Short-Term Load Forecasting with an Ensemble Model Using Densely Residual Block and Bi-LSTM Based on the Attention Mechanism

Short-term load forecasting (STLF) is essential for urban sustainable development. It can further contribute to the stable operation of smart grid. With development renewable energy, improving STLF accuracy has become a vital task. Nevertheless, most models based on convolutional neural network (CNN) cannot effectively extract crucial features from input data. The reason that fundamental requirement adopting space invariance, which be satisfied by received data, limiting performance. Thus, this paper proposes an innovative ensemble model comprises densely residual block (DRB), bidirectional long short-term memory (Bi-LSTM) layers attention mechanism, and thinking. Specifically, DRB adopted potential high-dimensional different types such as multi-scale temperature calendar extracted are Bi-LSTM layer. Then, mechanism assign various weights hidden state focus factors. Finally, proposed two-stage thinking improve generalization. experimental results show produce better performance compared existing ones, almost 3.37–5.94%.