A study on long-term industrial electricity consumption and carbon emission forecasting methods based on TimesNet models

Abstract Aiming at the prediction problem of industrial electricity consumption and carbon emission, a TimesNet model is established to improve the accuracy of industrial electricity consumption and carbon emission prediction. Five significant influencing factors were screened out by grey correlation analysis; based on the energy consumption dataset of an iron and steel enterprise in 2023, the dataset was preprocessed and inputted into the TimesNet model, the prediction results were integrated and summarized, and the prediction results of the DLinear, LightTS, PatchTST, and iTransformer time series prediction models were compared and analyzed. The results show that the predicted values of the TimesNet model fit the actual values, and its mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), and mean absolute percentage error (MAPE) are 0.31, 0.36, 0.6, and 1.12, respectively, which are better than those of the other models. It proves that the model can predict industrial electricity usage and release of carbon more accurately, providing a practical method for industrial electricity usage and release of carbon prediction.