Intelligent Fault Location of New Distribution Automation Terminal Based on Machine Learning

After a fault occurs in the distribution terminal, the diverse types of faults, rapid reclosing, and complex transient situations make real-time fault location in the distribution network challenging. Existing fault location methods are vulnerable to factors such as fault type, transition resistance, and branch lines. Traditional distribution terminal fault location methods are difficult to solve this problem, so there is an urgent need for an efficient distribution terminal fault location method. Based on the theory of graph convolutional neural networks, this paper proposes a new method for solving the problem of distribution terminal fault location. The distribution network is treated as non-Euclidean spatial graph data. After abstracting the distribution network as graph data, the feature information on the nodes and edges of the distribution network is jointly used as input to the graph convolutional neural network. At the same time, this paper compares the different types of feature data on the nodes and edges of the distribution network, and there can be more choices based on the degree of automation. For the two forms of data, nodes and edges in the distribution network, an input layer that merges edge features is designed on the node. Then, through the hidden layer, node features are passed to achieve feature abstraction from local to global. The output layer adopts the design of edge aggregation, which transforms the distribution network fault location problem into a graph edge classification problem. The method proposed in this paper achieves a detection accuracy of 98.2%.