Distinguishing Between Sensor Performance Degradation and Structural Damage Using Generalized Likelihood Ratio Test (GLRT) and Statistical Process Control

Abstract

This paper proposes a method which can distinguish between sensor performance degradation and structural damage in a vibration-based structural health monitoring system. This method utilizes the fact that the sensor performance degradation is local, while structural damage is global. In this paper, the generalized likelihood ratio test (GLRT) is firstly used to detect and localize a change in the system; and then the percentage of the extreme value of the largest principal component scores of the generalized likelihood is calculated, and finally a quality control chart is obtained which can be utilized to distinguish between sensor performance degradation and structural damage. On the other hand, if the source of the novelty is from sensor degradation, the extreme values are out of the bounds; else the extreme values are not out of the bounds. A simply-supported steel beam is experimentally performed to validate the proposed method. The results show that the proposed method can effectively detect the abnormal response and reduce the possibility of misclassifying sensor performance degradation and damage, as well as false alarming to some extent.