A Fault Detection and Isolation Scheme for Dual-channel Speed Sensors

Abstract

High-speed train dual-channel speed sensors are often influenced by factors like dusts, vibration, temperature changes, which easily cause false or miss alarm, even wrong isolation. The experimental comparison method commonly used in practical engineering, this method cannot detect faults in time or accurately isolate multiple faults. In this paper, a novel fault detection and isolation (FDI) scheme for dual-channel speed sensors is proposed by using the improved principal component analysis (PCA) and the improved reconstruction-based contribution plots (IRBCP). Take the dual-channel speed sensor of high-speed train as an example, experimental results show that the proposed scheme can satisfy FDI requirements of speed sensors, and is accurate and effective.