Construction and Use of Surrogate Models for the Dynamic Analysis of Multibody Systems

Data-driven simulation for general-purpose multibody dynamics using Deep Neural Networks

“…For example, a Bayesian formulation [6,17,26] in combination with Markov random field approximation, a Kalman filter, a Gaussian process modeling [21], or a particle filter has been applied to various multibody dynamics (MBD) problems to handle noise data effectively in real-life applications, generate reliable modeling with efficient computational cost, estimate multibody systems in a probabilistic sense, and identify nonlinear parameters in governing equations. ML approaches [3,11,18,19,27] such as regression methods, reinforcement learning algorithms, and surrogate models have also been employed. There are many different types of regression methods that can be performed in ML.…”

“…Previously proposed methods have method has enhanced accuracy of prediction, especially in the long time scales, and increased computational efficiency in simulating the dynamic response of multibody systems. Moreover, neural networks [3,7,9,16,20] have been suggested as effective alternatives to multibody dynamics simulation in comparison with conventional algorithms such as recursive formulations [13] and reduced-order modeling techniques [1,14]. The approaches have been proved to be fast and reliable to describe and predict characteristics of multibody systems.…”

Data-driven simulation for general purpose multibody dynamics using deep neural networks

“…For example, Bayesian formulation [4,13,15] in combination with Markov random field approximation, Kalman filter, or particle filter has been applied to various multibody dynamics (MBD) problems to handle noise data effectively in real-life applications, generate reliable modeling with efficient computational cost, estimate multibody system in probabilistic sense, or identify nonlinear parameters in governing equations. ML approaches [14,16,19,17,18] such as regression methods, reinforcement learning algorithms, and surrogate models have also been employed. Regression methods have many different types that can be performed in ML.…”

“…The proposed method has enhanced accuracy of prediction, especially in the long time scales, and increased computational efficiency in simulating dynamic response of multibody system. Moreover, neural networks [19,20,21,22,23] have been suggested as effective alternatives to multibody dynamics simulation in comparison with conventional algorithms. The approaches have been proved to be fast and reliable to describe and predict characteristics of multibody systems.…”