基于机器学习方法的磨浆过程打浆度软测量及其实现

Beating Degree Soft-sensing Based on Machine Learning

打浆度是衡量磨浆过程质量达标的一个重要因素,针对其难以在线测量,引入基于BP网络和支持向量机的打浆度软测量方法。仿真结果表明,支持向量机比BP网络的泛化能力强,此方法更能反映打浆度的真实值。

Considering the beating degree importance in pulp quality assessment,a beating degree soft-sensing method based on BP network and support vector machine（SVM） was proposed,the simulation results show that SVM generalization ability outperforms that of the BP network,and this method can reflect the real values of beating degree.