论文部分内容阅读
影响高炉铁水硅含量的因素往往复杂多变,影响程度不一。采用鱼骨分析法收集所有可能对硅含量产生影响的因素,经过相关分析和特征选择,最终选取6个参数作为模型的输入参数。采用改进的粒子群优化算法对支持向量机(SVM)中的参数进行优化,提出基于变邻域粒子群(VNPSO)优化SVM的铁水硅含量预测模型。通过钢厂的实际生产数据进行验证,平均相对误差达到0.69%,平均绝对误差达到3.4×10~(-3),模型具有很高的预测精度。同时,绘制铁水中硅含量控制图,分析硅含量波动情况,并依此模型给出硅含量稳定性控制措施。
The factors that affect the silicon content of blast furnace hot metal are often complex and changeable, with varying degrees of influence. Fishbone analysis was used to collect all the factors that may affect the silicon content. After correlation analysis and feature selection, six parameters were finally selected as the input parameters of the model. An improved Particle Swarm Optimization (PSO) algorithm was used to optimize the parameters in Support Vector Machine (SVM), and a predictive model based on variable particle swarm optimization (VNPSO) to optimize SVM was proposed. Through the actual production data of steel mills, the average relative error is 0.69% and the average absolute error is 3.4 × 10 -3. The model has high prediction accuracy. At the same time, draw the control chart of silicon content in hot metal, analyze the fluctuation of silicon content, and give the control measures of silicon content stability according to this model.