摘要
在烟丝加工过程中,水分分布受到温度、湿度多个因素的影响,控制系统无法准确反映整体水分情况;为全面提高加工型香烟的质量水平,设计基于LSTM神经网络的烟丝水分恒定控制系统;部署Profibus控制总线,并在线路体系中连接水分检测仪与水分恒定器,完成烟丝水分恒定控制系统的硬件设计;在系统软件设计方面,构建LSTM神经网络单元,根据烟叶吸湿能力分析条件,求解具体的水分分布模型,实现基于LSTM神经网络的烟丝水分模型建模;分别计算烟叶出口湿度与出口温度,并联合传递函数逼近参量与恒定时滞参数,完成对控制参数的整定处理,再联合相关应用部件,实现基于LSTM神经网络的烟丝水分恒定控制系统设计;实验结果表明,LSTM神经网络模型作用下,生丝含水量被稳定控制在13%~18%数值之间,不会因水分过量问题而导致香烟质量水平无法达到实际加工标准。
During the tobacco processing,its moisture distribution is affected by multiple factors such as temperature and humidity,the control system can not accurately reflect the overall moisture situation.In order to comprehensively improve the quality level of processed cigarettes,a tobacco moisture constant control system based on a long short-term memory(LSTM)neural network is designed.this paper deploys Profibus control bus,connects the moisture detection instrument and moisture constant device in the circuit system,and completes the hardware design of the tobacco moisture constant control system.In terms of the system software design,the LSTM neural network unit is constructed.Based on the moisture absorption ability of tobacco leaves,the specific moisture distribution model is solved,and the tobacco moisture model based on the LSTM neural network is achieved.The outlet humidity and outlet temperature of tobacco leaves are calculated respectively,and the transfer function approximation parameters and constant time delay parameters are jointly calculated to calibrate the control parameters.Then,combined with the related application components,the tobacco moisture constant control system based on the LSTM neural network is realized.Experimental results show that under the action of the LSTM neural network model,the moisture content of raw tobacco is stabilized between 13%to 18%,which will not cause the quality level of cigarettes not meeting actual processing standards due to excessive moisture.
作者
王海龙
王新辉
张志勇
朱岩
栾松年
WANG Hailong;WANG Xinhui;ZHANG Zhiyong;ZHU Yan;LUAN Songnian(Qingdao Cigarette Factory,China Tobacco Shandong Industrial Co.,Ltd.,Qingdao 266101,China)
出处
《计算机测量与控制》
2024年第11期177-183,189,共8页
Computer Measurement &Control
