Diagnosis of speed sensor faults in an induction machine based on a robust adaptive super-twisting observer

Authors: MOHAMMED ZAKARIA KARI, ABDELKADER MECHERNENE, SIDI MOHAMMED MELIANI, IBRAHIM GUENOUNE

Abstract: The present paper aims to determine a robust sensor fault-tolerant controller based on fuzzy logic using a robust adaptive super-twisting observer for the control of an induction machine and an inverter set by a state estimation method. The speed sensor is considered in the present case. The modular structure of the fault-tolerant control (FTC) scheme allows integrating this sensor within the existing closed-loop system, and the observer can therefore be designed independently. This article presents a new method to develop a fuzzy decision system that provides faulttolerant control. This paper also aims at detecting the mechanical speed sensor faults. The proposed approach allows the automatic reconfiguration of the system in the event of a speed sensor failure. The defective fuzzy detection system makes a transition between the speed sensor and the robust observer based on a super-twisting algorithm that ensures the continuity and stability of the system; the fuzzy detection and transition system is required to be robust to parametric variations, and must be fast enough in order to locate the defect and eventually make a transition in the event of a fault. The output of the fuzzy block is injected into the control block to ensure super-twisting speed regulation. The performance of the proposed strategy also the robustness against parameter variation are assessed by simulation thanks to Matlab/Simulink software.

Keywords: Sensorless control, fuzzy supervisor, fault detection and isolation, Liyaponov theory, induction machine

Full Text: PDF