Passivity control with adaptive feed-forward filtering for real-time hybrid tests

Abstract

Real-time hybrid testing combines the reliability of experimental testing with the convenience of numerical simulation. The system to be tested is split into a physical substructure and a real-time numerical simulation which are coupled using actuators and sensors to transfer data at the interface in real-time. In order to achieve stable and accurate hybrid testing representative of the true system, high fidelity control is required at the substructure interface. However, actuators have a response lag which results in tracking errors and potential instability in hybrid tests. This paper investigates the effectiveness of a combined compensation strategy based on passivity control and adaptive feedforward filtering to improve stability, robustness and tracking performance in real-time hybrid testing. The combined strategy is adaptive and requires no prior information of the actuator dynamics unlike conventional transfer dynamics compensators in real-time hybrid testing. Moreover, the scheme requires no extra hardware making it inexpensive and applicable to a wide range of systems. Experimental results on a single degree of freedom nonlinear real-time hybrid test show the potency of the scheme in synchronizing​ substructure displacements while improving stability. The scheme was also found to restore stability of hybrid tests inherently unstable due to actuator delay whilst phase lags of up to 58 degrees have been successfully mitigated in a lumped parameter mechanical oscillator system.