Model Order Reduction for Real-Time Hybrid Simulation: Comparing Polynomial Chaos Expansion and Neural Network methods

Abstract

Hybrid simulation is a method used to investigate the dynamic response of a system subjected to a realistic loading scenario by combining numerical and physical substructures. To ensure high fidelity of the simulation results, it is often necessary to conduct hybrid simulation in real-time. One of the challenges arising in real-time hybrid simulation originates from high-dimensional nonlinear numerical substructures and, in particular, from the computational cost linked to the computation of their dynamic responses with sufficient accuracy. It is often the case that the simulation time-step must be decreased to capture the dynamic behavior of numerical substructures, thus resulting in longer computation. When such computation takes longer than the actual simulation time, time delays are introduced and the simulation timescale becomes distorted. In such a case, the only viable solution for doing hybrid simulation in real-time is to reduce the order of such complex numerical substructures. In this study, a model order reduction framework is proposed for real-time hybrid simulation, based on polynomial chaos expansion and feedforward neural networks. A parametric case study encompassing a virtual hybrid model is used to validate the framework. Selected numerical substructures are substituted with their respective reduced-order models. To determine the robustness of the framework, parameter sets are defined to cover the design space of interest. A comparison between the full- and reduced-order hybrid model response is delivered. The attained results demonstrate the performance of the proposed framework.