基于投影寻踪回归的铜熔炼过程中关键工艺指标预测模型研究

Projection Pursuit Regression-Based Key Process Indictors Predicting Model for Copper Flash Smelting Process

提取金属铜的主要方式是闪速熔炼,由于铜矿资源复杂多变,为保证熔炼效果铜闪速熔炼过程需适应工况。目前铜闪速熔炼过程较难在线检测出冰铜温度、冰铜品位和渣中铁硅比,针对这个难题,提出了基于投影寻踪回归理论的铜闪速熔炼过程三大指标预测策略。首先构建三大工艺指标的投影寻踪回归子模型。并依据实数编码混沌伪并行遗传算法理论优化求解子模型参数。通过实验数据验证模型的预测误差比较小,能够满足实际生产需要。

The main way to extract copper metal is fash smelting. Because of the complexity and variety of copper resources, copper flash smelting process needs to adapt the working conditions to ensure the smelting effect. The copper flash smelting process is diffcult to measure the matte temperature, matte grade and ratio of iron and silicon in slag. For this problem, the three indexes prediction strategy of copper fash smelting process for projection pursuit regression theory based is put forward. First, the projection pursuit regression model of three major technological indexes is constructed. Based on the real coded chaotic pseudo parallel genetic algorithm, the sub model parameters are optimized. The experimental data show that the prediction error of the model is small and can meet the needs of actual production.