The efficient modeling, simulation, and complexity reduction of dynamical systems in the field of process and chemical engineering have become pivotal tools in today's industrial landscape, especially with the ever-growing digitalization trend for modernizing plants and units in the age of Industry 4.0. In this direction, these tools warrant an efficient integration of digital twins to augment the physical domain, to contribute to optimizing processes and workflows, and to offer sustainable and efficient real-time solutions. In this work, we address the challenge of reduced-order modeling of dynamical systems in the field of process engineering, by employing scientific machine learning and complexity reduction techniques. These are especially important when a large number of evaluations are needed from a complex/large simulation model, which is required to be continuously updated based on changes in parameters and in the operating conditions. The method of interest is operator inference (OpInf) [1], a non-intrusive data-driven method for learning dynamical systems from time-domain data. The physical process that is modeled and used as a test case is CO2 methanation, in the Power-to-X (P2X) framework. P2X plays a key role in the storage of energy from renewable sources. Consequently, CO2 methanation, i.e., the conversion of CO2 to methane, is a key aspect of this framework, facilitating the recycling of CO2 emissions. The numerical results show the ability of the reduced-order models constructed with OpInf (using the recent developments in [2]) to provide a reduced, yet accurate surrogate solution. These results come as a continuation of [3], and could potentially represent a milestone towards the successful implementation of fast and reliable digital twin architectures, i.e., for deployment in a green carbon processing plant. [1] B. Peherstorfer and K. Willcox, Data-driven operator inference for nonintrusive projection-based model reduction. Comp. Meth. Appl. Mech. Eng., Vol. 306, 196–215, 2016. [2] P. Goyal, I. Pontes Duff and P. Benner, Guaranteed stable quadratic models and their applications in SINDy and Operator Inference. arXiv preprint, arXiv:2308.13819, 2023. [3] J. Bremer, P. Goyal, L. Feng, P. Benner, K. Sundmacher, POD-DEIM for efficient reduction of a dynamic 2D catalytic reactor model. Computers & Chemical Engineering, Vol. 106, 777–784, 2017