LMI-based aircraft engine sensor fault diagnosis using a bank of robust H∞ filters

This paper proposes a sensor diagnostic approach based on a bank of H_∞ filters for aircraft gas turbine engines, taking into account real-time and anti-disturbance requirements to run on an aircraft on-board computer. First of all, by defining an H_∞ performance index to describe the disturbance attenuation performance of a dynamic system, the robust H_∞ filter design problem has been formulated in the linear matrix inequality (LMI) framework and been efficiently solved using off-the-shelf software. Then we make use of a bank of H_∞ filters for the sensor fault detection and isolation based on the logic status of a group of residuals. Finally, an illustrative simulation has been given to verify the effectiveness of the proposed design method. The benefit of this paper is to bridge the gap between the H_∞ filter theory and engineering applications for reliable diagnostics of aircraft engines.