Power Load Forecasting Based on Deep Learning

Aiming at the problem that the prediction methods of statistical analysis and machine learning are difficult to take into account the timing and nonlinearity of load data at the same time, the convolutional neural network is selected to optimize the bidirectional long short-term memory network model of attention mechanism, and the CNN-BiLSTM-Attention short-term power load prediction model is established. The establishment of the bidirectional long-term and short-term memory network model of the attention mechanism can reduce the loss of historical information and increase the impact of important information. On this basis, the convolutional neural network model is used to optimize it, which can reduce the error caused by human intervention, thereby improving the accuracy of short-term power load forecasting. The experimental results show that the RMSE value of the model used in this paper is 46.8022, the MAE value is 33.7006, the value is 0.9899, and the MAPE value is 2.8227 %. The CNN-BiLSTM-Attention model has better effect and higher accuracy on short-term power load forecasting, in order to provide some reference for forecasting research in other fields.