- Averaged Navier-Stokes to Direct Numerical Simulation Bridging Methodâ€, J. Appl. Mech., Vol. 73(3), pp. 413−421, (2006).
Keywords: computational fluid dynamics, PANS, prediction, turbulence modeling
The study of turbulence is one of the paramount areas of research in physics & engineering; in fact, it remains one of the seven unsolved Millennium Problems of the Clay Mathematics Institute. Turbulence is intrinsically chaotic and composed of a wide range of spatio-temporal scales. These flow motions naturally arise from laminar (regular) states that have a relatively simple spatio-temporal structure. However, flows typically undergo state changes, i.e., bifurcations, as the input parameters are varied. In this scenario, the flow states increase their spatio-temporal complexity successively, which leads to chaotic behaviour and ultimately fully-developed turbulence characterized by large flow fluctuations. It is thus common to find a transitional regime that involves a competition between these two extreme flow configurations. This regime is significantly important as it is widely encountered in a diverse range of fundamental and applied fluid dynamics problems, such as wall-bounded flows (e.g.,
Poiseuille systems), boundary layers (e.g., flow tripping in aerodynamics), and shear layers (e.g., formation of clouds over the ocean). The transition from laminar to turbulent flow can occur spontaneously, and it is a critical event that can affect the overall behavior of the fluid flow. In particular, during the transition, the flow may exhibit intermittent bursts of turbulence that are rare and unpredictable, and can have a significant impact on the system being studied. For example, coherent structures during the transition can cause large fluctuations in the drag and lift forces on objects in the flow, which can affect the stability and performance of aircraft, ships, and other vehicles. Therefore, understanding the mechanisms and dynamics of laminar and turbulent regimes, and the transition between both, is crucial for predicting and
controlling fluid flow. To this end, this mini-symposium is devoted to computational methods and data-driven tools for picturing the transitional and fully-turbulent regimes based on concepts and theories from dynamical systems. This mini-symposium will enable researchers to (i) extract fundamental insight from computational methods for the analysis of chaos and turbulence, and (ii) explore state of the art data-driven tools for investigating turbulent flows.
Keywords: Computational Methods, Data-Driven Tools, Dynamical Systems, Turbulence
Lately, model order reduction (MOR) techniques have emerged as an effective and helpful tool to alleviate the computational costs related to complex simulations in computational fluid dynamics (CFD). This field features an increasing need for low-dimensional models capable of providing fast and accurate real-time responses to better analyze flow problems and to accelerate time-consuming related studies such as parameter estimation, uncertainty quantification, and control.
Applying MOR techniques to CFD problems might be challenging. Differently from other branches, classical reduction techniques show several complications in this field, among which slow Kolmogorov n-width decay, instabilities, failure in extrapolatory regimes, etc. That is why it is of paramount importance to carefully design reduced order models (ROM) that are able to stably, accurately and efficiently approximate CFD simulations.
The mini-symposium aims at providing novel insights on MOR in the context of CFD, moving from the state-of-the-art to the latest advances in the field, opening a fruitful discussion on its challenges and future directions.
The main objectives of the mini-symposium are:
* Exchange ideas on developing and implementing novel MOR-based strategies, highlighting their advantages and limitations in diverse CFD applications.
* Showcase innovative approaches and robust techniques to exploit in advanced decision-making frameworks such as data parametric uncertainties and control.
* Discuss the effectivity of MOR strategies in complex CFD settings such as turbulent flows, multiphysics simulations, and coupled systems.
Keywords: Model Order Reduction
Recent years have witnessed a renewed interest in developing high-speed flight technology . For instance, commercial aircraft manufacturers are heavily investing in hypersonic technology that can dramatically reduce intercontinental passenger flight time making same-day global round trips a reality. The wide range of atmospheric and flow conditions corresponding to different altitudes and Mach numbers that a hypersonic vehicle undergoes presents a huge challenge to aerodynamic design, propulsion, and flow control. At lower altitudes, where the flow is regarded as continuum, the effect of turbulence is critical, whereas at higher altitudes, flow rarefaction due to low-density effects becomes important. This warrants high-fidelity numerical models that can reliably simulate these conditions, at all altitudes, accurately.
This mini-symposium will target and communicate on recent progress in CFD for supersonic and hypersonic aerodynamics at all altitudes. Topics of interest of this mini-symposium include, but are not limited to:
- High-performance computing of high-speed flows
- Numerical schemes, models, and algorithms
- Supersonic/hypersonic wall bounded flow and boundary layer control.
- Shock-wave/boundary layer interaction
- Turbulence modelling for high-speed flows
- Fluid–structure interaction in high-speed flows
- Application of CFD to supersonic/hypersonic engineering
- Direct simulation Monte Carlo (DSMC) and hybrid CFD-DSMC models for high-altitude flows
- Alternate models like Method of Moments, gas kinetics
- High temperature effects related to chemistry, ablation, etc.
Keywords: Computational Fluid Dynamics.
Computational Fluid Dynamics, High-Fidelity Modelling, Rarefied Gas Dynamics, Supersonic/Hypersonic Aerodynamics
Multiphase flows with liquid-vapor transition such as cavitating flows, sprays, boiling and flashing flows are found in numerous engineering applications and science fields. These flows are characterized by complex multiscale phenomena involving the dynamic formation of phase interfaces and inter-phase thermodynamic transfers. Important advances have been made in computational methods for the simulation of these multiphase flows, based on various mathematical and physical models and different numerical approaches. Yet there are many open challenges towards the accurate prediction of these flows in realistic configurations. First, there is a need for models and methods providing a more precise description of the flow physics and thermodynamics. For instance some difficulties concern the modeling of non-equilibrium phenomena in heat and mass transfer processes such as metastability and the description of nucleation mechanisms. In some problems additional multiphysics and multiscale effects should be taken into account, for example surface wettability in boiling and evaporation processes, or the possible presence of multiple species in the phase changing flow. This entails additional difficulties in the design of accurate and efficient numerical algorithms. Furthermore, the simulation of realistic problems demands time-affordable computational tools applicable to multi-dimensional complex geometries and to a large range of Mach number regimes. The aim of this minisymposium is to bring together scientists working on computational models for multiphase flows with liquid-vapor phase change to share and exchange ideas, discuss challenges and innovative methods in the field. The minisymposium will be open to a broad spectrum of modelling techniques and numerical approaches.
Keywords: boiling, cavitation, computational fluid dynamics, liquid-vapor transition, Multiphase flows
Over the last decades, Computational Fluid Dynamics (CFD) has become a widespread research and design tool in many fields of enormous societal impact, including biomedicine and renewable energies, among others – either alone, or as part of broader (coupled) multi-physics models. The driving force behind this growth is the continuous progress of high performance computing (HPC) systems, in conjunction with the development of more and more efficient numerical techniques and physics-based turbulence models. Furthermore, the quest for increasingly detailed results in multiple fields (e.g. precision manufacturing, personalized medicine, etc.), and the cost/difficulties associated with experimental campaigns, are pushing the CFD community towards enabling predictive eddy-resolving simulations of complex flows in fast turnaround times. This paradigm shift requires a wise combination of numerical robustness, physical fidelity and algorithmic efficiency, a set of conflicting requirements that continues to thrill developers and practitioners alike.
This mini-symposium aims at providing a forum to discuss current challenges in CFD towards the solution of increasingly complex multi-scale and multi-physics problems, with a particular focus on (but not limited to) the following three interconnected areas:
- physical fidelity: physics-compatible numerical methods (e.g. invariant-preserving schemes), physics-inspired implicit/explicit large-eddy simulation (LES) models, and the application of these approaches in scenarios of industrial or biological interest;
- numerical robustness: traditional stabilization strategies (e.g., numerical dissipation, filtering, etc.) vs. physics-compatible approaches; algorithms for complex geometries (e.g. unstructured vs. Cartesian immersed-boundary methods);
- algorithmic efficiency: low- vs high-order spatial/temporal methods, wall modeling, efficient time stepping (explicit vs. implicit), reducing the memory bandwidth and operation count of numerical schemes, adaptive mesh refinement strategies.
Contributions focusing on open-source implementations are also especially welcome.
Keywords: computational efficiency, high-fidelity CFD, physics-compatible schemes
COMPUTATIONAL KINETIC THEORY
Kinetic models have emerged as a comprehensive and versatile multiscale modeling paradigm in many applications in science and engineering to simulate transport in and out of equilibrium, e.g. rarefied-gas dynamics, radiative transfer, plasma physics, dispersed-particle flows, galactic dynamics, etc. Kinetic theories typically lead to high-dimensional equations, in which interactions manifest themselves as complicated non-linear integro-differential operators. The high dimensional setting of kinetic problems implies that the corresponding computational complexity is prohibitive, particularly when attempting to simultaneously resolve both equilibrium and nonequilibrium flow phenomena.
Recent advances in the numerical approximations for kinetic theories exploit fundamental structural properties of kinetic equations related to trends toward equilibrium and corresponding hydrodynamic (continuum) limits. Exploiting such properties enables the derivation of efficient and robust methods that are capable of adaptive resolution of equilibrium and non-equilibrium effects. However, computationally scaling such methods to complex geometries remains an outstanding challenge.
The aim of this mini-symposium is to assemble researchers in the area of multiscale modeling, analysis and simulation of kinetic theories, in order to discuss recent developments, to explore open issues, and to foster cross-fertilization. The envisaged range of topics spans (but is not limited to):
• Moment theories for kinetic equations
• Hydrodynamics and continuum limits
• Scale-bridging and adaptive methods
• Structure-preserving discretization techniques
Keywords: Direct Simulation Monte Carlo, Hyperbolic Equations , Moment Methods, Multiscale Fluid Dynamics, Multiscale Methods, Plasma Physics, Radiative Transfer, Rarefied Gases, Computational Kinetic Theory
Microfluidic multiphase flows are crucial for many engineering systems, such as Lab-On-a-Chip, inkjet printers, fuel cells, microreactors, oil-gas/water transport and CO2 sequestration in porous media. In most such cases, the liquid-gas system forms a contact line with the solid surfaces. Tending to reach its equilibrium configuration, the three-phase system exhibits a dynamic behavior, which is particularly determined by its physicochemical properties. The multiphase behavior, in particular, contact line dynamics in many of these microfluidic systems still needs to be fully understood. New numerical methods that can accurately handle contact line dynamics in complex geometries are under active development. Theoretical models fall into three main categories: those focused on the microscopic scale, in particular molecular dynamics, mesoscopic descriptions like phase field models and those developed at the continuum level. Combining the outcomes of these three categories can improve modeling and physical understanding. Some recent works tend to enrich the existing modeling approaches with the use of data-driven methods. Development of the Machine Learning approaches has opened new horzions for tackling challenges such as curvature approximation, model discovery for contact line dynamics, and optimization of complex microfluidic systems. Data-driven methods can also be a means for improving the efficiency of the numerical modeling of multiscale phenomena. We will discuss the latest research on how data-driven methodologies can improve both microfluidic system designs and the understanding of complex multiphase flow dynamics. Overall, the present mini-symposium will bring together researchers working on the different types of micro/meso and macroscopic modeling and simulation of multiphase flows in microfluidic applications as well as the emerging application of data-driven methods in multiphase microfluidics.
Keywords: bubbles, droplets, multiphase, surface tension, wetting
The Mini-Symposium on "From the Nose to the Lung: Fluid Dynamics of the Upper Airways" aims to explore the diverse aspects of fluid flow in the human airways, with a particular focus on numerical an related experimental studies. This symposium will provide a platform for researchers to present their work on airway fluid mechanics, covering a wide range of topics, including macroscopic flow mechanics, mucus rheology, patient-specific surgery planning for flow-related pathologies or targeted drug/aerosol deposition, etc.
Understanding the complex fluid dynamics in the airways is of importance for diagnosing and treating of various pathologies of the respiratory system. The symposium will showcase cutting-edge research and advancements in this field, shedding light on the complex interplay between airflow patterns, mucus properties, and associated pathological conditions. By addressing experimental and computational approaches, this symposium will foster interdisciplinary collaboration and promote a comprehensive understanding of fluid dynamics in the airways. Potential objectives of the minisymposium are:
 
• macroscopic fluid dynamics within the airways (naso-pharynx and upper airways): Characterizing airflow patterns, pressure gradients, and unsteady flow/turbulence effects in different regions of the respiratory system. Investigations into the effects of airway geometries, such as constrictions or bifurcations, on flow behavior will also be explored.
• mucus rheology and its impact on airflow and aerosol generation: Addressing the complex viscoelastic properties of mucus, its role in airway clearance, and its influence on flow characteristics.
• patient-specific surgical planning for flow-related pathologies: Highlighting the use of computational modelling and simulation techniques to predict airflow patterns and optimize surgical interventions.
This minisymposium will provide ample opportunities for networking and knowledge exchange. Interactive discussions will encourage collaborations among researchers from various disciplines, including biomedical engineering, fluid dynamics, and various clinical disciplines. It aims to advance our understanding of fluid dynamics within the airways and contribute to the development of innovative diagnostic and therapeutic strategies for respiratory disorders.
Keywords: Fluid Dynamics, Naso-pharynx, Respiratory System, Upper Airways, Airway Pathologies, CFD
Multiphase flows are found in many mechanical, process, chemical, maritime, civil, and biomedical applications. Their features include heat and mass transfer in bubbles, droplets, films, and sprays, potentially in a reacting environment; turbulence modulation and drag reduction in bubbly flows, fluid-structure interactions in free surfaces or cavitation in rotating machinery, among others [1]. The assessment of such critical features requires describing fundamental physical phenomena including bubble growth, detachment, dispersion, deformation, coalescence, and collapse; film instability and breakage; jet atomization, phase change, Marangoni convection, or electro-wetting, among others. All these physical phenomena share the core role of the two fluids interface in their underlying mechanisms.
To gain a deeper understanding of such multiphase flow physics, numerical simulation is an invaluable tool, particularly using the one-fluid approach. Nonetheless, it requires resolving a moving, deformable, two-phase interface; treating potentially huge differences in physical properties, and including interfacial phenomena itself (like surfactants surface diffusion). Consequently, the numerical simulation of multiphase flows is still a rich field of research with several open questions, as proven by the coexistence of several techniques.
The numerical treatment of multiphase flow physics is then challenged to develop better numerical schemes [2] that improve conservation (mass, momentum, energy), surface tension, interface reconstruction (surface area, normal vector, curvature), interface transport; and computational techniques for time-stepping, variable coefficient Poisson equation, Adaptive Mesh Refinement, or numerical instabilities, among others.
In this mini-symposium, we want to gather practitioners of interface-resolved multiphase flows from different techniques (e.g.: F-T, VOF, (C)LS, PF, etc) and different application areas to exchange their experiences in solving the numerical challenges highlighted above and discuss pros and cons of each technique.
REFERENCES
[1] D. Lohse, Bubble puzzles: From fundamentals to applications. Physical Review Fluids. 3 (2018)
[2] S. Popinet, Numerical Models of Surface Tension. Annual Review of Fluid Mechanics. 50 (2018)
Keywords: (Conservative) Level Set, Adaptive Mesh Refinement, conservation, Front-Tracking, immersed boundary method, interface capturing, Interface problems, interface tracking, Multiphase flows, Phase Field, surface reconstruction, surface tension, Volume of Fluid
Electrohydrodynamic (EHD) and magnetohydrodynamic (MHD) systems have gained significant attention in various fields due to their potential for improving performance and enabling novel applications. Active flow control using systems based on EHD and MHD principles is a rapidly developing field within fluid mechanics, aims to enhance drag reduction, lift increase, mixing enhancement, and noise reduction. Techniques such as plasma actuators, laser energy deposition, MHD- and EHD-guided flows have emerged as prominent examples of systems operating on the principles of EHD and MHD. Understanding the fundamental mechanisms, optimizing these actuators, and exploring new applications are crucial areas of research. The optimization and improvement of electrohydrodynamic systems heavily rely on comprehensive numerical modeling and robust computational tools. The computational fluid dynamics (CFD) modeling of EHD systems represents a challenging multiphysics problem where Maxwell's equations are coupled with the governing equations of fluid flow. Such problems typically involve various temporal and spatial scales that necessitate specialized numerical considerations. This mini-symposium aims to bring together researchers working on computational modeling aspects of EHD and MHD systems. While the primary focus is on flow actuators, secondary applications of these devices, such as ice control systems, sensory systems, and micropropulsion systems, will also be considered. The symposium seeks to provide a platform for sharing novel research, discussing advancements in computational modeling techniques, and exploring potential applications of EHD and MHD systems beyond flow control.
Keywords: magnetohydrodynamic, Active flow control, Computational Fluid Dynamics
This mini-symposium aims to provide a forum for sharing and discussing research works in the general area of multiphase flow and non-Newtonian fluid, especially those inter-disciplinary studies that cross the traditionary boundary between solids and fluids.
Many natural and industrial processes involve dynamic motions of both fluids and solids, forming a complex multiphase flow which is often further complicated by non-Newtonian / viscoelastic / viscoplastic behaviours, phase transitions, chemical reactions, and the presence of porous media.
Examples include welding and casting, 3D printing, polymer injection moulding, fresh concrete placement, oil and grease lubrication, debris flow, sedimentation, dust storm, and many more.
This session welcome, but is not limited to, the following topics:
• Physical and mathematical models of multiphase systems and processes
• Numerical modelling of multiphase flow and non-Newtonian fluids
• Computer simulation of complex systems involving multiple fluids and solids
• Numerical and experimental studies of materials and processes involving phase transition and/or chemical reaction
• Applied studies that cross the traditional boundary of solids and fluids: 3D printing, injection moulding, concrete placement, debris flow, sandstorm, welding and casting, food processing, glass forming, and others.
Keywords: Multiphase Flow, non-Newtonian Fluid, Numerical;, Particle-laden Flow
The flow of single-phase fluids and suspensions through porous media is of importance in a number of applications, such as filtration and geomechanics, among others. The general focus of the proposed mini-symposium is directed toward modeling and computing such flows. The more specific focus is on porous media configurations in which the pores are slender. In such cases, the porous media can be modeled as a network of connected slender pores, therefore allowing for the use of the range of tools emerging from network theory and computational topology to help correlate the porous media structure and its functionality. One important application of such types of porous media is in the case of filtration, however many other applications could benefit from asymptotic and computational methods developed in the filtration context.
The proposed mini-symposium will focus on bringing together researchers working on modeling network-based porous media in a variety of different contexts. Our expectation is that by bringing together researchers from different fields of application, we will encourage the flow of information between often disjoint fields of research. We hope that further development of asymptotic models and related computational techniques will allow for the formulation of new and efficient computations that will find relevance in a variety of important applications.
Keywords: asymptotic and network models, fluid dynamics, Porous media flow
The lattice Boltzmann method (LBM), known for its computing efficiency and adaptability in modelling fluid flows, has received broad popularity in a variety of scientific and engineering disciplines. This mini-symposium aims to provide a dynamic venue for LBM academics and practitioners to exchange their ideas, findings, and innovations. This event will encourage constructive talks and develop cooperation in the following important areas, with a major emphasis on innovative applications and cutting-edge techniques:
1. Advanced LBMs - Contributors will reveal ways for boosting simulation accuracy, stability, and efficiency by emphasizing recent methodological breakthroughs in LBM.
2. Multiphysics and complex systems - The numerous uses of LBM in modelling complicated multiphysics phenomena and complex systems will be covered in this theme area. Attendees may expect to hear about multiphase flows, fluid-structure interactions, and beyond.
3. High-performance computing - Experts will discuss how to optimize code for contemporary architectures and use parallel computing approaches.
4. Emerging applications - New applications emerge on a regular basis. Explorations of LBM's application frontiers, including biological systems, renewable energy, and beyond, will be discussed.
5. Validation and benchmarking - Maintaining the reliability of simulation findings is crucial. Contributors will explain effective techniques for verifying LBM simulations and benchmarking against analytical solutions and experimental data in this section.
Keywords: computational fluid dynamics, Modelling and simulation
Computational Fluid Dynamics (CFD) has been a powerful tool for the comprehension of complex fluid dynamics. However, traditional CFD methods, grounded in continuum assumptions, may falter in scenarios with increased Knudsen numbers, e.g., rarefied gas dynamics and microfluidics. Kinetic-Based Computational Fluid Dynamics (KCFD) offers a more proper approach, transcending the conventional continuum regime. Instead of relying on the Navier-Stokes equations, KCFD employs the Boltzmann equation from kinetic theory to compute fluid dynamics from the moments of fluid particle distribution functions. Operating at the mesoscopic level, KCFD forms a fundamental framework adaptable to a broad spectrum of Knudsen numbers, spanning from continuous hydrodynamics to rarefied gas dynamics. Rooted in the Boltzmann equation, KCFD naturally models various phenomena, such as interfacial dynamics in multiphase flows, non-Newtonian effects, fluid-structure interaction, rarefication, magnetohydrodynamics (MHD), and more. Two well-established KCFD methodologies are the lattice Boltzmann method (LBM) and the gas kinetic scheme (GKS). Notably, both exhibit compelling advantages, including suitability for GPU parallelization, ease of programming, and direct integration of physical effects into the modeling process. We aim to bring together researchers and practitioners from diverse fields in this minisymposium (MS) to explore the capabilities and applications of KCFD in tackling complex flows. The primary objective is to provide a platform for interdisciplinary discussions, knowledge exchange, and the dissemination of recent developments in KCFD. We aim to (1) showcase the latest advancements in KCFD methodologies, algorithms, and applications, (2) foster collaboration and networking opportunities among researchers working on KCFD from various disciplines, and (3) discuss challenges and opportunities in implementing KCFD in real-world applications. We invite abstracts that encompass a wide range of topics related to KCFD, including but not limited to methodologies and algorithms, validation and verification, uncertainty quantification, applications, GPU parallel computing, interdisciplinary research, and future directions. This MS will serve as a vibrant forum for researchers to explore the cutting-edge developments and applications of KCFD. We invite participants to join us in advancing our understanding of complex fluid dynamics through the lens of kinetic-based simulations.
Keywords: gas kinetic scheme, GPU parallel computing, kinetic-based computational fluid dynamics, lattice Boltzmann method
The Minisymposium on Offshore Wind Structures presents a unique platform for researchers, engineers, and industry professionals to discuss recent developments and cutting-edge solutions in the field of offshore structures and wind energy. This multidisciplinary event explores topics such as wind turbine technologies, monopiles and jackets foundations, the emerging field of floating wind turbines, dynamic vibration control systems, vibration absorption configurations, seismic isolation, nonlinear FE dynamics and advanced CFD simulations.
