Research Projects

The proposed project aims to develop a self-propellant crawler inline inspection tool for pipelines using nondestructive evaluation (NDE) techniques and artificial intelligence (AI) to ensure the safety and reliability of pipeline infrastructures. NDE techniques offer the advantage of examining pipelines without causing damage, allowing for early detection of potential defects like corrosion, cracks, or material thinning. This proactive approach helps prevent catastrophic failures, environmental damage, and loss of life. The NDE sensors are mounted to the self-propellant crawler to inspect the pipelines, which allows one to access the region that can not be inspected using conventional inspection methods. This will result in real-time pipeline monitoring that provides quantitative and volumetric information, which will be the first step in developing a reliable system. However, Traditional manual inspections are time-consuming and labor-intensive. Crawler-based inspection offers efficiency but faces challenges in capturing reflected ultrasonic signals in large-diameter pipelines due to the slower crawler movement than wave propagation. This limitation can increase inspection time and cost. Therefore, using the near-real-time AI simulator conditional generative adversarial networks (C-GAN) to obtain the ultrasonic signal data directly instead of using time-consuming experimental measurements is a new and novel approach for the 3D thickness mapping of the ultrasonic signal for the pipelines. Using AI algorithms for automated defect recognition (ADR), defect detection and classification, and real-time control is a novel pipeline systems inspection. Our ADR system, trained with simulation-assisted C-GAN-generated data, automates pipeline defect detection and classification. This innovative approach reduces human intervention, improves inspection speed, and enhances safety and reliability. This project can drive the creative development of ultrasonic Testing (UT) modules, probes, and calibration procedures, self-powered crawler technologies for 12 inches pipes, UT software platform for signal recording and imaging, AI for 3-D mapping of the ultrasonic signals for the entire pipeline, and 3-dimensional convolutional neural networks for ADR. The recorded NDE data can be analyzed using machine learning algorithms to gain valuable insights into pipeline performance and degradation patterns, enabling informed decision-making. This technology development compliance regulatory industry standard ensures adherence to stringent pipeline safety regulations and industry best practices. These developments and implementation will improve the accuracy and efficiency of pipeline safety.

Nondestructive Evaluation, Pipeline Inspection, Inline inspection tool, Ultrasonic Testing, Generative AI, Machine learning