The Application Of Nondestructive Pre-testing For Several Critical Process Points In Pixian Soybean Paste

The Pixian soybean pastes’ main raw material is high quality bean and pepper,and pass through natural inoculation fermentation and night dew,finally formed the unique semi-liquid viscous or semi-solid condiment.Because of lacking of process monitoring of quality indicators,and still have a strong dependence on the experience of the worker,lead to the low yield,uneven product quality and unable to satisfy the demands of the market and consumers.Because of its wide application scope,fast detection speed,and no damage to sample integrity,nondestructive testing method can effectively judge the internal quality information of samples that cannot be obtained from the appearance.Therefore,this article through studying of the correlation of Aspergillus oryzae and electronic nose,studying the regularity of Pixian soybean paste fermentation pool by nondestructive testing method and studying the changes of moisture content in the drying process by near infrared spectrum,in order to use scientific statistical method to replace worker’s subjective experience,and provide beneficial guidance to actual production of Pixian soybean paste.The main experiment results were as follows:(1)The rapid detection method of Aspergillus oryzae was established by real-time quantitative PCR method as the standard value of nondestructive testing.Using PEN3 electronic nose to detect the volatile compounds of Aspergillus oryzae during the starter propagation process.Using the pattern recognition methods such as radar chart,principal component analysis(PCA),linear discriminant analysis(LDA)and support vector machine(SVM)to explore the feasibility of electronic nose identify the growth of Aspergillus oryzae in starter propagation process.It can be reflected in the radar chart that the sensors have different responses to volatile gases and can reflect the change of volatile metabolites.The PCA analysis can only partially distinguish the growth of Aspergillus oryzae in different starter propagation time.The LDA combine with Mahalanobis distance and Euclidean distance analysis can be used to identify the growth of Aspergillus oryzae;QPCR can accurately measure the quantity of Aspergillus oryzae,and can directly reflect the growth of Aspergillus oryzae;The electronic nose combine with the SVM pattern recognition can predict the growth of Aspergillus oryzae in unknown starter propagation time and the accuracy is 100%.(2)View the Pixian soybean paste in fermentation period as the research object,using electronic nose and the near infrared spectrometer to analysis the samples of 1,2,3,4,5,and 6 months of fermentation time.Principal component analysis,linear discriminant analysis and support vector machine algorithm were carried out on the obtained data.The electronic nose can detect the characteristic gas composition in the sample of Pixian soybean paste.The PCA analysis can monitor the fermentation maturity of liquid samples,but cannot monitor the fermentation maturity of solid samples.Through LDA analysis,the fermentation maturity of all the samples was monitored.For the near infrared spectrometer,due to the rough surface of the solid samples,the effect of the stray light will be particularly big,so the liquid samples are selected for the study.Through the analysis of support vector machine algorithm,the established model is very good at distinguishing and predicting the different fermentation time.(3)Using near infrared spectrum to detect the moisture content in drying process of Pixian soybean paste.Analyzing the same product samples of different years of fermentation and the same product in different sites of samples showed that the near infrared spectrum combined with SVM could predict different years and sites of the unknown fermentation samples,and the accuracy is higher.It can be seen from the model that the correlation coefficient of the model is relatively high,and the accuracy of the model is higher too.