Study On Viscosity Prediction Of Different Varieties Of Tomato Paste Based On Artificial Neural Network

Tomato is the largest processed vegetable in China,and its processed product-tomato paste,is an important export advantage agricultural product in China.For tomato paste,viscosity is an important index to evaluate the grade and quality of the product,and the raw material properties of processed tomato are the important factors affecting the viscosity of tomato paste.At present,there is little evaluation of the processing suitability of different varieties of tomato sauce in China,and a statistical method has not been proposed to predict the quality traits of tomato raw materials to the viscosity of tomato paste.Therefore,in this study,the raw material quality traits and tomato paste flow quality of 44 kinds of processing tomatoes were deeply studied.Finally,the prediction models of different varieties of tomato paste viscosity based on artificial neural network were established.the specific results are as follows:?1?The study on the flow quality of tomato paste showed that there were great differences in Bostwick consistency,apparent viscosity,yield stress,consistency coefficient and stability of different varieties of tomato paste.Through the study of steady-state shear rheology,it is found that 44 kinds of tomato paste samples are shear thinning non-Bingham?Bingham?pseudoplastic fluid,that is,the viscosity of ketchup samples decreases with the increase of shear rate.Through the cluster analysis of 44 kinds of tomato paste texture indexes,it was found that the consistency,apparent viscosity,yield stress and stability of the third kind of samples were higher,so comprehensive consideration was made.The flow quality of this kind of ketchup?NDM2272,Zaohong,Xinfan 45,H9888,H9780,H8504 and H2401?is better than that of the first and second types.?2?There were differences in the quality traits of the raw materials of 44 kinds of processed tomatoes.The coefficient of variation of p H,magnesium,potassium and soluble solids were less than 10%,indicating that the differences among the four indicators were small among different samples.The coefficient of variation of the remaining 16 indicators was greater than 10%,indicating that the quality indicators of different processed tomato samples differed greatly.?3?Based on five core indexes of ketchup flow quality,the evaluation model of ketchup flow quality was established:Y=0.242×A₁+0.228×A₂+0.202×A₃+0.106×A₄+0.222×A₅?A₁?A₅represents apparent viscosity,centrifugal sedimentation rate,suspension stability,yield stress and consistency coefficient,respectively?.The comprehensive scores of flow quality of 44 kinds of tomato paste prepared by processing tomato were obtained by this model.?4?Nine characteristic indexes of tomato raw materials closely related to the flow quality of tomato sauce were screened out,which were centrifugal sedimentation rate of tomato juice,viscosity of tomato juice,calcium,iron,cell wall substance,total sugar,WSP,area average particle size and critical particle size,respectively.Through the artificial neural network algorithm,the above nine indexes are taken as the input layer of the model,and the comprehensive score corresponding to the tomato paste viscosity is taken as the output layer.The prediction model based on the tomato raw material character for the tomato paste viscosity is more accurate.The error degree is small,which indicates that the artificial neural network model can accurately predict whether different varieties of tomato raw materials are suitable for processing into tomato paste.