基于WA和LS-SVM的净化过程钴离子浓度预测

Prediction of Cobalt Ion Concentration in the Purification Process Based on WA and LS-SVM

针对硫酸锌溶液净化过程具有多变量耦合、强非线性、大时滞性等特点和过程数据具有高噪声、非平稳等特性,提出了一种结合小波分析和最小二乘支持向量机方法的净化过程钴离子浓度预测方法;该方法通过小波分解,将钴离子浓度序列分解得到不同的高频和低频序列,对分解得到的各序列重构并采用LS-SVM方法进行预测,最后将各预测结果合成得到钴离子浓度的预测值;仿真结果表明,该模型具有较高的预测精度,能为生产操作提供有益的指导。

Aiming at the purification process feature of multi-variable coupling, non-linear and large time delay and the high noise and non-stationary of the process data, a prediction model based on wavelet analysis （WA） and least squares support vector machine （LS-SVM） is proposed to enhance the prediction precision of cobalt ion concentration. Firstly, the original data series of cobalt ion concentration is decomposed to different series by WA. Then, each decomposed series is reconstructed and predicted by LS-SVM. Finally, the prediction result is obtained by reconstruction of the prediction results of LS-SVM models. The simulation results show that the predictive results are close to the practical values. The model possesses higher precision and practicability.