Research On Edible Oil Ingredients By Near Infrared And Rama Spectroscopy

Near infrared spectroscopy technology has become an advanced detection method in the field of modern detection because of its high efficiency and low cost.It has been used in many fields,such as biological pharmacy,chemical industry,food safety and so on.Quantitative analysis of saturated fatty acids,acid value and other pivotal physical and chemical indicators that reflect the quality of edible oil has been a hot research in food safety field,the traditional methods of detecting have many disadvantages,such as high cost,lack of environmental pollution and so on,so it is unable to meet the requirements of relevant law enforcement sector that detect oil quality rapidly and accurately.To make up for the deficiencies of traditional methods,the near infrared spectrum detection means are proposed to achieve detecting oil ingredients rapidly.In this paper,by using near infrared spectroscopy for quantitative analysis of edible oil quality components,puts forward the cooking oil rapid analysis method for the determination of saturated fatty acids,acid value,and explores the Rama spectroscopy feasibility and research of the saturated fatty acid cooking oil grease detection software module replacement,quantitative research main content includes:(1)In view of the long detection time of saturated fatty acids by gas chromatography,the sample pretreatment is complex and the cost is high,the method of near infrared spectroscopy is proposed to predict the saturated fatty acids.43 oil samples were collected,collecting oil by near infrared spectrum data;The pretreatment of three different ways to the spectral data was carried out;Respectively establish GS-SVR and PSO-SVR regression model,the comparison of the two models predicted results,select the grid search method(GS)as optimization algorithm;Three different pretreatment methods of GS-SVR model were analyzed,and the standard normal variable transformation and the de-trending(SNV-DT)were selected as the best pretreatment methods.(2)Aiming at an acid value of oil detection time-consuming and solvent-consuming and more complicated steps and other issues,study and explore a method of rapid application of near infrared spectroscopy to predict the acid value of edible oil.37 oil samples were collected,and the near infrared spectrum data were collected;The GS-SVR regression model was established by using the standard normal variable transformation(SNV),the multiple scatter correction(MSC)and the orthogonal signal correction(OSC)of three different pretreatment methods;In MATLAB,the results forecast,analysis and comparison of the prediction performance of the three models,select the OSC as the best pretreatment method.(3)Along the near infrared spectrum method to predict technical route of the saturated fatty acid,further explore Rama spectroscopy quantitative prediction of saturated fatty acids.The research results show that the correlation coefficients of the calibration set and the prediction set of the PSO-SVR regression model are both low,and the model does not have the ability to predict;Therefore,it is necessary to make a further research on the method of pretreatment for the quantitative prediction of saturated fatty acids by Rama spectroscopy.(4)The research group developed near infrared oil testing software,from the function can be divided into qualitative and quantitative detection.In the qualitative aspect,the software has been more mature,but in the quantitative detection of nutritional components in edible oil,the software can only detect the fatty acid content of low carbon.In order to expand oil and other nutrients quantitative composition,in the early development of software based on,also increased other algorithms,try to lower carbon number fatty acid model replacing saturated fatty acid model to quantitative software module of multi function.