By addressing the challenges associated with offshore structures and wind energy, this minisymposium aims to foster knowledge exchange, collaboration, and innovation towards sustainable and resilient offshore energy systems. The minisymposium encompasses a wide range of themes, including:
i. Offshore Wind Turbines: Advances in design, operation, and maintenance of offshore wind turbines, focusing on robustness, structural integrity, and reliability.
ii. Floating Wind Turbines: Novel approaches and design strategies for floating wind turbines, enabling the harnessing of wind energy in deeper waters, unlocking vast offshore wind resources.
iii. Monopiles and Jackets: Exploration of the latest developments in monopile and jacket designs, installation techniques, and foundation optimization to support offshore wind turbines.
iv. Vibrations Mitigation: Techniques and methodologies to reduce vibrations in offshore structures, enhancing their performance, longevity, and safety in harsh marine environments.
v. Seismic Isolation: Innovations in seismic isolation techniques for offshore structures to withstand earthquakes and enhance their resilience, minimizing potential damage and downtime.
Attendees will have the opportunity to engage in discussions, share their research findings, and explore potential collaborations. Moreover, the event will provide a platform for researchers, industry experts, and technology developers to exchange knowledge and experiences, accelerating the deployment of sustainable offshore energy systems. We invite researchers, engineers, and industry professionals to participate actively in this minisymposium, contributing to the collective understanding of offshore structures, offshore wind energy, vibrations mitigation, seismic isolation, and the transformative potential of next-generation wind turbines.
Keywords: Dynamic Vibration Control, Floating Structures, Monopile SSI, Vibration Mitigation, CFD , Offshore Wind Turbines
The objective of this Mini Symposium is to discuss progress and recent advancements in the numerical computation of fluid-structure-interaction problems, with an emphasis on new innovative formulations, methods and algorithms leading to faster, more accurate predictions and improved software design. The envisaged range of applications spans (but is not limited to) aero-elasticity, hydro-elasticity, biomechanical FSI and noise/structural acoustics. In particular, we welcome contributions in the vanguard of:
• error estimation;
• adaptive methods;
• immersed and unfitted methods;
• multiscale models;
• reduced order models and methods;
• artificial intelligence and machine learning;
• novel iterative solution techniques;
• shape optimization and inverse methods;
• software engineering.
In addition, this Mini Symposium is intended as a platform for other state-of-the-art developments in FSI, such as those pertaining to FSI with auxiliary-field interactions, e.g. FSI problems with (massive) self contact, FSI problems with fracture (e.g. hydraulic fracturing, blast-induced FSI, etc.), and elasto-capillary FSI.
Keywords: adaptive methods, auxiliary-field interactions, error estimation, immersed methods, iterative solution methods, multiscale models, reduced-order modeling
This minisymposium aims at gathering mathematicians and engineers working on the mathematical and numerical aspects of incompressible fluid-structure interaction. Topics may include, but are not limited to, fluid-structure interaction with contact, poro-elasticity, multi-dimensional 3D-1D coupling, unfitted mesh approximations, etc.
Keywords: Incompressible fluid-structure interaction, mathematical analysis, numerical analysis
Currently available computational capabilities make the coupling of multiple physical systems an intriguing option in the pursuit of scientific prediction or engineering design. Multiphysics models allow to relax the assumptions of decoupling and to provide insights on the coupling mechanisms themselves. Fluid-Structure Interaction (FSI) is a problem in which a multiphysics approach can lead to fascinating insights, both for biological systems and engineering scenarios. The complexity of this problem gave rise to broad variety of computational solutions, complying with partitioned or monolithic approaches, finite-differences or finite-elements methods, immersed or boundary-complying techniques. In this connection, each strategy has been developed to properly address a specific sub-class of FSI problem.
The aim of this mini symposium is to gather recent advancements for the comprehension of multiphysics systems, where FSI plays a crucial role. We warmly encourage submissions covering a wide range of application fields, including cardiovascular and respiratory problems, locomotion (swimming/flying) of animals and bio-inspired devices, transport of particles and vesicles, hydraulic fracturing of immersed structures, to name a few. A debate among complementary expertise is fostered, highlighting advantages, drawbacks, potentialities, and limitations of cutting-edge research for FSI applications.
Keywords: complex flows , elastic structures, immersed-boundary, interface-resolved FSI
The aim of this long-standing Mini-Symposium is to bring together specialists in computational mechanics, mechanics and micromechanics of materials, applied mathematics, continuum mechanics, materials science, physics as well as mechanical, automotive and aerospace engineering to discuss advances in computational analysis of relationships between the microstructural features of advanced materials and their local and global behaviour as well as the effect of microstructure on performance of components and structures.
The topics of the Mini-Symposium include, but are not limited to, the following:
• mechanics of advanced materials and structures;
• effect of microstructure on properties and performance of advanced materials;
• prediction of deformational behaviour and life-in-service of structures and components made of advanced materials;
• computational models of biomaterials and biomedical materials;
• simulation of deformation, damage and fracture processes in materials at the small scale;
• computational methods for analysis of modern visco-elastic composite and nanocomposites materials;
• mechanics of composite materials with relaxation and phase transitions;
• simulation of failure mechanisms and damage accumulation processes in advanced materials;
• computational analysis of cutting of advanced materials;
• optimization problems in mechanics of advanced materials and structures.
Keywords: Advanced Materials, Microstructure, Properties and Performance
In the civil and military, land, naval and aerospace transportation sectors as well as in the building and energy sectors, crash, collision, blast and impact are typical loading cases to be accounted for when designing or verifying engineering structures regarding accidental overloading or terrorist attack induced potential failure.
From the computational perspective, the challenge is to solve space and time multi-scale, multi-phase and multi-physic initial-boundary value problems, and accordingly to develop or adapt
* methods of multi-media interaction or equivalent loading condition,
* advanced rate and temperature dependent models applicable to high strain rate or/and high pressure,
* methods of space and time discretization, including FEM, FVM, DEM, SPH and their combination,
* numerical simulation involving crash, collision, blast or impact.
This mini-symposium aims to provide a forum for discussing new scientific and industrial challenges and developments in the field of computational mechanics in high strain rate and impact dynamics.
Keywords: Blast, Constitutive modeling, Impact, Numerical methods
The progress of Additive Manufacturing (AM) techniques in the fabrication of structural components and the major improvement in their quality call for structural optimisation methods that take into account the unique characteristics of the process. While AM allows for exceptional geometrical design freedom and leads to significant reduction of the component weight, the fabricated components remain vulnerable to damage and fracture. The prediction of the post-manufactured properties of AM components and their failure mechanics is a challenging aspect. Commonly faced anisotropy and its relation to stress concentration effects require further study.
The goal of this mini-symposium is to gather contributions that explore aspects of AM techniques that are pertinent to the mechanical behaviour of AM structural components, notably, but not solely, from the viewpoints of damage, fatigue, and fracture behaviour. Studies of all facets of computational damage and fracture mechanics of polymeric, metallic and concrete additively manufactured components by different AM processes are welcome. Papers on failure analysis of the process, performance and buildability of AM components are also encouraged. Damage and fracture mechanics of AM components from different perspectives: continuum, classical finite element, multiscale, etc., may be contemplated.
Keywords: 3D printing, Computational damage, Failure mechanics, Finite element method, Additive Manufacturing, continuum, Homogenization, Multiscale, Reduced Order Modeling
This minisymposium is dedicated to discuss recent advances in computational methodologies especially for the simulation of fracture processes. We welcome contributions on topics including (but not limited to):
 efficient discretization techniques for phase-field fracture
 enriched finite element methods
 virtual element methods for crack simulations
 eigenerosion and eigenfracture techniques
 advanced cohesive zone elements
 meshfree methods for fracture
 peridynamics for fracture
Keywords: Crack nucleation and propagation, Damage Mechanics, Brittle and cohesive fracture, Computational Mechanics
The use of fiber reinforced polymer composites as structural materials has seen a significant growth in the past decades on a wide range of industries, from aerospace to construction. This has been mainly motivated by their high strength-to-weight ratio and corrosion resistance. On the other hand, their brittle failure modes have been pointed out as one of their main weaknesses. Despite their linear-elastic behaviour, which could seemingly indicate easier analyses, due to their composite and anistropic nature, determining the strength of composite structures and components is a complex problem, often requiring onerous experimental tests. In fact, while failure initiation criteria for anisotropic materials have been developed since the 1950s, their application without coupling of damage progression models leads to overly conservative strength predictions in composite structures. In fact, in any moderately complex structural component, the failure index of those initiation criteria may be reached in localized stress concentration areas without affecting the global behaviour and for load levels much lower than those leading to collapse.
In this context, computational methods are a promising cost-effective alternative to experimental testing, and several researchers have focused on the development of composite damage progression models to be applied with structural analysis computational methods. In fact, some studies have been successful in predicting the failure behaviour of very complex problems (e.g., bolted connections), although, generally, at the expense of high computational costs.
This mini-symposium focuses on the use of computational methods to model the failure and post-failure behaviour of composite materials, components and structures. The validation and comparison of the computational results with experimental data is highly valued, particularly concerning the predictions of strength and failure modes, as well as the applicability of the proposed methods to real design situations and their suitability for practitioners.
Keywords: Damage, Damage Progression, Strength, Fiber Reinforced Polymers (FRP)
Research efforts focusing on soft materials such as polymers and hydrogels, to name a few, have witnessed an upsurge in the recent times. Such materials can undergo extremely large deformations prior to failure, hence their ubiquitous use in current technologies. Their application in the bioengineering field allows for synthetic structures mimicking biological tissues and functional structures, e.g., soft robots. The rapid growth in this research field has motivated detailed investigations on their failure mechanisms. Applications such as soft skin patches, meshes for wound closure, and bioadhesive skin sensors require an in depth understanding of failure mechanisms in soft matter to design materials with improved fracture behaviour. To this end, in silico strategies constitute a cutting-edge approach to explore physical mechanisms that govern the propagation of cracks and pre-existing flaws. Among other approaches, phase-field models are especially attractive, primarily due to their flexibility and ease of implementation [1]. The combination of experimental insights and computational approaches allows for a greater understanding of the intricacies in the design of soft materials, eventually fostering technological progress [2,3]. The goal of this mini-symposium is to bring together researchers from experimental and modeling communities to discuss recent advancements and new directions in the field of soft fracture. Topics of interest include, but are not limited to: modeling of fracture in soft materials using discrete or/and diffuse approaches, analytical approaches, and experimental investigations on failure mechanisms in soft materials, composites in particular.
REFERENCES
[1] Moreno-Mateos MA, Hossain M, Steinmann P, Garcia-Gonzalez D. “Hard magnetics in ultra-soft magnetorheological elastomers enhance fracture toughness and delay crack propagationâ€, J. Mech. Phys. Solid., Vol. 173, 2023.
[2] Lee S, Pharr M. “Sideways and stable crack propagation in a silicone elastomer.â€, Proc. Natl. Acad. Sci. U.S.A., Vol. 116, 2019.
[3] Shrimali B, Lopez-Pamies O. “The “pure-shear†fracture test for viscoelastic elastomers and its revelation on Griffith fracture.â€, Extreme Mech. Lett., Vol. 58, 2023.
Keywords: Computational techniques, Experimental mechanics, Finite strains, Soft fracture
Many modern materials, such as additively manufactured materials, architected and meta materials, composites, wood and foams have a pronounced microstructure which dominates their constitutive behaviour. It allows their use in specialised applications in mechanical and civil engineering, but also medicine. One particularly desired property of the microstructure is the resistance against fracture and failure, which, however, cannot be described with standard models.
Phase-field and, in general, gradient damage approaches are highly suitable to model fracture in these complex structures due to their flexibility, which allows to account for a wide range of phenomena including crack nucleation and propagation in arbitrary crack patterns. In this framework, the extension to behaviours such as plasticity, rate-dependent constitutive laws and anisotropy are straightforward as well as the coupling with environmental conditions, especially in terms of temperature or diffusion/reaction processes. Moreover, the effects of cyclic or dynamic loadings can be easily included, making thus the phase-field method an appealing technique to model complex materials.
A characteristic of heterogeneous materials often is their inherent multiscale nature, which greatly influences their fracture behaviour. Indeed, effective fracture resistance and crack paths are governed by the microscopic fracture behaviour, whose inclusion in a computational framework comes at the price of the computational burden. This is further increased by complex geometries and/or loading histories. Therefore, efforts have been made to reduce the computational cost with adaptive meshing strategies and new approaches to solve the system of coupled equations.
The topics of this MS of interest include, but are not limited to the following:
• Phase-field and gradient damage models for complex fracture behaviour, including viscous and ductile fracture
• Material failure under various loading conditions (e.g., dynamic and cyclic loading)
• Modelling of coupled multi-physics problems, including thermo-, chemo- and electro-mechanical systems in fracture mechanics
• Simulation of fracture in microstructures, taking into account the effects of texture and anisotropy
• Multi-scale approaches and macroscopic effective fracture properties
• Advances in algorithms and solver technologies (e.g., staggered/monolithic schemes, reduced order models and meshing/discretisation techniques) to reduce computational time
Keywords: coupled problems, damage, fracture, microstructure, Phase-field
ABSTRACT
Hydrogen-assisted fracture phenomena have been a longstanding concern in various industrial and energy sectors. With the increasing interest in hydrogen as a future energy source, it is crucial to understand the risks it poses to the structural integrity and fatigue life of components. The transport of dissolved hydrogen atoms within a component’s bulk and their entrapment in microstructural defects leads to dramatic reductions in the material’s ductility, fracture toughness and fatigue crack growth resistance, through a phenomenon termed hydrogen embrittlement.
This mini-symposium aims to explore the latest advancements in multi-physics modelling approaches for predicting hydrogen-assisted fracture phenomena. By bringing together researchers and industry experts, we seek to foster collaboration and exchange ideas to address the challenges associated with understanding and mitigating hydrogen-assisted cracking phenomena, across relevant scales and operating conditions. Particular emphasis is placed on the use of advanced computational fracture mechanics techniques such as continuum damage models, cohesive zone approaches and phase field modelling (see, e.g., [1,2] and Refs. therein). Additionally, the mini-symposium highlights the noteworthy advancements in innovative testing methodologies aimed at facilitating the assessment of hydrogen-assisted fractures in technologically-relevant scenarios, and their combination with advanced modelling techniques.
REFERENCES
[1] E. MartÃnez-Pañeda, A. Golahmar, C.F. Niordson, “A phase field formulation for hydrogen assisted crackingâ€, Comput Methods Appl Mech Eng., Vol. 342, pp. 742ï€761, (2018).
[2] C. Cui, R. Ma, E. MartÃnez-Pañeda, “A generalised, multi-phase-field theory for dissolution-driven stress corrosion cracking and hydrogen embrittlementâ€, J Mech Phys Solids, Vol. 166, 104951, (2022).
Keywords: Fracture analysis , Hydrogen-assisted cracking, Numerical modelling
The phase field method (PFM) for fracture originated in the late 1990s. Francfort and Marigo [1] proposed a linearly elastic variational principle for enhancing Griffith’s energy theory, i.e., the requirement of a well-defined crack path and previously existing crack. Bourdin et al. [2] proposed a regularization for enabling the numerical simulation of this theory, later called the PFM. As a continuous approach to fracture, instead of tracking crack surfaces, PFM represents crack using a scalar field (phase field). Fracture processes including crack nucleation, propagation, joining, and branching are naturally predicted as the result of the minimization of its energy functional without any extra fracture criteria.
PFM can incorporate various material properties including anisotropy, elastoplasticity, viscoelasticity, hyperelasticity, piezoelectricity, etc, thermodynamically consistent. Through enhancing the energy functional, PFM is successfully applied to not only classical mechanical problems, e.g., fatigue, composite delamination, functionally graded materials, rock fracture, large strain fracture of polymer, and interfacial fracture of concrete, but also many multi-physics problems, including hydrogen embrittlement, cement hydration, stress corrosion, Li insertion, coupled fluid–structure fracture, and polymer oxidative aging, etc.
We initiate this symposium to report and discuss recent progress in various aspects of PFM.
Topics of interest include but not limit to:
- Multi-physics problems of phase field method for fracture
- High performance computing strategies in phase field method for fracture
- Multi-scale phase field modeling of fracture
- Complex fracture problems with phase field method for fracture
- Experimental validation and comparison for the phase field theory
REFERENCES
[1] Francfort G A, Marigo J J. Revisiting brittle fracture as an energy minimization problem. Journal of the Mechanics and Physics of Solids, 1998, 46(8): 1319-1342.
[2] Bourdin B, Francfort G A, Marigo J J. Numerical experiments in revisited brittle fracture. Journal of the Mechanics and Physics of Solids, 2000, 48(4): 797-826.
Keywords: Computational fracture mechanics, Phase field approach to fracture
This minisymposium aims at gathering recent advances in the mathematical and numerical modeling of polymer mixing technologies.
Polymer manufacturing processes are ubiquitous in many industrial sectors. The development of accurate and efficient numerical simulation tools has greatly improved the manufacturing capability, control and industrial productivity of such processes. Further improvements can be obtained exploiting new emerging numerical technologies.
The topics covered in this mini-symposium include (but are not limited to) conforming and non-conforming discretization methods, mesh generation techniques, particle-based methods, machine-learning techniques, mixing indices, rheological modeling, turbulent and multiphase flows.
Contributions from both academic and industrial research centers are encouraged.
Keywords: : Polymer mixing, non-Newtonian rheologies
Supercomputers have made available to researchers an unprecedented amount of computing power. But "power without grip" is useless: the availability of thousands of processors to compute must be accompanied with a steep evolution in software development based on High-Performance Computing (HPC) techniques, to open a completely new way of facing the most complex simulation problems of Computational Physics and Engineering. Especially in technology niches such as in industrial, energy, environmental or biomechanical applications, treatment of complicated or coupled phenomena of fluid and solid motions, which require a huge amount of computing resources, pose extreme challenges. Another benefit of the availability enormous HPC resources is the ability to generate big data through simulations. In recent years, using neural networks, attempts have been made to speed up calculations using surrogate models and reduced models, to reconstruct flow fields from incomplete information, and to explore design spaces using reinforcement learning.
Thus, the objective of this Mini-Symposium is to communicate and discuss issues and perspectives of HPC simulation and/or AI techniques, targeting industrial applications, which cover fields, such as bio, automotive, aerospace, pharmacology, energy, environmental, etc. The expected topics should include algorithms, simulation strategies, and programming techniques for the kind of complex simulations of fluid/solid phenomena (usually including coupled multiphysics) requiring massively HPC environments to solve. Parallel issues, such as the robustness and performance analysis, and introduction of pre- and post-processing techniques, e.g., CAD integration, mesh generation, or visualization are also welcome.
Keywords: Computational Fluid Dynamics, Computational Mechanics, Industrial Applications, Artificial Intelligence, High-Performance Computing
Modeling pharmaceutical processes poses significant challenges due to the complexity and diverse nature of the involved steps. Creating something as seemingly simple as a tablet consists of many processing stages, including wet or dry granulation, drying, blending, tableting, and tablet coating, with the additional complexity of batch or continuous forms. The complexity further intensifies when considering different forms of drug delivery. Consequently, pharmaceutical applications of DEM simulations are often scattered across various sessions, addressing specific aspects such as granular flow, fundamentals, feeding and blending, and coupled problems.
This mini-symposium aims to overcome this fragmentation and consolidate the pharmaceutical applications of DEM simulations into one integrated session. By bringing together researchers from diverse groups working in the pharmaceutical field, the event seeks to facilitate collaboration and knowledge exchange. All applications utilizing pure or coupled DEM simulations to model pharmaceutical processes are welcome, encompassing many scenarios beyond the earlier examples.
Emphasis will be placed on presentations that link DEM results to critical quality attributes of intermediate and final products. The session will explore how DEM simulations can provide insights into cohesive pharmaceutical powders, complex particle shapes, granulation, drying, blending, and tablet coating, as well as their impact on the quality of pharmaceutical products.
This mini-symposium aims to foster a comprehensive understanding of the latest advancements in modelling pharmaceutical processes using DEM simulations by uniting these disparate applications under one session. Attendees will gain valuable insights into the challenges, opportunities, and potential improvements in drug manufacturing, ensuring the production of high-quality pharmaceutical products that meet stringent industry standards.
Keywords: Batch Manufacturing, Continuous Manufacturing, Feeding, Mixing, Powder Processing, Process Engineering, Discrete Element Method
The design process in engineering applications is currently experiencing a change in paradigm, away from experience-based design to numerical design. In many such engineering applications, flows of complex fluids are encountered, posing the challenge of understanding, describing, computing, and controlling these flows. In this spirit, this mini-symposium aims at providing a forum for questions concerning both numerical and optimization methods specific to fluid flow. On the modeling side, it covers the issues related to complex, non-Newtonian flow phenomena, such as the choice of model or appropriate stabilization. Furthermore, in the area of simulation, novel numerical methods, ranging from discretization methods to both free-boundary problems and deforming domain problems, are considered. In all cases, the flow solution may serve as the forward solution of a shape optimization problem. To this end, this mini-symposium will cover novel techniques for shape representation as well as new methods for an efficient evaluation of the design.
Topics of this mini-symposium include, but are not limited to:
- Non-Newtonian fluid models describing shear-thinning or viscoelastic properties.
- Simulation methods, including stabilization schemes, interface capturing, and interface tracking.
- Methods related to shape optimization in fluid flow, in particular geometry
representation, reduced order models, and development of objective functions.
- Methods particular to specific applications.
Keywords: Model Order Reduction, Moving Boundaries, Non-Newtonian Fluids, Shape Optimization
Design optimization in structural mechanics and fluid dynamics is nowadays widely applied in academia and industry. However, most applications in industry still utilize linear models on the simulation side. Optimization of highly dynamic and/or nonlinear phenomena, the consideration of Multiphysics and the treatment of uncertainties is still a challenge. These challenges result for instance from the need for mesh regularization, less reliable convergence of optimization algorithms, and the very high computational cost associated with nonlinear transient and/or uncertain problems.
This mini-symposium aims to bring together experts in the field of shape or topology optimization with applications in fluid dynamics, structural mechanics and dynamics as well as coupled problems. The goal is to exchange ideas and discuss the latest achievements on efficient optimization strategies for nonlinear problems, dynamical systems, optimization under uncertainty, and for novel approaches to CAD-free shape optimization methods.
Keywords: Computational Mechanics, Robust Design, Uncertainty, Fluid Dynamics., Optimization
Many problems in various fields of science and engineering are posed mathematically as inverse problems. In contrast to the more standard “forward problemsâ€, where the model describing the physical system is fully described, and the goal is to find the response due to given sources or known material and topology of the structure, the objective of an inverse problem is to find missing information on the model, based on some given information (acquired by measurements from sensors or by a priori design) on the response in space and/or time or to find optimal parameters or functions. Examples include, among many others:
(a) identifying damage in a structure;
(b) locating the epicenter of an earthquake from measurements on the ground;
(c) finding the optimal topology of an object that yields a desired function;
(d) finding the optimal melt cooling strategy (in space and time) in a crystal-growth process;
(e) identifying the main acoustic sources in a city;
(f) finding the optimal model and material parameters of a solid from experiments.
