借鉴最小二乘支持向量机求解的思路,文中提出了辨识多输入-单输出Hammerstein—Wiener模型的方法.引入共线性约束假设,将辨识问题转换为有约束的优化问题,从而辨识出Hammerstein—Wiener模型的参数.基于Hammerstein—Wiener模型,我们建立了一个多输入-单输出的滚动预测模型,对铜转炉造渣S2期吹炼所需总氧量进行了预测,其相对均方根误差为12.1%.仿真结果表明,该模型预测准确、具有较好的应用价值.
The identification method for a multi-input single-output Hammerstein-Wiener model is proposed by using the solving method of the least-squares-support-vector machine. The identification problem is converted into a constrained optimization problem by assuming collinear constraints so that the parameters of Hammerstein-Wiener model can be identified. Based on the Hammerstein-Wiener model, a multi-input single-output receding-horizon prediction model is developed for predicting the total oxygen quantity required by a copper converter in slag making S2 stage. The relative root-mean- square error (RRMSE) is 12.1%. The simulation research shows that this model provides accurate prediction and is with desirable application value.