论文部分内容阅读
连铸钢坯定重切割,对提高棒材轧制成材率,降低钢坯原材料浪费及能耗具有重要意义。首先针对特定钢流,实验研究了影响钢坯重量的因素,确定了钢坯拉速这一关键影响参数。其次建立了钢坯重量的支持向量机(SVM)回归预测模型,以钢坯拉速作为特征输入量,对钢坯重量进行预测。最后,采用小生境粒子群优化(PSO)方法,优化SVM模型参数,得到PSO优化SVM参数的钢坯重量预测模型。并通过建立神经网络钢坯定重预报模型A及最小二乘钢坯定重预报模型,进行对比研究,研究结果表明,采用小生境粒子群优化后的SVM预报模型可显著提高钢坯定重预报精度。
Continuous casting billet weight cutting, to improve the bar rolling rate, reduce the raw material waste and energy consumption is of great significance. First of all, for a specific steel flow, the factors affecting the weight of the billet were experimentally studied, and the critical influencing parameter of billet pulling speed was determined. Secondly, a support vector machine (SVM) regression prediction model of the billet weight was established, and the billet weight was predicted based on the billet pulling speed as the characteristic input. Finally, PSO (Niche Particle Swarm Optimization) method was used to optimize the parameters of SVM model, and the billet weight prediction model with PSO optimized SVM parameters was obtained. And through the establishment of neural network models of billet weight forecasting model A and least square steel billet weight forecasting model, the results show that using niche particle swarm optimized SVM forecast model can significantly improve the billet weight forecast accuracy.