Application of three principal component analysis techniques in sensor faults detection

Abstract

The sensor is the front-end component of the bridge structural health monitoring system. In this study, three fault diagnosis methods based on principal element analysis are applied to a bridge health monitoring system. The diagnostic results show that significant differences among the operating conditions will deteriorate the robustness of the diagnostic methods. The applicability of the sensor fault diagnosis method is verified based on the principal element analysis of a practical health monitoring system.