Improving the performance of grid‐connected doubly fed induction generator by fault identification and diagnosis: A kernel PCA‐ESMO technique

Abstract

Summary In this manuscript, a novel Kernel PCA‐ESMO technique is proposed for protecting the system and diagnosing the exact fault occuring in the DFIG. The proposed method is joined implementation of Kernel principal component analysis (Kernel PCA) and enhanced Spider Monkey Optimization SMO (ESMO) technique, and, hence, it is named Kernel PCA‐ESMO approach. Here, two phases are considered for fault analysis; they are fault identification and diagnosis. Primarily, the first phase identifies the system fault conditions of DFIG in grid‐connected system using Kernel PCA approach. After that, the ESMO classifies the type of fault that has occurred in the DFIG. The major scope of the proposed Kernel PCA‐ESMO method is to guarantee the system with less complexity for fault identification and diagnosis for enhancing the power quality of whole system. The implementation of proposed model is made at MATLAB/Simulink, and implementation is evaluated by existing techniques. The statistic analysis of proposed and existing systems of mean, median, and SD is analyzed. The efficiency of proposed and existing technique is also evaluated. The obtained values for percentage recall and precision that are enhanced using proposed technique are 97.8% and 98%. Consequently, the simulation outcome indicates that the efficiency of proposed technique and implementation of proposed strategy is compared to existing systems.

Study on the Analysis of Fault Data During Intelligent Operation of Power Equipment With Cluster Analysis

“…Sharan et al [5] designed a method based on spectrum analysis to monitor open-circuit faults in grid-connection and verified the effectiveness of the method by simulation analysis. Stallon et al [6] designed a method called kernel principal component analysisenhanced spider monkey optimization to achieve the analysis of grid faults and found through experiments that the method obtained 98% accuracy. Badr et al [7] proposed a support vector machine (SVM) based method for photovoltaic array fault monitoring and demonstrated the performance of the method through experiments in MATLAB/Simulink.…”