Fault Diagnosis Method Based on Kernel Fuzzy C-Means Clustering with Gravitational Search Algorithm
Biyuan Wu, Xiangshun Li · 2018 IEEE 7th Data Driven Control and Learning Systems Conference (DDCLS) · 2018
The main drawback of the traditional fuzzy C-means clustering algorithm (FCM) is the randomness of the initial clustering center, which usually leads to the local optimal solutions and have a great influence on the clustering results. It also has to mention the FCM cannot deal with the non-linear data effectively. In this paper, gravitational search algorithm (GSA) is proposed to solve the randomness of the clustering centers. In addition, kernel fuzzy c-means clustering (KFCM) is introduced, which can improve the clustering result of the fuzzy c-means clustering for non-linear data. Finally, the proposed improved algorithm are verified with the three-tank system, and the results show that the concurrent faults can be diagnosed effectively.