Prediction Method of Top Oil Temperature for Distribution Transformers Based on ARMA-SSA-BP

The top oil temperature of transformers serves as a crucial indicator for assessing their performance and operational health status. Accurate prediction of this temperature can effectively monitor transformer operation, identify potential faults in advance, and ensure the safe and stable operation of power equipment. A combined forecasting method based on the autoregressive moving average (ARMA) model, sparrow search algorithm (SSA), and back propagation (BP) neural network (ARMA-SSA-BP) is adopted to predict the top oil temperature of distribution transformers. Initially, ARMA is used to perform time series prediction on the top oil temperature of transformers, generating preliminary predictions. Subsequently, the actual predicted values, three-phase currents, and ambient temperature are taken as input features, and a BP neural network optimized by the SSA is introduced for deep learning and optimized prediction of historical data. Finally, simulation analysis is conducted using actual operational data from a substation area in Jiangning. It has been verified that this model can effectively improve prediction accuracy.