- This event has passed.
Ph.D. Defense: Huda Aldosari
November 11 @ 1:00 pm - 2:00 pm EST
HTitle: New Architectures using Machine Learning to Detect and Predict Underwater Oil and Gas pipelines Defects
Ph.D. Candidate: Huda Aldosari
Major Advisor: Dr. Reda Ammar
Associate Advisors: Dr. Sanguthevar Rajasekaran, Dr. Song Han, and Dr. Raafat Elfouly ( External Associate Advisor)
Review Committee Members: Dr. Yufeng Wu, and Dr. Sheida Nabavi
Date/Time: Thursday, Nov. 11th, 2021, 1:00P.M. – 2:00P.M.
Meeting number: 2623 994 0992
Join by phone: +1-415-655-0002 US Toll
Access code: 2623 994 0992
Underwater pipelines are widely being used all over the world for the transportation of hydrocarbon fluids over millions of miles. Pipeline structures are designed to combat the environmental variations and loading conditions for ensuring reliable and safe distribution from the production venue to the distribution depot. However, the risk of terrible accidents due to the leakages in the pipeline system is always present. Failure in the pipeline networks can cause severe financial loss, human casualties, and ecological disasters. Although various researchers have done the right amount of work by proposing different monitoring architectures to make oil and gas pipelines more reliable and defectless, these pipelines usually suffer from some corrosions or defects which require an immediate response from an onshore control station to avert any expected catastrophe. A reliable and robust monitoring architecture is required to detect and localize the leakage in underwater oil & gas pipelines helping the authorities inspect the faulty point accurately in real-time and maintain a reliable infrastructure of the pipelines network to avoid such disastrous incidents. In this study, an oil and gas dataset based on the concept of magnetic flux leakage signals has been used for designing robust and authentic monitoring architectures to make sure safe underwater oil and gas transmission. Strategies based on ML-machine learning have been proposed for localizing the faults in pipelines.
Moreover, the optimization of the ML algorithm for real-time efficient application has also been considered in this work. The presented system has exhibited encouraging expected outcomes indicating that this automated monitoring system augmented with ML-machine learning techniques can successfully be implemented for precise localization, detection, and prediction of a fault in underwater oil and gas pipelines in real-time.