Research On Key Technologies Of COD Test Based On UV-Vis Spectroscopy

Water is the basis of human life and the indispensable resource guarantee of human production and life.In recent years,with the rapid economic development and urbanization process in China,water pollution,especially urban surface water pollution,is becoming more and more serious.All kinds of water safety accidents or water pollution incidents occur frequently,which seriously endangers people’s health and sustainable development of economy.Therefore,through the excellent water testing instruments and systems for water accurate and efficient test has become particularly important and significant.In this paper,a large number of methods and instruments for the test of chemical oxygen demand(COD)of water have been investigated.Aiming at the problems of the traditional water test instrument,such as the tedious test process,the time-consuming,the need to consume chemical reagents and cause secondary pollution to the environment,this paper focuses on the research of the new water testing system based on ultraviolet visible spectroscopy(UV-Vis spectroscopy).Due to the interference of many factors on UV-Vis spectroscopy,the accuracy of the system is not enough.In order to improve the test accuracy of the test system,it is necessary to research the key technologies that can improve the test accuracy of UV-Vis spectroscopy,in order to maximize the elimination of interference and improve the test accuracy of the system.The main contents of this paper are as follows:(1)The spectral denoising technology based on UV-Vis spectroscopy method for COD test is studied.A spectral denoising method based on EEMD decomposition and improved general threshold filter(EEMD-Based)is proposed.The proposed denoising method is used to denoise the simulation signal and the actual spectral signal.Both show good denoising effect.Finally,we use Gaussian smoothing,Fourier transform,wavelet transform and EEMD-Based denoising methods to denoise 240 spectral curves collected by the equipment,which proves the superiority of the proposed EEMD-Based denoising method.(2)The modeling method of COD test based on UV-Vis spectroscopy is studied,and an improved modeling method based on Bagging algorithm is proposed.The decision tree(DT),Bagging and Improved-Bagging prediction models for COD test of water were established by using full spectrum data(processed by different denoising methods)and spectral feature dimensionality reduction data(processed by different feature dimensionality reduction methods).The results show that the best prediction effect is the Improved-Bagging prediction model based on EEMD-Based denoising and SCARS wavelength screening.Its R~2 on the prediction set reaches 0.9317,and RMSEP reaches 5.39mg/L.The effectiveness of the proposed EEMD-Based denoising method and Improved-Bagging modeling method is fully verified.(3)The turbidity compensation technology based on UV-Vis spectroscopy for COD test is studied,and a full band turbidity compensation method based on filter difference method is proposed.After compensation by the proposed turbidity compensation method,through the compensated spectral curve and the COD prediction results of the model,both found that this compensation method can effectively eliminate the influence of turbidity on COD test.The modeling results show that the best prediction effect is the COD prediction model based on Improved-Bagging algorithm,which is established after EEMD-Based denoising,turbidity compensation method proposed in this paper and SCARS dimension reduction combined pretreatment.R~2 of the model in the prediction set reaches 0.9862,RMSEP reaches 1.81mg/L.(4)The compensation technology for the influence of other environmental parameters of COD test based on UV-Vis spectroscopy was studied,and the influence mechanism of p H,temperature and conductivity on UV-Vis absorption spectroscopy was analyzed through experiments.The advantages of data fusion technology in data processing are used to comprehensively compensate the impact of environmental parameters on COD test,and improve the test accuracy of COD.The modeling results show that the best prediction effect is the COD prediction model based on Improved-Bagging algorithm,which is based on EEMD-Based denoising,turbidity compensation method proposed in this paper,SCARS dimension reduction and weighted superposition of three environmental parameters.R~2 of the model in the prediction set reaches 0.9982,RMSEP reaches 1.03mg/L.(5)The development and performance test of the water test system based on UV-Vis spectroscopy are carried out.The hardware system of water monitoring node and remote test software system are developed,and the prototype is developed.On this basis,the performance test of the water test system prototype based on UV-Vis spectroscopy was carried out.The test results show that the performance indicators of the whole machine meet the requirements of national standards,and the COD test indicators of water meet the relevant national and industrial specifications.The relative error of the actual water sample is less than or equal to5.5%;the repeatability is less than or equal to 1.7%.The water test system based on UV-Vis spectroscopy method developed in this paper realizes the rapid and accurate test of COD,which verifies the validity of the key technologies studied in this paper.The appearance of this system makes up for the defects of low precision and insufficient reliability of domestic water testing instruments.It will make an important contribution to the further establishment of real-time test and monitoring of surface water.