Discharge Moisture Content Prediction And System Development Of Perfuming Process In Tobacco Production Line

Tobacco processing is the key link of cigarette processing,which directly affects the quality of tobacco and plays a decisive role in the cost.The processes of tobacco production process are complex and influence each other,and the coupling between process parameters is serious.This also determines how the key to the stable operation of tobacco production is to improve the predictability and autonomy of multiple factors in the production process,and to solve the problem of precise prediction of tobacco production quality and rapid response of production system.At present,the research on quality prediction of tobacco making process mainly focuses on statistical analysis model,machine learning model and combination model.In the actual tobacco production process,it is rare to construct a prediction model based on the time series and correlation characteristics of tobacco production,deeply explore the characteristics between process parameters,and develop a prediction support system to support the prediction of tobacco production process quality,which makes it difficult to predict the actual production process.In this thesis,the perfuming process in the process of tobacco production was studied.Aiming at the problem of quality stability in the perfuming process,a process parameter prediction model based on VMD-GRU-TAM-BPNN was constructed,and a perfuming process prediction support system was developed.Firstly,an S-MIC-RFE feature selection method was designed through detailed analysis of the perfuming process,which realized the feature screening of process parameters.Secondly,aiming at the characteristics of process sequence,correlation coupling and random complexity among the process parameters of perfuming,the technical basis of quality prediction modeling of production process was explored,and a VMD-GRU-TAM-BPNN combined model was established to achieve the prediction of discharge moisture content.In order to ensure the effectiveness of the prediction model,based on the established combination model,the error between the true value and the predicted value of the quality index was considered.The GWO optimization algorithm was used to optimize the parameters of the prediction model,which further improved the goodness of fit of the prediction model.Finally,based on the above model basis and technical basis,the accurate prediction of the quality of the flavoring process was developed.Finally,based on the above model basis and technical basis,a precise prediction system for the quality of the perfuming process was developed to achieve‘observable’ and ‘measurable’ and improved the stability of production quality.The research results of this thesis have practical guiding significance for the production practice of perfuming production line.It provides guidance for improving the quality control of the perfuming process,lays a technical foundation for parameter optimization and precise prediction of perfuming process,and provides reference for quality prediction modeling method and its application.