Research Article
Research on Fault Diagnosis Technology of Bidirectional LLC Resonant Converter Based on CSA-DBN
@INPROCEEDINGS{10.1007/978-3-032-33662-0_41, author={Jianyu Liu and Chengzhao Shan and Zhuoming Li and Shuai Yuan and Xiaoying Zhang and Yi Lou}, title={Research on Fault Diagnosis Technology of Bidirectional LLC Resonant Converter Based on CSA-DBN}, proceedings={Communications and Networking. 20th EAI International Conference, ChinaCom 2025, Xi’an, China, November 8--9, 2025, Proceedings}, proceedings_a={CHINACOM}, year={2026}, month={10}, keywords={Fault diagnosis Machine learning Deep belief networks Crow search algorithm Bidirectional LLC resonant converter}, doi={10.1007/978-3-032-33662-0_41} }- Jianyu Liu
Chengzhao Shan
Zhuoming Li
Shuai Yuan
Xiaoying Zhang
Yi Lou
Year: 2026
Research on Fault Diagnosis Technology of Bidirectional LLC Resonant Converter Based on CSA-DBN
CHINACOM
Springer
DOI: 10.1007/978-3-032-33662-0_41
Abstract
Aiming at the bidirectional LLC resonant converter topology, and combining the advantages of the Deep Belief Network (DBN) in feature self-learning, this study proposes a structural fault diagnosis scheme based on the Crow Search Algorithm (CSA) optimized DBN. The scheme first undergoes hierarchical unsupervised pre-training, and then optimizes the network structure and parameters using CSA. The final experimental results demonstrate that compared with traditional machine learning fault diagnosis methods, the improved model with crow search algorithm significantly enhances the generalization ability and powerful representation ability of DBN for complex data by adaptively adjusting the number of hidden layer nodes and learning rate, thereby improving the learning and classification performance of the network and significantly improving the accuracy of fault diagnosis.