Multi-variable Water Quality Monitoring Based On UV-Visible Spectroscopy

Water is the source of life and an important part of life and matter.The quality of water determines the quality of life.However,due to the active nature of water,most of the water in nature has been contaminated more or less.According to investigation,there is no absolute pure water in nature.With the improvement of people's material living standards,the water in nature is far from reaching people's drinking water standards.In water resources and their scarcity,research on water quality monitoring has become an important part of safeguard-ing water security in the process of improving water quality.At present,China has developed a series of indicators to ensure water quality safety.The indicators include:Chemical Oxygen Demand(COD),Biochemical Oxygen Demand(BOD),and Total Organic Carbon(TOC).Total Organic Carbon),Turbidity.The main monitoring methods at present are chemical methods and optical methods.Chemical methods are divided into potassium permanganate method and potassium dichromate method,which is character-ized by good repeatability,high measurement accuracy;but there are also many things that can not be simultaneously monitored,the measurement cost is too high,the measurement period is long,the measurement reagent pair Pollution caused by the environment and other limita-tions.The optical method is mainly UV-visible spectroscopy,which has the advantages of low measurement cost,short measurement period,and no pollution.Due to the development of the era,real-time performance has become a consistent demand in the field of water quality de-tection.Therefore,this article studies the COD monitoring based on the real-time UV-Visible Spectroscopy method.At present,in the research of water quality monitoring by UV-Visible Spectroscopy,there are mainly two issues that are worth discussing:First,the selection of characteristic wavelengths of COD,and second,the influence of different substances on the characteristic wavelength variation of COD.For the first question,this paper uses the chemical oxygen demand monitoring of water as an example.Based on the overlap interval partial least squares method(OIPLS),the whole wave band is divided into equal-length and overlapping intervals,and each interval is estab-lished.The least squares model,which preserves the interval with small RMSEP of prediction RMSE,is used as the initial wavelength subset of the genetic algorithm(GA).After iterating its genetic algorithm,the model has better accuracy and goodness of fit.Above 0.99,the RMSEP value is less than 0.02.The method not only guarantees the speed of operation,but also avoids the missed selection of the wavelength variable with high correlation and the erroneous selec-tion of unrelated wavelengths,thereby ensuring the convergence speed and model accuracy of the wavelength selection.For the second problem study,turbidity was used as a material for influencing the chemical oxygen demand measurement.Based on the quantitative analysis of the absorbance of differ-ent concentrations of turbidity solution at the characteristic wavelength of chemical oxygen demand,a nonlinear model was established for chemical analysis.The oxygen demand charac-teristic wavelength compensates.The research shows that the accuracy of the nonlinear model is higher,and the mixed liquors with different concentrations in the laboratory are combined with the actual water samples for empirical analysis.The average relative error rate between the model prediction data and the actual data is 0.035.This compensation method is applied to the turbidity The analysis of the effect of oxygen demand has better accuracy.