Unknown parameter excitation and estimation for complex systems with dynamic performances
Journal
Journal of Mechanical Design, Transactions of the ASME
Journal Volume
143
Journal Issue
9
Date Issued
2021
Author(s)
Abstract
Simulation models play crucial roles in efficient product development cycles, therefore many studies aim to improve the confidence of a model during the validation stage. In this research, we proposed a dynamic model validation to provide accurate parameter settings for minimal output errors between simulation models and real model experiments. The optimal operations for setting parameters are developed to maximize the effects by specific model parameters while minimizing interactions. To manage the excessive costs associated with simulations of complex systems, we propose a procedure with three main features: the optimal excitation based on global sensitivity analysis (GSA) is done via metamodel techniques, for estimating parameters with the polynomial chaos-based Kalman filter, and validating the updated model based on hypothesis testing. An illustrative mathematical model was used to demonstrate the detail processes in our proposed method. We also apply our method on a vehicle dynamic case with a composite maneuver for exciting unknown model parameters such as inertial and coefficients of the tire model; the unknown model parameters were successfully estimated within a 95% credible interval. The contributions of this research are also underscored through multiple cases. Copyright ? 2021 by ASME
Subjects
Sensitivity analysis; 95% credible intervals; Dynamic model validation; Dynamic performance; Estimating parameters; Global sensitivity analysis; Hypothesis testing; Optimal excitation; Product development cycle; Parameter estimation
SDGs
Type
journal article