EMD-CPO-GRU-based Transformer Oil Temperature Prediction DOI

Pengbo Han

Published: Feb. 27, 2025

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

Language: Английский

EMD-CPO-GRU-based Transformer Oil Temperature Prediction DOI

Pengbo Han

Published: Feb. 27, 2025

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

Language: Английский

Citations

0