There is a rich literature on computational methods for inverse problems. Known methods include Linear Sampling Method, Arrival Time Imaging (Kirchhoff Migration), Time Reversal, Parameter/Topology-sensitivity based analysis, Stochastic (e.g. Bayesian) approaches, Full Waveform Inversion, the latter often relying on a gradient-based optimization and on the adjoint method, and more. Research in this area, aimed at devising new methods and improving existing ones, is very active, since inverse problems are notoriously hard; not only are they usually strongly nonlinear, but they are almost always ill-posed. In order to be effective, methods for the solution of inverse problems must be robust, efficient, and perform well even in the presence of noisy or uncertain data. Additional interesting and important challenges arise from the need for computationally-intensive solution methods. Examples include reduced order and surrogate models, and linear/nonlinear solvers.
Keywords: Detection, Identification, Inverse Problem, optimal
The purpose of this session is to bring together researchers who successfully develop techniques for solving structural and multidisciplinary optimization problems, focusing on both single and multi-objective problems. We encourage new research into all aspects of the optimal design of structures as well as multidisciplinary design optimization where the involved disciplines deal with the analysis of solids, fluids or other field problems.
This special session aims to provide a forum for researchers from theoretical developments for single or multi-objective optimization as well as their application for solving real engineering problems. Submissions presenting novel developments or critical reviews are welcome.
Contributions are invited on the following topics:
• Application of structural optimization in nanotechnology
• Application of structural optimization in computational design of materials
• Application of structural optimization for the design of mechanisms
• Application of structural optimization with peridynamics
• Optimization of composite and smart structures
• Sampling and surrogate modeling in design optimization
• Numerical optimization techniques
• Experimental optimization techniques
• Shape and topology optimization
• Surrogate-based optimization
• Multidisciplinary optimization
• Multi-objective optimization
• Robust and reliability-based design optimization
Keywords: Experimental optimization techniques, Multi-objective optimization, Multidisciplinary optimization, Numerical optimization techniques, Shape and topology optimization, structural optimization, structural optimization in computational design of materials
Computational Fluid Dynamics (CFD) tools, methods and workflows are an integral part of modern aerodynamic design and analysis of aerospace systems over a significant range of the flight envelope. Despite the immediate availability of these tools and significant progress since their inception, the full potential of RANS CFD for informing early design stages can still be limited by it’s labour and computational demands. Therefore, high-fidelity CFD tools are not deployed at the most critical point in the design process.
Current technological inflexion point for aircraft propulsion and the pressure to reduce time to marker, requires manufactures to rapidly assess new and radical configurations. Therefore, further advances are necessary to enable rapid and accurate aerodynamic analysis, capable of capturing the relavant physical phenomena, for new and unconventional configurations at a cost suitable for conceptual and preliminary design studies. This represents a significant challenge due to the lack of detailed knowledge available and wide design space available to assess the potential of a new vehicle.
This symposium offers a platform for researchers developing fast numerical methods that can be adopted for early design of transonic, fixed wing aircraft, to communicate the latest advancements in the field. Applications of interest include, among others, rapid assessment of unconventional configurations, prediction of flight loads and wave drag assessment. New methodologies for multi-fidelity, reduced-order models, efficient generation of surrogate models including AI, robust fluid-structural coupling, etc., all of which expand the use and/or accuracy of fast aerodynamic predictions tools are of particular interest.
Keywords: aerodynamic analysis, CFD, Model Order Reduction, Multi-objective optimization, Multidisciplinary optimization, Numerical optimization techniques, Reduced Order Modeling, Robust Design, Surrogate models
Coupled systems as they appear often in the form of e.g. fluid structure interaction or control of complex systems pose particular challenges already in solving the forward problem, as they often require the coupling of discretisations, solution algorithms, and software. These challenges are even larger if inverse problems like identification are tackled, or if one wants to perform an uncertainty quantification or optimisation for such a system, or if it is intended to design a control algorithm to achieve some desired optimal outcome. To reduce the computational burden, in such cases reduced order models (ROMs) or proxy models, sometimes combined with machine learning --- lately often in the form of deep neural networks --- are used. Due to the intended use in optimisation, uncertainty quantification, control, or deterministic or stochastic identification, such models are necessary parametric models, often involving large numbers respectively dimensions of parameters.
For coupled systems, the use of such ROMs is even more desirable, but they are often produced for each system component separately, and the problems of coupling then transfers to these ROMs. The questions which arise and should be addressed in this mini-symposium are then on how to produce such parametric reduced order models, and how to couple or combine models from different sub-domains to perform one of the afore mentioned tasks for the whole system.
Of particular interest are new fast and computationally effective and accurate algorithms, as well as contributions to their formulation and analysis, computational procedures, understanding, error estimation.
The mini-symposium is to bring together researchers in these fields and offer a look at the problems alluded to above, contribute to the computational, physical, engineering and mathematical insight, and offer vistas and perspectives at the formulation, analysis, and computational solution.
Keywords: Coupled Systems, Identification, Optimisation, Reduced Order Models
This MS is in the area of gradient-free and gradient-based (mostly discrete and continuous adjoint) optimization methods as primarily used in fluid and solid mechanics. Support by reduced order models and Machine Learning techniques for these kinds of applications is also in the scope of the MS. Papers in either Single- or, particularly, Multi-Objective Optimization, including also design under uncertainties (and, thus, uncertainty quantification) are welcome. Emphasis is laid on the use of the above methods in the optimal/innovative designs for a climate neutral aviation. However, papers in other areas that make use of similar tools are welcome.
The topic of climate neutral aircraft has been selected as the key application area since sustain air traffic growth while reducing the environmental footprint produced by CO2, NOX emissions and aviation-induced cloudiness (such as contrails) represents a crucial challenge for the evolution of the aeronautical sector. To reach this very ambitious objective, new technologies must be developed, matured, and integrated into disruptive aircraft architectures. The development of innovative fast design tools and optimization methods for aerodynamics, structure and propulsion is therefore crucial to this goal. Unlike conventional aircraft architectures and engine systems, for which an exhaustive return on experience is available, the design of breakthrough technologies and new concepts requires enhanced simulation approaches, able to easily integrate high fidelity modelling, ingest real data as well as to account for tighter multi-physics couplings and uncertainties. Although advanced simulation models can account for epistemic uncertainty (through full resolved physics and data assimilation), additional sources of uncertainty are unavoidably introduced by manufacturing processes and the variability of operating conditions across the lifespan (also due to progressive deterioration or damage). Therefore the development of more efficient robust optimization capabilities is of utmost importance to avert showstoppers and reduce all potential barriers to the entry into service of future green aircraft and engine.
Keywords: adjoint, gradient-based, gradient-free, machine learning, multi-disciplinary optimization, robust optimization
The conception, design and operation of complex systems entail computationally-intensive procedures, such as inverse problems, uncertainty quantification, data assimilation, real-time monitoring and optimisation.
Efficient sampling and surrogate modelling techniques, as well as multi-fidelity approaches, are among the key enabling technologies to make such problems treatable. These techniques should not be understood as mutually alternative options, but rather as complementary tools. For instance, surrogate models allow for substantial speed-up of computations but are crucially grounded on the ability to efficiently sample the parameter space. Moreover, when sampling expensive models, multi-fidelity approaches become essential, since they allow to appropriately allocate the computational effort on models with different degrees of accuracy, depending on the specific needs of the target application and on the region of the parameter space under investigation (exploration vs. exploitation).
This minisymposium aims to gather contributions on state-of-the-art techniques in scientific machine learning (understood in a broad sense, from reduced-order modelling and statistical computing to deep and reinforcement learning) and to foster the cross-dissemination of ideas towards the development of multi-fidelity and multi-disciplinary digital twins. Particular attention will be devoted to the challenges of blending experimental data, physical models and mechanistic knowledge of different fidelities. This is especially critical in the context of data-driven engineering, where information stemming from different sources, with different cost and reliability, is available. Application-oriented contributions tackling problems related to industry, sustainability, smart cities and environmental issues are particularly welcome.
Keywords: Digital Twins, environmental application, industry, inverse problems, mathematics of planet earth, multi-fidelity, Optimisation, sampling, Scientific Machine Learning, smart cities, surrogate models, sustainable development, uncertainty quantification
Analysis and management of engineering systems require decision-making under uncertainty. Such uncertainty stems from the intrinsic variabilities in the system and manufacturing processes, ambiguity in the computational model due to lack of precise knowledge of the governing physics as well as noisy measurements/data. Optimization methods have widespread application in engineering decision-making. Accounting for the uncertainty in engineering models during optimization often involves dealing with high-dimensional uncertain inputs, which poses additional computational challenges. Recent advances in the fields of uncertainty quantification and machine learning provide effective tools for optimization under uncertainty. This mini-symposium aims at bringing together researchers, academics and practicing engineers concerned with the various forms of optimization in the presence of uncertainties. We seek contributions discussing novel optimization algorithms and methods, as well as applications to practical problems. Areas of interest include, but are not limited to, (multi-objective) design optimization of engineering systems under uncertainty, robust optimization, performance-based optimization, reliability-based optimization, stochastic optimization, risk management and optimization, optimal decision-making in presence of uncertainty, development and application of (machine learning) surrogate models for optimization under uncertainty, reduced order modeling, multi-level and multi-fidelity formulations and data-driven optimization.
Keywords: Computational Models, Design, Machine Learning, Optimization, Uncertainty Quantification
Computational models play a crucial role in predicting the behavior of complex physical and mechanical systems. However, the parameters within these models are often influenced by uncertainties, making the determination of the model responses troublesome. Bayesian inference emerges as a potent statistical tool for inferring these uncertain model parameters using available data and information. To assure that the numerical model is trustworthy enough, one must take into account the prior uncertainties of these input parameters or fields, and improve these models by data (e.g. of model responses) gained from some higher fidelity numerical model or from physical experiments, or even by the observation of the actual system.
The problems of uncertainty quantification and Bayesian updating have gained significant attention in the field of computational science and engineering due to their potential to enhance decision-making, optimize system performance, and provide valuable insights into complex phenomena. Bayesian inference methods typically rely on sampling algorithms. When dealing with computationally expensive models, such exploration demands substantial computational resources. As a result, recent research endeavors concentrate on developing efficient Bayesian inference methods through surrogate modelling and novel mathematical formulations and advanced sampling and filtering techniques. Methods facilitating a sequential update of the model parameters based on continuously observed responses have also enabled the “digital twinning†concept. This chimes very well with the view that machine learning is in fact a Bayesian estimate of a conditional expectation; and as such machine learning procedures for digital twins are also in the scope of the mini-symposium.
This mini-symposium serves as a platform to discuss the latest developments and foster collaboration among researchers working on Bayesian inference. The scope encompasses a range of topics, including but not limited to methodological advancements or novel applications, advanced methods for Bayesian analysis, machine learning, inference techniques accounting for spatial or temporal dependence of uncertainty, recursive or online Bayesian inference, hierarchical models, optimal experimental design, Bayesian inference utilizing surrogate models, accelerated Bayesian inference leveraging machine learning, and high-performance computing.
Keywords: Bayesian Inference, Digital Twins, inverse problems, Machine learning, model updating, probabilistic optimization, Surrogate Modelling
Composite materials have become increasingly attractive to modern industries due to their exceptional properties and diverse applications. The advent of new technologies, such as automated fibre placement, has opened unexplored possibilities for composites, allowing tailored structures to meet specific design requirements. Despite these advancements, several challenges persist in designing and verifying composite structures, primarily due to the lack of appropriate methodologies and analysis tools.
The proposed Mini-symposium, titled "Multi-scale Advanced Modeling and Design of Variable-Stiffness Composite Structures" aims to provide a comprehensive overview of the current state-of-the-art and prospects in simulating modern composite structures. This collaborative forum seeks to unite scientists and researchers to share innovative ideas and recent findings related to modelling and design for both classical laminates and variable-stiffness composite structures.
The Mini-symposium will delve into various topics, including the development of composite beam, plate, and shell models. Additionally, it will explore multi-scale methodologies to bridge the gap between micro- and macro-scale modelling, homogenization techniques, and capturing complex interactions at different length scales.
A crucial focus will be on addressing the challenges of uncertainty analysis to account for variations in material properties, manufacturing defects and loading conditions and how those affect design optimization. Other topics of interest include, but are not limited to, composite structures failure and damage modelling, delamination, cohesive mechanics, plasticity, impact, vibration, robust design, surrogate modelling and AI-based optimization techniques.
Through this Mini-symposium, we aim to foster interdisciplinary discussions and encourage knowledge exchange, providing a platform for researchers to explore innovative ideas and approaches. This collaborative event will create a network of experts dedicated to driving progress in composite materials research and promoting their adoption in real-world engineering challenges.
Researchers from academia and industry are welcome to contribute their work and expertise in advancing the field of composite materials. The Mini-symposium offers a unique opportunity to build valuable connections, leading to the development of novel methodologies and tools for the design and optimization of variable stiffness composite structures.
Keywords: Additive manufacturing, Advanced theories, Automated fibre placement, Design, Laminates, Multi-scale modelling, Optimization, Robust design , Tow-steered composites, Uncertainty, Composite structures
Soft tissues play a critical role in the functioning and integrity of the human body. Accurate knowledge of their mechanical properties is crucial for a wide range of biomedical applications, including surgical simulations, medical device design, and tissue engineering. However, soft tissues' complex and nonlinear behaviour poses significant challenges in obtaining precise material data through traditional experimental approaches. The advent of inverse methods has opened new avenues for efficiently estimating these material properties, offering promising opportunities to revolutionize medical research and clinical practice. Research on inverse methods combines techniques from mathematical analysis, differential geometry, numerical analysis, machine learning, image processing, scientific computing, and computer science.
This mini-symposium aims to provide a platform for researchers and practitioners to present their latest work and insights on inverse methods for determining the material properties of soft tissues, shape design and/or unknown external forces. We encourage submissions that encompass experimental, computational, and combined approaches and contributions that address advancements in related fields, such as medical imaging and biomechanical modelling.
Keywords: Computational Modelling, Optimization, Biomechanics, Material Parameters Identification
In the last decades, topology optimization (TO) has been developing at a fast pace, reaching remarkable advancements, also thanks to the employment of cutting-edge computational methods and enhanced production technologies. In particular, the synergy between the design optimization phase, with the involved spatial scales and materials, and the fabrication method, especially when resorting to Additive Manufacturing (AM), has become crucial to foster the application of TO towards diverse fields.
The recent research in TO is engaged both in the methodological formalization of the optimization problem and in the hands-on implementation aspects. Specifically, novel methods for topology optimization are complementing the already established ones, in order to attack some of the challenges that the TO community is facing. For instance, special interest has been attracted by methodologies that permit to accurately and smoothly describe the topological changes occurring during the optimization process and to provide designs that can be immediately integrated in CAD software (e.g., Computer Aided Design (CAD)-compatible density-based algorithms, Moving Morphable Components, Moving Morphable Voids, and Geometry Projection). On the implementation side, many efforts are deployed to devise algorithms that allow to efficiently tackle the complexity arising in diverse case studies, possibly involving multi-physics, multi-scale, multi-material, multi-constrained and multi-objective scenarios.
In this mini-symposium, we aim to gather recent contributions on different topics regarding TO and the related manufacturing processes. With the aim of providing a stage for innovative ideas and fruitful collaborations, we encourage the presentation of original results on:
• Emerging methods for topology optimization;
• Multi-scale and multi-material topology optimization techniques, with emphasis on the manufacturing phase via traditional or AM fabrication processes;
• Efficient and scalable numerical methods for large-scale multi-objective and multi-physics topology optimization;
• Topology optimization for industry-driven case studies.
Keywords: Computer Aided Design, Manufacturing processes, Multi-objective optimization, Multi-physics modeling, Multi-scale modeling, Topology Optimization
Spurred by the dramatic increase in the cost of raw materials and the pressing necessity of realizing energy savings, shape and topology optimization has been a thriving field of research for applied mathematicians, physicists and engineers over the last decades, leading to the development of manifold numerical algorithms.
The main issue at stake — that of finding the optimal domain with respect to a given criterion about its physical performance — naturally arises in a whole gamut of situations: beyond its historical applications in structure mechanics and exterior aerodynamics, optimal design has made remarkable forays in acoustics, electromagnetism, chemistry, to name just a few recent examples. The algorithmic developments of shape and topology optimization have been greatly influenced by the advent of high performance computing and machine learning, and, from the mechanical viewpoint, by the rich perspectives heralded by additive manufacturing techniques.
This mini-symposium aims to gather experts in shape and topology optimization and to discuss recent advances in this discipline, ranging from the theoretical introduction and analysis of new, promising methods, to applications in the context of concrete, challenging or emerging physical situations.
Keywords: optimal control, Partial Differential Equations, Shape and topology optimization
Inverse problems and structural optimization are cornerstones in the application of computational mechanics and physics to discover solutions to challenging problems in various engineering disciplines. The growing complexity of real-world applications needs innovative computational methodologies for efficient and effective optimization. This mini-symposium aims to explore the potential of machine learning (ML) in the realm of inverse problems, optimization, and optimal design. The symposium endeavors to bring together researchers and practitioners to discuss the interplay between traditional optimization techniques and innovative ML approaches. Specific sub-topics include, but are not limited to:
1. Surrogate-based Optimization: Highlighting the role of surrogate models, such as Gaussian Processes and Neural Networks, in reducing the computational overhead in iterative optimization tasks [1].
2. Topology Optimization: An in-depth exploration into the application of ML in optimizing material distribution within a predefined design space, contributing to lightweight and robust designs [2]. Specific emphasis on the usage of gradient based techniques including adjoint methods is encouraged.
3. Constraint Handling in ML approaches: A discourse on methodologies for efficiently managing constraints in optimization problems solved by ML techniques [3,4].
4. Real-world Applications: Case studies from illustrating the efficacy of ML in complex design optimization and inverse problems [5].
The goal of this mini-symposium is to discuss ML advancements in inverse problems and optimization applications. Submissions including dynamic behavior of materials, shock physics, or multi-physics are strongly encouraged. We will encourage discussion highlighting the advantages and disadvantages of these methods.
[1] Forrester, A., Sobester, A., & Keane, A. (2008). Engineering design via surrogate modelling: a practical guide. John Wiley & Sons.
[2] Bendsoe, M. P., & Sigmund, O. (2003). Topology optimization: theory, methods, and applications. Springer Science & Business Media.
[3] Coello, C. A. C. (2007). Evolutionary algorithms for solving multi-objective problems. springer. com.
[4] Ji, W., Chang, J., Xu, H. X., Gao, J. R., Gröblacher, S., Urbach, H. P., & Adam, A. J. (2023). Recent advances in metasurface design and quantum optics applications with machine learning, physics-informed neural networks, and topology optimization methods. Light: Science & Applications, 12(1), 169.
Keywords: Computational Mechanics, Deep Learning, Fluid Dynamics, Inverse Problems, Material Dynamics, Optimization, Shock Physics
The field of materials manufacturing processes has evolved considerably in recent decades, driven in particular by the numerical, energy and ecological transitions: with the support of numerical approaches, processes are targeted to consume less energy and even less raw material. This is the case of additive manufacturing processes (metallic, ceramic or polymer) which make it possible to obtain complex shaped parts, or the manufacture of composites with long or short, natural or synthetic fibres that make structures lighter. All these processes involve the dynamics of several types of fluids (or/and fusion zones), generally multiphase, giving rise to rather complex phenomena such as capillarity, the understanding of which is essential to the process optimisation. This requires high-fidelity, multi-physics and multi-scale simulations, with limited computation time, and even reaching real time. The most widely used numerical strategies which are still the subject of intensive research, include (stabilised/enriched/embedded) finite elements, phase-field and/or level-set approaches, mesh adaptation, parallel computing or computational homogenization (see e.g., Ref. [1]) among others. Nowadays, these techniques are increasingly coupled with machine or deep learning methods, allowing for example, via an off-line step, to simply generate data, reduce computation times (through reduced order model methodologies, meta-models or physically informed neural networks), or to statistically quantify the intrinsic variability of upscaling descriptors (see e.g., Gaussian processes / kriging in Ref. [2]).
The objective of this mini-symposium is to review the latest advances in the computational mechanics community in high-fidelity modelling and simulation of flows in manufacturing processes. This includes the development of numerical methods and multi-scale approaches to enrich physical models, as well as the coupling between simulations and advanced learning techniques.
REFERENCES
1 M. Shakoor, C.H. Park. Computational homogenization of unsteady flows with obstacles. Int J Numer Meth Fluids, 95(4): 499–527, 2023.
2 A. Geoffre, N. Moulin, J. Bruchon, S. Drapier: Reappraisal of upscaling descriptors for transient two-phase flows in fibrous media. Transp Porous Med, 147: 345–374, 2023.
Keywords: flows, multi-physics, multi-scale, multiphase, AI, Manufacturing processes
High-performance components produced using Additive Manufacturing (AM) technologies are a reality in many industrially-relevant applications. Nowadays, complex structural parts can be produced using AM technologies with almost unlimited design freedom and locally varying material properties. Therefore, in recent years, the so-called metamaterials, e.g., architected cellular or lattice structures, have known an increasing interest due to the possibility to design structural parts with tailored mechanical properties and performance.
Despite their huge potential, widespread adoption of metamaterial structures has been hindered by concerns about their structural integrity. In fact, considering the complexity of the manufacturing process, many potential process-induced defects can be present in the final component, limiting their reliability and applications.
Numerical models can thus be a crucial tool to shed light on the complex process-structure-property-performance relationships occurring in AM metamaterials. In fact, only a deep understanding of these relationships will allow us to fully control process errors and their effects on the mechanical properties and performances of these kinds of structures.
In the present mini-symposium, the following topics, related to AM metamaterials, will be considered:
 Cell-based modeling for periodic structures
 Instability and structural behavior
 Material modeling, calibration, and validation
 Effects of defects
 Image-based analysis
 Machine learning techniques
 Process simulations
 Fatigue analysis
 Engineering applications
 Computational ontology
Keywords: Lattice structure, Material modelling, Metamaterials, Additive Manufacturing
Numerical modelling of welding and WAAM processes becomes a decision making tool used
to speed up the development and qualification of welding and repair techniques. Both
computation welding mechanics which concerns the modelling of welding effects in the base
metal and in the weld at solid state (temperature field, microstructure, stress and strain
distribution…) and multiphysics simulation which focuses on arc/plasma and weld pool
modelling can be implemented for this purpose.
This minisymposia is organised by the Scientific and Technical Committee on Numerical
Welding Simulation with the help of the French Association of Mechanics (AFM). The goal is
to make an update of the progress made in welding and WAAM concerning:
ï‚· modelling of the process what can we model today and with what accuracy, what are
the couplings effects to be taken into account;
ï‚· behaviour laws (metallurgy, hardening recovery, viscous effects, simplified methods,
...);
ï‚· real-life size structures (life time, match computation time with industrial needs, ...).
These elementary bricks will be helpful to characterize the overall welding process in order to
numerically simulate the behaviour of a structure (distortions, fatigue resistance, damage),
while relying on cases validation tests (calculation / test comparison).
Topics of the minisymposia on modelling and simulation of welding and wire arc additive
manufacturing processes in the broad sense will include:
Very large structures, thick components, how to simulate the very large number of passes?
ï‚· Performance and process control: multiphysics advances for the simulation of welding
processes (molten bath and arc) allowing high quality welding.
