Design Of Fringing Electric Field Sensors And Study On Parameter Estimation Algorithm

Fringing electric field (FEF) sensors based on capacitance edge effects are widely used in non-destructive measurement of product characteristics in industrial process control, including thickness, damage, moisture, porosity, viscosity and density, with the advantages of single-side access, adjustable signal strength and tomography capability.This thesis analyzes characteristic indices and structural parameters of FEF sensor based on the finite element model. Combining two-dimension and three-dimensional model, the method of orthogonal optimization design and regression analysis is proposed to find optimal structural parameters. The simulation results show that the characteristic of the sensor with optimal parameters is better.The single-wavelength FEF sensor and multi-wavelength FEF sensor is developed using printed circuit board according to the optimum structure parameters. And then, sensors are calibrated with Lab View-based measurement system, the standard capacitance is used to correct the value of the feedback capacitance of the measurement circuit to improve the accuracy of the measurement system. Then repetitive calibration and crosstalk analysis of the fringing electric field sensor is carried on the corrected circuit. The experiment results show that the measurement repeatability of the FEF sensors based on PCB technology is good, and there is no crosstalk between channels of three-wavelength sensor.Parameter estimation algorithm for fringing electric field (FEF) sensors is studied in this thesis. Linear regression parameter estimation method is proposed for the single-wavelength sensor, and multivariate nonlinear regression parameter estimation method proposed for three-wavelength sensor to establish a functional relationship between transcapacitance value and the sample permittivity.Experiments for power cable insulation testing, PCB plate flaw and thickness measurement are done. The results show that the single-wavelength FEF sensor can detect sample damage and measure the thickness of the samples, and the three-wavelength sensor is able to distinguish dielectric properties of the sample layers effectively.