Abstract
With the continuous expansion of power system scale and advancements in intelligence, the accuracy and timeliness of busbar fault diagnosis-an essential component of the power system-are crucial for ensuring the safe and stable operation of the grid. This paper presents a method for busbar fault diagnosis and analysis that combines the weighted mean of vectors (INFO) algorithm with the Random Forest (RF) model. Building on the accurate identification of busbar fault types, the method further predicts fault resistance. A simulation model of a dual-busbar power system is first established, and key electrical quantities such as differential current, bus tie current, and voltage are extracted to quantify fault features using Root Mean Square (RMS) values. The RF model is then used to predict fault types and fault resistance, with the INFO algorithm iteratively optimizing the hyperparameters of the RF model to further improve prediction accuracy. Experimental results show that the INFO-RF model achieves an accuracy of 98.472% on the test set, significantly outperforming traditional methods such as BP neural networks, GRNN, and decision trees. This method not only accurately identifies busbar fault types but also predicts fault resistance, providing strong support for fault location and maintenance in power systems.