FNN approximation-based adaptive control for suppressing chatter in nonlinear milling with piezo-actuators

Abstract

To suppress the regenerative chatter in milling, an active control strategy with utilization of an adaptive controller and two piezoelectric actuators (PZTAs) is developed, where the nonlinear cutting forces, uncertain dynamics of the milling, hysteresis effect of the PZTAs and time-delayed effect of the chatter are considered. Fuzzy neural networks (FNNs) are employed to approximate the unknown function of the dynamics of the milling. The Prandtl-Ishlinskii (PI) model is adopted to describe the hysteresis nonlinearities of the PZTAs and its inverse model is applied to alleviate the hysteresis effect. In order to achieve the expected tracking trajectories with prescribed performances, two performance functions are used. By employing two PZTAs as active control elements and by combining the dynamic surface control (DSC) method, the FNNs, the Lyapunov-Krasovskii functionals and the performance functions, the boundedness of all signals of the closed loop and the chatter stability are obtained. The validity of the developed control strategy is revealed by simulations.