The Research Of Warning Model And Algorithm On Ethyl Carbamate In Chinese Rice Wine

Ethyl carbamate?EC?in Chinese rice wine is a potential carcinogenic substance,thus attracting widespread attention.Although the current domestic study of EC in Chinese rice wine has some achievements,a lot of countries in Europe have made the limitation standard of EC content in their various wines.At present,China's government has yet to make limited provision of EC content in Chinese rice wine,so it is imminent to build related early warning model of EC in Chinese rice wine.Through the early warning model,it can not only provide some scientific guidance and advice of effectively reducing and controlling the EC content in Chinese rice wine,but also establish certain theoretical basis for settling the limitation standard of EC content in Chinese rice wine and provide certain technical support for the healthy development of Chinese rice wine.EC in Chinese rice wine was the object of the study.Firstly,HPLC-FLD method was used to establish the determination of urea,urea was a precursor substance of EC in Chinese rice wine.At the same time,the urea content in Chinese rice wine was studied.Then,this section focused on the relationship between the urea,temperature,the physical and chemical indicators and EC in Chinese rice wine to build early warning model.However,the routine method of determining the physical and chemical indicators in Chinese rice wine is complicated and time-consuming,so we developed the rapid detection of the physical and chemical indicators in Chinese rice wine by using near infrared spectrum method.Finally,this study was based on the rapid detection of physical and chemical indicators by using near infrared spectrum method,the physical and chemical indicators of Chinese rice wine before selling as the breakthrough point,using Logistic regression and Fisher discriminant analysis by SAS 9.1 software to build warning models of EC content more than 150 ug/L in Chinese rice wine.The results are as follows:HPLC-FLD method was used to establish the determination of urea,urea was a precursor substance of EC in Chinese rice wine.The results showed that the method R²was 0.9998,the average recovery rate was 97.7%,the relative standard deviation was3.5%,detection limit was 0.025 mg/L,the limit of quantification was 0.09 mg/L.This method with high precision and good accuracy,was suitable for the accurate determination of urea content in Chinese rice wine.Real-time tracking the 4 batch of different raw material ratio of Chinese rice wine fermentation,the urea content in the fermentation process was determined.The results showed that different materials have less effect on final urea content,each batch of end of urea content in the fermented liquid was between12 mg/L and 15 mg/L.Survey results showed that the average content of urea in Chinese rice wine was 10.0 mg/L.This section focused on the relationship between the urea,temperature,the physical and chemical indicators and EC in Chinese rice wine to build early warning model.We had established EC generation model in the process of simulation fried wine,the model was EC=0.01772*urea²+0.00257*temperature².We had established the model of the urea,year,the physical and chemical indicators and EC in Chinese rice wine,the model were EC=2.08908*urea-56.14792*year+113004,EC=21.87765*year-72.85838*alcohol-72.07986*acid-6.79169*solid+1470.88861*volatility ester-0.77979*urea+45634,the72h potential EC accelerating generation model was EC=12.041*urea-1.8107.Research results of the constructed model showed that R²>0.98,Pr>F<0.0001,the model had significant meaning.Year,urea and temperature are the potential risk factors of EC in Chinese rice wine.It could do a preliminary exploration and research on the shelf life prediction of pre-finished Chinese rice wine by early warning model.On the basis of near infrared spectroscopy of Chinese rice wine and chemical stoichiometry,the near infrared model of alcohol degree,sugar,acid and amino nitrogen had been established by using partial least square.Coefficient of determination?R?,cross-validation mean square?RMSECV?and relative analytical error?RPD?were used to evaluate the prediction accuracy and stability of the four models,respectively.The Rvalues were 0.9994,0.9989,0.978,0.9603,respectively,the RMSECV values were 0.0891,0.698,0.0898,0.0414 respectively,and the RPD values were 27,21,4.8,3.58,respectively.The RPD values were more than 3 which indicated good modeling effect.Randomly drew12 Chinese rice wine samples to verify this model,the deviation between measured and predicted results will be less than 10%and it did not have significant differences?P>0.05?.In this paper,the models of Logistic regression and Fisher discriminant analysis were built by SAS 9.1 software.The result showed that the Logistic regression model was (P=1/(1+e^-z),Z=50.1057-1.6721*alcohol degree-6.0497*acid-0.1352*sugar+72.1687*amino nitrogen+0.5531*solid-15.0515*pH-9.4160*color.It had realized the content of EC more than 150 ug/L accurate early warning,the accuracy of the model reached 95.08%.To sum up,this study had built the early warning model of EC in Chinese rice wine,the algorithm of model were studied.It could provide certain technical support to quick warning of food safety about EC in Chinese rice wine,further improve the food safety quality control system of Chinese rice wine.