烘丝工艺参数对出料水分的影响研究——基于分位数回归的实证分析

Study on the Influence of Drying Process Parameters on the Moisture Content of the Discharged Material—Empirical Analysis Based on Quantile Regression

摘要: 为精准表征烘丝段工艺参数对出料水分的影响关系，针对出料水分不具备正态性且存在右偏的问题，以制丝过程全批次数据为研究对象，构建重要分位点的分位数回归模型，并与普通最小二乘回归模型进行比较分析。研究表明，切叶丝水分、工艺气速度和筒壁温度是影响出料水分的重要因素。同时，各因素对出料水分影响程度随出料水分分位点的变化而变化，切叶丝水分的影响随分位点增高呈现“U”型变化趋势，工艺气速度的影响随分位点增高有明显的下降趋势，而筒壁温度的影响随随分位点增高而上升。5%分位点的分位数回归模型可表征出料水分在生产过程技术标准中心附近各变量的影响程度，为出料水分精准控制的参数设置提供了重要的理论依据。

Abstract: To accurately characterize the influence of process parameters in the tobacco drying section on discharge moisture content, this study addresses the non-normality and right skewness of discharge moisture content by constructing quantile regression models at significant percentile points using complete batch data from the tobacco processing. Comparative analysis was conducted with ordinary least squares regression models. The research reveals that cut tobacco moisture content, process air velocity, and cylinder wall temperature are critical factors affecting discharge moisture content. Moreover, the impact of each factor varies with different percentile points of discharge moisture content: the influence of cut tobacco moisture content demonstrates a U-shaped trend as percentile points increase, process air velocity shows a significant decreasing trend with higher percentile points, while cylinder wall temperature exhibits an increasing influence with ascending percentile points. The quantile regression model at the 5th percentile effectively characterizes variable impacts near the technical standard center of production process parameters, providing crucial theoretical support for precise parameter settings in discharge moisture control.