Optimization of Planar Array Electrostatic Sensor for Metal Surface Defect Detection

Abstract

Static electricity is usually generated in the damage area of the metal surface due to contact friction, and the charge distribution density can reflect the size, shape, and relative position of the damage area. Based on this, a planar array electrostatic sensor is designed to detect metal surface defect in this paper, and the shielding method, number of electrodes, electrode shape, and arrangement of the sensor are optimized taking account of the induced charge value, the uniformity of sensitivity and the image correlation coefficient. Different image reconstruction algorithms (e.g. Landweber algorithm, conjugate Gradient algorithm, Tikhonov regularization and primary dual interior point method) are utilized to evaluate the performance of the designed electrostatic sensor. The results demonstrated that the sensor with hexagonal electrode shape, integrated shielding, a new arrangement, a duty cycle of 80%, and a peripheral shielding electrode, has better image quality for all the tested damage models. When using the PDIPA algorithm for image reconstruction, the image correlation coefficient can exceed 0.9.