An improved principal component analysis scheme for sensor fault detection and isolation : Application to a three tanks system

In this paper a sensor fault detection and isolation procedure based on principal component analysis (PCA) is proposed to monitor a three interconnected tanks system. The PCA model is built to maximize fault detection sensitivity using a new index. The localization procedure is carried out using two methods. The first is based on the variables contribution to the fault index. The second is based on the reconstruction approach.