关键词: 煤炭消费预测, 煤炭消费结构, 组合模型, 消费趋势

Abstract: To study coal consumption of China and that of thermal power, metallurgy, building materials and chemical industry, combined forecasting models of major industries are built based on GM, ARIMA, ANN and LOGISTIC. To predict the coal consumption trend of major industries in 2020-2030, the optimal models of the combined forecasting models are selected through parameter evaluation of R2, MAE, MAPE and RMSE. The results show that parameter evaluation of R2, MAE, MAPE and RMSE of the optimal models are better than those of the single models; The combined model GM-ARIMA of China coal consumption with weight of (0.32,0.68), GM-LOGISTIC-ARIMA for thermal power industry with weight of (0.28,0.14,0.58), GM-LOGISTIC of metallurgy industry with weight of (0.40,0.60), ANN-ARIMA of building materials industry with weight of (0.32,0.68) and ANN-ARIMA of chemical industry with weight of (0.79,0.21) are built respectively. It is predicted that the total consumption of coal and the consumption of thermal power industry will increase slightly in the future, reaching 4.167 billion tons and 2.210 billion tons in 2030; the consumption of metallurgy and building materials industry will be stable, reaching 660 million tons and 504 million tons in 2030; the consumption of chemical industry will increase rapidly, reaching 378 million tons in 2030.