ï‚· What are the benefits of integrating the manufacturing history to justify the lifetime.
ï‚· Effect of welding on the service behaviour of welded joints (low cycle fatigue, stress
corrosion, fracture ...).
ï‚· New models of welding simulation to improve the controllability of structures and make
NDT diagnostics more reliable.
ï‚· Simulation of heterogeneous welding.
ï‚· Special processes (reloading, repair, FSW, resistance, Hybrid, ...).
ï‚· Residual stresses and distortions, control of the risks of cracking during welding.
ï‚· State of modelling materials for welding simulation.
ï‚· Wire Arc Additive Manufacturing process.
ï‚· Research and experimental computational tools.
Keywords: Computational Mechanics, cracking risks, distortions, manufacturing, repair, residual stress, WAAM, Welding process, Wire Arc Additive Manufacturing
Various additive manufacturing (AM) techniques including 4D printing have been developed to manufacture complex-shaped components with well-controlled precision. Sophisticated AM techniques often require systematic modelling and simulation efforts during the design stage and for the purpose of part qualification/certification. The objective of this minisymposium is to provide a platform to discuss recently developed modelling and simulation techniques for AM, including experimental calibration and validation efforts for the process. The topics include (but are not limited to):
Simulation and optimization of the manufacturing process to predict heat transfer, residual stress/distortion, surface topology, composition, and microstructure including defects at multiscale length and time scales.
• Process simulation and optimization at micro and macro scales
• Data-driven approaches for simulation acceleration
• Combined simulation and in-situ monitoring for rapid build qualification
• Effects of microstructure and defects on mechanical properties
• Feedback control for minimizing defects and residual stress in as-built structures
Computational modelling and simulation for any AM processes (e.g. laser power bed fusion, electron beam melting, form deposition modelling, stereolithography, binder jetting) and materials (e.g. metals, plastics, ceramics and their composites as well as biological materials) are welcome.
The importance of advanced computational models in developing revolutionary materials and structures cannot be overstated, especially in light of the transformative impact of additive manufacturing on the production of intricately detailed, precision-based small-scale features. This technological shift brings forth new design challenges for creating innovative materials and structures. The primary objective of this mini-symposium is to gather researchers from diverse disciplines to facilitate in-depth discussions on state-of-the-art analytical models, numerical tools, and experimental approaches for the design, optimization, characterization, and fabrication of advanced materials and structures. The ultimate goal is to seamlessly integrate materials design and digital manufacturing into a unified workflow, an endeavor of paramount significance for advancing the realms of digital twin technology and Integrated Computational Materials Engineering (ICME) during the fourth industrial revolution.
The potential topics for discussion in this symposium include, but are not limited to:
1. Design, simulation, and optimization of advanced materials:
Acoustic, mechanical, thermal, and electromagnetic metamaterials; Architecture materials; Hierarchical materials; Nano-materials; Soft materials; Bio-inspired materials; Composite materials;
2. Computational methods for materials and structures design:
Stochastic modeling and uncertainties; Isogeometric Analysis; Molecular dynamics; Artificial intelligence / Machine Learning; Homogenization Methods/Inverse Homogenization; Topology optimization; Multiscale algorithms
3. Additive manufacturing simulation and experimental studies related to materials and structures design:
Multi-scale and multi-physics modeling in additive manufacturing; Microstructure and defects analysis; Topology optimization considering manufacturing constraints; Uncertainty quantification and propagation in additive manufacturing; Combined experimental and numerical studies in additive manufacturing; Fracture and fatigue behavior of additively manufactured materials and structures; Machine learning techniques in additive manufacturing
Keywords: Artificial Intelligence, Multi-scale modelling, Topology Optimization, 3D printing, Finite element method, Materials Design
Smart soft materials are a class of soft materials whose physical properties can be controlled by one or more external stimuli such as temperature, pH, light, electric or magnetic fields. They include, for instance, phase transforming soft materials, stimuli-responsive polymers and hydrogels. Due to their unique properties these materials can be applied in numerous applications, from automotive to medical and robotics. Additive Manufacturing, often denoted as 4D Printing, has also emerged as a technological frontier in the advancement of complex parts using multi-functional soft materials.
To further accelerate this technology towards widespread adoption, computational modelling, simulation, and design optimization are of particular importance. However, they still represent a fundamental, but challenging and developing topics, due to the close connection between the manufacturing process, material(s) functionalities, and the final design.
This Minisymposium welcomes contributions on the latest advances in modelling, simulation, and optimization of additively manufactured smart soft materials. Contributions will outline the application problem of interest and will demonstrate the theoretical, numerical, and experimental work being conducted to validate the proposed solution method. Areas of interest will include, but will not be limited to:
• Mechanical and multi-physical constitutive modelling of smart soft materials at different scales
• Modelling and control of smart soft materials and composites
• Computational simulation and discretization methods, including additive manufacturing process modelling
• Topology and design optimization for smart soft materials and composites
• Additive manufacturing and 4D printing technologies for smart soft materials
• Experimental characterization and validation methods
• Computer-aided design for smart applications
Keywords: 4D printing, Additive Manufacturing, computer-aided design, experimental characterization, material modeling, smart materials, soft materials
The term “metamaterial†indicates, in a broad sense, an engineered material with effective properties arising from a tailored design of its architecture, either at the nano/microscale or at the macroscale. This architecture is conceived to attain exotic properties that may not be found in nature, as enhanced strength and stiffness for given density, negative stiffness, negative or zero Poisson’s ratio, enhanced energy absorption, and frequency-dependent mechanical properties to mention only a few, which open new and unforeseen opportunities for mechanical and acoustic applications. Moreover, a growing number of studies is currently devoted to designing metamaterials with programmable properties in space and time and to developing materials with multifunctional properties, including mechanical, thermal, piezoelectric and optical. In this context, the mini-symposium aims to gather the most recent theoretical and computational studies on elastic and acoustic metamaterials, with the purpose of stimulating a fruitful discussion among experts in this broad field. Topics covered include, but are not limited to:
• Locally resonant metamaterials
• Metasurfaces
• Origami/kirigami-based designs
• Bioinspired designs
• Hierarchical designs
• Multiphase metamaterials
• Time-varying metamaterials
• Topological metamaterials
• Design, size/shape optimization, topology optimization
• Fabrication processes
Keywords: Computational Mechanics, Material modelling, Materials Design, Metamaterials
Mechanical metamaterials are engineered materials with unconventional mechanical behavior that originate from artificially programmed microstructures along with intrinsic material properties. One of the rapidly emerging trends in this field is to couple the mechanics of material behavior and metamaterial architecture with different other multi-physical aspects such as electrical or magnetic fields, and stimuli like pneumatic pressure, temperature, light or chemical reactions to explore the scope of programming on-demand mechanical responses [1, 2]. This mini-symposium aims to concentrate on such active programmability in metamaterials along with physical and artificial intelligence across the length scales (including nano-scale metamaterials). The interest in this context would include (but not limited to) the evolving trends and challenges concerning the notions of real-time reconfigurability and functionality programming, nano-scale metamaterials, artificial intelligence and machine learning in metamaterials, inverse design and topology optimization, multi-physical origami/kirigami, soft and conformal metamaterials, intuitive understanding in metamaterial design, and computational additive manufacturing.
Keywords: Active metamaterials, AI and ML in metamaterial design, Mechanical metamaterials, On-demand property modulation, Programmable matter
TRACK NUMBER 1600
Multiscale computational homogenization methods refer to a class of numerical homogenization techniques for determining the effective behavior of complex and highly heterogeneous materials, and for computing the response of structures composed of these materials.
This minisymposium focuses on the developments and applications of either multiscale computational homogenization methods, including all pending challenges in this area, or on modeling and simulation methods at the scale of heterogeneous microstructures with an implicit or explicit connection to another scale. Particular emphasis is given on complex models to incorporate specific phenomena at a given scale and related simulation challenges (complex morphologies, large models, lack of deterministic description of constituents, presence of interfaces…) and emergent behavior (effective behavior not described by individual constituents).
The topics covered include (but not limited to):
• FE2 methods and alternatives (e.g. FE-FFT);
• Machine-learning/artificial intelligence techniques and surrogate modeling for multiscale analysis
• Advanced algorithms for reduction of computational costs associated with multiscale algorithms (model reduction, parallel computing…)
• Data-driven multi-scale mechanics
• Numerical or virtual material testing across the scales;
• Emergent behavior through upscaling
• Scientific computing and large data in multiscale materials modeling
• Coarse-graining of nano- and micromechanics
• Numerical modeling of materials based on realistic microstructures, e.g. provided by high resolution 3D imaging techniques;
• Computational homogenization of heterogeneous, linear, time-dependent and nonlinear heterogeneous materials, including material dynamics and metamaterials;
• Heterogeneous materials with coupled multi-physics behavior (phase change, chemo-mechanics, nonlinear thermo-mechanics...), including extended homogenization schemes;
• Multiscale damage modeling, capturing the transition from homogenization to localization;
• Computational homogenization including size effects, higher-order gradients or lack of scale separation;
• Numerical modelling of the macroscopic behavior of microstructures with complex interfaces, microcracking, instabilities or shear bands;
• Integration of stochastic microscopic models and its multiscale treatment
Keywords: Multiscale
Cementitious materials, such as concrete, cement, and mortars, play a pivotal role in modern society as the foundation of infrastructure development. With an increasing rate of urbanization and the continuous ageing of existing infrastructure worldwide, there is an urgent need for improving the sustainability, durability and performance of these materials. Virtual laboratories can harness modern computational methods to not only provide deeper insights into the complex behavior of the material during production, processing and service but also help design durable and high-performance materials while reducing carbon emissions e.g. during cement production. However, attempts at developing an accurate and comprehensive virtual laboratory for these materials is still at an early stage. Cementitious materials such as concrete are multiphase materials whose properties in the fresh and hardened state are governed by chemical and physical (multi-physics) properties and processes that range over multiple size scales (µm – cm) and additionally change with time (seconds - years). Thus, even the state-of-the-art models and simulation methods are limited in scope due to this immense complexity. The current focus of the scientific community is not only in incorporating additional physics (multiscale, multiphase, multiphysics) but also using reduced order strategies and materials informatics. This mini-symposium will focus on recent advances, challenges, and ongoing research in the computational modelling and simulation of building materials. Among others, the following topics will be covered by the mini-symposium:
• Multiscale and multilevel models (continuum micromechanics, computational multiscale models)
• Reduced-order modeling strategies
• Data-driven methods, materials informatics, AI and Machine Learning tools for building materials
• Methods for simulating damage, fracture, transport and physico-chemo-mechanical processes (creep, shrinkage, chemical dissolution, chemically expansive processes)
• Rheological modelling and classification of fresh cementitious composites
• Simulation of concrete flow and additive manufacturing processes (e.g., DEM, SPH, PFEM, CFD etc.)
Keywords: Cementitious materials, Concrete, Concrete flow, Multiscale models, Virtual Lab
In recent years, the industry's demand for materials has increased enormously. As a result, the focus of current research efforts is on the utilisation of complex materials with tailored properties and superior performance under given multi-physical loading conditions. However, due to their large influence on the individual material behaviour, the consideration of the underlying microstructure and lower-scale processes in numerical simulations is almost indispensable. The heterogeneity of the microstructure in general poses a particular challenge with regard to numerical modeling. Microstructure changes induced, amongst others, by phase transformations, dynamic recrystallization and dislocation movement are of particular importance to make reliable predictions of the material behaviour. Similarly, the consideration of multi-physics phenomena such as thermo-mechanical, electro-mechanical or electro-chemical couplings poses challenges while the efficiency of the numerical models plays a crucial role so as to save time, costs and resources.
Against this background, this mini-symposium deals with the description of microstructural changes in the context of multiscale material modeling. Particular focus is on but not limited to:
- Multiscale modeling of evolving microstructures (plasticity, phase transformation, ...)
- Multiscale modeling of coupled problems (mechanical, thermal, electrical, ...)
- Efficient solution approaches for unit-cell problems (FFT, model order reduction, ...)
- Characterisation of material properties resulting from microstructural changes
- Generation of Representative Volume Elements
- Experimental validation
Keywords: Coupled Problems, Multiscale Simulations, State-of-the-art Solution Approaches
Understanding the hygrothermal behavior of porous building materials such as concrete, earthen materials and masonry is a crucial matter considering the threats imposed by climate change and its impact on our cultural heritage and built environment [1,2]. Moreover, despite numerous advancements, the influence of temperature and moisture on the mechanical behavior of building components and structures has not been fully well understood yet [3,4,5]. Multi-physics problems in civil engineering constitute complex challenges that require interdisciplinary research and advanced computational and experimental techniques [6]. Understanding the behavior of materials and structures under various physical phenomena is crucial for ensuring the safety, durability, and sustainability of infrastructure systems. Further research in this area will contribute to the advancement of civil engineering and the development of innovative solutions to tackle the challenges faced by the field. Current directions of research include coupled modeling, multiscale analysis, consideration of nonlinear and time-dependent behavior, uncertainty quantification, computational techniques, monitoring methods, and experimental validation.
This mini-symposium will cover a range of topics related to multi-physics research on porous building materials. This includes, but is not limited to, modeling of the microstructure of concrete and other quasi-brittle materials and their mechanical response under different loading conditions, simulation of fluid flow and transport phenomena in porous media, and investigation of the impact of environmental factors on the durability and sustainability of porous building materials. The session will provide a platform for researchers and practitioners to share their latest findings, discuss challenges and opportunities in the field, and explore new paths for research and collaboration. This mini-symposium is targeted towards researchers, practitioners, and students in civil engineering and related fields. Submissions may encompass theoretical discussions, numerical simulations, and/or experimental studies, all of which are equally encouraged. The session will provide a valuable opportunity for participants to engage with experts in the field, learn about the latest advances in multi-physics research, and contribute to the development of innovative solutions for designing and constructing sustainable civil engineering structures.
Keywords: Coupled Problems, experimental characterization, multi-physics, numerical simulation, porous building materials
’Can materials act as machines?’ is one of the most pressing questions among material scientists and engineers since the last decade of the twentieth century. Machines consisting of a set of materials are usually designed to perform some specific tasks such as generating motion or lifting an object. Hence, one of the most active fields of current research is syntheses, experiments, modelling, and designs of responsive materials that can integrate within machines or act as machines. Responsive materials are smart and innovative substances that can be activated under the application of external or internal stimuli including electric field, magnetic field, pH, light, temperature, humidity or combinations of two or more of them. One of the most promising features of these materials is their ability to undergo large deformations upon the (remote or contactless) application of active fields. Their multifunctional properties make them outstanding candidates for innovative technical applications ranging from large-displacement actuators over smart sensing devices to synthetic soft tissues in flexible electronics. Most of the smart materials have unique microstructures which can be tuned to further enhance their properties. In case of magneto- and electro-active composites, these are usually composed of a soft matrix and embedded inclusions. From a theoretical and computational viewpoint, this calls for the development of homogenization schemes to help at conceptualizing customized composite's effective properties. Moreover, recent advancements in additive manufacturing (3D printing) provide ample opportunities to intricately design these materials from the micro and nanoscale to “program†their macrostructural response. At the same time, the advance of experimental techniques allowing for precise and reliable validation and testing is paramount. Thanks to advances in almost all areas of soft multifunctional materials, innovative structures can be designed in the form of thin and slender components with the potential to undergo structural instabilities (i.e., buckling) in certain loading ranges. The resulting phenomena could, for example, be harnessed to arrive at very large deformations under rather small applied fields, making materials ready for even more efficient actuation and sensing purposes. The goal of this minisymposium is to bring together researchers from experiment, modeling and simulation in order to discuss recent advancements in the area.
Keywords: 3D printing, 4D printing, active materials, electro-active materials, light-driven materials, magneto-active materials, smart materials
Multiscale problems are inherently high-dimensional and typically require (excessively) high resolution discretizations. This poses major computational problems and often renders the problem infeasible to solve within a satisfying time frame. To alleviate these problems, some form of dimensionality reduction is required. Often the larger scales in the problem are still computed directly (for instance by filtering the governing equations), yet the effect of the small scales can only be approximated by so-called closure models or empirical constitutive equations. While traditionally these were based on approximate physical models, current research focuses on machine-learning (ML) based closure models which could offer better accuracy.
In the machine learning community so-called latent-space models are an active area of research. The general idea is that most problems, while posed in a high-dimensional space, actually lie on a low-dimensional manifold. The task at hand is then to identify an optimal latent representation. In this regard, deep learning has shown to be an excellent framework due the capabilities of handling high-dimensional and non-linear problems. In this mini symposium will we focus on the application ML-based latent space methods to the problem of multiscale (closure) modelling. There are still many open questions to be answered. For instance, how do we know when we have an optimal latent-space representation? Should the latent-space ML model be deterministic or stochastic? How do we ensure that the coupling between latent-space ML model and the large-scale physical model remains stable? How do we efficiently incorporate information from the high-fidelity space into the latent space? We welcome talks from both a theoretical and a broad multiscale application perspective.
Keywords: Multi-scale modelling, Scientific Machine Learning
This mini-symposium focuses on recent developments in computational strategies to predict fracture and damage processes in multiphysics environments across multiple length scales. Predicting material degradation and failure presents significant challenges in computational mechanics, particularly when these phenomena are governed by the coupled effects of multiple physical processes, such as changes in temperature, moisture, electric fields, magnetic fields, and chemical potentials. Additionally, many material systems exhibit an inherently multiscale nature, where the interaction between material behaviour at different length scales triggers the fracture response at the macroscopic level.
The relevance of these computational strategies extends to a wide variety of applications, ranging from technical engineering scenarios (e.g., the chemo-mechanical corrosion of cementitious materials and the electro-chemo-mechanical fracture of Li-ion batteries) to, surprisingly, cultural heritage materials (e.g., the hygro-chemo-mechanical degradation of historical paintings). In this mini-symposium, the underlying common features between these different classes of applications, which are typically poorly appreciated and exploited, will be strongly emphasised. This will provide the opportunity to gather and profit from the advances made in different research communities.
The mini-symposium will cover a broad range of research areas related to computational strategies for fracture and damage processes in multiphysics simulations, possibly exploring the incorporation of multiscale models and numerical techniques capable of capturing different length scales. This includes (but is not restricted to) the following topics:
• Thermo-hydro-chemo-mechanical problems
• Diffusion-reaction processes
• Chemically induced corrosion
• Active magneto-electric materials
• Thermodynamically consistent models
• Application of machine learning in multiphysics and multiscale modelling
• Integrated numerical-experimental strategies for multiphysics processes
Keywords: Computational Mechanics, damage, fracture, multi-physics, Multi-scale modelling
The macroscopic, observable behavior of advanced materials is governed by the structure at different scales. Non-exhaustively, these scales include atomistic and mesoscopic levels. Of utmost interest, understanding failure and evaluating strength and fracture toughness requires multiscale approaches. Advanced materials include nanocomposites, polymer blends, inorganic amorphous materials, smart materials, and hierarchical materials. Advanced materials can also be extended to biological materials, which display complex multiscale features.
A possible classification categorizes the required multiscale approaches into sequential and concurrent methods. Sequential methods obtain findings on the fine scale, which are then applied to the coarse scale in a separate simulation. In contrast, concurrent methods simultaneously consider the coarse and fine-scale in hierarchical or partitioned-domain approaches. In hierarchical methods, both scales are evaluated in the entire simulation domain, while the partitioned-domain strategies only resolve the regions of interest, e.g., the vicinity of fillers in nanocomposites at the fine scale.
Commonly, multiscale strategies do not only bridge scales but also methods and disciplines, which is the scope of this mini-symposium.
Keywords: Atomistic-to-Continuum Coupling Methods, Multi-Physics Coupling, Multifunctional Materials, Multiscale Modeling, Numerical Homogenization
The minisymposium focuses on classical petrol-based and novel bio-based polymers with particular interest in modelling, simulation, and links to experiments. Depending on the problem of interest, microscale, macroscale as well as scale-bridging approaches may be suitable means to identify and reproduce the material behaviour appropriately. This minisymposium invites contributions in the field of computational treatment of polymers and their composites based on profound theoretical knowledge and / or experimental evidence. In particular, contributions addressing structure-property relations and coupled multi-physics problems covering, e.g., chemical reactions, biological processes, electromagnetism, or phase transformations, are highly welcome. Furthermore, aspects of uncertainty quantifications related to the aforementioned fields are of specific interest.
Possible contributions may discuss materials (thermosets, thermoplastics, elastomers, gels, liquid crystal elastomers, and bio-inspired materials as well as composites and nanocomposites thereof), structures (bulk polymers, membranes, fibres including muscle fibres), mechanical properties (viscoelasticity, plasticity, damage, creep, fracture, adhesion, instability), links to experiments (testing techniques as DMA, DMTA, etc. and strategies to characterise the material behaviour), engineering applications (additive manufacturing, 3d printing techniques, smart materials and (in-situ) sensors, materials for energy storage purposes, life-cycles issues including processing conditions and production methods, service conditions, long-term performance, aging, recycling, and sustainability), physical states (melts, solids, semi-crystalline and amorphous polymers as well as their evolution including polymerization, curing, and crystallization during processing), coupled problems (piezo-elasticity, electro-elasticity, magneto-elasticity, flexo-elasticity, photo-elasticity, magneto-rheology, crystallization, effects of physical aging and chemical degradation on the mechanical behaviour), and interfacial phenomena (surface and confinement effects, interfaces, and interphases).
Keywords: Bio-based Materials, Polymer Composites, Polymer Modelling, Polymer Simulation, Polymers
Architectured metamaterials have obtained significant attention in recent years due to their unique properties and potential applications across various industrial sectors. These materials are developed with specific internal structures to achieve desirable mechanical, acoustic, electromagnetic, and thermal behaviours not found in natural materials. Therefore, the optimal design of architectured metamaterials is critical for opening their full potential and realizing their practical applications.
This Mini-Symposium will focus on innovative theoretical, numerical, and experimental approaches to model and analyse architectured metamaterials. Attention will be given to homogenization approaches, multiscale and multi-physics techniques, multi-field coupling phenomena, and innovative computational techniques tailored for predicting the unique behaviour of these materials and exploring the key aspects of optimizing their design. Researchers will present advancements in simulation methodologies to tune the properties of architectured metamaterials, facilitating their integration into several engineering applications. Participants will explore how these materials have revolutionized areas ranging from mechanical, civil, naval, aerospace, and biomedical to robotics, and sports engineering.
The topics of the Mini-Symposium are but not limited to: 1) innovative manufacturing techniques for architectured metamaterials, such as 3D printing and nanofabrication; 2) computational methods and simulation techniques for analyzing architectured metamaterials under complex loading scenarios; 3) wave propagation in metamaterials; 4) case studies and real-world applications showing the successful integration of architectured metamaterials into practical engineering solutions; 5) local and nonlocal constitutive modelling approaches; 6) parametric and topological optimization methods for material design and performance enhancement; 7) multi-field problems involving coupled physical phenomena.
