A Novel Actuation Dynamics Adaptive Compensation Strategy for Real-Time Hybrid Simulation Based on Unscented Kalman Filter

To address the varying time-delay problem in real-time hybrid simulation (RTHS), an unscented Kalman filter (UKF)-based adaptive time-delay compensator (UKF-ADC) is proposed in this study. UKF-ADC comprises of two main parts: a feedforward controller with adjustable parameters and a parameter estimator of UKF. The former is constructed with the inverse model of the control plant, which is represented by a first-order transfer function, whereas the latter is used to estimate and update the parameters online in the feedforward controller using displacement commands and measurements. To fully explore the performance of UKF-ADC, the effects of the initial parameters in UKF and factor alpha in the compensator are investigated through a prescribed displacement command adopting a nonlinear actuator model. As a result, the performance of UKF-ADC is mainly affected by factor alpha and robust to the initial parameters of UKF. In addition, a series of virtual and actual RHTSs are performed, employing the validated compensator. The results reveal that UKF-ADC can effectively improve the accuracy of RTHS and exhibit strong robustness.