High Energy Load Prediction of Industrial Parks Based on TCN-BiGRU-Attention

The high power demand of industrial parks poses significant pressure on the grid, especially the volatility and uncertainty of high energy-consuming loads, which poses challenges to grid stability. In order to cope with this challenge and improve the stability and accuracy of power grid operation, accurate short-term power prediction is particularly important. To this end, we propose an innovative short-term power prediction model that combines Temporal Convolutional networks (TCN) and bidirectional Gated recurrent units (BiGRU) with an attention mechanism. Through TCN, we can effectively mine the association between multi-dimensional time series data and extract high-dimensional time series features. At the same time, the residual module is used to speed up the convergence of the model. In order to further enhance the model's ability to mine bidirectional temporal relationships, we establish BiGRU and introduce a temporal attention module at its output, which helps to improve the influence of important historical temporal information. The experimental results show that our model achieves significant results in industrial park power prediction. Specifically, the mean Absolute error (MAE) is 221.5, the Mean Absolute percentage error (MAPE) is 3.160/0, and the coefficient of determination (R-squared) is 0.676. Compared with other single models and combined models, our model shows higher accuracy, which verifies its efficiency in short-term power prediction. This achievement is of great significance for improving the stability and accuracy of power grid operation.