Keywords: Design Criteria, Multiscale Homogenization, Nanoscale Architectures
Cauchy continuum-based theories typically employed to model conventional solids may not be able to capture the complex or exotic behaviour of certain materials. In particular, materials exhibiting size effects or atypical mechanical behaviour, like architected materials, metamaterials, and materials undergoing rather complex microscopic phenomena, require models that include additional information concerning their microstructure. For instance, generalised continua theories and multi-scale approaches may be employed to predict the behaviour of this type of materials.
This mini-symposium intends to provide a place for discussion and exchange of ideas regarding the modelling, design and analysis of materials, taking into account their microstructure and their (possibly) non-classical behaviour at different scales. On the one hand, recent advances on the numerical description of this class of materials are foreseen, with focus on the multi-scale modelling through homogenisation schemes, techniques for optimal design of macro or microstructure and constitutive modelling based on generalised continua (Cosserat, micromorphic, strain gradient, ...). On the other hand, there is also place to share the application of this sort of techniques to specific classes of materials, like multi-phase materials, metamaterials, fibre reinforced composites, polycrystalline materials, biological structures, and architected materials, not to be exhaustive.
Contributions addressing but not limited to the topics listed in what follows are welcomed:
• Multi-scale models based on second-order homogenisation, micromorphic homogenisation, and homogenisation of generalised continua;
• Analysis of size effects across the scales;
• Data-driven and reduced-order-models for generalised continua;
• Constitutive modelling and parameters calibration in second-gradient continua, micromorphic continua or Cosserat continua;
• Multi-scale design and topology optimisation of high-performance materials and metamaterials;
• Analysis of the influence of strain gradients in materials behaviour;
• Numerical methods to solve generalised continua and multi-scale problems.
Keywords: Generalised Continua, Homogenisation, Metamaterials, Size Effects
The rising global demand for engineering infrastructure and the urgency to minimize the carbon footprint pose great scientific challenges regarding the development of sustainable concretes. As far as computational methods are concerned, models are needed to link the chemical/microstructural composition of cementitious materials to the behavior of engineering structures made from plain and reinforced concrete. Such models help to understand and quantify how physico-chemical processes occurring at nanoscopic and microscopic scales interplay with the macrostructural behavior of engineering infrastructure. Interdisciplinary approaches are expected to be particularly well suited to bridge all relevant spatial and temporal scales. The objective of this minisymposium is to discuss recent advances in computational modeling of concrete and concrete infrastructure. Computational models addressing various length and time scales and physical phenomena relevant for the behavior of concrete and concrete infrastructure subjected to different environmental and loading conditions are welcome. Innovative approaches providing insight into complex phenomena, predictive models increasing safety, durability, and sustainability in practical applications, and models leading to new design concepts in the field of structural engineering science are especially encouraged. Contributions linking different fields of research, e.g., physical chemistry, material science, multiscale and probabilistic mechanics, as well as structural engineering science are also particularly welcome.
Keywords: engineering science, multiscale mechanics, physical chemistry, probabilistic mechanics, reinforced concrete structures, cementitious materials
Materials made up either of porous fibrous networks or of matrices reinforced by fibres and appearing, at the macroscopic scale, as layers – such as paper sheets, nanopaper films or various types of epoxy-resin-based composites – are typically lightweight, strong, and often renewable as well as biodegradable; characteristics which, in times of heightened environmental awareness and distress, offer them a competitive edge against other engineering materials. Accordingly, new (often layered) structures incorporating those types of materials – such as paper- or nanopaper-based transistors, sensors, or batteries – are emerging and reshaping our technological landscape. Yet, despite the many advances, the mechanical and physical behaviour of such materials and structures is still poorly understood. This has several implications, a major one being that research and development activities on related products are still very much experiment-, trial-and-error-based, so that innovation potential remains largely untapped. Moreover, when innovation does take place, there is still considerable uncertainty surrounding the long-term performance of such products over their entire life cycle. This is true, even for century-old materials and structures such as paper sheets or paperboards, for which there is still huge room for improvement in terms of their controlled production and use. Cutting-edge theoretical, computational, and experimental research has the capability to address these challenges and to lead us to new generations of products and applications based on such layer-like, fibrous materials and structures based thereon.
This symposium is a forum for scientists and engineers working in the field of mechanics and physics of layer-like, fibrous materials and structures based thereon. The submitted contributions should address recent theoretical, computational, and/ or experimental advances.
Topics of interest include: • Experimental determinations on hierarchical, multiscale organization (e.g., high-resolution images of cellulose fibrils) • Experimental determinations on mechanical response (e.g., time-independent and -dependent elasticity or elastoplasticity), • Experimental determinations on physical response (e.g., thermal conductivity, electrical conductivity, dielectric function, or water vapour diffusivity), • Experimental determinations on coupled physical response (e.g., piezoelectrical), • Theoretical modelling and computational implementation ...
Keywords: Fibrous Materials, Layer-like, Mechanics, Physics, Structures
Deep learning, i.e., using deep neural networks for regression and classification, has been very successful in many different contexts in science and engineering. These include image analysis, natural language understanding, game intelligence, and protein folding. As deep neural networks are universal function approximators, employing them as ansatz spaces for solutions of ordinary and partial differential equations is natural, paving the way for their successful use in scientific computing. An incomplete list of examples where deep learning is used for the numerical solutions of differential equations includes the solution of high-dimensional linear and semi-linear parabolic partial differential equations and many-query problems such as those arising in uncertainty quantification (UQ), PDE constrained optimization, and (Bayesian) inverse problems.
Another interesting use of deep learning is data-driven computing, that is to combine available data with existing physical laws to more accurately predict events as diverse as the spread of pandemics, the weather forecast, the behavior of a vehicle, a structure, or a drug, among others. Deep learning algorithms can be used to solve all these problems. Moreover, they have been the best performers in the last decade and are being massively adopted by various industries and institutions.
Previous problems can be recast as parametric partial differential equations, and the use of deep neural networks in their solution is explored for elliptic and parabolic PDEs, transport PDEs, and hyperbolic and related PDEs, and as operator learning frameworks. All the aforementioned methods are of the supervised learning type, i.e., the underlying deep neural networks have to be trained on data, either available from measurements or generated by numerical simulations.
Another interesting application of deep learning algorithms is the so-called Physics Informed Neural Networks (PINNs) which collocate the PDE residual on training points of the approximating deep neural network. PINNs have been successfully applied to simulate various forward and inverse PDE problems.
This mini-symposium accepts contributions that employ deep learning techniques to predict the result of physical systems often governed by ordinary and/or partial differential equations and are possibly complemented with measured data. We encourage works ranging from the mathematical analysis of these methods to their industrial applications.
Keywords: Computing, Deep Learning, Partial Differential Equations
The use of numerical methods for the approximate solution of nonlinear partial differential equations (PDEs) is fundamental to modern science and engineering. There has been an increased interest in developing higher-order methods and reduced-order models that are as robust as low-order methods typically used in industry. For example, in the context of fluid mechanics problems, much effort has been dedicated to constructing provably-stable methods over the last decade (e.g., schemes which are entropy stable or invariant-domain preserving). In this minisymposium, the focus is on the mathematics that enable the use of discretization techniques (such as higher-order methods, adaptive methods, or reduced-order models) which preserve key properties of nonlinear PDEs.
Keywords: discontinuous Galerkin methods, finite element methods, finite-difference methods, high-order methods, mesh adaptation, model reduction, nonlinear stability
Simulation of the wave propagation phenomena in realistic, complex heterogeneous and engineered, scenarios ranging from industrial facilities to civil structures or biological environments, involves challenges not only from a mathematical modelling point of view, but also from the point of view of the design of accurate and reliable numerical methods. These should be capable of dealing with coupled phenomena, heterogeneous media, nonlinear material responses, or multiscale phenomena.
This mini-symposium is focused on the novel advances of numerical and algorithmic methodologies, which introduce new perspectives in the approximation of classical wave phenomena problems (e.g. exterior problems, fluid-structure interaction systems or aeroacoustics), but also on the design of new numerical procedures to recent challenges in sophisticated mechanical and vibroacoustic systems such as (but not limited to) composites, porous materials, noise barriers, sonic crystals, metamaterials, and underwater environments. Those contributions merging classical numerical techniques with other data-driven procedures are very welcome.
Keywords: Complex media, Numerical methods, Wave propagation
The European Commission has high expectations for the development of a hydrogen based economy in the coming years.
In 2022, hydrogen accounted for less than 2% of Europe’s energy consumption. But the European Commission has proposed to produce 10 million tonnes of renewable hydrogen and to import 10 million tonnes by 2030 [1]. This objective requires efficient and vast storage systems to allow this hydrogen production and to control the energy recovery.
This minisymposium is oriented to the numerical and experimental analysis of the different massive underground hydrogen storage systems (caverns, old depleted oil/gas reservoirs, aquifers, etc.). The analysis from a computational mechanics point of view of crucial aspects like gas properties, mixing processes, gas purity and chemical reactivity, repository stability and capacity, operability and retrieval efficiency, and induced seismicity are the main topics of this minisymposium. Computational mechanics, solid mechanics, structural mechanics, geotechnical stability, computational fluid dynamics, porous media flow, reactive transport, strategic storage operation, optimization, economic viability and other related aspects regarding computational mechanics for underground hydrogen storage are welcome. Experimental analysis and field observations will be also considered for validating numerical models and characterization of properties in hydrogen storage facilities.
Contributions regarding this recent challenging topic are welcome in this minisymposium.
REFERENCES
[1] “The EU strategy on hydrogen“ European Commission, July 2020, (https://ec.europa.eu/commission/presscorner/detail/en/fs_20_1296)
Keywords: Computational Mechanics, Hydrogen storage, Underground storage
Composite lightweight structures are used in a wide range of applications in mechanical, marine, and aerospace engineering. They have several advantages, such as exhibiting high buckling loads and being lightweight. However, for some of these structures, compressive stresses might lead to unstable post-buckling behaviours influenced by an elevated imperfection sensitivity and consequent dangerous load capacity reductions.
On the other hand, the geometrically nonlinear behaviour of lightweight composite structures does not necessarily represent an adverse structural response. In fact, for some applications where a structural shape change is required, this behaviour can be exploited to achieve complex shape reconfigurations and realise structures that can be quickly deployed.
In all those cases, using methods to evaluate the geometrically nonlinear response of lightweight composite structures with high accuracy and short times is crucial.
In light of these premises, this mini-symposium aims to bring together scientists worldwide working on advanced methods for tracing and optimising the geometrically nonlinear response of lightweight composite structures. Therefore, contributions may involve the following:
• Enhanced computational methods to evaluate the stability behaviour of lightweight composite structures as path-following methods or reduced order models.
• Novel computational methods to study the structural stability considering other physical phenomena, such as inelastic deformation, multiphysics systems (e.g. magneto-electro-thermo-mechanical problems, fluid-structure interactions).
• Multi-level and multi-scale analysis of geometrically nonlinear composite structures.
• Discretisation methods as: strong formulations (i.e., collocation and differential quadrature method, inverse differential quadrature method), weak formulations (i.e., finite element method, boundary finite element method, Galerkin methods, isogeometric analysis).
• Optimisation algorithms to solve convex and non-convex problems with single or multi objectives for composite structures dominated by a geometrically nonlinear behaviour.
Keywords: geometrically nonlinear analysis, numerical methods, structural stability, composite structures, Numerical optimization techniques
Surrogate models have become the de facto method of approaching engineering problems for which simulation or experimental data is very expensive to obtain. Such models are typically designed to require the smallest possible number of data points to obtain a useable model, and can then be used to investigate a multitude of scenarios. Two types of problem are particularly difficult to approach using classical techniques – optimization with high numbers of variables, and yield estimation and optimization.
While both these areas have seen widespread research over the past decade, the fields of microwave and antenna engineering have long been focused mostly on rigorous mathematical surrogate models, where the aim is to fit a model as accurately as possible to a set of data. Recently however, approximate models, and models incorporating random behaviour, have started to appear in the field. Two examples are the design of a very large antenna array [1], and yield estimation of microwave filters with up to 70 randomly varying dimensions [2]. Both these examples made extensive use of surrogate models, to achieve solutions not otherwise possible. Currently, the level of activity in this field of microwave and antenna engineering has risen dramatically, with a number of groups across the world currently active in this field.
Microwave and antenna problems offer very specific problem types, (even in electronic engineering, this is a unique field) due to the underlying fundamental network and electromagnetic theory, and therefore the field calls for very specific models, and very specific implementation of models, as the aim in all cases is to exploit the physics as much as possible to reduce the required data sets. It therefore differs substantially from other engineering fields, and have become a sub-type of its own.
The aim of this mini-symposium is to create a forum to discuss the newest developments. The work is of very high interest to industry, as real-world design problems typically fall in these categories. The mini-symposium will consist of one session, with a keynote talk, and 5 other presentations. It will cover various aspects regarding state-of-the-art techniques in the use of surrogate models of varying types, in optimization and yield estimation of microwave and antenna structures.
Keywords: Optimization, Surrogate Modelling, Yield
Multiscale multiphysics problems are ubiquitous in various science and engineering applications such as hydrology, environmental sciences, petroleum engineering, and biomedical engineering. Mathematical modeling and numerical simulations for these problems are of great importance and have attracted much attention of mathematicians and computational scientists to develop advanced numerical methods capable of efficiently and accurately capturing the different physical processes and the wide range of scales in both space and time inherent in the simulations. The goal of this minisymposium is to present recent progress in the development of decoupling algorithms with such features, such as domain decomposition and time-splitting methods, for flow and transport in (fractured) porous media, surface-subsurface flow coupling, fluid-structure interaction, fluid-poroelastic structure interaction, and related models. Topics of interest include, but are not limited to, stability and accuracy of the numerical solutions, convergence and preconditioning of the iterative algorithms, a posteriori error estimation, and adaptivity.
Keywords: Coupled Systems, Domain Decomposition, Fluid Dynamics, multi-physics, multi-scale, Numerical methods, Time-Splitting
The coupling of numerical simulations for fluid dynamics with machine learning techniques, such as neural networks, reinforcement learning, autoencoders, etc. is emerging as a powerful approach for enhancing the accuracy and efficiency of computational fluid dynamics (CFD) simulations [1]. This integration leverages the strengths of both disciplines to tackle complex fluid flow problems that are difficult to solve using traditional computational methods alone.
Neural networks (and variants such as CNN, LSTM, GANs, etc.), with their ability to learn complex patterns and relationships, have been successfully employed to approximate fluid flows, model turbulence and accelerate CFD applications (e.g., [2]). Reinforcement learning, on the other hand, offers a unique framework for optimizing control strategies in fluid dynamics. Through trial and error interactions with simulated environments, reinforcement learning agents can learn optimal policies that govern fluid flow behaviors, leading to improved turbulence models [3] or new flow control paradigms [4]. This dynamic approach has the potential to revolutionize engineering design processes and enhance the performance of flow applications. Finally, autoencoders [5,6] have been employed in the context of fluid dynamics to extract meaningful low-dimensional representations of high-dimensional flow data. By encoding and decoding fluid flow variables, autoencoders can effectively compress and reconstruct flow fields, facilitating efficient data analysis, dimensionality reduction, and anomaly detection. Such techniques enable rapid exploration of large datasets, identification of flow features, and the development of reduced-order models for real-time simulations.
This mini-symposium highlights the growing importance and potential of coupling numerical simulations for fluid dynamics with machine learning techniques. The synergy between these fields promises advancements in fluid flow understanding, optimization of flow control systems, and the development of more efficient and accurate computational methods. This mini-symposium will offer a venue for sharing leading edge research into the technical, methodological and theoretical aspects of coupling machine learning and CFD.
Keywords: high order methods, Machine Learning, Neural Networks, reduced order models, Reinforcement Learning
The numerical simulation and prediction of flows around wind turbines and wind farms are difficult. This difficulty is caused, on the one hand, by the rotating blades, which leads to unsteady flow features and dynamic effects, and to the high Reynolds number turbulent atmospheric flow in which these devices operate. Both the blade movement and the high turbulence lead to costly numerical simulations. Large eddy simulations (LES) of farms are common in academia [1] and have helped to study energy extraction or wake interactions. The choice of simulation approach depends on the desired level of accuracy, with large eddy simulation being the preferred method for capturing the most detailed information. However, when conducting parameter studies, (unsteady) RANS is required, while more analytical approaches are pursued to increase the physical understanding of the flow.
Various simplified techniques exist to mimic the effect of rotating blades (e.g., actuator discs or actuator lines [2]), which are used to reduce cost. These techniques impose sectional forces on the fluid (calculated/measured for 2D airfoils) to mimic the effect of individual blades (actuator line) or, if azimuthally averaged, mimic the entire rotor (actuator disc). These techniques can efficiently replicate the effect of the energy extracted from the flow and flow swirls arising from rotors in the far field but face limitations when detailed flow features need to be considered for fine design or aeroacoustic predictions since some effects are neglected (e.g. trailing edges, 3D effects on lift/drag, dynamic stall).
This colloquium will unite wind turbine and wind farm simulation experts to explore cutting-edge simulation strategies. In addition, we will discuss how the valuable insights gained from these simulations can be leveraged to enhance low-order and analytical models, driving meaningful advancements in our understanding of the fluid mechanics of wind turbines and wind farms. With active wind energy research communities across Europe, North and South America, and China, we anticipate a diverse and engaging group of participants worldwide to join us in this colloquium.
[1] Breton, et al, A survey of modelling methods for high-fidelity wind farm simulations using large eddy simulation. Philosophical Transactions of The Royal Society A Mathematical Physical and Engineering Sciences.
[2] Sorensen et al.. Numerical modeling of wind turbine wakes, J. Fluids Eng,
Keywords: actuator discs, actuator lines, Computational Fluid Dynamics, sliding meshes, wind farms, immersed boundary method, Wave propagation, Wind turbines
Since their introduction 30 years ago [1], the FFT-based homogenisation method has established itself as a particularly useful numerical tool in the field of micromechanics and the prediction of effective properties.
The main advantages of this full-field calculation method are its ease of implementation, its simplicity of use (the description of the microstructure under study consists of an image of it) and its numerical efficiency. The aim of this symposium is to report on (but not limited to):
• Latest advances in the method itself, whether in terms of new, more efficient algorithms or the extension of the method to new materials and material behaviors,
• applications of the method to problems involving crystalline materials and composite materials, including damage, dynamic deformation, thermal effects etc.,
• coupling with other methods such as FE-FFT multiscale computations, machine learning techniques, reduced order models, etc.,
• synergy of the method with advanced characterization techniques (CT, EBSD, 3DXRD, etc.) ,
• efficient numerical implementation of the method (high-performance computing, including GPUs, etc.),
and above all to promote discussions between scientists involved or interested in the method.
[1] H. Moulinec and P. Suquet, “A fast numerical method for computing the linear and nonlinear properties of compositesâ€, C. R. Acad. Sc. Paris, II, 318, pp. 1417-1423, 1994
Keywords: Full-field Method, Homogenization, Scientific Computing
The objective of this symposium is to discuss new advances in numerical methods for linear and non-linear time-dependent and time-independent partial differential equations used in mechanics. Topics of interest include, but are not limited to: new space and time discretization methods; high-order accurate methods with conforming and unfitted meshes including finite, spectral, isogeometric elements, finite difference methods, fictitious domain methods, meshless methods, and others; special treatment of boundary and interface conditions on irregular geometry; new time-integration methods; adaptive methods and space and time error estimators; comparison of accuracy of new and existing numerical methods; application of new numerical methods to engineering problems; and others.
Keywords: conforming and unfitted meshes, PDEs, Numerical methods
Many science and engineering applications feature differential equations with terms that are partially unknown. For example, in neural ODEs, one tries to train a neural network to model a dynamical system. Similarly, in (non-intrusive) reduced order models (NIROMs), operators are derived in such a way to closely match (projected) snapshot data. Additionally, recent years have seen a big push to replace empirical and often ad hoc subgrid-scale process models (e.g., closure terms in turbulence modelling, constitutive models in structural mechanics, physics process representations in climate models) in high fidelity modelling toolchains with machine learned representations. Many existing approaches try to learn these terms in an ‘offline’ fashion, e.g. through supervised learning, and then substituted back into the differential equation in order to give predictions in an ‘online’ setting [1]. An alternative to this “operator-fitting†approach is known as “embedded model learningâ€, “solver-in-the-loopâ€, or “trajectory-fitting†[2,3]. In this approach, one learns a model in such a way that, upon embedding in the solver, it results in accurate predictions of the solution trajectory. This has the promise to lead to more stable models, but comes at the price of increased computational costs associated with differentiating through the entire differential equation solver (e.g. by using fully differentiable solvers or adjoints). Such fully differentiable solvers are actively being developed in the Scientific Machine Learning community [4].
In this minisymposium we bring together researchers working on learning models for various science and engineering applications (e.g., computational fluid mechanics, structural modelling, climate modelling), either with operator fitting or with embedded learning. We welcome contributions on the topic of learning turbulence models, reduced order models, and other types of ‘closure’ models that appear in partial differential equations.
[1] A. Beck and M. Kurz, A perspective on machine learning methods in turbulence modeling, GAMM-Mitteilungen, vol. 44, 2021.
[2] B. List, L.-W. Chen, N. Thuerey, Learned Turbulence Modelling with Differentiable Fluid Solvers, ArXiv: 2202.06988, 2022.
[3] J. F. MacArt, J. Sirignano, J. B. Freund, Embedded training of neural-network subgrid-scale turbulence models, Phys. Rev. Fluids 6, 050502, 2021.
[4] C. Rackauckas et al. Universal Differential Equations for Scientific Machine Learning, arXiv:2001.
Keywords: closure model , computational fluid dynamics, Dynamical systems, model discovery, neural ODE, Reduced Order Models, Scientific Machine Learning, turbulence modeling
The primary objective of this minisymposium is to explore recent advancements in polytopal methods and their application in the context of coupled problems. Although not limited to these topics, the symposium will primarily focus on coupled poromechanics modeling geological materials and biological tissues, non-isothermal flows and deformation processes, bulk-surface couplings, fracture and contact mechanics, and fluid-structure interaction problems. Moreover, it will also encompass discussions on the utilization of polytopal methods in other biological or industrial applications.
Keywords: Interface problems, Multiphysics problems, Polygonal/polyhedral grids
For a large class of partial differential equations (PDEs) arising in engineering science, e.g., singular perturbation, Helmholtz, nearly incompressible elastostatics problems, and coupled problems, the corresponding finite element (FE) discretizations suffer from a loss of stability. Certain FE discretizations
provide relief from the loss of stability by constructing FE approximations of the corresponding weak formulation to the PDE that satisfies the discrete inf-sup condition. Furthermore, recent work in data-driven methods can also be incorporated to develop stable FE schemes. To create discussions around
these issues, we invite contributions with a focus on the following:
• Stable discretization schemes for stationary and transient linear and non-linear problems,
• Residual minimization techniques such as the least squares FE method and the discontinuous Petrov-
Galerkin method and their analysis.
• A posteriori error analyses and estimates for stable discretization schemes leading to error indicators and adaptive mesh refinement strategies.
