现代信息科技2025，Vol.9Issue(21)：34-38,5.DOI:10.19850/j.cnki.2096-4706.2025.21.007
基于改进双深度Q网络在变压器不平衡样本故障诊断的研究
Research on Transformer Unbalanced Sample Fault Diagnosis Based on Imbalanced Classification Double Deep Q Network
王锦1
作者信息
- 1. 国网西安供电公司,陕西 西安 710032
- 折叠
摘要
Abstract
To address the problem of model bias caused by unbalanced distribution of sample categories in transformer fault diagnosis,this paper proposes an Imbalanced Classification Double Deep Q Network(ICDDQN).By designing a dynamic category weight reward function and combining KL divergence to construct a sample distribution compensation mechanism,the contribution of different categories of samples is dynamically adjusted during the model training phase.Experiments are carried out based on a self-built dataset containing 3 200 sets of oil chromatographic data.The comparison results show that the F1-score convergence accuracy of ICDDQN reaches 99.25%,which is superior to the DDQN baseline,and provides technical support for power equipment diagnosis under class imbalance conditions.关键词
变压器/故障诊断/深度强化学习/神经网络Key words
transformer/fault diagnosis/Deep Reinforcement Learning/Neural Network分类
信息技术与安全科学引用本文复制引用
王锦..基于改进双深度Q网络在变压器不平衡样本故障诊断的研究[J].现代信息科技,2025,9(21):34-38,5.
