改进型支持向量回归预测模型的轧机轧制力预测

Rolling force prediction of rolling mill based on improved support vector regression prediction model

对轧机轧制力预测模型进行研究。使用人工鱼群优化算法对支持向量回归(SVR)参数选取进行最优的参数组合,将粒子群优化算法引入到常规人工鱼群算法中,并对其进行改进,提高了人工鱼群算法的性能。研究结果表明:Ekelund模型的轧制力计算结果误差较大,超过了10%,常规SVR预测模型的轧制力预测精度低于10%,而本文研究的改进SVR预测模型得到的轧制力误差低于5%,说明通过人工鱼群算法优化SVR算法模型的参数能够提高预测模型的预测精度,并且预测消耗时间在3种预测模型中是最短的。

Rolling force prediction model is studied. Use artificial fish swarm optimization algorithm for SVR parameters selection of the optimal parameters combination and the particle swarm optimization algorithm is introduced to conventional artificial fish swarm algorithm, and it is improved, to improve the performance of the artificial fish swarm algorithm. Research results show that the rolling force calculation results error of Ekelund model are larger than 10 %. The rolling force prediction precision of conventional SVR forecasting model is less than 10 % ,and the rolling force error obtained by the improved SVR prediction model is lower than 5 %. Through the artificial fish swarm algorithm, the parameters of the SVR algorithm model can improve the prediction precision of the prediction model, and the consumption time is the shortest in the three prediction models.