Study On The Method Of Measuring Water Carbonate And Bicarbonate By Electrode Method Based On BP Neural Network

Carbonate(CO₃^2-)and bicarbonate（HCO₃^-）are the main forms of carbon in aqueous solution.In the environment,the content of CO₃^2-and HCO₃^-in natural water is an important basis for characterizing the geochemical carbon behavior and carbon cycle.Medically,CO₃^2-regulates human acid-base balance and maintains environmental homeostasis.In industry,CO₃^2-and HCO₃^-are the main anions in crude oil separation water.In agriculture,when CO₃^2-is one of the components of soil soluble salt content,its content will directly affect the germination and growth of crops.However,at present,for the measurement of water quality CO₃^2-and HCO₃^-,it is usually necessary to take samples in the field,and then send them to the laboratory for determination by titration,chromatography,and spectroscopy.These methods not only large,complex operation and CO₃^2-,HCO₃^-concentration in the solution with water,factors such as pH,temperature,atmospheric pressure changes,once samples collected from the mother,the carbonate concentration change,lead to test result is not representative of maternal itself CO₃^2-and the concentration of HCO₃^-,and ask the more long,with this deviation.Therefore,it is necessary to establish a method to quickly measure the concentrations of CO₃^2-and HCO₃^-in water quality.In this paper,aiming at the defects of large measurement error of carbon dioxide（CO₂）and difficult determination of equilibrium constant of carbonic acid dissociation in indirect measurement of water quality CO₃^2-and HCO₃^-concentration by double electrode method,CO₂ measurement method and double electrode method were studied respectively.For the CO₂measurement method research,first of all,according to the characteristics of designed electrode output signal,signal modulation circuit and printed circuit board（Pcb）routing strategy,then according to the requirements of on-site rapid measurement applications,such as electric potential point temperature compensating model is established and electrode temperature compensation model,at last,through the experiment on the two models to improve and comparative analysis,it is concluded that electrode live full compensation model is suitable for the rapid measurement of temperature,stability is good,the measurement precision is better than 2%;For double electrode measurement method research,respectively established carbonic acid dissociation equilibrium constant fitting model and genetic algorithm to optimize the BP neural network,through analysis and comparison,the existing double electrode method need only on the basis of pH and concentration of CO₂ is always add temperature variables to establish three input two output after genetic algorithm to optimize the BP neural network,a good ignores the carbonic acid dissociation equilibrium constant,the influence of double electrode method improves the measuring precision,the experimental results show that this method CO₃^2-measurement precision is better than that of 0.5 mg/L,HCO₃^-measurement precision is better than 6.2%,has some reference value for related design.