Status: Active, 2018-2021
Role: institutional PI (40% share)
Degradation of the pipeline health is susceptible to hazard due to failure. To prevent such failures, a major challenge for the maintenance crew to detect and repair corrosion still prevails due to difficult and expensive accessibility during scheduled maintenance. The proposed method will focus on the development of novel structural light-based imaging for internal corrosion detection, which simplifies the detection process while achieving superior spatial resolution. The proposed approach will develop an endoscopic structured light scanning tool that is based on phase measurement profilometry (PMP). The developed system will be simple to fabricate and easy to be used by maintenance personnel with minimal skillset due to its intuitive scans. The structured light system will be developed to generate high-resolution reconstructed images representing surface texture with high accuracy. Based on the images, additional processing capabilities developed using Bayesian updating technique will give the capability of automatic classification and identification of different types of precursors. A convolutional neural network based corrosion detection method will provide automated detection, which further minimizes the operator involvement. The uncertainty quantification technique will be integrated to enhance the probability of detection and to quantitatively determine the damage size and location.