Research on Failure Rate Prediction of Power Equipment Based on Random Forest

To effectively estimate the failure rate of medium-voltage distribution network equipment, this study proposes a Random Forest-based model for fault rate estimation. Utilizing equipment ledger and operational data as the foundation, a Random Forest model was developed and optimized through parameter tuning, selecting the optimal number of decision trees. The experimental results demonstrate that the Random Forest-based model achieves an accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{\boldsymbol{2}}=\boldsymbol{0.968}$</tex> and RMSE = 0.095. To further validate its stability, a 1:1 ratio of training and validation data was used for additional modeling. Despite a slight decrease in precision, the Random Forest algorithm continues to exhibit robust predictive performance and stability. Compared to traditional methods, this model provides more accurate predictions of equipment failure rates under specific conditions, offering precise reference data for maintenance and replacement decisions in power equipment.