• Development of new FE basis functions leading to stable schemes.
• Implementational aspects and issues surrounding stable FE methods in modeling physical phenomena.
• Development of solvers for the linear system of equations resulting from stable discretization schemes.
• Application of stable FE methods to large-scale, complex problems in engineering science.
• Integration of data-driven techniques with FE technology.
Keywords: Adaptivity, Computational Methods, Data-driven methods, Stability
This mini-symposium aims to showcase and discuss recent advancements in mesh reduction methods, specifically focusing on boundary element methods (BEM), mesh-free methods, and particle-based methods such as smoothed particle hydrodynamics (SPH) and discrete element methods (DEM). The primary objective is to address complex problems associated with multi-scale phenomena, multi-interaction scenarios, and the interaction between solids and fluids.
The mini-symposium will cover a wide range of topics, including but not limited to hybrid approaches that combine BEM, mesh-free methods, particle-based methods, and molecular dynamics for multi-scale simulations. Sophisticated frameworks based on BEM for studies of the micro-mechanics of heterogeneous materials. Multi-scale couplings between BEM and other approximations such as molecular dynamics (MD) or coarse-grain models. Development and enhancement of SPH and DEM for modelling solid-fluid and fluid-fluid interactions, with a focus on high-strain morphology changes, soft materials and complex fluids. Mesh reduction techniques specifically designed for simulating multi-scale phenomena in coupled solid-fluid systems using molecular dynamics . Modelling multi-interaction scenarios involving particles, interfaces, and fluid-solid boundaries, particularly addressing fluid-fluid capillary interfaces and surface deformations. Parallel computing and optimisation techniques for accelerating simulations in mesh reduction methods, like in the BEM open source software. Also, in in-house developments on SPH and DEM, enabling efficient 3D modelling of different phenomena. Applications in various fields including, additive manufacturing, granular materials, bio-mechanics, microfluidics, and fluid dynamics.
We invite researchers and practitioners actively working on these mentioned applications and methods to present their latest research findings at this mini-symposium. The aims are to provide a platform for meaningful discussions, exchange of ideas, and potential collaborations among experts in the field. Also, address challenges, explore novel methodologies for future developments.
Keywords: boundary elements, Mesh reduction methods, mesh-free, particle-based
Cardiovascular (CV) modeling is a critical interdisciplinary research field that enhances the understanding of physiological and flow-field characteristics, particularly in patient-specific contexts [1, 2]. However, the high computational demand associated with these simulations hinders their applicability in clinical real-time scenarios. To address this challenge, Model Order Reduction (MOR) techniques have emerged as a promising approach to obtaining precise results within a competitive time frame by approximating the full-order model, while retaining the essential dynamics of the system [3].
This minisymposium aims to explore recent advancements in MOR techniques
specifically designed for patient-specific simulations of CV flows, highlighting their
applications and potential clinical implications. Reduced simulations face various challenges, such as complex geometries, multiphysics and multiscale modeling, accurate boundary conditions, blood rheology, and other factors that necessitate the integration of numerical methods with sophisticated MOR techniques. Therefore, we invite discussions on a wide range of topics, including but not limited to advanced discretizations and numerical methods, surrogate modeling, reduced basis methods, proper orthogonal decomposition, non-linear MOR techniques, data assimilation, inverse problems, optimal flow control, and data-driven approaches. These topics should be explored in the context of their applications to cardiac electromechanics, blood fluid dynamics, and cardiovascular fluid-structure interaction (FSI) problems. By fostering collaboration and sharing novel insights, we aim to collectively advance the state-of-the-art in both MOR techniques and their application in cardiovascular research.
REFERENCES:
[1] Africa PC et al., “lifeX-fiber: an open tool for myofibers generation in cardiac computational modelsâ€, BMC Bioinformatics, Vol. 24 (1), pp:143, (2023).
[2] Rathore S et. al., “Numerical computation of blood flow for a patient-specific
hemodialysis shunt modelâ€, Jpn J Ind Appl Math, Vol. 38 (3), pp: 903-919, (2021).
[3] Girfoglio, M et al., “Non-intrusive data-driven ROM framework for hemodynamics
problemsâ€, Acta Mechanica Sinica, Vol. 37, pp: 1183-1191, (2021).
Keywords: cardiovascular modeling, data-driven approaches, model order reduction
Computational electromagnetism plays important roles to design the electric facilities and to assess the influence of electromagnetic fields; for example, transformer, motor, and hyperthermic potentiation. However, as the ability of computers progresses and the demand of more precise approximation, the number of Degrees Of Freedom (DOF) of computational models derived from conventional discretizations becomes larger even in case of adaptive mesh refinements. In this minisymposium, we discuss on the accuracy and efficiency of novel numerical schemes of electromagnetic field problems from both mathematical and engineering points of view. We have some possibilities of novel numerical schemes, which are discussed in this minisymposium. First, we discuss efficient numerical schemes to compute directly such large scale computational models within the required computational costs, for example, based on Domain Decomposition Methods (DDM) with parallel computations; see Takei-Ogino-Sugimoto, IEEE Trans. Magn., 54 (2018). Second, we discuss efficient numerical schemes reducing the problem size without deteriorating accuracy, which can be, for example, realized by Model Order Reduction (MOR) methods; see Sato-Igarashi, IEEE Trans. Magn., 52 (2016). Third, we welcome to discuss novel schemes based on other strategies not mentioned above.
Keywords: numerical analysis, computational electromagnetism, novel numerical scheme
Progress in multidisciplinary analysis and optimisation capabilities is key to meet today’s challenges in the design of complex industrial products such as aircraft, engines, ground/water transport vehicles or components thereof. High numerical resolution requires advances in algorithmic and parallel efficiency of HPC computing frameworks, e.g. for coupled CFD-CSM evaluations. Gradient-based solution and optimisation algorithms need access to exact and efficiently computed sensitivity derivatives. Moreover, robust design plays a vital role, e.g. in conjunction with uncertainties in parameters or operating conditions. Usability is a success factor in the context of multiphysics and high fidelity for complex industrial applications.
This minisymposium aims to review the state of the art in integrating the disciplinary advances into large scale, complex, and/or coupled industrial scenarios.
In particular contributions are invited on
- advanced frameworks and algorithms for scalable MDAO
- geometry modelling coupled to simulation and optimisation
- robust design and uncertainty treatment
- data-driven techniques for MDAO such as surrogate and reduced-order models
Keywords: coupled adjoints, Data-driven methods, MDAO frameworks, multidisciplinary analysis and optimisation (MDAO), Uncertainty quantification (UQ)
In recent years, there has been a growing focus on the design of (arbitrarily) high-order methods for hyperbolic differential problems. In fact, such methods are well known to better approximate the solution and to provide higher accuracy for coarser discretizations, often resulting in strong computational advantages.
However, increasing the order of accuracy brings along its own set of challenges. Simulations using these methods often encounter difficulties such as instabilities, blows-ups, violations of physical properties, as well as technical issues when applied to complex models. Moreover, the computational costs of arbitrarily high order methods can easily increase because of time step restrictions or the number of stages involved. Hence, current research focuses on building (arbitrarily) high order schemes that are also (nonlinearly) stable, robust and efficient.
Nowadays, as high order methods are becoming well established and mature, the need to solve these outstanding issues is more and more pressing. With this minisymposium we try to answer the following question: how to develop effective high order methods for hyperbolic problems?
In this minisymposium, we envision the following topics to be addressed:
arbitrary high-order methods for hyperbolic PDEs with any discretization technique (finite difference, finite volume, finite elements, discontinuous Galerkin, summation by parts, etc.),
structure preserving methods (positivity, entropy, energy, divergence-free, etc.),
shock capturing methods,
high-order schemes for multi-dimensional and multi-physics hyperbolic systems,
asymptotic preserving methods,
efficient implementation of high-order hyperbolic solvers,
adaptive refinement strategies,
parallel implementations, parallel algorithms and GPU computing for high-order methods,
error estimation, stability analysis, and convergence properties of high-order schemes,
applications of high-order methods in engineering and scientific problems.
The objective of this minisymposium is to bring together researchers and practitioners from different disciplines to discuss novel techniques, applications, and theoretical developments in the field, encouraging discussion, knowledge sharing and collaborations among researchers.
Keywords: Arbitrarily High Order, Balance Laws, Efficiency, Hyperbolic PDEs, Stability, Structure Preservation
This mini symposium deals with the development of safe and reliable intelligent systems through the combination of physics-based and data-driven approaches. Several main challenges of non-deterministic approaches are related to the limitations of the two main modelling strategies available: physics-based (white-box) or data-driven (black-box). Black-box models are fast and able to capture very complex behaviors but they are incapable of accurate forecasting, limiting their usage in the context of large uncertainties. The Physics-based models can have great accuracy on vast domains of validity but their computation burden often makes them unusable for non-deterministic approaches. A current trend in the uncertainty quantification community is to explore ways of integrating both approaches in highly efficient methods, sometimes referred to as grey-box models. These approaches can be tailored for a large variety of applications depending on data availability, black-box architecture, type of uncertainty, industrial application, or even the stochastic problem to be solved (e.g. design optimization, sensitivity analysis, reliability analysis or process monitoring).
Therefore, this symposium is aimed at gathering contributions that discuss new theoretical developments and advanced applications related to the efficient combination of computationally intensive numerical simulation codes with efficient surrogate models and/or data-driven black-box approaches, as well as grey-box and other hybrid combinations of machine learning and numerical modelling.
Keywords: Machine learning, Physics-Informed Machine Learning, Surrogate models, Uncertainty quantification (UQ)
Forecasting the rapid changes in the Earth’s climate is one of the biggest challenges of our times. Fast and accurate weather/climate/ocean dynamics forecasts need state-of-the-art numerical and computational methodologies due to the high computational complexity of solving systems described by partial differential equations. This minisymposium is about the development and application of computational approaches in the field of geophysical fluid dynamics [1, 2, 3, 4, 5].
The focus is on efficient numerical techniques, including high-fidelity finite elements, finite volumes, discontinuous Galerkin and spectral elements methods as well as reduced order models to deal with complex phenomena arising in ocean and atmospheric flows. Examples of these phenomena are turbulence, compressibility and multi-phase interfaces. Efficient and accurate numerical methods for real world applications have undergone fast development during the last decade and have become a new frontier in scientific computing. This minisymposium will discuss the most recent development and identify new directions and perspectives.
REFERENCES
1. M. Girfoglio, A.Quaini, and G. Rozza, Validation of an OpenFOAM-based solver for the Euler equations with benchmarks for mesoscale atmospheric modeling, AIP Advances, 13, p. 055024, 2023.
2. M. Girfoglio, A. Quaini and G. Rozza, A novel Large Eddy Simulation model for the Quasi-Geostrophic Equations in a Finite Volume setting, Journal of Computational and Applied Mathematics, 418, p. 114656. 2023.
3. M. Girfoglio, A. Quaini and G. Rozza, A linear filter regularization for POD-based reduced order models of the quasi-geostrophic equations, Comptes Rendus Mècanique, 351, p. 1-21, 2023.
4. N. Clinco, M. Girfoglio, A. Quaini and G. Rozza, Filter stabilization for the mildly compressible Euler equations with application to atmosphere dynamics simulations, https://arxiv.org/abs/2305.12978, 2023.
5. M. Girfoglio, A. Quaini and G. Rozza, GEA: a new finite volume-based open source code for the numerical simulation of atmospheric and ocean flows, https://arxiv.org/abs/2303.10499, 2023.
Keywords: atmospheric flows, geophysical fluid dynamics, ocean flows
Optimization problems are central to many fundamental research areas within applied mathematics and computational mechanics, with significant applications in limit analysis, shape and topology optimization, and material design. On the one hand, limit analysis allows a reliable estimation of the ultimate load-bearing capacity of masonry, reinforced concrete, or steel structures and is also employed in geotechnical problems such as assessing soil slope stability and foundation bearing capacity. On the other hand, boosted by the development of additive manufacturing technologies, which allow the realization of objects with arbitrarily complex geometries at affordable production costs, shape and topology optimization techniques can be exploited to enhance the mechanical performances of materials and structures and lead to innovative advancements in multiple domains. The universality of optimization problems has prompted the emergence of various numerical methods that include conventional approaches, such as gradient-based or gradient-free optimization algorithms, and modern meta-heuristic approaches or data-driven techniques exploiting neural networks and deep learning.
This mini-symposium aims at inspiring the scientific community to tackle the numerous challenges associated with optimization problems in both established and new, rapidly growing fields of computational mechanics. While contributions in all aspects related to the aforementioned fields are invited, some of the featured topics in this MS will include:
• limit analysis of structures under ordinary or extreme actions
• yield design
• microstructure optimization
• homogenization
• material design
• optimal structural design
• shape and topology optimization
• multi-objective optimization with uncertainties.
Exploring and advancing research in these areas will enhance the understanding and further increase the impact of optimization techniques in computational mechanics.
Keywords: Materials Design, Optimal microstructures, Structural optimization, Limit analysis
The study of brain function is an active field of clinical research, yet many questions about the physical processes underlying it are still open. In this regard, computational multiphysics models allow the investigation of the complex interplay between processes such as electrophysiology [1], fluid dynamics [2], and tissue mechanics [3]. Moreover, diagnostic data can be integrated in these models, to study the patient-specific development of pathologies like epilepsy [1] or neurodegenerative diseases [4], among others. This minisymposium aims at gathering experts on the mathematical and numerical modeling of brain function, to discuss recent advancements in both the development of numerical methods and their application to clinical problems.
Keywords: Brain, Computational Fluid Dynamics, Electrophysiology, Neurodegenerative diseases, Numerical methods, PDEs, Poromechanics
Many real world porous media applications are governed by multiphysics models formulated on complex geometries, introducing a range of spatial and temporal scales. For instance, fractured and faulted systems play a significant role in the modeling of geological carbon storage and geothermal energy, which are governed by porous media flow coupled to geomechanics taking into account contact. In addition, strong coupling between the different potentially highly nonlinear and non-smooth subproblems often arise. Both poses challenges to the development of accurate, efficient and robust simulation technology - on the other hand such are required. The focus of this minisymposium is to gather recent advances of simulation technology tailored to multiphysics problems and complex porous media as tailored block-partitioned solvers and preconditioners, tuning of solver parameters, field-scale simulations and related topics.
Keywords: contact mechanics, iterative solvers, multi-physics, Poromechanics, porous media, preconditioners
Large-scale simulations of wave propagation are essential for solving real-world problems representing critical societal challenges in various areas including energy, space, environment, and health. Examples of these current challenges are geothermal energy, helioseismology, CO2 storage monitoring, medical imaging, etc. For several decades, increasingly advanced numerical methods have been developed to efficiently solve large-scale problems with high accuracy and significant progress have been made with the aid of supercomputers equipped with modern architectures for high-performance computing. Nowadays, with the advent of exascale machines, new approaches are being proposed to make the most of the capabilities of supercomputers under energy sobriety constraints.
This mini-symposium will provide an opportunity for engineers, scientists, and applied mathematicians to share their new ideas and approaches they have been recently developing for solving large-scale wave propagation problems as well as their findings on wave problems including, direct, inverse, and eigenvalue problems.
Topics of interest include but are not limited to:
- high-order methods,
- time-stepping schemes,
- non-reflecting boundary conditions,
- reduced order modeling
Results of cross-pollination between pure Science and Engineering are recognized by both communities. Due to long-standing collaboration programs, Engineers are now able to create optimal parts, to optimize existing parts, to precisely predict the failure of a given part and even efficiently solve large combinatorial problems. Theory unification percolated to Engineering, with a remarkable example of this being the merging of the themes of fracture, topology optimization, image segmentation and distance determination in contact [1]. Intense interdisciplinary work with the areas of pure and applied Mathematics, especially (abstract) Algebra and Analysis have greatly improve the scope and effectiveness of the Engineering solutions. Significant advances in Applied/Continuum Mechanics, such as constitutive analysis and formalism, stability analysis, proofs of existence and uniqueness of solutions and classification of problems were achieved with the contributions from pure and applied Analysis. Results from Mathematical and Computational Physics have significantly contribute to recent developments in Computational Mechanics. In addition, Computational Mechanics has greatly evolved with the contributions from Computer Science with the emergence of deep learning as an important tool. In this minisymposium, we welcome the following themes:
• Contributions in the areas of contact mechanics, frictional contact mechanics and cohesive fracture without ad-hoc solutions.
• Established algorithms which can be relevant for Computational Mechanicians
• Computational Physics applications and developments relevant for Mechanicians, including consistent Multiphysics.
• Computational Materials Science developments.
REFERENCES
[1] P.Areias, N.Sukumar, J. Ambrósio “Continuous gap contact formulation based on the screened Poisson equation. Computational Mechanics, 2023. in press
Keywords: Computational Material Science, Interdisciplinary advances, Applied and Computational Mechanics
The evolution in hardware technologies enables scientific computing to reach further aims. Nowadays, the use of massively parallel supercomputers is rather common in the solution of both industrial and academic scale problems. Unfortunately, the literature reveals that the performance of numerous scientific computing applications is often limited to tiny fractions of the system’s peak performance. While the peak performance has been growing 9 and 4.5 times faster than network and main memory bandwidth per decade over the past 30 years, many algorithms employed in scientific computing have very low operational intensity. Numerical simulation codes are increasingly memory-bounded, making processors suffer from serious data starvation. Moreover, distributed memory systems pose an additional bottleneck: network bandwidth and latency. The relative weight of data exchanges grows as parallel applications are distributed among a greater number of processors, making it advantageous to increase the problem size per node. However, hardware accelerators with high-bandwidth memory, which are major contributors to the increased computational power of individual nodes, have smaller memory capacities than traditional processors. Application bottlenecks becoming more severe exacerbates the already difficult job of architects, programmers, and researchers. In this context, this mini-symposium aims to bring together people working on advanced, cutting-edge parallel methods for solving extreme-scale simulations with a special focus on efficiency, portability, and sustainability. Works on domain-specific languages, parallel linear solvers, sparse algebra kernels, eigenvalue problems, parallel-in-time methods, load balancing, adaptive mesh refinement, and memory footprint reduction techniques, among others, are welcomed.
Keywords: Distributed Memory, Extreme-Scale Simulation, Heterogeneous Systems, Linear Solvers, Parallel Computing, Sparse Algebra
The aim of this minisymposium is to present recent numerical methods for solving contact mechanics problems and associated variational inequalities problems. This includes the introduction of new algorithms and associated numerical methods (penality, generalized saddle point, AMR, HPC, immersed boundaries…). Theoretical results on convergence, error estimates and stability are also welcomed. We are also interested in the use of machine learning for model order reduction and surrogate models in general for contact problems, when algorithms do not forget scientific knowledge in contact theory. This concerns hybrid approaches to scientific machine learning.
Keywords: error estimates, numerical methods, reduced order models, scientific machine learning, Contact mechanics
Mixed-dimensional PDEs represent coupled problems involving lower-dimensional manifolds or sub-manifolds of one or many domains. From a mathematical approximation perspective, mixed-dimensional PDEs (MD-PDEs) are connected to reducing the complexity of modeling and computations by utilizing model order reduction techniques. These techniques exploit the heterogeneity of scales present in different components or regions of the overall problem. In simpler terms, MD-PDEs often arise when applying dimensional or topological model order reduction. This modeling approach has been applied to real problems of paramount importance in the sciences, including materials science, geo-sciences, and life sciences. For instance, one successful example of MD-PDEs is the simulation of blood flow and transport in complex vascular networks spanning different spatial scales, from small perforating arteries and arterioles, through capillaries, up to venules [1]. The computational modeling framework needs to account for the specificities of the scales involved with sufficient detail while being efficient and accurate. Other notable examples in the context of biomechanics include the simulation of biomedical devices or the enhancement of advanced imaging techniques such as magnetic resonance elastography. In this mini-symposium, we aim to address the forefront of current research on MD-PDEs, focusing on the mathematical formulation and the approximation of advanced applications in biomechanics.
Keywords: Biomechanics, Coupled Problems, Dimensionality Reduction
Unfitted finite element methods have received increasing attention during the past ten to fifteen years. Their central principle is to embed the domain of interest into a larger computational domain that is easy to mesh. On this computational domain, a finite element type computation is performed. Under the denotation ‘fictitious domain’, ‘embedded domain methods’, or ‘immersed boundary methods’, the central principle has been followed already since the 1960ies. The recent new interest results from innovative and efficient algorithmic developments, from mathematical analysis showing optimal convergence despite the presence of cut elements, the possibility to efficiently link these methods to various types of geometric models including the embedding of domains with a large dimensionality gap, and from many new engineering applications. Many variants of Unfitted Finite Element Methods have been developed, like CutFEM, the Finite Cell Method, Unfitted Discontinuous Galerkin Methods, the Shifted Boundary Method, and Trimmed Isogeometric Analysis, just to name a few.
This mini-symposium will focus on Unfitted Finite Element Methods and their various aspects that make them successful in addressing complex problems, namely: mathematical analysis, a priori and a posteriori error estimation and adaptivity, advanced numerical integration procedures, data structures and parallel scaling of algorithms, integration with CAD models and non-standard geometric representations, and applications. The scope of this mini-symposium is to be as broad as possible in terms of applications, such as, but not limited to: problems in solid mechanics, heat transfer, CFD, fluid/structure interaction, and any other types of domain coupling including mixed-dimensional problems, e.g., coupled surface-bulk problems. Topics will also include computational homogenization techniques and the connection between Immersed Boundary Methods and meta-algorithms, such as the ones used in Uncertainty Quantification, Reduced Order Models, Machine Learning and Artificial Intelligence, Direct and Inverse Problems, and Topology Optimization, just to name a few.
Keywords: Embedded Domain Methods, Fictitious Domain Methods, Trimmed Isogeometric Analysis, Immersed Boundary Methods
Nonlinear structural mechanics problems are encountered in many areas of engineering science. Their understanding, analysis, numerical modelling, and computational implementation are challenging tasks, as the computational strategy strongly depends on the type of nonlinearity involved and the specific nature of the problem being studied.
The proposed MS will focus on recent scientific developments in nonlinear computational dynamics, covering all aspects of numerical modelling of nonlinear phenomena, including numerical methods and computational algorithms involving industrial problematics. The MS will also discuss applications of numerical modelling to nonlinear phenomena, with a focus on new emerging methodologies.
The MS will focus on the following scientific topics:
• Solid computational mechanics
• Detuning and mistuning in turbomachinery
• Fluid-structure interactions and coupled problems.
• Nonlinearities in material constitutive equations
• Geometric nonlinearities
• Localized friction and contact nonlinearities
• Nonlinear reduced-order models
• Uncertainty quantification in nonlinear computational models
• Machine learning and probabilistic learning techniques for stochastic inverse problems
Keywords: Computational mechanics, Data science, Nonlinear structural mechanics, Uncertainty quantification (UQ)
The ubiquity of new technologies in our daily lives is steadily increasing, transforming our societies into interconnected and pervasive systems. In this context, numerical modeling emerges as a powerful instrument to comprehend real-world problems, providing qualitative and quantitative insights into their spatio-temporal evolution. By synergizing numerical modeling with information technology, novel strategies for sustainable development and innovation systems can be devised, thus optimizing industrial workflows and resource utilization.
