摘要
为了提高板带热连轧轧制力预测精度,建立了利用粒子群优化的混合核最小二乘支持向量机(PSOLSSVM)的预测模型,并将最小二乘支持向量机模型与传统数学模型进行组合,得到组合模型,以进一步提高预测精度。通过采集现场数据,对模型进行训练并离线仿真。结果表明,PSO-LSSVM有更强的学习能力和泛化能力,预测精度得到很大提高,该方法在实际应用中具有很大潜力。
The paper establishes a prediction model,which uses particle swarm optimization optimized least square support vector machine( PSO-LSSVM) with mixed kernel function,to improve precision of rolling force prediction in hot strip rolling,and then a combined model which combined least square support vector machine and tradition mathematical model mathematical model is built to further improve the prediction accuracy. By collecting field data,the model is trained and an off-line simulation is performed. The result shows that PSO-LSSVM possesses stronger learning ability and generalization ability,the precision of prediction is greatly improved,and the method has a strong application potential.
引文
[1]孙一康.冷热轧板带轧机的模型与控制[M].北京:冶金工业出版社,2010.
[2]王秀梅,吕程,王国栋,等.轧制力预报中的神经网络与数学模型[J].东北大学学报,1999,20(3):97.
[3]丁敬国,胡贤磊,焦景民,等.变异PSO算法协同神经元网络在轧制力预报中的应用[J].钢铁研究学报,2007,19(12):56.
[4]VLADIMIR N Vapnik.The Nature of Statistical Learning Theory[M].New York:Springer-Verlag,1999.
[5]陈治明,罗飞,黄晓红,等.基于混沌优化支持向量机的轧制力预测[J].控制与决策,2009,24(6):808.
[6]刘俊,杨荃,何安瑞,等.贝叶斯框架下的LS-SVM在热连轧机轧制力预报中的应用[J].钢铁研究学报,2011,23(6):5.
[7]SUYKENS J A K,VANDEWALLE J.Least squares support vector machineclassifiers[J].Neural Processing Letters,1999,9(3):293.
[8]汪海燕,黎建辉,杨风雷.支持向量机理论及算法研究综述[J].计算机应用研究,2014,31(5):1281.
[9]宋勇,苏岚,荆丰伟,等.热轧带钢轧制力模型自学习算法优化[J].北京科技大学学报,2010,32(6):802.
[10]郭韬.热连轧精轧轧制力预报模型研究及优化[J].冶金自动化,2014,38(2):48.
[11]魏立新,李兴强,李莹,等.基于自适应遗传算法的冷连轧轧制力模型自学习[J].轧钢,2010,27(3):7.
[12]唐奇,王红瑞,许新宜,等.基于混合核函数SVM水文时序模型及其应用[J].系统工程理论与实践,2014,34(2):521.
[13]CLERC M,KENNEDY J.Particleswarm-explosion,stability and convergence in a multidimensional complex space[J].IEEE Transactions on Evolutionary Computation,2002,6(1):44.