Control allocation based fault tolerant control of descriptor system with actuator saturation

Abstract

This paper deals with the design of active fault-tolerant control (FTC) system based on control allocation (CA) technique for the uncertain descriptor system (DS) with actuator saturation. This work assumes that a fault detection unit is available to estimate the actuator fault. The proposed CA technique is responsible to redistribute the control effort to the healthy actuators taking the advantage of actuator redundancy of the DS. The so-called virtual control law is designed based on a variable gain super twisting sliding mode algorithm (VGSTSMA) to meet the admissibility criteria of continuous control input for the DS. This VGSTSMA based control strategy is designed based on the generalized regular form (GRF) of DS ensuring the robustness against the actuator faults, error in the fault estimation, and the external uncertainties. The CA scheme developed here is applicable to the DS even with a full rank input matrix which is different from the traditional design where it generally deals with the system with rank deficient input matrix. Furthermore, a static control redistribution mechanism is introduced to handle the actuator saturation which is responsible to re-allocate the excess control efforts to the available healthy actuators in the face of actuator saturation. Finally, a dual-pipe heat exchanger unit, modeled as DS, is simulated with proposed CA-based FTC such that the system can follow the time varying and constant reference trajectories in a situation of actuator faults and saturation. The simulation results are compared with the CA-based FTC where virtual control is designed based on super-twisting algorithm (STA) and with the traditional pseudo inverse-CA-based FTC to establish the superiority of the proposed scheme.