| The safety of drinking water is related to the national economy strategies and the people’s livelihood.In recent years,water quality organic contamination happen from time to time,resulting in the health and safety of the people being affected,so timely and effective detection and early warning of organic pollutants in water have important practical significance.Three-dimensional fluorescence spectroscopy has the features of high sensitivity,high selectivity and real-time,and it is suitable for rapid qualitative and quantitative detection of part of organic matters with fluorescence characteristics.Combined with the National Natural Science Foundation of China,this paper carries out the research on the qualitative determination of organic pollutants in drinking water based on three-dimensional fluorescence spectroscopy.Some key technical issues are focused on in the paper,including the interference of instrument and external noise,background fluctuation of water quality and difficult detection of low concentration samples and so on.The organic contamination samples and normal drinking water samples are effectively distinguished to verify the organic pollutants qualitative discrimination performance based on three-dimensional fluorescence spectroscopy.The main work and innovation of the thesis are as follows:(1)The three-dimensional fluorescence spectroscopy pretreatment and qualitative discrimination method for detecting organic matters in drinking water is researched.Firstly,the solvent blank subtraction method and interpolation method are used to remove the Raman and Rayleigh scattering of the three-dimensional fluorescence spectroscopy.Then,to reduce the influence of random noise caused by instrument and external disturbance on the detection of drinking water organic pollutants,the Savitzky-Golay polynomial fitting method is applied to the smoothing of the three-dimensional fluorescence spectra.Finally,the method of Alternating Trilinear Decomposition(ATLD)is used to extract the features of the three-dimensional fluorescence spectra of normal drinking water samples,and the qualitative discrimination of the test sample is achieved combined with the threshold method.(2)Considering the background fluctuation of water quality,qualitative discrimination of drinking water organic pollutants based on three-dimensional fluorescence spectroscopy is researched.Firstly,the causes of water quality background fluctuation are discussed and the characteristics of drinking water three-dimensional fluorescence spectra under the influence of water quality fluctuations are analyzed.Aiming at the organic pollutants whose fluorescence peak overlaps with the peak of normal drinking water,the qualitative discrimination method of water quality based on Partial Least Squares Regression(PLSR)and threshold method is proposed.The method of PLSR is used to construct the spectral peak prediction model,and combined with the threshold method for qualitative discrimination.The feasibility of detecting organic pollutants in drinking water based on Partial Least Squares Regression method and threshold method is proved by experimental data.However,the detection of low concentration samples has some limitation.(3)The qualitative discrimination method of three-dimensional fluorescence spectroscopy based on the prior knowledge of pollutants in the drinking water is studied.In order to solve the problem of low concentration samples detection,the qualitative discrimination method based on alternating residuals tri-linearization(ART)feature extraction and classifier is proposed.The results show that the detection limit of the model can be effectively lower by extracting the three-dimensional fluorescence spectral characteristics of contaminants as the prior knowledge of pollutants combined with the classifier model,and the detection of low-concentration pollutant samples can be better realized.This paper focuses on the several key issues during the qualitative analysis process of drinking water organic pollutants based on three-dimensional fluorescence spectroscopy.The methods overcome the influence of random noise and background fluctuation,and improve the detection limit of pollutants combined with the prior knowledge of pollutants to realize the effectively qualitative discrimination of drinking water organic pollutants.It provides technical support for drinking water monitoring and early warning system based on three-dimensional fluorescence spectroscopy. |