纸浆洗涤过程的多目标进化优化研究  

Research on Optimization of Pulp Washing Process Based on Multi-objective Evolutionary Algorithm

作　　者：于晓明[1] 王孟效[1] 柏松[1] 

机构地区：[1]陕西科技大学电气与信息工程学院,陕西西安710021

出　　处：《中国造纸学报》2009年第2期92-95,共4页Transactions of China Pulp and Paper

基　　金：咸阳市科研计划项目(xk06015-8)

摘　　要：应用最小二乘支持向量机建立了纸浆洗涤过程中黑液残碱和波美度数学模型,利用非劣排序遗传算法对残碱和波美度进行最小二乘支持向量机模型和BP神经网络模型寻优,获得进浆浓度、进浆流量、清水用量的优化值。其结果用于某造纸厂纸浆洗涤过程的优化,发现最小二乘支持向量机模型多目标进化算法的寻优结果明显优于基于BP神经网络模型的多目标进化算法,并且最小二乘支持向量机模型的多目标进化算法更好地解决了纸浆洗涤过程多个目标的冲突问题。This article applied the least square support vector machine （LS-SVM） to establish the residual alkali and Baume degree mathe- matical model. By using NSGA- Ⅱ to optimize the model of residual alkali and Baume degree based on LS-SVM and ANN, the optimized values of the inlet stock flow and consistency, fresh water inlet flow were obtained. The optimization of pulp washing process based on the results in a paper mill indicated that the model based on LS-SVM is more efficient than based on method of ANN. Consequently, the multiobiective evolutionary algorithm based on support vector machine can well solve the conflict of more than one aim in pulp washing process.

关 键 词：多目标进化算法 最小二乘支持向量机 纸浆洗涤 

分 类 号：TS744[轻工技术与工程—制浆造纸工程]