针对具有强动态变化特征但过程信息有限可知的铜转炉吹炼过程,提出一种基于有限数据信息的吹炼动态过程智能集成建模方法.依据冶金反应动力学原理,建立了描述吹炼过程反应体系变化的非线性动力学模型:引入动力学系数修正因子,基于有限的数据信息和龙格一库塔公式,构建了动力学系数修正因子的优化模型;结合智能决策生成的典型样本集,提出了基于微粒群算法和模式搜索法的混合智能算法确保有效获得最优修正因子,最终形成吹炼过程的动态模型.用实际生产数据仿真实验,模型预测的最大相对误差小于5%,仿真结果验证了模型有效性.
For the copper-converting process with significant variation in dynamics and limited information of process, we propose an intelligent integrated modeling method based on its limited data information. First, we develop a nonlinear reaction kinetic model for the copper-converting process by the metallurgical reaction principles; second, we create a model for optimizing its kinetic coefficients based on the limited data and by using the Runge-Kutta formula. Finally, by employing particle swarm optimization and pattern searches with the typical sample set generated by intelligent decision, we put forward a hybrid intelligent algorithm to acquire the optimal kinetic coefficients for the dynamic model of copper converting process. The maximum relative error is less than 5%, when comparing the simulation result of our dynamic model with the real value of the Peirce-Smith(PS) converter in a copper smeltery. This shows the effectiveness of the proposed model.