基于RBF-ARX模型的烘丝机出口含水率优化控制方法  

RBF-ARX Model-based Optimal Control Method for Moisture Content in Output Dried Cut Tobacco

作　　者：王小飞[1] 彭晓燕[2] 杨玉波[1] 彭辉[3] 阮文杰[3] 

机构地区：[1]秦皇岛烟草机械有限责任公司 [2]湖南大学机械与运载工程学院,长沙市岳麓山410082 [3]中南大学信息科学与工程学院,长沙市左家垅410083

出　　处：《烟草科技》2014年第1期26-30,共5页Tobacco Science & Technology

基　　金：科技部国家国际科技合作计划项目"若干典型应用系统的优化控制关键技术研究"(2011DFA10440);湖南省科技计划项目"复杂工业系统建模与优化控制方法研究"(2006GK3158)

摘　　要：烘丝机出口烟丝含水率采用PID算法以及前馈补偿等控制方案,各控制回路之间相对独立,协调性差,控制精度低,难以实现完全的闭环自动控制。采用RBF-ARX模型对烘丝过程动态特性进行建模,提出了一种基于模型预测的烘丝机出口含水率控制方法。根据生产过程及生产工艺模式,将烘丝过程分为干头、中间及干尾3个阶段分别采用基于模型的优化设定或预测控制方法进行控制,并综合考虑了筒温、排潮等工艺参数同时变化时对出口含水率的影响,以获得工艺参数的最佳配置。应用效果表明:①实现了烘丝过程的自动化控制,避免了在生产过程中人工干预。②头尾段出口含水率在设定值±0.5%内,中间段出口含水率在设定值±0.2%内且标准偏差≤0.10%,可快速抑制各种干扰引起的含水率波动。③头部干烟丝量小于来料流量的0.27%,尾部干烟丝量小于来料流量的0.28%,降低了烟丝浪费和造碎。基于模型预测控制方法能够获得比筒温和排潮独立调节模式更好的控制效果,满足了工业应用的要求。In PID and feedforward control systems, the control loops for moisture content in dried cut tobacco are relatively independent of each other, low in control precision, poor in coordination and hard to be integrated into a full closed-loop automatic control. A model-based predictive control method was proposed, the model of the dynamic characters of drying process was created by RBF-ARX model. According to production process and technology, drying process was divided into 3 stages （start, middle and finish stages）, and controlled by model-based optimized setting or predictive control method separately. To obtain optimal technical parameters, the combined influence of cylinder temperature and moisture exhaust acted at the same time on the moisture content of tobacco was taken into account. The results of application showed that： 1） The drying process was automatically controlled, and interference became unnecessary. 2） The moisture content in tobacco at start and finish stages was controlled within the set value±0.5%, at middle stage was controlled within the set value±0.2% with the standard diviation of no more than 0.10%, and the fluctuation of moisture content caused by various disturbances could be restrained quickly. 3） The amounts of over-dried tobacco at start and finish stages in a drying operation were less than 0.27% and 0.28% of throughput; the waste and breakage of tobacco reduced. The model-based predictive control method was better than controlling cylinder temperature and moisture exhaust independently, and better fit the requirements of industrial application.

关 键 词：烘丝机 烟丝含水率 RBF—ARX模型 预测控制 干头干尾 

分 类 号：TS432[农业科学—烟草工业]