摘要
变压器作为电压转换设备,一旦发生故障,直接影响设备性能。然而在设备运行过程中,变压器系统大多处于正常状态,故障发生频率较低,所监测到的正常状态数据远远多于故障状态数据,存在样本不平衡问题。在变压器系统故障诊断技术和不平衡样本处理技术的基础上,研究了基于Wasserstein生成对抗网络与决策树相结合的WGAN-DT故障诊断技术。结果表明,样本平衡时,采用WGAN-DT模型在测试集上的故障诊断准确度高达96.00%。
As a voltage conversion device,a transformer can directly affect the performance of the device if it malfunctions.However,during the operation of the equipment,the transformer system is mostly in a normal state,with a low frequency of faults.The monitored normal state data is far more than the fault state data,resulting in sample imbalance issues.On the basis of transformer system fault diagnosis technology and imbalanced sample processing technology,WGAN-DT fault diagnosis technology based on Wasserstein generative adversarial network and decision tree is studied.The results indicate that the fault diagnosis accuracy of using the WGAN-DT model on the test set is as high as 96.00%when the sample is balanced.
作者
王锦
WANG Jin(School of Electronics and Information,Xi'an Polytechnic University,Xi'an 710048,China;State Grid Xi'an Gao Ling District Power Supply Company,Xi'an 710299,China)
出处
《现代信息科技》
2023年第12期43-47,共5页
Modern Information Technology
关键词
变压器
故障诊断
不平衡数据
生成对抗网络
决策树
transformer
fault diagnosis
unbalanced data
generative adversarial network
decision tree
