Efficient and Accurate Real-Time Simulations of Coupled Mechanical Systems Using the Universal Hydraulic Surrogate

Abstract The real-time simulation of coupled multibody and hydraulics systems can be challenging due to the complexity, accuracy, and non-linear behavior of the modeling methods and the computational requirements. To this end, this study introduces the universal hydraulic surrogate approach instead of the conventional lumped fluid theory to solve the dynamics of hydraulically driven systems. The universal hydraulic surrogate is prepared with an approximated force model and trained with the behaviors of lumped fluid theory using the Covariance Matrix Adaption Evolutionary Strategy (CMA-ES). The universal hydraulic surrogate is used to solve the dynamics of the hydraulically actuated four-bar mechanism. The performance of the universal hydraulic surrogate is compared with the lumped fluid theory in terms of numerical accuracy and computational efficiency. The maximum percent normalized root mean square error (PN-RMSE) in the states of the four-bar mechanism using a simple force model compared to the lumped fluid theory is 2.06 %. The simulation with the universal hydraulic surrogate approach is up to 45 times faster than the lumped fluid theory in the studied case. The proposed study will enhance the computational efficiency of the coupled multibody systems and enhances their use in artificial intelligence, neural networks, and machine learning applications for complicated industrial systems.