Research On Capacitive Sensor Of Soil Moisture Based On RC Steady-state Response And Its Application System

Soil moisture detection is of great significance for precision irrigation and intelligent agriculture.The development of low-cost and high-precision soil moisture sensor is an urgent bottleneck for the modern intelligent agriculture.With the emergence of high-quality and low-cost high-frequency oscillator,capacitive sensor has become the focus of domestic and foreign scholars.It has the advantages of relatively cheap capacitance probe,different geometry which can meet the needs of different applications,and online rapid automatic detection easily.However,the capacitance sensor still has the problems of unreasonable structure,imperfect signal processing circuit,and easy to be affected by soil type and conductivity,which leads to the poor applicability of the detection object and low detection accuracy.In order to solve this problem,this paper analyzed the research status of soil moisture sensors at home and abroad,and improved the structure of the capacitive sensor probe.Based on the sensitive response mechanism of capacitive sensor to soil moisture,the amplitude detection method of soil moisture capacitive sensor based on RC series steady-state response was proposed.We prepared a low-cost capacitive sensor and proposed a new two-step calibration method to constructed the relationship model of soil moisture,dielectric constant and sensor output.Finally,a modified model was presented based on conductivity compensation,and developed an application system of soil moisture detection was developed,which included soil moisture sensor,acquisition recorder and PC terminal.The main research contents and results are as follows:（1）Research on soil moisture capacitive sensor and detection circuit:the structure of the sensor probe was improved.The two plates has no slot ensures the good insulation and strength of the probe,so as to avoid the influence of the slotted gap on the measurement reliability;based on the theoretical calculation of capacitance value of single plane scattering field,we used ANSYS to simulate the two-dimensional electric field of capacitance sensor probe,and obtained the relationship between probe size,sensitivity and sensitive depth.The most suitable size of probe width and space is6 mm;based on RC series steady-state response amplitude detection method,we added a slow charging and fast discharging circuit in the traditional first-order RC series circuit.It has the advantages of improving the sensitivity and anti-conductivity of the capacitance sensor.（2）Performance test and calibration of soil moisture capacitance sensor:firstly,the sensor was made by PCB four layer technology,and its performance was tested.The results show that:the dynamic response time of the sensor is within 540ms,and the suggested switching time of the sensor should be more than 600ms;the added voltage stabilizing circuit makes the output of the sensor not affected by the power supply voltage,and the maximum working current is only 4mA;through the test of sensitive area,the sensitive dimensions of the sensor in L1,L2 and L3 directions are3.8 cm,2.5 cm and 7.2 cm respectively,and the sensitive soil volume of the sensor is about 214.78 cm³;the output of the sensor has good consistency.The maximum single point repetition error is 1.0mv,and the maximum interchangeability error is 1.5mV.There is no significant difference between the same type of sensors.It provides the basis for using the same calibration model for the same sensor;the working frequency of the sensor was set to 133MHz.Compared with the working frequency of 100MHz and80MHz,the anti-conductivity performance of 133MHz is the best.In addition,by adding a fast discharge circuit,the influence of conductivity on the sensor was obviously reduced.（3）Research on calibration and compensation of soil moisture capacitive sensor:firstly,aiming at the low accuracy of the existing two-step calibration method,we proposed a new two-step calibration method.The first step was carried out in standard solution,and the exponential relationship model between sensor output and dielectric constant was obtained,with R² of 0.983 and RMSE of 0.039mV;the second step established the third-order polynomial relationship model between soil moisture（volumetric water content VWC）and soil dielectric constant in soil samples,with R²of 0.996.Soil samples were prepared again to verify the detection performance of the new two-step calibration model.The conclusion is that the maximum soil VWC detection error of the model is-0.014 m³·m^-3,and the RMSE is 0.010 m³·m^-3.Compared with the general model,the detection accuracy of has been significantly improved.Then,we analyzed the influence of temperature and conductivity on the sensor.The temperature characteristic test results show that the maximum drift error of output voltage caused by temperature in the air is 8.2mV and the maximum error is 0.007%m³·m^-3in the range of 0～50℃.The temperature has low influence on the sensor.The conductivity characteristic experiment was carried out in the standard solution with conductivity of 0～1.95 dS/m.It shows that the maximum detection error of soil VWC caused by conductivity is less than 0.043 m³·m^-3.Finally,we used the binary regression model to compensate the conductivity effect,and the final calibration equation with conductivity compensation was established.The compensation model is effective and the maximum soil VWC deviation is reduced from 4.3%m³·m^-3 to 0.53%m³·m^-3.（4）Design and experiment of soil moisture detection application system:firstly,based on the design of sensor signal processing and acquisition circuit,we designed a portable soil moisture acquisition recorder.And a soil moisture detection application system was designed based on wireless transmission technology.The data acquisition recorder can realize the function of parameter detection through the key control.The upper terminal of computer and lower terminal of acquisition recorder can communicate through wireless technology.The upper terminal can realize the functions of real-time display,alarm and query of soil moisture and temperature of PC terminal upper computer.Then,the application experiment of the detection system was carried out,and the results show that the absolute value of the maximum soil VWC detection error is 0.89%m³·m^-3 in loess samples with conductivity from 0 to 1.98dS/m,and the detection accuracy can meet the requirements of engineering application.