Multi-Scene Water Quality Detection Technology Based On UV-Vis Spectroscopy

With the continuous improvement of human production and quality of life,the number of humans,animals and plants are constantly approaching the environmental load of nature,and the protection and efficient use of natural resources have also attracted more and more attention.Urban sewage treatment plants are the key infrastruction for humans to impact the quality of water environment,more and more attention is paid to their operation efficiency and treated water quality.In order to meet the increasingly stringent sewage discharge standards,more than required energy and chemical consumption were applied at WWTPs,which to exchange for achieving water quality standards.As a result,a lot of resources and investment were wasted,which also brings troublesome to followup treatment and reuse.The application of intelligent technology in sewage treatment and intelligent control of treatment would be helpful to achieve the goals of improving quality and efficiency,saving energy and reducing emission.Real time on-line ultraviolet visible(UV-Vis)spectrum water quality detection method is a kind of water quality real-time detection technology with many advantages,which has been developed in the field of online water quality detection.In this study,the feasibility of using UV-Vis spectroscopy to detect the concentration of organic pollutants in secondary effluent was investigated;Furthermore,the prediction models of COD are constructed by using statistical method,and the optimization procedure of the model are discussed in detail;Finally,the method foe online monitoring of antibiotic water is proposed by combining the principle of UVVis spectrum detection of organic matter in water and the the COD models’ construction procedure,which lays the foundation for the development of intelligent water quality detection in many fields.More details of this study are:1.Feasibility study on the detection of organic matter concentration in secondary effluent by UV Vis spectrometry.(1)The spectral characteristic absorption of nine kinds of organic compounds with one or several kinds of active functional groups detected by UV-Vis spectrometry was studied.The results showed that glucose,methanol and formaldehyde with hydroxyl or aldehyde groups,which are not directly linked with each other,had no characteristic absorption;Phenol,benzoic acid,sodium acetate,sodium propionate,sodium butyrate and p-benzoquinone,which have two kinds of active groups directly linked,have certain characteristic absorption in UV-Vis band.For simple structure organic compounds with two or more active functional groups,especially when the functional groups are directly connected,shows higher UV-Vis absorption characteristics.(2)The lowest instrument detection limits of UV-Vis spectrometry for the detection of organic compounds were calculated.The order of detection limits from low to high was: p-benzoquinone(0.404 mg/L)< benzoic acid(0.776 mg/L)< phenol(1.354 mg/L)< sodium acetate(35.87 mg/L)< sodium propionate(55.70 mg/L)< sodium butyrate(57.69 mg/L).2.Construction of water quality COD prediction model by UV-Vis spectroscopy(1)The COD prediction models of 260 groups of effluent from secondary sedimentation tank and tertiary sedimentation tank in full wave band and four characteristic wavelengths were established.The results show that the difference between RMSE and RMSE is small,and the maximum difference is only 0.721 mg / L;The minimum value of internal RMSE is 4.991 mg / L,and the relative error with the average value of COD in tertiary effluent is 16.3%.It can be seen that the prediction ability of the prediction model based on secondary sedimentation tank effluent and tertiary total effluent needs to be further improved.(2)Compared with all the water samples,the COD prediction model constructed by 90 groups of three-stage total effluent has higher accuracy.At the same time,the best prediction ability was obtained by using the model of COD,nitrate and turbidity characteristic wavelength.The goodness of fit of the model was 0.992,and the RMSEP was 1.979 mg/L.It can be seen that the wavelength modeling with negative coefficient in the regression equation of the full band model in the characteristic bands of nitrate and turbidity can calibrate the model,reduce the influence of spectral absorption of non prediction target water quality parameters on the model,and further improve the prediction accuracy of the model.3.Study on the feasibility of detecting antibiotic wastewater by UV-Vis spectroscopy.(1)Tetracycline,ofloxacin and streptomycin were obtained from water in laboratory in the UV-Vis band,tetracycline has three characteristic absorption wavelengths: 217.5 nm,275 nm,357.5 nm,corresponding to the highest absorbance:169.69(1/m),188.32(1/m),183.31(1/m);Ofloxacin has four characteristic absorption wavelengths,which are 227.5 nm,255 nm,287.5 nm and 330 nm,respectively.The corresponding maximum absorbances are 378.38(1/m),339.44(1/m),609.77(1/m)and276.15(1/m),respectively;Streptomycin only has a characteristic absorption at 375 nm,corresponding to a maximum absorbance of 12.12(1/m).(2)The lowest detection limits of ofloxacin,tetracycline and streptomycin were0.68 mg/L,1.57 mg/L and 47.22 mg/L,respectively.The detection limits of ofloxacin reached the microgram level.(3)This paper proposes the following ideas for the construction of the prediction model of pharmaceutical wastewater containing antibiotics: using the data-driven model to build the UV-Vis spectrum antibiotic prediction model,using the statistical analysis method to find the wavelength to measure the concentration of other organic compounds for calibration model,Combined with practical factors(such as reference to the manufacturing process parameters of pharmaceutical factory),the prediction accuracy of antibiotic prediction model was further improved.The composition of antibiotics in medical wastewater is more complex,and it is difficult to detect one of them.One or more characteristic wavelengths can be found by stoichiometry method to construct a model to characterize the degree of wastewater treatment.