基于ANN和LSSVR的造纸废水处理过程软测量建模

Soft Sensor Modeling of Papermaking Wastewater Treatment Processes Based on ANN and LSSVR

中文关键词:  人工神经网络  最小二乘支持向量回归  造纸废水处理  软测量建模  粒子群优化算法

Key Words:artificial neural network  least squares support vector regression  papermaking wastewater treatment  soft sensor modeling  particle swarm optimization

中文摘要:

      针对造纸废水处理系统的时变性、非线性和复杂性等特点,将人工神经网络(ANN)和最小二乘支持向量回归(LSSVR)分别用于造纸废水处理过程中的软测量建模,实现造纸废水处理过程中出水化学需氧量和出水悬浮固形物浓度的预测。ANN采用误差反向传播算法建模,LSSVR通过粒子群优化算法进行模型参数优化。结果表明,与ANN模型预测结果相比,LSSVR模型预测结果的均方根误差降低了50%以上,相关系数提高了近10%,表明LSSVR模型在造纸废水处理过程中的预测精度高于ANN模型。

Abstract:

      Concerning the time-varying, nonlinear, and complex characteristics of papermaking wastewater treatment systems, soft sensor modeling methods based on artificial neural network (ANN) and least squares support vector regression (LSSVR) were used to predict effluent chemical oxygen demand and suspended solids in a papermaking wastewater treatment process. ANN model was established by using error back propagation algorithm. The particle warm optimization was used to optimize model parameters in the LSSVR model. The results showed that the root mean square error of LSSVR model reduced by more than 50% compared with that of ANN model, and the correlation coefficient of LSSVR model increased by about 10% compared with that of ANN model. These results indicated that the LSSVR model had better prediction performance and higher accuracy compared to the ANN model in papermaking wastewater treatment process．