Header image

AI-based Auto-classification of PE plastic pipe butt fusion joints using custom PAUT inspection system

Tracks
Advanced UT Presented by TPAC
Wednesday, October 23, 2024
11:30 AM - 12:00 PM
127

Speaker

Mohammad Marvasti
Ultrasonic Application Development Engineer

AI-based Auto-classification of PE plastic pipe butt fusion joints using custom PAUT inspection system

Presentation Description

A custom Phased Array Ultrasonic Testing (PAUT) inspection tool, comprising both hardware and software, has been developed for the inspection of polyethylene (PE) pipe butt fusion joints. This system is designed for use by gas pipeline welding technicians with no prior knowledge of ultrasonic testing. The inspection technique was qualified using a set of butt fusion joint samples containing real, embedded defects. Scan results from these samples were then employed to train autoencoder models which are used for automated classification of butt fusion joints (flawed vs. unflawed) using the PAUT data. This presentation will cover the system's design, the inspection methodology, the AI training process, and the results of the qualification tests.

Biography

Mohammad is an Ultrasonic Application Development Engineer with 10 years of experience specializing in developing custom ultrasonic inspection solutions. His focus is on utilizing advanced imaging techniques to create practical, field-deployable systems for challenging inspection applications. His expertise spans a range of key ultrasonic technologies, including Phased Array Ultrasonic Testing (PAUT), Phase Coherence Imaging (PCI), Full Matrix Capture -Total Focusing Method (FMC/TFM), and Time of Flight Diffraction (ToFD). He is also proficient in CIVA modeling, imaging and signal processing, Probability of Detection (PoD) studies, and the development of inspection and analysis procedures for field applications.
loading