Executable Digital Twin (xDT) represents a paradigm shift in the industrial world, offering virtual replicas synchronized with physical assets through real-time data exchange. In this context, the implementation of Model Order Reduction (MOR) techniques in the time domain and the seamless integration of multiphysics simulations emerge as indispensable technologies for manifesting the concept of Digital Twin, ensuring advanced predictive capabilities across domains. High-order adaptive Finite Element Methods (FEM) have demonstrated improved efficiency in vibroacoustic simulations compared to conventional methods. Moreover, the inclusion of Flexible Infinite Elements is notable for its effectiveness in enforcing non-reflecting conditions, accommodating for arbitrary convex-shaped computational domains. To further alleviate the computational burden, the implementation of Krylov-based MOR techniques is employed. Leveraging input/output information, Krylov-based MOR guarantees high accuracy while offering a significant level of automation. This approach aims to substantially reduce computational cost and package models into a more condensed form, enabling online simulations. Several industrial applications necessitate coupling acoustics with multibody dynamic simulations. Current approaches handle these domains separately, performing time domain flexible multibody simulations followed by frequency domain vibroacoustic analysis. However, this sequential process displays computational inefficiency and substantial procedural complexity. To tackle this challenge, the proposed methodology concurrently solves motion and acoustics in the time domain through a floating frame of reference (FFR) formulation, capturing motion and sound radiation simultaneously. Validated against separate motion and vibroacoustic full-order models in the time domain, this unified approach demonstrates enhanced computational efficiency and a streamlined workflow. We showcase the advantages of employing the proposed methodology in simulating gearbox sound radiation while emphasizing its potential applicability across diverse domains, including the analysis of noise emissions in electric motors. This approach bears profound implications in the realm of multiphysics analysis, setting a foundational framework to optimize the design of low-noise systems and to enable online monitoring of complex mechanical systems.