The fruitful combination of numerical modeling, information technology, and data analysis reveals to be an impactful tool to address critical challenges in various domains, such as precision agriculture, energy management, predictive monitoring, and transportation.
This mini-symposium aims to present and discuss exemplary experiences within the aforementioned areas, emphasizing mathematical modeling and efficient computational methods for simulating and analyzing complex real-world problems. Additionally, the focus will be on methodologies related to sensor data analysis and automated decision-making processes. Particular emphasis will be devoted to the tangible impact on sustainability resulting from the application of such cutting-edge approaches.
Indicative topics for presentations and discussions include, but are not limited to:
- Data-driven modeling and analysis;
- Numerical modeling of vegetation problems;
- Simulation of material deterioration dynamics;
- Machine Learning applications;
- IoT for sustainable practices;
- Smart Agriculture;
- Multi-Criteria Decision Aiding;
- Green and digital transformation through IoT;
- Recommender Systems for sustainable choices;
- Green computing and sustainability in ICT;
- Efficient solutions for sustainable data centers;
- Sustainable supply chain management.
By fostering interdisciplinary discussions and knowledge exchange, this mini-symposium aspires to advance the field of sustainable development and inspire further research in the integration of mathematical modeling, computational methods, and information technologies for a more sustainable future.
Keywords: Data Analysis, Mathematical Modeling, Sustainable Design
The main aim of this Symposium is to foster discussion and collaboration among environmental and industrial scientists and professionals on cutting-edge computational modelling technologies for AI-based predictive models. In particular, it will cover fundamental research areas on reduced-order methods, machine learning algorithms, sensitivity and uncertainty methods, optimisation and data assimilation methods for single/multi-physics problems. Contributions are sought on, but not limited to, the following topics:
(a) Computational structural/fluid/radiation dynamics;
(b) Predictive methods for sensitivity and uncertainty quantification and data assimilation;
(c) Coupling models for multi-physics problems;
(d) Parallel numerical algorithms for large scale simulations.
Keywords: Artificial Intelligence, uncertainty quantification
This series of talks will discuss recent developments for accelerated high-performance simulations in solid mechanics, fluid dynamics, and coupled problems. We aim at bringing together researchers and practitioners from academia and industry to discuss optimal code design and implementations with a special focus on multi-core graphics processing unit (GPU) and central processing unit (CPU) accelerated simulations. Applications of artificial intelligence (AI) and reduced order models (ROMs) to enhance computational efficiency are also welcomed. Examples of applications to be discussed include (but are not limited to): rock mechanics and civil engineering, climate modelling, subsurface and energy engineering.
Keywords: coupled problem, fluid dynamics, Artificial intelligence (AI), civil engineering, climate modelling, CPU, energy, GPU, rock mechanics, ROMs, solid mechanics, subsurface
A variety of discretization techniques have been proposed for describing evolving, coupled interfacial phenomena using finite elements. These methods, including Generalized/eXtended FEM, Interface- and Discontinuity-Enriched FEM, cutFEM, Shifted Boundary Methods, and Conformal Decomposition FEM have been shown to be effective at capturing static and evolving interfacial discontinuities.
As these methods have matured, attention has been given to the conditioning, robustness, and performance of the methods and the associated coupling algorithms. Fully transient phenomena are being addressed with dynamic discretizations. Complex, coupled interfacial phenomena are being addressed with a variety of monolithic and iterative coupled strategies.
Software designs are being proposed for reducing the complexity and code development costs for implementing these advanced discretization techniques. Mixtures of commercial, open-source, and research codes are being developed and adapted to provide the end-user with cutting edge simulation and modeling capabilities not available previously.
This mini-symposium aims to bring together engineers, mathematicians, computer scientists, and national laboratory and industrial researchers to discuss and exchange ideas on new developments, applications, and progress in the discretization methods and algorithms for transient, coupled interfacial phenomena. While contributions to all aspects of these methods and their implementation are invited, topics of particular interest include:
• verification and validation; accuracy, computational efficiency, convergence, and stability of FEM discretizations and coupling algorithms for moving interfaces.
• new developments for immersed boundary or fictitious domain problems, flow and fluid-structure interaction, among others.
• applications to industrial problems exhibiting multiscale phenomena, localized non-linearities such as fracture, damage, or contact, and non-linear material behavior.
• acceleration techniques for coupling algorithms.
Keywords: interfaces, coupled problem, Enriched finite element methods
Neural networks and learning algorithms have garnered substantial attention among researchers engaged in the numerical approximation of partial differential equations (PDEs). Notably, there are well-established methodologies for employing these tools in solving PDEs [1]. Additionally, a significant overlap exists between the machine learning and computational modeling communities in the realm of data-driven reduced order models [2]. However, this field of research remains dynamic, with numerous novel concepts emerging. These encompass the utilization of learning algorithms to expedite the resolution of linear systems, the creation of adaptive computational meshes, the discovery of optimal approximation spaces, and even the acquisition of insight into underlying operator structures.
This mini-symposium aims to provide a comprehensive platform for researchers to delve into the dynamic synergy between numerical methods for PDEs and emerging techniques in deep learning. This session will spotlight the intersection of these domains, focusing on innovative approaches that integrate solvers, preconditioners, and approximation methods with the power of deep learning. Discussions will also revolve around the incorporation of deep learning methodologies to enhance error estimation and adaptive mesh generation, enriching the understanding of PDE solutions and accelerating convergence rates of classical algorithms for the approximation of PDEs. Participants will have the opportunity to explore how the fusion of traditional numerical techniques and cutting-edge deep learning strategies can lead to novel breakthroughs and foster advancements in both fields.
Keywords: approximation methods, error estimators, solvers and preconditioners, adaptive mesh generation, numerical methods for PDEs
The electric excitation of the myocardium is a pivotal mechanism for sustaining proper cardiac function. Related diseases such as arrythmias or atrial fibrillation have major impact on heart failure and infarction, and thus are a significant source of mortality in developed countries. Numerical simulation of cardiac excitation mechanisms helps understanding of disease onset and progression, and is therefore an invaluable tool for improving diagnosis and therapies.
Established models such as the bidomain and monodomain models homogenize the myocardium for describing the electrical activation in terms of a reaction-diffusion system. While very successful in reproducing many effects and mechanisms, they do not capture some aspects caused by heterogeneous microstructure of the myocardium. This includes heterogeneity on the cellular scale as present in fibrosis, spatial distribution of ion channels and gap junctions, or shape and interconnection of myocytes.
For that reason, detailed models representing the discrete cellular structure of the myocardium have recently gained attention. The most prominent example is the EMI model, representing all myocytes as individual subdomains, and treating the nonlinear dynamis of ion channels only on the membranes. These models are also used, with different ion dynamics, for describing neural activation with subcellular resolution. Further modeling directions are towards detailed models of single myocytes, the impact of cellular resolution on mechanics, or gap junction function.
Such models pose new challenges for simulation: significantly increased size due to increased resolution, dimension-heterogeneous structure, and only piecewisely continuous solutions. They call for new or adapted approaches for numerical simulation with a perspective towards high performance computing.
This minisymposium will present a forum for presenting new results and methods for simulating EMI and similar models in relevant areas such as space and time discretization approaches, adaptivity, heterogeneous model coupling, iterative solvers, or domain decomposition preconditioners. Potential contributions include, but are not limited to, new finite volume schemes for EMI models, boundary element and static condensation approaches, stabilized explicit Runge-Kutta schemes, spatial adaptivity with SDC time stepping, BDDC and block preconditioners adapted to the EMI geometry, or high-performance GPU implementation of preconditioned Krylov solvers.
Keywords: domain decomposition, finite elements, time stepping, cardiac electrophysiology
The past few decades have witnessed a notable rise in natural disasters that are extreme and multi-hazard. Based on climate change forecasts, this trend is expected to escalate in the coming years. Many of these hazards are driven by hydrological processes, such as floods, mudslides, landslides, avalanches, and tsunamis.
Recent developments and improvements in numerical methods, together with the increase in computing power, have encouraged the application of computational tools for simulating natural hazards.
Shallow water models, advanced finite elements and finite volumes schemes, and in particular Particle-Based methods (e.g., the Smoothed-Particle Hydrodynamics (SPH), the Discrete Element Method (DEM), the Material Point Method (MPM), the Particle Finite Element Method (PFEM), etc.), can be used and coupled to simulate such complex scenarios and to evaluate the impact of these extreme events. Moreover, some of these methods' good CPU or GPU parallelization makes them suitable for large-scale 3D simulations.
The aim of this thematic session is to present and discuss recent developments in numerical simulations of the initiation and dynamics of natural hazards. Additionally, the event seeks to facilitate collaboration among experts in this field to promote productive discussions around this critical topic. Although the focus of this thematic session is primarily on hydrological hazards, we also welcome numerical methods applied to other types of natural events, including geological and meteorological phenomena. In particular, all those numerical methods that analyze multi-hazard events (e.g., landslides triggered by earthquakes or tsunami waves generated by landslides) will be appreciated. To account for possible interactions with civil constructions, contributions within the framework of fluid-structure or fluid-soil-structure interactions will also be appreciated.
Keywords: DEM, fluid-structure interaction, MPM, PFEM, shallow water, coupled problem, SPH
The last decades were characterized by a rapid growth of natural hazards involving not only large mass movements on complex terrain such as landslides, debris flows, and mud flows, but also floods and tsunamis. In this scenario, the modeling and simulation of these physical phenomena have become essential in many areas of applied and industrial sciences. Shallow Water Equations represent a fundamental mathematical tool for comprehending and modeling this kind of phenomena. The simulation of the above events still represents a big challenge for different reasons: the need to deal with large strain regimes, the intrinsic multiphysics nature of such events, and the difficulties induced by the geometric complexity of the terrain under consideration. These complexities necessitate the development of advanced numerical methods to be able to obtain an accurate and reliable numerical solution. Classical finite element methods are well established and widely used in many engineering fields both in academia and industry, but they can show some limitations. This is particularly true when dealing with problems where large deformation occurs, like, e.g., hypervelocity impact, crack propagation, multi-phase interactions, and free surface simulations. In recent years, possible alternatives have been proposed and developed to overcome this drawback.
In this Minisymposium, we will delve into the connection between the SWE model and its numerical solution, with a focus on the recent advances in the available numerical methods. This Minisymposium aims at covering the state-of-the-art in the mathematical and computational framework for solving the shallow water model to stimulate interdisciplinary research in applied mathematics and to foster interactions among the scientific community.
Keywords: shallow water equations, computational fluid dynamics, free surface models
Numerous physical phenomena are modelled in unconventional geometries, encompassing ramified spaces, unions or intersections of manifolds, fractals, networks, etc.
One example is given by transport and growth models, where interconnected structures such as roads, pipelines, and vessel networks serve as domains for both linear and nonlinear partial differential equations. These equations, often hyperbolic or hyperbolic-dominant, exhibit low degrees of regularity, evident for example in Conservation Laws and Hamilton-Jacobi equations. This limited regularity is further complicated by the non-Euclidean structure of the domain. Consequently, the treatment of such equations requires a focused understanding of solutions and their numerical processing.
Another example can be found in the modeling of flow, transport and mechanical deformation processes in porous fractured media. Fractures are thin structures embedded in a much larger domain, such that their geometrical reduction to planar interfaces is of great importance to reduce the computational cost of simulations. After such dimensional reduction, the problem is rewritten as a system of coupled PDEs on a hybrid dimensional domain: the 3D bulk domain, the 2D fractures and the 1D fracture intersections.
In some applications (e.g., the growth of roots in soil, the formation of new blood vessels in tissues,…) a direct coupling of a 3D problem with a 1D problem might appear as a result of geometrical reduction. This is the case, for example, of thin and elongated structures embedded in a 3D domain, that can be conveniently reduced to line interfaces. In this case, however, the geometrical simplification results in an ill-posed mathematical problem in the standard approximation spaces and ad-hoc formulations and numerical schemes need to be developed.
The principal aim of this mini-symposium is to convene international experts in this field, providing a platform for the presentation of the most recent advances and fostering discussions on the current state of the art in overcoming these challenges.
Keywords: 3D-1D coupling, 3D-2D coupling, modelling, networks, aeronautical applications, numerical methods on complex geometries
The aim of this mini-symposium is to discuss and to share recent ideas from numerics (w.r.t. discretization and solver techniques), scientific computing (w.r.t. computational, algorithmic, and implementational aspects) and data science (w.r.t. machine learning and neural network approaches) for the highly efficient treatment of partial differential equations (PDEs) that arise in the simulation of problems from computational fluid dynamics (CFD) and computational solid mechanics (CSM). The presented approaches shall particularly address new ideas regarding future high-performance computing (HPC) environments which will be in the exascale range and which will include massively parallel, heterogeneous architectures together with specific accelerator hardware (GPUs, TPUs, FPGAs) including reduced arithmetic precision. The mini-symposium will concentrate on methods and their foundations and will highlight the interplay of these aspects with computational and algorithmic tools and particularly their realization in simulation software. We shall discuss, for instance, aspects regarding hardware-oriented numerics, energy-efficient and extremely scalable numerical approaches, scientific machine learning techniques together with artificial neural networks, numerical cloud computing, and massively parallel solvers exploiting parallelism in time. Other aspects to be included are nonlinear domain decomposition methods and extremely scalable numerical homogenization methods.
Keywords: CFD, CSM, Exascale algorithms, HPC, Mathematical software, Scientific Computing
Multiscale systems, characterized by interactions between different scales in space and time, offer valuable insights into diverse scientific disciplines, including physics, biology, materials science, and engineering. With advances in hardware technologies enabling simulations of larger and more complex systems, the development of robust and scalable numerical methods becomes crucial to fully unlock the potential of modern High-Performance Computing (HPC) architectures.
This minisymposium aims to address key topics in this field, including multilevel solvers for partial differential equations, nonlinear optimization, and multi-fidelity techniques. A fundamental, but usually neglected, part of the solution procedure is the choice of the hardware and the representation of infinite dimensional objects such as irrational numbers. Because of this, we also welcome contributions related to hybrid computing, matrix-free solvers and mixed-precision algorithms.
Keywords: high-performance computing, iterative solvers, preconditioners
The need for describing, understanding, and predicting the dynamics of complex systems has led to a multitude of numerical methods across several disciplines over the last decade, blending physics-based and data-driven techniques to different extents. Recent advances in deep learning, among many algorithmic methods of machine learning, have allowed to overcome several bottlenecks often hindered by high dimensionality, significant complexity, and chaotic behaviors, opening new horizons for data-driven predictive modeling. However, new techniques with general applicability to complex systems in science and engineering are still needed promptly.
Relying on well-established paradigms of reduced order modeling and dimensionality reduction through, e.g., autoencoders and sparsity-preserving techniques, data-driven modeling and discovery is an extremely active research area, nowadays integrating Bayesian and kernel methods for uncertainty quantification, multi-fidelity methods for data fusion, and deep reinforcement learning for controls, to name a few.
This minisymposium will gather a broad spectrum of contributions in this very vibrant research area, covering the theoretical analysis, computational techniques, and practical use of data-driven methods for the model reduction and discovery of dynamical systems, all towards efficient and accurate predictions in applied sciences and engineering.
Keywords: Deep Learning, Digital Twins, Dynamical systems, model discovery, Neural Networks, Partial Differential Equations, Reduced Order Modeling
1800 - SCIENTIFIC COMPUTING
N. R. Franco (PoliMi), F. Pichi (EPFL) and S. Fresca (PoliMi)
ABSTRACT
The solution of differential equations using full-order models (FOMs) incurs significant computational costs, especially in real-time simulations and multi-query routines. To address this challenge, reduced order models (ROMs) have emerged as a crucial framework for generating efficient and reliable approximations essential for simulations in basic sciences and industrial applications. The increasing demand for efficient non-intrusive methods, coupled with the availability of large amounts of data from measurements or simulations, has spurred the development of new techniques for complexity reductions based on deep learning in computational science. These approaches leverage state-of-the-art machine learning algorithms capable of extracting previously unseen patterns inherent in the data.
Integrating data-driven techniques with physics-based approaches enhances the modeling capabilities and interpretability of the models, enabling consistent and accurate predictions even for complex systems. This integration has given rise to several research lines that combine traditional ROMs with scientific machine learning.
This mini-symposium aims to bring together researchers actively involved in the theories, methods, and applications of deep learning-based techniques for complexity reduction. The areas of interest include, but are not limited to, deep-learning-based reduced order modeling, approximations using neural operators, physics-informed deep learning, multi-fidelity methods, as well as the approximation and mathematical properties of neural networks.
REFERENCES
[1] Brunton, S.L. and Kutz, J.N. (2022) Data-Driven Science and Engineering: Machine Learning, Dynamical Systems, and Control, Cambridge University Press. doi: 10.1017/9781009089517.
[2] Franco, N.R. et al. (2023) ‘Approximation bounds for convolutional neural networks in operator learning’, Neural Networks, 161, pp. 129–141. doi: j.neunet.2023.01.029.
[3] Fresca, S. and Manzoni, A. (2022) ‘POD-DL-ROM: Enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition’, CMAME, 388, p. 114181. doi: j.cma.2021.114181.
[4] Pichi, F., Moya, B. and Hesthaven, J.S. (2023) ‘A graph convolutional autoencoder approach to model order reduction for parametrized PDEs’, arXiv:2305.08573.
Keywords: parametrized PDEs, Reduced Order Modeling, Scientific Machine Learning
Abstract
SU2 is an open-source software for the analysis of (coupled) partial differential equations (PDEs) and (multi-objective) PDE-constrained optimization problems on unstructured meshes with state-of-the-art numerical methods. The availability of a shared code base facilitates the collaboration of engineers and scientist on a global level and grants access to industry-standard analysis tools. Thus, SU2 fosters a rapid dissemination of advances in numerical methods for (coupled) simulations, and (shape) design optimization (in multiphysics context) for the online community of users and developers.
This mini-symposium invites presentations from engineers and researchers who develop methods within SU2 or use SU2 for their engineering design optimization problems. Topics include, but are not limited to:
1. (Coupled) adjoint methods and algorithmic differentiation.
2. Multidisciplinary design optimization methods and applications.
3. Surrogate-based design optimization methods.
4. Engineering design under uncertainties.
References
[1] Economon, Thomas D., et al. "SU2: An open-source suite for multiphysics simulation and design." AIAA Journal 54.3 (2016): 828-846.
[2] Albring, Tim A., Max Sagebaum, and Nicolas R. Gauger. "Efficient aerodynamic design using the discrete adjoint method in SU2." 17th AIAA/ISSMO multidisciplinary analysis and optimization conference. 2016.
[3] Burghardt, Ole, et al. "Discrete adjoint methodology for general multiphysics problems: A modular and efficient algorithmic outline with implementation in an open-source simulation software." Structural and Multidisciplinary Optimization 65.1 (2022): 28.
Keywords: Adjoint methods, Engineering design under uncertainties;, Multidisciplinary design optimization;, Open-source project;
This mini-symposium aims to assemble both developers and users of open-source scientific software libraries for fluid and solid mechanics. It will encompass discussions on software design and performance, which includes topics such as data structures, the implementation of approximation methods, adaptive hp-refinement, and high-performance computing. Additionally, the mini-symposium will delve into linear solvers for large systems of equations, meshing tools, and various pre- and post-processing tools. Intended as a forum for developers, contributors, and users of open software libraries, the mini-symposium will facilitate the sharing of knowledge and experience about modern infrastructures, emerging technologies, and algorithms. It will also cover performance profiling, testing, and development. The symposium will provide insights into the process of building and managing communities of developers and users, drawing from experiences with successful and widely-used open-source codes such as FEniCS project, deal.II, MFEM, PETSc, MoFEM, and others.
Keywords: algorithms, Fluid Mechanics, high performance computing, open-source software, solid mechanics
Non-linear systems exhibit much more intricate and complex behaviors compared to their linearized counterparts. In numerous research and industrial scenarios, understanding these behaviors often requires computationally intensive numerical solutions using finely discretized models. Unfortunately, this approach leads to significant memory and time requirements, which can impede analysis, design, and optimization processes. To address these challenges, model order reduction (MOR) techniques come into play, facilitating the aforementioned tasks by creating lower-order approximations of the high-fidelity model. 
Substantial improvements have been achieved in this field over the past years. Interpolation methods for non-linearties and hyper-reduction approaches [3] have been proposed by the reduced basis community. Reduction of specific classes of non-linear input-output maps, such as bilinear control system [2] and quadratic-in-state system [1], have been successfully addressed in the system theoretic framework. Despite this, there are several challenges when dealing with reduced order models approximating non-linear systems. One of the challenges is error certification, as ensuring the accuracy of these reduced models is a difficult task [4]. Additionally, another challenge is the lack of generalization. Some model reduction techniques may perform well for specific non-linear systems but struggle to be effective across different types of non-linear problems. 
The aim of this minisymposium is to bring together early-stage researchers to present recent advancements of MOR for non-linear systems. In particular we encourage participation from both the reduced basis and system-theoretic communities.
1. P. Benner, T. Breiten, “Two-sided projection methods for non-linear model order reductionâ€, SIAM J. Sci. Comput., Vol. 37, pp. 249-260, (2015).
2. P. Benner and T. Damm, “Lyapunov Equations, Energy Functionals, and Model Order Reduction of Bilinear and Stochastic Systemsâ€, SIAM J. Control Optim., Vol. 49, pp. 686-711, (2011).
3. J.A. Hernández, M.A. Caicedo, A. Ferrer, “Dimensional hyper-reduction of nonlinear finite element models via empirical cubatureâ€, Comput. Methods Appl. Mech. Eng., Vol. 313, pp. 687-722, (2017).
4. M. Yano, A.T. Patera, “An LP empirical quadrature procedure for reduced basis treatment of parametrized nonlinear PDEsâ€, Comput. Methods Appl. Mech. Eng., Vol. 344, pp. 1104-1123, (2019).
Keywords: Model Order Reduction, Nonlinear MOR, Nonlinear Systems, parametrized PDEs
Machine Learning has demonstrated transformative potential in diverse scientific fields, with deep learning standing out for its exceptional performance [1]. However, the black-box nature of these deep learning models has raised concerns about their lack of physical interpretability.
In an endeavor to bridge this interpretability gap, sophisticated methodologies such as Physics-Informed Neural Networks (PINNs) [2], DeepONet [3], symbolic regression [4], Scientific Machine Learning (Sci-ML) [5], and differentiable programming [6] are an attempt to solve this issue. This symposium seeks to bring together experts for an in-depth discussion on approaches that integrate domain-specific physical laws and principles directly into machine learning algorithms, thereby fostering a more interpretable and reliable modeling framework.
[1] LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444.
[2] Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2019). Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378, 686-707.
[3] Lu, L., Meng, X., Mao, Z., & Karniadakis, G. E. (2021). DeepONet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators. Nature Machine Intelligence, 3(5), 389-405.
