基于支持向量机的轧机压下系统辨识

通过对轧机压下液压控制系统的介绍,分析和计算压下系统响应频率,设计相应的Butterworth滤波器对轧制力进行高频去噪声处理,并利用最小二乘支持向量机进行轧机轧制力的非线性建模.在Matlab仿真环境中,利用轧制力的实测数据进行仿真与分析.仿真结果表明,基于最小二乘支持向量机的轧制力模型预测精度可以控制在5%范围内.

Rolling force model in rolling mill based on support vector machine

Based on the hydraulic screw-down control system of rolling mill,the response frequency of the screw-down system is analyzed and computed,and the Butterworth filter is designed to filter the noise of the input and output signals.Moreover,this paper presents the system identification for hydraulic screw-down control system in rolling mill to predict the rolling force based on least squares support vector machine(LS-SVM).The comparisons of the simulations by LS-SVM and the real values in the Matlab simulation indicate that the force prediction error is limited to 5%,and the system identification by LS-SVM has good performances and extensive application portents in engineering.