Research On The Quality Diagnosis Of Tobacco Primary Process Based On Neural Network Diagnosis Of Tobacco Primary Process Based On Neural Network

With the development of the society and the improvement of living standards,people demand higher requirement in the quality of cigarette products.To stabilize the cigarette quality and to satisfy the market segment,tobacco silk technology increasingly tend to be more subtle and complex,and it also makes the tobacco industry pay more attention to control the quality of cigarettes.The process of silk are influenced by many and the quality of cut tobacco inevitably fluctuate.In order to improve the wirer of high efficiency,high quality must be guaranteed the stability of production process in a controlled state,so the quality of the production process monitoring is very important.On the other hand,the implementation of digital silk process greatly improves the silk workshop of informatizationn and automation and silk process condition monitoring.If it just relies on the human factors,it would be a serious impediment to the tobacco industry and the improvement of quality management automation.Aiming at these two problems,for one,this paper focuses on the combination of SPC method and BP neural network technology and applying it in the quality control of tobacco silk.For another,silk process quality intelligent monitoring technology is studied.First of all,according to the characteristics of the process of silk,the SPC method is applied in the quality control of silk.Through the analysis of the key process of silk production technology as well as the important quality index,we will determine the SPC control object,put forward the application scheme of control chart,and select the mean-poor control chart in quality control.Finally,some examples are given to illustrate the application method of control chart in silk process.Secondly,this paper studies the control chart pattern recognition problems based on the BP neural network.SPC control chart can be divided into normal first,step up,step down,rising trend,step upward,downward trend,cycle six basic mode,and establish the corresponding mathematical model.And the article then puts forward the overall scheme of the control chart pattern recognition and studies the basic model of control chart recognition method and abnormal patterns of key characteristic parameters estimation method.On the basis of theoretical research,with the improved BP algorithm,this paper establishes a control chart pattern recognition network,three anomaly model parameter estimations,the estimation step mode amplitude,respectively cycle mode amplitude and cycle length,trend model of slope.Then,it will be based on the theory of SPC and six aspects,including human,machine,material,method,ring and measuring.As for each kind of abnormal patterns,it will analyze and summarize the causes and the adjustment of the typical abnormal and suggests that abnormal diagnosis knowledge base is established.Finally,this paper develops process of the silk quality monitoring and diagnosis based on SQL Server2000 database design silk quality management and abnormal diagnosis knowledge base,based on Visual c + + 6.0 and Matlab7.0 platform system.It realizes the silk quality data management,control,drawing process state recognition and feature parameter estimation,abnormal cause analysis,diagnosis,and other functions,and a test example is given.