Short-Term Power Load Forecasting Based on an EPT-VMD-TCN-TPA Model

Accurate short-term load forecasting is the key to ensuring smooth and efficient power system operation market dispatch planning. However, nonlinear, non-stationary, time series nature of sequences makes difficult. To address these problems, this paper proposes a method (EPT-VMD-TCN-TPA) based on hybrid decomposition sequences, which combines ensemble patch transform (EPT), variational modal (VMD), temporal convolutional network (TCN), pattern attention mechanism (TPA). In which, trend component (Tr(t)) residual fluctuation (Re(t)) are extracted using EPT, then Re(t) decomposed into intrinsic function components (IMFs) different frequencies VMD. The Tr(t) IMFs fused meteorological data predicted separately by TCN-TPA prediction model, finally, results each reconstructed superimposed obtain final value load. addition, experiments after reconstructing IMF according fuzzy entropy (FE) values discussed in paper. evaluate performance proposed paper, we used datasets from two Areas 9th Mathematical Modeling Contest China. experimental show that predictive precision EPT-VMD-TCN-TPA model outperforms other comparative models. More specifically, had MAPE 1.25% 1.58% Area 1 2 test sets, respectively.