Research On Temperature Compensation Method Of Eddy Current Sensor

The eddy current sensor is widely used in industrial production due to its high sensitivity,non-contact measurement and being not affected by the interference of smoke and dust.It can achieve high-precision measurement at room temperature,but in high-temperature environments,the measurement accuracy of the eddy current sensor will deteriorate.This is mainly due to changes in the ambient temperature T of the conductor under test and excitation frequency f,which changes the magnetic permeability ? of the conductor under test,thereby affecting the self-parameters of the probe coil of the eddy current sensor(equivalent inductance L and quality factor Q),so that the measurement accuracy of the eddy current sensor is reduced or even normal measurement cannot be continued.Aiming at the practical problems of the above eddy current sensor detection,the corresponding research is carried out in this paper,the specific content is as follows:The working principle of eddy current sensor,the factors affecting the magnetic permeability,etc.are analyzed.It is found that the excitation frequency,the ambient temperature of the conductor under test and other factors have a very complicated effect mechanism on the eddy current sensor probe coil's own parameters(L and Q).The relationship between the quality factor,equivalent inductance value,excitation frequency of the eddy current sensor probe coil,the displacement measured by the sensor,and the ambient temperature of the conductor under test cannot be obtained by analytical methods or simple nonlinear fitting methods.So put the eddy current sensor probe in the temperature control box,start from room temperature,the gradient(?T=20?)is heated to 400?,after each temperature gradient is kept for a sufficient time(40 minutes).Then using the scan Mode of LCR,when the excitation frequency f changes from 100kHz-1MHz(?f=10kHz),scan and record the data of inductance L and quality factor Q of the eddy current sensor probe coil with detected displacement x(detected displacement range 0-30mm,?x=2mm)change,and save the collected temperature T,detected displacement x,inductance L,excitation frequency f,quality factor Q.Then use the relevant data saved in the above experiment,and use the neural network method to establish a model between the quality factor,equivalent inductance value,excitation frequency of the sensor coil of the eddy current sensor probe,the displacement measured by the sensor,and the ambient temperature of the conductor being measuredIn this paper,the LSTM network model is established.The input variables of the model are the equivalent inductance L,excitation frequency f,and quality factor Q of the probe coil of the eddy current sensor,and the output variables are the displacement x measured by the eddy current sensor and the ambient temperature T of the measured conductor;Then,the simulation experiment was conducted in Python,and the LSTM network model was compared with the traditional BP,RBF,MOSVR network model.The experimental results show that the LSTM network model can effectively predict the displacement x measured by the eddy current sensor and the ambient temperature T of the measured conductor,and is superior to other prediction methods.Finally,the BP and LSTM network models are implemented on the STM32H743?T6 microcontroller,which has important practical significance for the future temperature compensation of the eddy current sensor and the simultaneous detection of the temperature and displacement of the smart sensor under high temperature conditions.