基于预测模型和遗传算法的配煤优化研究

焦化企业配煤和炼焦过程是存在诸多不确定性、无法用数学模型描述的复杂工业过程,传统控制方法难以实施控制。因此要实现配煤成本的最优控制是个比较复杂的问题。本文在焦化理论和实际生产所获的数据基础上,以神经元网络为指导,建立焦炭质量预测模型;利用单种和混合煤中各组分的关系,建立起混合煤的质量预测模型。在以上两个模型的基础上,把炼焦过程中配煤成本最小化的问题转化成为带约束的最优化问题。再利用遗传算法可以比较方便地求得近似最优解。本文利用实际数据和仿真实验,验证方法的可行性。

Research on Coal Blending Based on Prediction Model and Genetic Algorithm

There are lots of uncertainties in the process of coal blending and coking.It＇s hard to describe the complicated process using exact mathematical model and the traditional control method is difficult to be implemented in such process.To optimize the cost of coal blending is a complicated problem.A coke quality prediction model of neural networks based on coking theory and the analysis data is created.Another prediction model of blended coal quality is also put forward based on the quality analysis of the single coal and the blended coal.Based on the two prediction model,a genetic algorithm is employed to minimize the coal blending cost.The method is proved to be effective by applying the process analysis data to the simulation model.