




摘" 要:卷煙質量的好壞是衡量消費者滿意度的關鍵度量因素。對于空頭這種卷煙過程中出現的A類質量問題,如何在生產過程中實時控制,降低其缺陷率一直是卷煙企業高度關注的焦點問題。該文站在卷煙廠實際生產的角度,利用W市卷煙廠提供的實際生產數據,先后建立基于決策樹的最優參數模型,BP神經網絡預測模型以及情景分析模擬仿真,對該卷煙廠的空頭剔除量過高且生產不穩定的問題進行機臺參數控制區間尋優以及多情景計算機模擬仿真,并給出相應的解決方案。
關鍵詞:數字化轉型;設備最優參數調控;決策樹;BP神經網絡;情景分析
中圖分類號:F426.8" " " 文獻標志碼:A" " " " " 文章編號:2095-2945(2024)25-0068-05
Abstract: The quality of cigarettes is the key factor to measure consumer satisfaction. For the category A quality problems in the process of short cigarettes, how to control them in real time in the production process and reduce their defect rate has always been the focus of cigarette enterprises. From the point of view of the actual production of the cigarette factory, using the actual production data provided by the W city cigarette factory, this paper successively establishes the optimal parameter model based on the decision tree, the BP neural network prediction model and the scenario analysis simulation. The machine parameter control interval optimization and multi-scenario computer simulation are carried out for the problem of excessive short elimination and unstable production in the cigarette factory, and the corresponding solutions are given.
Keywords: digital transformation; optimal parameter control of equipment; decision tree; BP neural network; scenario analysis
卷煙質量的好壞是衡量消費者滿意度的關鍵度量因素。某種意義上來說,產品質量決定了企業的核心競爭力。隨著煙草企業信息化進程的不斷推進。煙草企業中積累的質量管理和生產數據迅速增加,企業目前缺乏對這些數據進行深層次的挖掘,缺少對卷煙機設備改造和檢測儀器改進[1]。汪云等[2]通過“總體方差未知且大樣本情況下的點估計”參數估計建立了煙支空頭檢測率評價方法,可用于卷接包數據系統實現在線動態追蹤。秦鵬等[3]則通過實驗對比2種平整器對煙支空頭剔除率的影響,得出在同樣降低了煙支空頭率的情況下,三深三淺型平準器具有更好的質量穩定性。最近云杰等[4]通過實驗利用8種不同規格的平準器研究其對煙支空頭、煙支指標標準、感官質量評吸的影響,得出8種不同規格平準器對卷煙感官質量沒有明顯影響?!?br>