一种多种群广义遗传CMAC的软测量模型  

A Soft Sensor Model by Multi-Population Generalised Genetic CMAC

作　　者：王华秋[1] 姜群[1] 

机构地区：[1]重庆理工大学计算机学院,重庆400050

出　　处：《控制工程》2011年第5期664-667,702,共5页Control Engineering of China

基　　金：重庆自然科学基金资助项目(CSTC2007BB2406);中国铝业贵州分公司技改项目(2008Q26);重庆市教委科学研究项目(KJ100805);教育部人文社会科学研究青年基金项目(10YJC870037)

摘　　要：对CMAC的惯性系数和学习率进行了优化,提出了基于广义遗传优化的小脑模型神经网络(CMAC)算法,提高CMAC的计算速度和精度以满足复杂动态环境下的非线性实时控制的需要。结合溶出预脱硅系统工艺优化的需求,提出了基于广义遗传优化的CMAC的溶出赤泥A/S比系统软模型,用于准确实时地预测溶出赤泥A/S比。试验说明了该模型在对化工软计算的预测精度和快速性上具有明显的优越性,在某氧化铝厂工艺优化系统中的应用,提高了溶出的生产效率和指标。The cerebella model articulation controller （CMAC） based on generalized genetic optimization to adjust the parameters of in-ertia factor and） learning rate is presented. The generalized genetic algorithm CMAC improves the speed and accuraey of calculation to meet the complex and dynamic demand in the non-linear environment for real-time control. Consider the combination of the demand of the process optimize of digesting pre-silicon systems, a soft sensor model of A/S ratio of red mud by the generalized genetic algorithm CMAC is presented to forecast the A/S ratio of red mud accurately and quickly. The quantity of recycled liquor is optimized in this application. Industry test shows that the accuracy and rapid of the soft sensor model has obvious advantages than other algorithm. And the model has been applied to certain alumina plant to optimize dynamically the recycled liquor quantity.

关 键 词：赤泥A/S比 软测量模型 广义遗传优化 小脑模型神经网络 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]