Study On Water Quality COD Detection Method Based On UV-Vis Spectroscopy

Water is an important part of all life and matter.In recent years,water pollution has become a major problem in the world's water resources and environmental protection work.Chemical oxygen demand(COD)is an important evaluation index for the degree of water pollution,and it is very necessary to carry out rapid and accurate detection.Ultraviolet-visible spectroscopy is a method for indirectly measuring the COD concentration of water based on the selective absorption of ultraviolet light by organic matter.Based on the UV-Vis spectroscopy method,the following research work was carried out on water quality COD detection:(1)The COD standard solution is prepared using potassium hydrogen phthalate.Firstly,the applicability of Lambert Beer's law in the range of COD concentration was studied by analyzing the correlation between UV absorbance and COD concentration.In order to get a modeling method with strong prediction performance,First-order derivative and smooth two spectral preprocessing methods were combined with principal component regression(PCR)and partial least squares regression(PLSR)to establish a prediction model of UV full spectrum.The experimental results show that the first-order derivative combined with the PLSR established UV full-spectrum model predicts the best performance.(2)Based on theoretical and experimental research,the effects on the detection of COD in water quality of two water quality parameters,temperature and turbidity was also studied.The experimental results show that a large range of temperature changes will have a significant impact on the UV absorbance data,while the turbidity parameter seriously interferes with the output accuracy of the detected COD model.In order to improve the accuracy of the model to detect COD,we established single-wavelength,visible-light full-spectrum and nonlinear turbidity compensation models based on PLSR with characteristic wavelength points.Experimental results show that the nonlinear compensation model effectively suppresses the influence of turbidity in mixed water samples,and is the best method for detecting COD under a single information source.(3)In order to enable the established COD prediction model to contain information on the impact of all parameters in the water sample.The multi-source information fusion algorithm was studied.A prediction model for the fusion of water quality parameters and spectral information was established,and the prediction effects was compared with the best model under a single information source.The experimental results show that the multi-water quality parameter fusion UV spectral information modeling method improves the evaluation parameters of the COD model and is more reasonable.The COD detection method studied in this thesis has important reference value for independent research and development of instruments for detecting COD in complex water samples and monitoring of multiple water quality parameters.