Transmission Line Fault Detection Based on Multilayer Perceptron

The complex structure and unit composition of the transmission line routing can operate stably in a balanced state. However, due to some natural or human factors, such as strong winds, earthquakes, animal activities and artificial misoperation, the circuit units of transmission lines are often affected by external factors and cause faults. Transmission line faults will bring hidden dangers. Under the tide of intelligence, it is necessary to develop a fault detection system for transmission lines based on intelligent algorithms in order to clear the line faults as soon as possible. Machine learning technology can effectively establish the fault detection system. In this paper, we develop a transmission line fault detection system based on a multi-layer perceptron, and compared it with a variety of machine learning algorithms to highlight the advantages of the multi-layer perceptron. This paper conducts experimental verification on a data set of more than 12,000 data with three-phase current and voltage as input. Experiments show that the multi-layer perceptron algorithm has higher accuracy than random forest, decision tree and SVM methods. Multi-layer perceptron can improve 0.6% compared to random forest, 2.7% compared to decision tree, and 1.9% compared to SVM