EMD-CPO-GRU-based Transformer Oil Temperature Prediction

To improve the accuracy of transformer oil temperature prediction and ensure the stability and safety of transformers during operation, this paper proposes an innovative oil temperature prediction method—an EMD-CPO-GRU hybrid model based on Empirical Mode Decomposition (EMD), Crested Porcupine Optimization (CPO) algorithm, and Gated Recurrent Unit (GRU). The method first decomposes the oil temperature data using EMD, effectively extracting the nonlinear and non-stationary characteristics of the signal, thereby providing more representative and effective features for subsequent predictions. Next, the CPO algorithm is applied to optimize the key hyperparameters of the GRU model, establishing efficient CPO-GRU sub-models for each modal component to improve the accuracy and robustness of the prediction model. Finally, the prediction results of each sub-model are weighted and integrated to obtain the final oil temperature prediction value. Experimental results show that the EMD-CPO-GRU model outperforms traditional prediction models and other hybrid models in transformer oil temperature prediction tasks. In terms of prediction accuracy, the EMD-CPO-GRU model achieves significant improvement, fully verifying its effectiveness as an efficient and precise transformer oil temperature prediction method. This approach not only provides a reliable basis for real-time monitoring and fault warning of power transformers but also offers new ideas and solutions for similar time-series prediction problems.