Effective Compensation of Nonlinear Actuator Dynamics Using a Proposed Linear Time-Varying Compensation

Abstract

Actuator tracking and compensation are important in the general field of experimental structural dynamics to effectively conduct vibration testing. Real-time hybrid substructuring (RTHS) is a method of vibration testing utilized to effectively characterize the system-level performance by physically testing a component of interest while numerically simulating the remaining support structure in real-time. The physical and numerical coupling is referred to as a transfer system, and actuators typically act as this system in RTHS. The inherent dynamics of actuator systems is a main cause of RTHS instability and inaccuracy. This work presents the methodology to achieve control of a six-degrees-of-freedom shake table. The corresponding system identification and model-based linear time-varying (LTV) compensation are robust enough to facilitate stable and accurate RTHS testing of mechanical systems at small- and large-amplitude excitations. When compared with a minimum-phase inverse compensation technique, the LTV technique was superior in linearizing actuator dynamics at varying excitation amplitudes. The LTV technique was also able to accurately command the three-dimensional (3D) displacements of the 2020 magnitude 6.4 Puerto Rican earthquake.