基于T-S模糊神经网络的新型烟叶烤房控制系统研究  

Study on new tobacco bulk curing barn control system based on T-S neuro-fuzzy approach

作　　者：李光尧 曹阳[1] 刘维岗[1] 舒成松 张大斌[1] LI Guang-yao;CAO Yang;LIU Wei-gang;SHU Cheng-song;ZHANG Da-bin(College of Mechanical Engineering,Guizhou University,Guiyang 550025,China)

机构地区：[1]贵州大学机械工程学院,贵州贵阳550025

出　　处：《甘肃农业大学学报》2018年第5期217-224,共8页Journal of Gansu Agricultural University

基　　金：中国烟草总公司贵州省公司科技项目(中烟黔科[2017]11号);贵州大学机械工程"技术众筹"研究生创新基地项目(贵大研CXJD[2015]003)

摘　　要：【目的】为了准确控制烟叶烘烤过程的温度和湿度,提高烟叶烘烤品质,设计了一种基于T-S模糊神经网络烟叶烤房控制系统.【方法】从能量守恒和水汽平衡定律出发分析了烟叶烘干过程,建立了双输入双输出的烤房温湿度耦合动态数学模型.结合模糊控制的专家知识表达能力和神经网络自学习特点,设计了基于T-S模糊神经网络控制器.将该控制器与常规PID控制器在Simulink环境下进行仿真对比和烘烤实验对比.【结果】仿真结果表明:T-S模糊神经网络控制器性能优异,其干球温度调节时间和超调量分别减小18s和2.96%,湿球温度调节时间减小了37s.烘烤实验表明:基于T-S模糊神经网络的新型烟叶烤房控制器相对普通控制器,其全过程的干球温度平均误差降低了71%,湿球温度平均误差降低了68%,烤后上等烟比例提高了2.86%,中等烟比例提高了2.46%.【结论】新型烟叶烤房控制系统比普通控制系统能够更好地跟随烘烤工艺曲线,烤后烟叶品质与其所产生的经济效益显著提升.【Objective】In order to control the temperature and humidity of tobacco baking process accurately and improve tobacco baking quality,a bulk curing barn control system was designed based on T-S neuro-fuzzy system.【Method】The drying process of tobacco leaves was analyzed based on the laws of energy conservation and water vapor balance for establishing the dual input and dual output dynamic coupling model of temperature and humidity inside the barn.Combining with the expert knowledge expression ability of fuzzy control and the self-learning characteristic of neural network,the controller based on T-S neuro-fuzzy was designed and compared with the PID controller in Simulink and baking experiment.【Result】The simulation results indicated that the T-S fuzzy neural network controller showed excellent performance,the adjustment time and overshoot of dry-bulb temperature are reduced by 18s and 2.96%,the wet bulb temperature adjustment time was reduced by 37s.Baking experiments showed that the new controller reduced the lower average error of dry-bulb temperature by 71%and that of wet-bulb temperature by 68%compared to ordinary controller.The proportion of top-grade and middle-grade baked tobacco was increased by 2.86%and 2.46%.【Conclusion】The new control system could better meet the baking technology curve and the quality of baked tobacco were significantly improved.

关 键 词：烤房控制器 温湿度耦合 建模 T-S模糊神经网络 

分 类 号：TS441[农业科学—烟草工业]