Study On Error Analysis And Compensation Method Of Electrode-Type Conductivity Sensor

Conductivity is an index that characterizes the conductivity of liquids,and is an important attribute of liquids.The conductivity index of liquid is widely used in many fields such as marine science,chemical industry,biology and medicine.At present,electrode-type conductivity sensor is the most widely used in many fields because of its simple structure and low cost.Therefore,the error analysis and compensation research of the electrode-type conductivity sensor has important practical significance for scientific research and practical production.This paper will analyze the error of the electrode-type conductivity sensor,and build the electrode-type conductivity sensor data acquisition system to obtain the measurement data under different external factors through experiments,and use the machine learning model to carry out the error compensation research based on the measurement data.Firstly,the measurement principle and error generation principle of the electrode-type conductivity sensor are studied.In the part of the measurement principle of conductivity,the electrode-type conductivity sensor with different structures and the signal detection circuit commonly used are mainly studied.The error interference items that affect the measurement results of the electrode-type conductivity sensor are analyzed separately.In addition to the error of the instrument itself,the environmental factors that affect the measurement of the electrodetype conductivity sensor mainly include temperature factors,pressure factors,electrochemical factors including polarization effects and capacitance effects,etc.,and the mechanism of error generation is studied separately.Secondly,in order to realize the error research of the electrode-type conductivity sensor,a conductivity sensor data acquisition system is built.In order to obtain multiple parameters including conductivity measurement value,conductivity true value,temperature,pressure,excitation frequency,and excitation amplitude at the same time,the seabird CTD and self-made conductivity measurement circuit are used for synchronous measurement.The self-made conductivity measurement circuit is based on the bipolar op amp method and uses the AduCM360 microprocessor as the control core to measure the three-electrode conductivity sensor.The hardware circuit design and the software design of the lower computer are respectively carried out on the measurement circuit.Then,based on the conductivity sensor data acquisition system design experiment,and set up an experimental environment to complete the required data collection work.The directly collected data requires certain data processing before it can be used.The main data processing process includes matching and splicing the data sent by the two devices according to the time stamp,and removing the error items and duplicates,and finally establishing the electrode conductivity Sensor error research dataset.Finally,a variety of methods are used to establish an error regression model on the dataset established above.Mainly used include polynomial regression,support vector regression machine(SVR),automatic machine learning(AutoML),using a variety of evaluation indicators to evaluate different methods,and finally determine the best error compensation model.