基于WNN的亚法糖厂澄清过程工艺指标建模  

Modeling of Process Index for Sugar Mill Clarification Based on WNN

作　　者：宋绍剑[1] 曾令丰[1] 林小峰[1] 

机构地区：[1]广西大学电气工程学院,广西南宁530004

出　　处：《计算机仿真》2013年第9期327-330,360,共5页Computer Simulation

基　　金：国家自然科学基金资助项目(60964002);广西自然科学基金项目(0991057)

摘　　要：研究亚硫酸制糖生产中澄清过程工艺指标的优化预测问题。对亚法糖厂澄清过程操作参数进行优化可以确保各项工艺指标满足生产要求,提高白糖的品质和产量,但目前各项工艺指标无法全部实现在线检测,且人工化验滞后时间长,难以及时对生产过程的操作参数进行优化。为了解决上述问题,根据澄清过程的生产工艺要求和大量的现场数据,建立用小波神经网络(Wavelet Neural Network,WNN)的澄清过程工艺指标预测模型,并与BP网络模型进行了性能比较分析。仿真结果表明:基于WNN的糖厂澄清过程工艺指标预测模型,可以达到对澄清工程工艺指标的预测作用,进行预测精度高、收敛速度快,优化预测效果比BP模型好。Clarification process is one of the most important processes of cane sugar produced in sulfitation process sugar mill,the optimization of the clarification process operating parameters can ensure that each technical index meets the production requirements,so that to improve the quality and yield of sugar.But it is hard to achieve operating parameters optimize immediately because not all index can be obtained online and artificial test takes long time.In order to solve the above problems,according to the principle of clarification process and the massive field data,the wavelet neural network （WNN） model was built,and compared with the BP one.Simulation results show that the WNN model has better precision and more fast convergence rate than the BP model.The WNN model can satisfy the sugar mill production requirements and attain clarification process technical index prediction effect.

关 键 词：澄清过程 预测模型 小波神经网络 工艺指标 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]