钢铁企业自备电厂是副产煤气的主要缓冲用户，在消纳富余煤气、减少煤气放散、实现煤气平衡方面发挥着极为重要的作用。充分考虑自备电厂煤气供入量特点，建立了HP-Elman-LSSVM预测模型，并根据自备电厂能源利用的特点，建立拟合模型求解自备电厂锅炉的经济运行负荷，在此基础上对供入自备电厂的煤气进行优化调度。将该模型应用于具体企业，实现了钢铁企业自备电厂煤气预测和优化调度。模型应用表明：所建模型对自备电厂煤气供入量30、45、60个点的预测平均相对误差分别为1.9%、1.4%、1.4%，能有效解决实际生产中自备电厂煤气供入量预测不准问题。并通过煤气优化调度，自备电厂可大幅度提升蒸汽产率，应用企业每年可多产蒸汽约8.1322万t，折合节约标煤9443.955t。

The iron and steel enterprise self-provided power plant is the main buffer user for by-product gas, which can play a very important role in consuming affluent gas, reduce gas emission and realize the gas balance. An HP-Elman-LSSVM for prediction the supply of self-provided power plant was proposed through fully considering the characteristic of gas supplying. According to the characters of self-provided power plant energy utilization, the economic operating load of the boiler was calculated and the optimal scheduling was carried out. The simulation results show that average forecast relative error of the gas supply of self-provided power plant 30, 45 and 60 points are 19%, 1.4% and 1.4%, respectively. The model can handle the issue of the inaccurate forecast with the high accuracy and smaller average error. The optimal scheduling for self-provided power plant can obtain high steam productivity and the steel enterprises can produce an extra steam 81322t per year, about 9443.955t standard coal.