Fault Analysis of Intelligent Substation Secondary System Based on Improved Random Forest

To solve the problem of low efficiency of manual analysis due to noise data in secondary system fault analysis of intelligent station, a fault data analysis method based on particle swarm optimization and improved random forest was proposed. Before the analysis, the fault data of the secondary system is collected first, and then the specific communication state of each loop and each terminal is determined by combining the secondary loop resolved by the SCD file, setting the fault feature values to attribute values, the analysis results to labels, using the improved random forest algorithm carries on the classification, combined with particle swarm algorithm optimize the key parameters, Weight is set for each tree according to the classification accuracy of each tree, and finally, the data is classified by voting. Through the test set verification, the improved method has higher accuracy and stronger anti-noise ability compared with the traditional random forest, SVM, and BP neural network, and improves the efficiency of fault analysis.