[4] Rudy, S. H., Brunton, S. L., Proctor, J. L., & Kutz, J. N. (2017). Data-driven discovery of partial differential equations. Science advances, 3(4), e1602614.
[5] Raissi, M., Yazdani, A., & Karniadakis, G. E. (2020). Hidden Fluid Mechanics: Learning Velocity and Pressure Fields from Flow Visualizations. Science, 367(6481), 1026-1030.
[6] Innes, M., Edelman, A., Fischer, K., Rackauckas, C., Saba, E., Shah, V.B. and Tebbutt, W., 2019. A differentiable programming system to bridge machine learning and scientific computing. arXiv preprint arXiv:1907.07587.
Keywords: Convolutional Neural Networks, Deep Learning, Fluid Mechanics, Transport Phenomena
In order to meet the requirements of tomorrow’s world, engineers and architects must design extremely efficient structures. Making engineering structures adaptive is a promising approach to reach that target. One approach to adaptive structures is to noticeably increase the load carrying efficiency of structures by the employment of sensors, actuators and control units. Hence, the active manipulation of the static and dynamic structural response (i.e., forces, deformations and vibrations) enables dramatic mass reduction of engineering structures while increasing their performance. Additionally, control and actuation allow adjustment to evolving requirements occurring during lifetime of a building. Deployable and retractable structures, with significant changes between the configurations are another application of adaptive structures. Here, not only the motion, but also the structure itself needs to be designed to meet the requirement of compliance.
This mini-symposium focuses on load-carrying adaptive structures in general within civil engineering as well as on compliant adaptive structures. It is devoted to new approaches in the computational design, analysis and optimization of such structures including (but not limited to):
- modelling and simulation
- form finding and optimization
- optimal strategies for sensor and actuator placement
- active and passive control strategies
- reliability and fail-safe design of adaptive structures
- criteria for the evaluation of adaptive engineering structures
- design of compliant structures and mechanisms
- design and analysis of deployable structures
Keywords: actuator placement, compliant structures, control strategies, deployable structures, flexibility, Modelling and simulation, optimization
Unreinforced Masonry (URM) constructions are widely spread, especially in European countries. They represent an invaluable part of the world's architectural and historical heritage, being distinguished by different geometric shapes, architectural details and mechanical properties. Due to the environmental actions, in many cases, masonry structures experienced severe deterioration processes which increased the structure vulnerability to seismic events. Several approaches have been proposed for the analysis of masonry structures, ranging from empirical to analytical and computational procedures. A suitable criterion for classifying masonry modelling approaches relies on the scale at which masonry is analyzed, distinguishing between micromechanical, macromechanical, and multiscale models, but also other criteria can be adopted.
Structural Health Monitoring (SHM) plays a paramount role for the preservation of both existing masonry constructions composing urban environments and cultural heritage buildings. The use of advanced techniques and innovative instrumentations allow for identifying important parameters characterizing the dynamic response of these structures throughout noninvasive and expeditive procedures. The correct identification of these parameters allows to derive additional important information about the presence, location and extent of possible damages.
Due to both the actual health state of masonry constructions and the very high vulnerability against seismic actions, different seismic retrofitting and strengthening approaches have been developed and implemented in the past few decades. The primary concept of retrofitting methods is a) to reduce the impact of external loads (ii) to upgrade the load-bearing capacity of structural parts and (iii) to upgrade the structural integrity. These targets can be pursued through different traditional and innovative consolidation and strengthening interventions.
Based on the above premises, in the current Mini-Symposia, the following objectives will be achieved:
- to compare and assess the advantages and disadvantages of each analysis method of URM structures to identify the most suitable method in different cases.
- to focus on methodological aspects, recent developments and applications of vibration-based assessment and SHM of ordinary and historic structures.
- to provide some helpful guidance for the researchers in choosing an appropriate technique for strengthening/retrofitting URM structures.
Keywords: masonry constructions , modelling, retrofit, Structural Health Monitoring
Functionally graded materials (FGM) are an advanced type of composites that may present enhanced behaviour performance when compared with other types of composites. Their continuously varying microstructure provides smooth properties’ variation without interfacial stress concentrations [1]. These materials were initially designed for a specific application in concrete to reduce the thermal stresses arising from the high temperatures of the metal and ceramic interfaces in a space shuttle project [2]. Besides this first goal of thermal barrier function, the ability to tailor materials mixture’s distributions according to specific operating requirements, without the disadvantages shown by traditional laminates, is progressing significantly in different Science and Engineering fields [3]. The FGM concept can be found in many natural materials, and it can also be extended to active/passive materials mixtures, thus providing adaptive capabilities to the structures they are built with.
This mini-symposium aims to highlight the recent advances in the broad scope of the modelling, analysis and optimization of functionally graded composites and structures.
[1] P.S. Ghatage, V.R. Kar, P.E. Sudhagar, “On the numerical modelling and analysis of multi-directional functionally graded composite structures: A reviewâ€, Composite Structures, Vol. 236,111837, (2020).
[2] Koizumi M. FGM activities in Japan, Composites Part B: Engineering, Vol.28,1–4, (1997).
[3] B. Saleh, J. Jiang, R. Fathi, T. Al-hababi, Q. Xu, L. Wang, D. Song, A. Ma, “30 Years of functionally graded materials: An overview of manufacturing methods, Applications and Future Challengesâ€, Composites Part B: Engineering, Vol. 201, 108376, (2020).
Keywords: Functionally graded composite materials, Materials and structures modelling, Passive and active materials, structural optimization
Soft robotics is a promising and innovative field that offers remarkable advantages due to its compliant and flexible nature, enabling safe interactions with humans and delicate objects. Moreover, integrating embodied intelligence further enhances soft robots, granting them the ability to learn, adapt, and interact intelligently with the environment, making them more versatile, responsive, and user-friendly. This combination of soft materials and embodied intelligence is promising for transforming industries, healthcare, and everyday life with innovative and human-centric robotic solutions.
In developing soft robotic medical devices, computational mechanics plays a vital role as it enables accurate modeling and simulation of the complex and deformable structures in soft robots. This precision is crucial for designing, optimizing and controlling of soft robotic systems, especially in medical engineering applications where soft robots can revolutionize minimally invasive surgeries, enhance rehabilitation, and provide personalized medical assistance, all while conforming to delicate biological tissues with enhanced safety and efficacy.
This mini-symposium aims to provide a platform to present latest research, exchange ideas, and address challenges in applying computational methods to soft robotics and embodied intelligence. Topics of interest include, but are not limited to:
- Computational modeling of soft robots by advanced numerical methods and state-of-the-art tools for modeling the complex behavior of soft robotic structures and materials.
- Optimization methods to improve the performance and design of soft robotic systems.
- Computational mechanics of soft sensors and actuators.
- Machine learning for computational soft robotics.
- Multiscale and multiphysics simulations for soft robotic systems.
- Interaction of soft robots with biological systems and the human body.
- Model-based control in soft robotic applications.
Keywords: Control Algorithms, Embodied Intelligence, Motion Planning, Physical Intelligence, Soft Actuators, Soft Sensors, Soft Robotics
This minisymposium focuses on both theoretical and practical aspects concerning the transient solution of structural dynamics problems in science and engineering. Particularly, novel numerical methods and solution strategies as well as discretization schemes in space and time for wave propagation, structural vibration, structural health monitoring, coupled problems (e.g., fluid-structure-interaction) and impact problems are of interest. This includes, but is not limited to the development or the application of
• isogeometric and high-order finite element methods (e.g., IGA, SEM, p-FEM, etc.),
• fictitious domain methods,
• meshfree methods,
• mass lumping and mass scaling techniques, or
• advanced time integration schemes (e.g., novel implicit and explicit time integration schemes, implicit-explicit or asynchronous time integration schemes, sub-cycling, parallel implementation, etc.).
Furthermore, contributions dealing with large-scale, industry-relevant applications are expressly welcome.
Keywords: Advanced Discretization Schemes, Advanced Time Integration Schemes, Computational Structutral Dynamics, Mass Lumping, Mass Scaling
Starting from the mid-1990s,1 the subject on vehicle-bridge interaction dynamics has been booming, partly due to the widespread construction of highspeed railways in the world. Major concerns in this area include: (1) efficient simulation of the interaction between the moving vehicle and bridge for the purpose of analysis, (2) optimal design of span length for highspeed railway bridges to reduce the bridge vibration, (3) riding comfort of passengers during the movement of the highspeed trains, (4) safety of the vehicle-bridge system under the external shaking of earthquakes and winds, (5) effect of infrastructure such as ballast, sleepers and fasteners on the track system, (6) detection of dynamic properties of the sustaining bridge by moving test vehicle, and so on. Particularly, the technique of using a moving a test vehicle to detect the dynamic properties (frequencies, mode shapes, damping ratios, damages, etc.) of sustaining bridges has become an interesting topic of research most useful to highway bridges,3 as well as on railway bridges. Such a technique has been referred to as the indirect method for bridge measurement, since it does not require the vibration sensors to be mounted on the bridge, but only a small number of sensors on the test vehicle. It was also renamed as the vehicle scanning method for bridges for its straight conveyance of the meaning involved.2,3 In this mini-symposium, we welcome all researches on the vibration and detection aspects of both highway and railways bridges, but not restricted to the items mentioned above.
REFERENCES
[1] Yang, Y. B., and Lin, B. H., Vehicle-bridge interaction analysis by dynamic condensation method, J. Struct. Eng., ASCE, 121(11), 1995, 1636-1643.
[2] Yang, Y. B., Lin, C. W., and Yau, J. D., Extracting bridge frequencies from the dynamic response of a passing vehicle, J. Sound & Vibr., 272(3-5), 2004, 471-493.
[3] Wang, Z.L., Yang, J.P., Shi, K., Xu, H., Qiu, F. Q., Yang, Y.B., Recent advances in researches on vehicle scanning method for bridges, Int. J. Struct. Stab. & Dyn., 2022 (15), 2230005, DOI: 10.1142/S0219455422300051.
Keywords: bridge, track-bridge interaction, vehicle
Nowadays, modelling and simulation is common practice in engineering and its results may be important for making decisions. Therefore, the credibility of the results obtained must be assessed. In the numerical solution of physics based models, which are the most usual in Computational Solid Mechanics (CSM) and Computational Fluid Dynamics (CFD), verification, validation and uncertainty quantification (VVUQ) are the activities that allow to establish the credibility of modelling and simulation. Several organizations, as for example ASME, AIAA and NAFEMS have published documents about best practices in VVUQ that can help engineering practitioners to apply VVUQ and understand its benefits.
The proposed Mini-Symposium will address many aspects of verification, validation and uncertainty quantification based upon the contributed abstracts. The following topics are suggested by the organizers:
• Standard techniques to perform Code Verification, Solution Verification, Validation and Uncertainty Quantification activities.
• Description and explanation of published VVUQ standards and guides.
• Application of VVUQ techniques to practical engineering problems.
Keywords: Modeling and Simulation, uncertainty quantification, Validation, Verification
The numerical modeling, simulation and identification of random heterogeneous materials play an ever-growing role in materials science and give rise to many appealing engineering and scientific challenges for the design of innovative metamaterials or the characterization of real-world existing materials, such as e.g. sedimentary rocks, natural composites, fiber- or nano-reinforced composites, some concretes and cementitious materials, some porous media, some living biological tissues, among many others.
This Minisymposium aims at presenting the recent developments on Uncertainty Quantification in random linear and nonlinear materials, their quantification and propagation through computational models as well as the statistical inverse identification of their probabilistic models using stochastic optimization methods. Data-driven scientific machine learning and probabilistic learning methods with applications in the scope of this Minisymposium are also welcome. Practical applications and demonstration problems may concern, among others, linear and nonlinear random material models, involving e.g. plasticity, damage or fracture mechanisms or other material non-linearities, in uncertain linear and nonlinear computational mechanics for (quasi-)static or dynamic analyses.
Keywords: Optimization Under Uncertaintites, Random Linear and Nonlinear Materials, Statistical Inverse Problems, Uncertainty Quantification
High-fidelity scale-resolving simulations of turbulent flows have an utmost importance for understanding the flow physics and achieving optimal engineering designs. Such simulation approaches which include DNS, LES, and hybrid RANS-LES [1] require (prohibitively) large computational resources. Moreover, their resulting quantities of interest are uncertain up to some extent due to various sources. Therefore, not only the accurate quantification of uncertainties for such simulations is vital, but also cost-effective techniques must be considered when addressing outer-loop problems where several flow realisations are required.
This minisymposium aims at gathering experts in the theoretical development and application of uncertainty quantification (UQ) and data-driven approaches for scale-resolving simulations of turbulent flows. The topics of interest include, but are not limited to, forward and inverse UQ problems, error estimation, Bayesian optimization, multi-fidelity/multi-level models, sensitivity analysis, predictive machine learning models, reduced-order and surrogate models [2]. A particular focus will be on the strategies capable of making the overall computational cost of the data-driven methods affordable while retaining high accuracy.
Keywords: Data-driven models, Machine learning , Multi-fidelity models, Uncertainty quantification (UQ), Turbulence simulation
Reliability analysis is essential for the development, design and assessment of engineering systems under uncertainty. Challenges in computing the probability of failure are associated with non-linear system behaviour, large numbers of uncertain parameters and failure/rare events inducing multiple, disconnected failure domains. We invite talks discussing efficient computational methods for simulating rare events and quantifying failure probabilities based on sampling, surrogate modelling and machine learning as well as approximation approaches. Relevant applications of these techniques are in the assessment of static and dynamical engineering systems, reliability-based design optimization, reliability-oriented sensitivity analysis, Bayesian updating of failure probabilities with real-time data and applications in digital twin models.
Keywords: Bayesian Updating, Digital Twins, Rare Events, Reliability Analysis, Reliability Sensitivity Analysis, Reliability-based Optimization, Surrogate Modelling
Most natural phenomena and engineering problems are intrinsically nonlinear involving potentially many complex physical phenomena. Indeed, the nonlinearities may come from large amplitude vibrations, nonlinear material behaviour, contact interfaces, fluid-structure interactions, multi-physics coupling etc. The management of uncertainties is essential for determining reliable prediction of the nonlinear dynamic response.
This mini-symposium aims at gathering current works and fruitful discussions around the development of numerical methods for uncertainty quantification, model calibration, inverse methods and optimisation of nonlinear dynamic systems, or more generally on how to deal with numerous uncertainties and varying parameters and/or how to integrate nonlinear features in such numerical approaches. A non-exhaustive list of relevant topics includes the following:
- Nonlinear vibrations and bifurcation analysis
- Surrogate modelling
- Advanced design of experiments
- Parametric and structural optimisation
- Optimisation under uncertainty
- Model calibration under uncertainty
- Inverse problems, model updating and identification methods
- Multi-scale and multi-fidelity uncertainty propagation
- Etc.
Any applications that involve nonlinear dynamics are relevant to the MS.
Keywords: inverse problems, model calibration, Nonlinear dynamics, optimization, Surrogate Modelling, uncertainty quantification
The main aim of this mini-symposium is to provide an overview of recent advances in numerical methods relevant to the stochastic mechanics of solids, fluids as well as their interactions. These numerical methods include the Finite and Boundary Element as well as Finite Difference Methods, and also some meshless techniques necessary in uncertainty quantification of civil, mechanical and aeronautical systems. Stochastic approaches of interest include but are not limited to polynomial chaos, various implementations of the Monte-Carlo simulation, stochastic perturbation techniques, fuzzy sets and Bayesian methods; their usage in conjunction with AI tools would be also welcome. These methods should demonstrate the impact of various nature random variables, processes and fields on spatial-temporal uncertain response of the analyzed engineering structures. Numerical aspects of these methodologies including probabilistic convergence, computational error estimation or some numerical comparative studies, as well as large uncertainties, would be very beneficial for this mini-symposium. The contributions devoted to computer simulations across different geometrical and/or material scales in composite or heterogeneous materials are of special importance, but traditional engineering problems of reliability assessment, durability analysis and life cycles prediction are equally welcome.
Keywords: Computational Mechanics, Modelling and simulation, Multiscale modelling, Numerical Homogenization, probabilistic optimization, Reliability Analysis, semi-analytical approach, Uncertainty quantification (UQ)
Cardiac modeling and simulation are emerging as powerful tools in clinical cardiology, providing insights into disease mechanisms. Through numerical simulations, cardiac models establish connections between microscopic and macroscopic quantities, enable the prediction of disease progression and response to therapy, and support the design of medical devices and personalized interventions. The development of these models, the associated numerical methods, and their patient-specific customization present many challenges arising from inter and intra-patient variability, multiscale and multiphysics processes, heterogeneous data, and computational complexity. This mini-symposium aims at bringing together novel approaches to address the simulation and the personalization of coupled models encompassing cardiac mechanics, electrophysiology and hemodynamics. Among the many topics are: advanced numerical methods for large-scale simulations in high-performance computing frameworks; machine learning and reduced-order modeling techniques for cardiac applications; integration of experimental and clinical data into multiphysics models; patient-specific modeling of cardiovascular diseases; model validation, verification and uncertainty quantification. By facilitating knowledge exchange, this symposium aims at advancing the field of cardiac modeling and simulation, paving the way for more detailed computational models and the effective integration of clinical data, towards the development of cardiac digital twins.
Keywords: Cardiac modeling, Computational medicine, Digital twin, Multiphysics, Multiscale
Masonry structures were built over a period of about seven thousand years and they represent the major part of the built heritage of European countries. As their structural conception was mainly based on resisting gravitational loads, these structures are particularly sensitive to extraordinary and environmental actions, suffering damage even with events of moderate intensity [1]. Thus, the in-depth knowledge of masonry mechanics is essential to guarantee the structural safety and the conservation of heritage buildings.
In the last fifty years, the scientific community focused attention to this topic and devoted a great effort to propose computational methods suitable for the evaluation of the mechanical response of historic masonry and structures. Different analysis strategies and computational approaches at different scales have been proposed using limit analysis-based solutions or incremental-iterative analysis procedures [1].
This mini-symposium will offer an opportunity to present and discuss the recent advances in the development of numerical models for masonry, including block-based models (or “micro-models†based on interface elements, contact, XFEM, DEM, DDA, etc.) [2], continuum models (no-tension models, anisotropic damage models, homogenization procedure and multi-scale approaches, etc.) [3-4], geometry-based models [5], and equivalent frame models (i.e. based on macro-elements).
Furthermore, the increasing use of composite materials for retrofitting masonry constructions, which has become prevalent in structural applications, requires computational tools capable to investigate both the local and global behavior of the strengthened members. While some promising computational approaches were proposed in this context [6-7], several open issues still require further investigation.
In line with these research developments, this mini-symposium will offer an opportunity for presenting and discussing the recent advances in this field. It will gather researchers with interest in computational mechanics within the context of masonry structures to share and compare their findings, and to explore new frontiers for reliable and efficient numerical simulations. The participation of young researchers will be particularly appreciated.
Keywords: Computational Mechanics, masonry mechanics, masonry strenghtening, non-linear mechanics
This minisymposium is organized by young investigators (all of which are members of the ECCOMAS Young Investigators Committee) for young investigators. The format, which has first been introduced at the ECCOMAS Congress 2016 with great success, is quite different from the regular minisymposia in order to particularly attract young researchers.
Keywords: Computational Mechanics
This session focuses on the use of computational techniques to better understand the mechanical behavior of wood and bio-based structural materials. The underlying mechanics of such materials is characterized by a rather complex nonlinear behavior that involves anisotropy, viscoelasticity, elasto-plasticity, damage, etc., which can get even more complex under environmental influences such as due to moisture and temperature variations [1]. Mainly simplified material models are used/replaced for simulations to describe the behavior of these materials. However, the models need to sufficiently cover the relevant phenomena for efficient simulation/optimization to support design decisions [2]. Therefore, this session aims to discuss recent advances in computational mechanics of wood and biocomposites and their engineering applications, focusing on modeling methodologies. Topics of interest for this session are:
• Time-dependent static, quasi-static and dynamic analysis
• Stochastic modelling and optimization
• Development of new material models
• Model order reduction and computational efficiency
• Damage and failure
• Validation and verification of the numerical problems
REFERENCES
[1] Yu T., Khaloian A., van de Kuilen J.W.G. (2022). An improved model for the time-dependent material response of wood under mechanical loading and varying humidity conditions. Eng. Struct., 259: 114116. https://doi.org/10.1016/j.engstruct.2022.114116
[2] Lukacevic, M., Füssl, J., Pech, S., Vida, C., and Eberhardsteiner, J. (2021). Computational mechanics concepts for wood-based products and timber structural elements. WCTE 2021: World Conference on Timber Engineering. Santiago, Chile. DOI=10.34726/1341.
Keywords: Bio-based Materials, Uncertainties, Wood, Computational Mechanics, Model Order Reduction
Masonry structures and infrastructures represent a large part of the European historic and architectural heritage, but also characterize the modern building constructions of some major countries. Scientific community devoted large efforts over the past decades to develop reliable models for masonry and deliver efficient and accurate procedures for the assessment of their structural safety, also aiming at designing repairing and strengthening interventions when required.
A huge number of approaches have been proposed and applied for the analysis of both ancient and modern masonry structures. However, this is still a challenging and interesting task deserving attention to improve existing procedures and propose new approaches characterized by optimized performances in terms of accuracy and computational costs, as well as capability to interact with material testing methods for the definition of model’s input parameters.
Computational modelling techniques are the most adopted today, thanks to the increasing availability of computational resources and the advancements in numerical tools for structural analysis. These are aimed at describing masonry structural response under static and dynamics loading conditions, such as earthquakes, ground settlements, environmental actions also due to climate change and so on.
Classical and enhanced finite element procedures have been widely developed, where nonlinear constitutive laws capable of reproducing the main nonlinear mechanisms characterizing masonry response have been introduced. However, other alternative approaches have been successfully adopted. Various criteria can be adopted for classifying masonry modelling procedures. As for example, the existing numerical strategies could be subdivided into the following four classes: block-based models, continuum models, geometry-based models, and macro-element models. Also, the scale at which masonry is analyzed can be used to distinguish between micromechanical, macromechanical and multiscale models.
The aim of the proposed mini-symposium is to collect the most recent research contributions to this field and discuss the current issues and future developments concerning masonry computational modelling.
Keywords: Finite element method, Limit analysis, masonry constructions, masonry mechanics, Multiscale Modeling
Advanced materials and Structures have an increasing role in engineering, in various industrial applications [1-3]. These structures operate in severe environments, and withstand complex multi-axial loading conditions. Fracture of advanced materials is also a major problem that may occur inside the structures consisting of different materials and at the interfaces between the different advanced materials.
Topics of interest include but are not limited to the Computational analysis of composite structures made from advanced materials, Failure of composite structures, Comparison of computational and experimental methods in composite structures from advanced materials, Computational analysis of interface problems in composite structures, Computer-aided design in composite structures, Computational study of constructions made of advanced materials.
Keywords: Advanced Materials, Computational Mechanics, Interface problems in composite structures. Composite Structures
