Research On Key Technologies Of Seawater Nitrate Sensor Based On Deep Ultraviolet Narrow-band Tunable Light Source

The modern human society is inextricably linked to marine ecological environment.However,marine ecological environment is deteriorating due to human social activities.Eutrophication of seawater leads frequent breakouts of red tide,which causes huge damage to society and economy.Nitrate is one of the three major nutrients in seawater,which is an important indicator of marine environmental change and marine research.Therefore,it is of great importance to research a scientific,feasible and convenient method to measure the concentration of nitrate in seawater.Currently,the nitrate measurement standard in China is the cadmium column reduction method.This method should be carried out in laboratory,which will easily cause the sample contamination and seawater condition variation.Therefore,it is necessary to measure the concentration of nitrate in situ.The most widely used nitrate sensors,such as SUNA V2 and OPUS,are based on ultraviolet(UV)spectrophotometry.The light source of these products is broadband,which does not meet the basic requirement of Beer-Lambert’s law.Therefore,based on the actual application requirements and existing technique problems,a novel nitrate measurement system with narrow-band UV light source is designed and optimized based on the existing nitrate sensor.Based on the spectral data of seawater samples in different sea areas,the prediction model is trained and established to verify the feasibility of the method.The main contents are as followed:(1)the measurement principle of the system,Beer-Lambert’s law,is introduced.By the transfer matrix,the central wavelength changing with incident angle is proved,which provides a theoretical basis for the tunable central wavelength of the light source.The calibration formula of the incident angle and the central wavelength is derived by using the multi-interference theory,which is convenient to determine the corresponding central wavelength at different incident angles of each filter.(2)Based on the system principle and the central wavelength tuning theory,the system is established.To realize the miniaturization of the system,the narrow-band light source and control circuit module are designed,and the other components are selected.(3)To realize the automatic data acquisition and analysis,the control system program and upper computer program are compiled.(4)The narrow-band spectrum processing algorithm is used to process the multiple narrow-band spectrum into absorbance spectrum.After that,the temperature and salinity correction algorithm is established.(5)Different seawater samples with different nitrate concentrations are configured.Support vector machine(SVM)is used as the nitrate concentration prediction method,and the prediction error with or without temperature and salinity correction is analyzed and compared.The innovations of this thesis are as follows:(1)a miniaturized central wavelength tuning light source is independently designed to filter the original broadband UV light into narrow-band light with good monochromaticity,which is in line with the requirement of Beer-Lambert’s law.(2)A narrow-band spectrum processing algorithm is proposed,and the temperature and salinity correction algorithm(TSC)for absorbance spectrum is established to reduce the prediction error.The experimental results show that the system realizes the central wavelength tuning of the light source and automatic spectral data acquisition.The calculation results show that the error ranges of broadband light source system with SVM and Seabird’s algorithm(ISUS)are-5.34～5.16 and-5.82～5.65,while with TSC combined with SVM(TSC-SVM),the error ranges of broadband light source system is-3.47～4.57.The error ranges of narrowband light source system with SVM and Seabird’s algorithm are-4.03～3.01 and5.22～5.61,while with TSC-SVM,the error ranges of broadband light source system is-3.01～2.99.It follows that the error ranges of the narrowband spectral data are less than those of the broadband data,and the error range predicted by the TSC-SVM is less than the error ranges predicted by the SVM and ISUS.The narrowband nitrate measurement system is therefore most accurate in conjunction with the TSC-SVM.