Development of an Image Analysis Protocol to Define Noise in Wet Magnetic Particle Inspection

Abstract

This study presents a novel method to quantify the effectiveness of wet magnetic particle inspection (MPI) when detecting possible defects. Wet MPI is an established method utilizing magnetic fields to locate possible areas of defects in ferromagnetic parts. The capability of this method has been evaluated in the past most notably using the probability of detection graphs. However, MPI requires a large amount of data and is subjective because it is based on human operators’ evaluation. The method proposed in this paper is an objective method to determine the effectiveness of the MPI test based on how well discontinuities can be delineated in the image. This approach utilizes the intensity of the particle illumination in the defect area and compares it to its surroundings. This analysis generates a value to objectively represent how well a discontinuity can be identified. This method was then used to validate the effect of surface roughness on the effectiveness of wet MPI using two experiments. The first experiment was conducted to test for the collection of particles on varying surface roughness levels, and the second experiment was used to evaluate the effect of surface roughness when detecting a subsurface discontinuity. Results indicate that there is a significant increase in particle collection as roughness increases, and as the surface roughness increases, the harder it is to locate discontinuities. This method provides a quantitative measure that could be used to aid parameter selection.