基于某厂实际铜闪速熔炼工艺和控制过程,对神经网络模型在铜闪速熔炼过程在线控制进行了研究.在分析影响溶剂率、熔炼氧单耗、反应塔总风量操作参数因素的基础上,提出一种基于BP神经网络的操作参数的预测方法,分别建立了输入向量只包含主要元素和考虑杂质元素的BP神经网络模型.网络的训练和测试结果表明,两种神经网络的输出值与实际值的最大相对误差均小于1.0％,输出值与实际样本值吻合得较好,模型输入参数中包括杂质元素时具有更高的计算精度.

Unsteady flowing about the flow around high temperature metallurgical slag particle of the three –dimensional was simulated, and the heat transfer and resistance characteristics of slag particles in different Streaming Reynolds number were studied. The results show that the heat transfer of slag particles is nonuniform when the slag particles are cooled in air, windward of the slag exchanges heat quickly, circumfluence of leeward enhances heat transfer, and the temperature is higher. The average Nusselt number (Nu) of slag particles increases with increasing of streaming Reynolds number (Re), reduces with decreasing of the particle diameter, and increases with increasing of air velocity. Under the same conditions, the effect of the diameter on average Nusselt number is strongen than that of air velocity. Streaming Reynolds numberis fitted as the nonlinear power function by the average Nusselt number (Nu), and the three dimensionless correlations can be gotten. The drag coefficient of slag particles increases with increasing of Streaming Reynolds number, andultimately tends to constant. Air velocity should be controlled well. The Streaming Reynolds number is fitted as the nonlinear power function by the drag coefficient, and the correlations of the drag coefficient are gotten.

由于合成炉余热锅炉的烟气处理能力是合成炉提产的主要限制因素之一,通过对合成炉余热锅炉进行热平衡测试与分析,提出一种热平衡迭代算法计算余热锅炉的热平衡特性.计算结果表明:在合成炉处于正常连续生产的稳定工况下,余热锅炉有效热利用系数和综合换热系数可以视为常数;对数平均温差与出口烟气温度关系可回归为线性关系;通过热平衡迭代算法,得出满足工艺要求时合成炉最大烟气量与余热锅炉出口烟气温度的关系,计算结果显示采用该算法能预测余热锅炉热力性能,有利于铜合成炉稳定生产,提产增效.