%0 Journal Article
%T 基于改进迁移学习的电能质量扰动分类方法<br>Power Quality Disturbance Classification Method Based on Improved Transfer Learning
%A 孙中海
%A 吴杰康
%A 彭和平
%A 莫文雄
%A 王勇
%A 栾乐
%A 许中
%J Smart Grid
%P 63-70
%@ 2161-8771
%D 2023
%I Hans Publishing
%R 10.12677/SG.2023.133006
%X 针对电能质量扰动信号的分类问题，本文提出一种基于格拉姆角场与多重迁移学习的电能质量扰动分类方法。首先利用格拉姆角场将一维电能质量扰动信号转化为GAF编码图像，然后构造三个ResNet子模型网络，选用具有代表性的信噪比为0 dB、20 dB、40 dB的扰动信号作为子模型的输入分别训练三个子模型，期间采用多重迁移学习的方法，将子模型的训练权重依次传递迁移，使得后一个模型的预训练权重继承自上一个模型的训练权重，并采用部分冻结与部分微调的权重处理方式保证模型具有最优的训练效果。最后将三个子模型的特征进行融合后训练全连接层分类器，最后获得完整的电能质量扰动分类模型。仿真验证该方法具有良好的分类准确度与抗噪性能，所提模型具有良好的鲁棒性与泛化性。<br />
Aiming at the classification of power quality disturbance signals, a power quality disturbance classification method based on gram angle field and multiple transfer learning is proposed in this paper. Firstly, the one-dimensional power quality disturbance signal is transformed into GAF coded image by using gram angle field, and then three RESNET sub model networks are constructed. The disturbance signals with representative signal-to-noise ratios of 0 dB, 20 dB and 40 dB are selected as the input of the sub model to train the three sub models respectively. During this period, the training weights of the sub models are transferred in turn by using the method of multiple transfer learning. The pre training weight of the latter model is inherited from the training weight of the previous model, and the weight processing methods of partial freezing and partial fine-tuning are adopted to ensure the optimal training effect of the model. Finally, the features of the three sub models are used to train the full connection layer classifier, and finally a complete power quality disturbance classification model is obtained. Simulation results show that the method has good classification accuracy and anti noise performance, and the proposed model has good robustness and generalization.
%K 迁移学习，ResNet网络，格拉姆角场，特征融合，电能质量扰动<br>Transfer Learning
%K ResNet Network
%K Gramian Angular Field
%K Feature Fusion
%K Power Quality 
Disturbance
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=68219