Identification of depth and size of subsurface defects by a multiple-voltage probe sensor: analytical and neural network techniques

摘要

A theoretical study is presented for a multiple-voltage probe sensor to detect nonconducting inclusions in a conducting media. We show that the multiple-voltage probe sensor is capable of carrying out precise quantitative measurements of submerged non-conducting objects if the surface voltage response has a standard two-peak form. The standard response is observed for well-localized non-slender single inclusions below the sensor surface. In this case, the peak separation distance is associated with the inclusion depth, whereas the peak magnitude is associated with the inclusion volume. The predefined form of the surface voltage response makes it possible to recognize inclusion responses at very high noise levels. This is carried out using a 2D neural network classifier, based on the probabilistic neural network. A reasonable recognition error of less than 20 percent is obtained if the signal to noise ratio is larger than or equal to 1/10.