Some Methods For Improving The Accuracy Of Spectral Analysis Of Complex Liquid System

Complex liquid system is a kind of mixed liquid with both absorption and scattering characteristics,and it exists in all aspects of production and life.Spectral analysis technology is widely used in the quantitative analysis of complex liquid system components due to its advantages such as rapidness,low cost,and no consumption of reagents.It has broad application prospects and great commercial value in the fields of food,medicine,industry,agriculture and so on.However,in the process of spectral analysis of complex liquid system,it is easy to be disturbed by many factors,including not only external factors,but also related to the complexity of the liquid itself,which seriously restricts the accuracy of spectral quantitative analysis.How to deal with it and put forward reasonable methods to restrain the influence of these factors is the key to improve the accuracy of spectral analysis of complex liquid system components.In this thesis,several interference factors in the measurement system are explored from three aspects: spectral measurement,processing and modeling.In view of these influencing factors,we proposed some methods to verify the effectiveness of these methods through the spectral analysis of isolated blood.The main research work of this thesis is as follows:(1)By constructing a RC circuit model and analyzing the working principle of the grating spectrometer CCD(Charge-coupled Device),it is found that the difference in the production process of CCD will lead to the nonlinear relationship between the spectral data and the integration time,and quantify the spectral analysis error caused by this nonlinear relationship.A modeling strategy insensitive to the change of integration time is proposed to restrain the spectral analysis error caused by this nonlinear relationship.(2)According to the strategy of "multi-optical path and multi-mode" in the "M+N" theory,it is proposed to use the nonlinearity of the spectrum to increase the amount of information in the modeling,increase the constraint equation in the model,and reduce the solution space,so as to improve the accuracy of spectral analysis.The effectiveness of this strategy is verified by the detection of hemoglobin in blood bag and the analysis of total cholesterol in plasma.(3)A sample selection method of calibration set considering the concentration of components to be tested and non-tested components is proposed,and the multi-dimensional space is constructed based on the concentration of components to be tested and non-tested components.The uniformity of the component space in which the sample is located is used as the basis for the selection of correction samples.This method has obtained satisfactory results in the analysis of serum bilirubin.(4)In order to realize the non-destructive analysis of the complex liquid system in the bag,a multi-path spectroscopy based on continuous sampling is proposed on the basis of multi-path spectroscopy.The non-destructive analysis of hemoglobin in the blood bag was used to test the effectiveness of this method,and the results proved that the multi-path spectroscopy based on continuous sampling can effectively restrain the influence of blood bag differences on spectral analysis and improve the prediction accuracy of the model.First of all,this thesis analyzes that changing the integration time in the stage of spectral measurement will affect the measurement accuracy;Secondly,the method of multi-optical path and multi-mode is used to increase the amount of information of modeling spectral data;Then,a calibration set sample selection method considering the concentration of components to be tested and non-tested components is proposed;Finally,a continuous multi-optical path acquisition method is proposed on the basis of multi-path spectrum to realize the non-destructive analysis of bagged blood component.This thesis demonstrates the feasibility of the proposed method from three aspects of theory,simulation and experiment,laying a foundation for the application of complex liquid system spectral analysis in medical field,and providing reference for follow-up research.