基于机器学习的气流烘丝出口水分的预测模型的建立与研究

Establishment and Research of Prediction Model of Air Flow Drying Wire Outlet Moisture Based on Machine Learning

摘要: 在卷烟制丝加工过程，气流丝、梗丝是卷烟配方中必不可少的填充物。气流烘丝设备作为制丝加工过程的重要工序点，主要是利用高温气流对回潮后的叶丝进行高强度处理，有效的降低香烟的焦油含量，和使烟丝能够有较好的膨胀作用和青杂气去除效果。整个气流烘丝的特点就是：工艺气体温度高，热量快速传递到物料中，干燥时间短，瞬间脱水膨胀，膨胀效果明显。所以说能够较好的控制烘丝出口水分对整个制丝过程有着举足轻重的作用。某卷烟工厂在制丝气流烘丝环节，为了满足气流烘丝出口的冷床出口水分，烘丝出口水分值的控制稳定性以及烘丝出口水分值的预测就显得尤为的重要，本文通过基于机器学习的气流烘丝出口水分的预测模型的建立，为生产过程提供准确的出口水分预测。

Abstract: In the process of cigarette silk making, air flow silk and stem silk are essential fillers in cigarette formulations. Air drying equipment, as an important process point in the process of silk processing, mainly uses high temperature air to carry out high strength treatment on the leaf silk after mois-ture return, effectively reduce the tar content of cigarettes, and make the tobacco can have a better expansion effect and green gas removal effect. The characteristics of the whole air drying wire are: High temperature of the process gas, rapid transfer of heat to the material, short drying time, instant dehydration and expansion, and obvious expansion effect. Therefore, the ability to better control the drying outlet moisture has a decisive role in the whole process of silk making. In the process of wire making and air drying in a cigarette factory, in order to satisfy the moisture at the cold bed outlet of air drying, the control stability of the moisture value at the wire drying outlet and the prediction of the moisture value at the wire drying outlet are particularly important. In this paper, the establishment of the prediction model of air drying outlet moisture value based on machine learning is used to provide accurate prediction of the moisture at the production process.