Detection And Analysis Method Of Colored Dissolved Organic Matter In Supply Raw Water By 3D Fluorescence Spectrum

The quality of urban raw water(surface water coming into the water plant)directly affects the water treatment process and the quality of drinking water,and then affects the physical and mental health of residents and social harmony and stability.Organic pollutants pose a great threat to the quality of raw water.Among them,colored dissolved organic matter(CDOM)is not only an important part of natural organic matter in raw water,but also has a close relationship with the generation of some disinfection by-products in water treatment process,which needs to be focused on to avoid affecting the quality of water supply.Therefore,rapid and effective detection and analysis of CDOM in raw water is of great practical significance.Three-dimensional(3D)fluorescence spectrum has been gradually used in the field of water organic matter detection and analysis due to its advantages of large amount of information,good selectivity,high sensitivity and low detection limit.In this thesis,combined with the national scientific research projects and the actual urgent needs,the research of 3D fluorescence spectrum detection and analysis method for CDOM in supply raw water is carried out.Aiming at the influence mechanism of environmental factors on 3D fluorescence spectrum,the weak feature stability and limited generalization ability of traditional 3D fluorescence spectrum analysis methods,as well as the problem of organic matter detection and analysis under the interference of overlapping fluorescence peaks,this thesis carried out the related research of 3D fluorescence spectrum method,such as the feature extraction,classification and identification,quantitative analysis and so on.Method validation was carried out through experimental design and analysis.The main work and innovations of the thesis are as follows:(1)The influence analysis and inhibition methods of the changes of environmental factors on 3D fluorescence spectrum are studied.Aiming at the effect of environmental factors on 3D fluorescence spectrum,the influence law and inhibition method of environmental factors on 3D fluorescence spectrum were studied.Through the combination of mechanism analysis and experimental verification of environmental factors,the 3D fluorescence spectrum influence law under the changes of temperature,pH and turbidity was studied.With the increasing of temperature,the fluorescence intensity of organic matter decreased gradually.Finally,a water bath controlled constant temperature was used to attenuate the temperature effect on the 3D fluorescence spectrum.when the pH vatue of the sample is in sirong acid or sirong aikan,the 3D fluorescence spectrum of organic matter will change greatly.Therefore,the method of controlling the pH threshold of the sample is adopted to reduce the effect of pH change.The influence of turbidity standard solution on fluorescence spectrum is mainly reflected in the interference phenomenon of formazide fluorescence peak,so the influence of turbidity standard solution can be suppressed by deducting formazide fluorescence peak in proportion.However,when the cause of turbidity is unknown,the unified suppression method is difficult to achieve,so the physical filtration method is used to weaken the influence of turbidity.Above measures can effectively weaken or inhibit the effect of environmental factors(temperature,pH and turbidity)on the 3D fluorescence spectrum,which provide a theoretical basis for the classification and quantitative analysis of organic matter based on the 3D fluorescence spectrum in the following chapters.(2)The methods of extracting special information from 3D fluorescence spectrum and detecting and recognizing characteristic organic matters were studied.Aiming at the problems of poor stability and limited generalization ability of the extracted features in 3D fluorescence spectrum analysis,a feature extraction and detection model based on BinBoost was proposed.BinBoost can extract the local specific information of the 3D fluorescence spectrum image effectively.According to its adaptability to the changes of target translation,scale and illumination,the specificity and stability of extracted features can be ensured when the position,shape and intensity of fluorescence peaks of organic matter change.Combined with ratio purification method and random sample consensus,by comparing the matching degree between the local specific information of the test samples and the various samples in the sample set,the organic matter classification model is established.(3)The suppression of overlapping interference of fluorescence peaks and the quantitative analysis of known characteristic organic compounds were studied.In order to solve the problem that it is difficult to detect and quantify the known characteristic organics under the interference of overlapping fluorescence peaks,a quantitative analysis method of known characteristic organics based on discrete cosine transform(DCT)was studied.The spectral signals of organic matter samples are projected to the frequency domain by DCT,the overlapped spectral signals in the time domain are separated,and the high-dimensional features were reduced by Zigzag scanning and energy thresholding method.The regression model between the spectral frequency domain features and the target component concentration is established by support vector regression.Suppression of overlapping interference of fluorescence peaks and quantitative analysis of characteristic organic compounds are accomplished.In general,this thesis focuses on 3D fluorescence spectrum feature extraction,detection and recognition,concentration quantitative analysis.The representative CDOM fulvic acid,tryptophan and tyrosine in raw water were used to verify the effectiveness of the method.The method improves the stability and generalization performance of spectral features,realizes the detection and recognition of characteristic organic compounds,suppression of overlapping interference of fluorescence peaks and quantitative analysis of known organic compounds.The research work provides technical support for the detection and analysis of CDOM in raw water by 3D fluorescence spectroscopy.