A Hybrid Algorithm for Fault Diagnosis using Fuzzy Clustering Tools

Subject Areas : Fuzzy Optimization and Modeling Journal

Adrián Rodríguez Ramos ¹ , Pedro Juan Rivera-Torres ² , Antônio José da Silva Neto ³ , Orestes Llanes-Santiago ⁴

1 - Departamento de Automática y Computación, Universidad Tecnológica de la Habana José Antonio Echeverría, CUJAE, La Habana, Cuba
2 - Departamento de Ciencias de Computos, Universidad de Puerto Rico, Recinto de Río Piedras, San Juan, Puerto Rico
3 - Instituto Politécnico da Universidade do Estado do Rio de Janeiro (IPRJ/UERJ), Nova Friburgo, RJ, Brazil
4 - Politécnico da Universidade do Estado do Rio de Janeiro (IPRJ/UERJ), Nova Friburgo, RJ, Brazil

Received: 2020-05-01 Accepted : 2020-08-25 Published : 2020-07-01

Keywords: Automatic learning, Online detection, Fuzzy clustering tools, Optimal parameters,

Abstract :

In this paper, a hybrid algorithm using fuzzy clustering techniques is proposed for developing a robust fault diagnosis platform in industrial systems. The proposed algorithm is applied in a fault diagnosis scheme with online detection of novel faults and automatic learning. The hybrid algorithm identifies the outliers based on data density. Later, the outliers are removed, and the clustering process is performed. To extract the important features and improve the clustering, the maximum-entropy-regularized weighted fuzzy c-means is used. The use of a kernel function allows achieving a greater separability among the classes by reducing the classification errors. Finally, a step is used to optimize the parameters m (regulation factor of the fuzziness of the resulting partition) and (bandwidth, and indicator of the degree of smoothness of the Gaussian kernel function). The proposed hybrid algorithm was validated using the Tennessee Eastman (TE) process benchmark. The results obtained indicate the feasibility of the proposal.