Research on Generator Test Data Processing and Fault Diagnosis

• Published in: 2022 China Automation Congress (CAC) pp. 5690 - 5695
• Main Authors: Lin, Ruping, Huang, Jing, He, Zhiguo, Song, Huishu, Huang, Xiaosheng, Lin, Yang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 25.11.2022
• Subjects: Automation; Fault diagnosis; Feature extraction; generator test system; Signal processing algorithms; Simulation; Spirals; support vector machine (SVM); Support vector machines; variational mode decomposition (VMD); whale optimization algorithm (WOA)
• ISSN: 2688-0938
• DOI: 10.1109/CAC57257.2022.10055187

In order to improve the automation of the generator test system, this paper proposes a Variational Mode Decomposition (VMD) signal processing method combined with an Improved Whale Optimization Algorithm and Support Vector Machine (IWOA-SVM) pattern recognition method to process test data and diagnose generator faults. VMD processes current signals to obtain several Intrinsic Mode Functions (IMFs) and extracts the energy entropy as feature vectors. Then, aiming at the shortcoming of WOA's insufficient global search ability, an improved WOA is proposed by introducing adaptive weight and a variable spiral position update strategy. IWOA is used to optimize SVM parameters to improve classification accuracy. The experimental results show that the proposed method can accurately extract the fault information and improve the accuracy of fault identification. The diagnostic accuracy based on the actual generator test data is 93.75%, which proves that the method is feasible.