Data Mining Based Power System Fault Type Prediction Method

Currently used power system fault diagnosis methods have the problem of low accuracy and efficiency while classifying and predicting the types of faults. In recent years, big data technology has been applied to power system failure classification and prediction. This paper proposes a predictive model of each fault type based on using different data mining algorithms three times: First, cluster the original failure data samples by using a K-means clustering algorithm and simplify them with a self-encoding method. Second, it then uses Apriori algorithm to extract some potential rules, and mine the data that meets the legal requirements to form training datasets. Third, the stochastic gradient descent algorithm is applied to train a prediction model for each fault type. Finally, a simulation example shows that the proposed method results in high accuracy of fault classification and prediction. Irrelevant data is eliminated by this method to reduce over-fitting.