A novel fault diagnosis approach based on improved possibilistic C-means clustering and fault vector

Fault diagnosis approach based on fuzzy clustering is an important data-driven diagnosis approach. It does not need accurate mathematical models and can use a large amount of system historical and online data to extract failure information and to realize fault diagnosis. However, there are still no mature methods and techniques to deal with new faults. This paper proposed a possibilistic C-means clustering and fault vector based fault diagnosis approach to handle this problem. Simulation results have verified the effectiveness of the proposed method.