Study On Qualitative Identification Methods Of Unknown Components With Fourier Transform Infrared Specturm

Fourier transform infrared(FT-IR)spectrum has been widely used in the quantitative analysis of complex and trace multi-component gases.In recent years,a variety of high-precision quantitative analysis methods for components based on FT-IR spectrum have been developed.In general,qualitative analysis of FT-IR spectrum is required before quantitative analysis using FT-IR spectroscopy.Traditionally,a certain auxiliary means,such as GC-MS,were used to make qualitative analysis to FT-IR spectra.Of course,sometimes,FT-IR spectra were used to make qualitative analysis directly with FT-IR spectral qualitative identification analysis algorithm.At present,the existing FT-IR spectral qualitative identification analysis methods have the disadvantages of poor qualitative identification analysis performance,and cannot be analyzed in real time and so on,which lead to the limitation of FT-IR spectrum analysis in many applications.The most typical application scenario is to use FT-IR spectroscopy to detect the pollution emissions of chemical parks in real time.Because of the inability to carry out real-time qualitative analysis,after obtaining FT-IR spectra,before making quantitative analysis to those FT-IR spectra,it needs to make qualitative analysis to FT-IR spectra firstly.So,it is hard to monitor emissions of chemical park.Similarly,in some cases of industrial production process,such as pharmaceutical production process,it needs to monitor or analyze the raw materials or some special components in it with FT-IR spectrum.In fact,FT-IR spectral qualitative identification analysis methods have not been a great substantive breakthrough.There are many assumptions and prerequisites for qualitative identification analysis using FT-IR spectrum.For example,few FT-IR spectral qualitative identification analysis methods take the nonlinear effect of FT-IR spectrum into account.This makes the existing FT-IR spectral qualitative identification analysis methods in practical application by various restrictions.In this paper,the FT-IR spectral qualitative identification model based on the variable selection method which base on the linear model of FT-IR spectrum and the genetic algorithm combined with the nonlinear least squares fitting algorithm which base on the nonlinear model of FT-IR spectrum respectively.In this paper,we first introduce the variable selection method developed in recent years,and construct the qualitative identification model of FT-IR spectrum by combining various variable selection techniques with variable selection criteria respectively.And the qualitative identification model of FT-IR spectrum based on variable selection method was evaluated from the perspective of recommendation system.The qualitative identification model of FT-IR spectrum was studied by simulation experiment.Finally,several sets of experiments were carried out and spectra data were required,and the FT-IR spectral qualitative identification model was validated.The experimental results show that the FT-IR spectral qualitative identification model based on the variable selection method is an effective model.Then the nonlinear generation model of FT-IR spectrum is developed,and the nonlinear least squares fitting algorithm is used to solve the mathematical model.Based on the genetic algorithm,the qualitative identification model of FT-IR spectrum based on genetic algorithm and nonlinear least squares fitting algorithm is developed.And qualitative identification models were analyzed by simulation experiments and practical experiments.In the simulation experiments,the qualitative identification performance of the qualitative identification model under different noise conditions is studied.Based on the analysis of the results of qualitative identification analysis of practical experiments and simulation experiments,some improvements are put forward,and the simulation experiments and practical experiments are carried out on the improved model.The results show that the FT-IR spectral qualitative identification method based on genetic algorithm and nonlinear least squares fitting method is an effective method for qualitative identification analysis of FT-IR spectrum.Considering that the genetic algorithm is a random search algorithm,the reproducibility of the results of qualitative identification method is verified from the experimental point of view.Finally,the simulated of pollution discharge experiment is carried out,and two kinds of FT-IR spectral qualitative identification methods are used to analyze the experimental data in real time,and the ideal results are obtained.