Pulse-modulation eddy current imaging and evaluation of subsurface corrosion via the improved small sub-domain filtering method

Abstract

Subsurface corrosion occurring in the conductive structures prone to long-term exposure to harsh and corrosive environments essentially leads to the reduction in the critical area of load carrying members, adversely aggravating the mechanical strength and structural integrity of in-service structures. This has prompted the high demand for non-invasive detection and evaluation of subsurface corrosion using non-destructive evaluation techniques. Pulse-modulation Eddy Current testing (PMEC) has been found to be superior to pulsed eddy current testing in terms of higher sensitivity to subsurface corrosion and accuracy in corrosion imaging. In this paper, a funnel-shaped PMEC probe is utilised for subsurface corrosion imaging along with the inspection sensitivity further enhanced. A novel signal feature of PMEC, namely the averaged peak value is proposed in a bid to improve the signal-to-noise ratio of the PMEC testing, based on which corrosion images are produced. In order to enhance accuracies in image-based profile identification and opening-area assessment of subsurface corrosion, an improve Small Sub-domain Filtering Method (SSFM) together with Canny algorithm is proposed and exploited in the image processing for corrosion-profile recognition. It has been indicated from the experimental results that the subsurface corrosion profile can be well identified using the proposed signal and image processing methods which could keep robust in the testing environment subject to extraneous noises. For the experimental specimens, the accuracy of the subsequent evaluation particularly regarding the corrosion-opening area reaches 94.5% at minimum.