Sensor fault detection and isolation for a class of nonlinear systems using sliding mode observer

Abstract

The present paper proposes a new scheme based on sliding mode observers (SMOs) for sensor fault detection and isolation for a class of uncertain nonlinear systems. By using the integral observer approach, sensor faults are treated as actuator faults. A state and output transformation is introduced to transform the original system into two subsystems. One subsystem is subject to system uncertainties but free from sensor faults. A sliding mode observer is designed for this system so that the effect of uncertainties can be eliminated. The second subsystem is only subject to sensor faults but system uncertainties. A Luenberger observer is designed for this system for the purpose of fault detection. The sensor faults can be further isolated by applying a bank of SMOs to the second subsystem. The sufficient condition of stability of the proposed SMOs has been derived and expressed as Linear Matrix Inequalities (LMIs). The design parameters of the proposed SMOs are determined by using LMI techniques. The effectiveness of the proposed SMOs in detecting and isolating sensor faults has been illustrated considering a numeral example. The result has been found to be satisfactory even for incipient sensor faults which are hard to be detected due to the nature of small in size and growing slowly during the initial phase.