Fault Line Selection and Location of Distribution Network Based on Improved Random Forest Method

Fault voltage and fault current have nonlinear and timing characteristics. To deeply study the importance of each characteristic variable for fault line selection and fault location of the distribution network, a fault line selection and fault location method based on a random forest (RF) algorithm was proposed. The location function is completed simultaneously based on line selection by improving the random forest sub-model. The proposed network amplifies the single label system to the multi-label task network, which facilitates the better expression of fault data features, improves the feature extraction ability of the network, and dramatically reduces the time of fault diagnosis. In this study, the fault data obtained by Simulink simulation is used as the training set, and the RF model is established based on the Scikit-Learn framework. The results show that this model has a high fault line selection rate and small location error. It can be used as an auxiliary means of distribution network fault diagnosis.