Asymmetric Bouc-Wen hysteresis modeling for MFC actuator via hybrid APSO-TRR identification algorithm

Abstract

Macro fiber composite (MFC) is widely used in active vibration and deformation control. The intrinsic asymmetric hysteresis nonlinearity of MFC affects the control accuracy. In this paper, a modified Bouc-Wen (MBW) model based on the sigmoid function is proposed to describe the asymmetric hysteresis characteristics of MFC, and its parameters are identified by a hybrid algorithm composed of the trust-region reflection method and asynchronous particle swarm optimization. The accuracy of the proposed MBW model is verified though hysteresis tests of MFC under different drive frequencies. The results show that the MBW model can accurately model the asymmetrical hysteresis of the MFC actuator, the modeling error is reduced by 72% compared with the classic Bouc-Wen model. The proposed hybrid parameter identification method saves 95% of the time compared with the particle swarm optimization and asynchronous particle swarm optimization algorithms. • A modified Bouc-Wen (MBW) model is proposed to describe the asymmetric hysteresis characteristics of MFC actuators. • A hybrid algorithm (APSO - TRR) with global and local search ability is proposed for parameter identification. • Performance of proposed APSO - TRR is compared with classical PSO、APSO and TRR algorithms. • Accuracy of proposed asymmetric Bou–Wen model is compared with classical Bou–Wen model.