基于DCAE-KSSELM的变压器故障诊断方法

为了充分利用变压器发生故障时产生的大量无标签样本，提高故障诊断精度，提出基于深度收缩自编码器(DCAE)与核半监督极限学习机(KSSELM)相结合的故障诊断方法。首先使用无标签样本对DCAE网络逐层训练，初始化网络参数，然后用有标签样本数据对网络参数进行微调，最后将有标签样本与无标签样本一起作为深度收缩自编码器与核半监督极限学习机(DCAE-KSSELM)混合网络的输入并完成故障诊断。实验结果表明，所提模型稳定性好，故障诊断精度高，鲁棒性强。

Transformer Fault Diagnosis Method Based on DCAE-KSSELM

In order to make full use of the large number of unlabeled samples generated during transformer fault and improve the accuracy of fault diagnosis,an innovative fault diagnosis method is proposed based on the combination of deep contractive autoencoder(DCAE)and kernel semi-supervised extreme learning machines(KSSELM).First,the unlabeled samples are used to train the DCAE network layer by layer and initialize the network parameters.Then the labeled samples are used to fine-tune the network parameters.Finally,the labeled samples and unlabeled samples are used as the inputs of the hybrid network of DCAEKSSELM to make the fault diagnosis.The experimental results show that the proposed hybrid model has good stability,high fault diagnosis accuracy and strong robustness.