Intelligent fault diagnosis framework of microgrid based on cloud–edge integration
Weidong Chen, Bin Feng, Zhiguang Tan, Ning Wu, Song Fen
Abstract
This paper proposes an intelligent diagnosis framework of microgrid based on cloud–edge integration. First, the digital twin model of the microgrid is established on the cloud server. Based on the model, the operation data of the microgrid in various conditions can be obtained. Then, the neural network-based fault diagnosis model is trained on the cloud server by using the data provided by the digital twin model. Next, the trained neural network is downloaded to the edge device for the offline fault diagnosis of the microgrid. The proposed method is implemented based on the well-known digital twin platform CloudPSS and test results demonstrate the effectiveness. Extensive tests have been conducted on this framework using fully connected neural network algorithms with an accuracy rate of over 95%.