材料期刊网

针对钢铁企业副产煤气系统产消量频繁波动，不平衡现象比较严重，供需之间的平衡程度对钢铁企业的生产成本、能源消耗情况影响较大，并且钢铁企业中工序、设备繁多，每道工序都涉及多种能源介质的问题，利用HP滤波、支持向量机分类（SVC）、最小二乘支持向量机(LSSVM)和Elman神经网络的特性建立了SVC-HP-ENN-LSSVM模型，并根据用能设备的能源利用特点和预测结果对副产煤气进行优化调度。模型应用表明：所建预测模型对煤气系统的预测平均相对误差小于4%，满足工业生产需要。根据预测结果进行的优化调度解决了煤气系统的不平衡问题，应用于钢铁企业典型工况，主工序可降低10%左右能耗，应用其自备电厂（一年按照330天计算），可多产蒸汽约104148 t，节能约9998208 kg标煤。

In iron and steel enterprises,the volume of byproduct gas system fluctuates frequently,the imbalance phe-nomenon is serious and the byproduct gas balance between supply and demand has enormous influence on the enter-prise's production cost and energy consumption. There are various processes and equipment relating to variety of energy medium. Combined the property with support vector machine classification,the HP filter,Elman neural network and least squares support vector machine were applied to establish the SVC-HP-ENN-LSSVM forecasting model,and the optimization operation was made according to the characteristics of the energy-using equipment,energy utilization and the predicted results. The application of the model showed that the predicted average relative error values of byproduct gas were under the 4% which can meet the requirement of industrial production. The forecast results of optimization scheduling solved the imbalance of gas system,and when it was applied to the steel business typical working,about 10% of main process energy consumes was saved. Assuming there are 330 days operation in a year,the self-provided power plant can produce more than 104 148 t steam which can save 9 998 208 kg standard coal.