Deep Learning Based Approach for High Voltage Circuit Breaker Mechanical Fault Diagnosis

• Published in: 2022 China International Conference on Electricity Distribution (CICED) pp. 268 - 272
• Main Authors: Zhuang, Zhijian, Yang, Qiuyu, Shi, Xinyong, Su, Liang, Ge, Liang, Gu, Xueming
• Format: Conference Proceeding
• Language: English
• Published: IEEE 07.09.2022
• Subjects: Circuit breakers; Deep learning; Fault diagnosis; high voltage circuit breaker (HVCB); High-voltage techniques; mechanical fault diagnosis; smart distribution network; Time-frequency analysis; Training; vibration time-frequency image; Vibrations
• ISSN: 2161-749X
• DOI: 10.1109/CICED56215.2022.9929160

12 kV high voltage circuit breaker (HVCB) is one of the most important equipment in the distribution network. Its reliability has tremendous impact on the safety of the network. The mechanical fault severity evaluation and diagnosis of HVCB is intensively investigated and studied in this article by means of vibration data processing and artificial intelligence methodology implemented. A deep learning based diagnosis approach is presented to overcome the shortcomings of traditional diagnosis models. We will benefit from this approach that single and compound fault are diagnosed with high detection rate and extremely low false alarm rate. The proposed method is purely data-driven and directly implemented by using time-frequency image of raw vibration data only. Experiments were conducted on a 12 kV switchgear with CB inside demonstrated the effectiveness of the proposed approach.