Industrial Consultancy & Sponsored Research (IC&SR) , IIT Madras
Technology Category/Market
AI based NDT datasets generating method & System.
Application: Automated Defect Recognition (ADR) System; Software for automated Defect Recognition, Visual/Surface/volumetric Inspection.
Market: The NDT testing software market is expected to reach $853.7 million by 2026, registering expansion at a CAGR of 11.1%.
Image Gallery
Problem Statement
- Non-destructive testing (NDT) is the process of inspecting, testing or evaluating materials components or assemblies for characteristics differences or welding defects, discontinuities etc. without causing damage to the serviceability of such material components or assemblies.
- The technical problem underlying the invention is stated that “how to provide accurate detection and classification of defects in NDT/NDE.”
- Present Invention provides the technical solution to the technical problem of the existing NDT method by integrating the artificial intelligence (AI) automation system for generating a large volume of NDT datasets.
Technology
Present invention describes an AI based time efficient method and system for generating synthetic non-destructive training datasets.
- The system determines a CAD model representing the actual physical defect sample based on the received geometrical features, further including critical statistical distribution parameters, and generates a synthetic NDT dataset based on training the AI model.
- The method comprises a few steps depicted in the figures. A smart chart shows herein below:
- Receiving real time experimental NDT datasets by the processor; & Performing numerical analysis on said dataset via numerical solution model.
- Training a deep convolutional generative adversarial network (DCGAN) by using the generated NDT datasets with flaw geometrical features.
- Receiving random number input vectors iteratively at the trained DCGAN; & Generating a synthetic NDT datasets for each of the received said input vector by the trained DCGAN.
Key Features/Value Proposition
Technical Perspective:
Present system provides AI driven NDT datasets, wherein the testing dataset includes dimensions of defective samples, expected defect morphologies, defect probabilities, the sensitivity of instruments, observation from experimental datasets, and noise from instrumentation.
Industrial Perspective:
- Claimed system facilitates an automated, robust, highly scalable, time efficient platform to generate a large volume of synthetic NDT datasets.
- The system reduces computational resources and time by a factor of N/n.
Questions about this Technology?
Contact for Licensing
Research lab
Prof. Krishnan Balasubramanian
Department of Mechanical Engineering
Intellectual Property
IITM IDF Ref.: 2124.
Patent Application No. 202141007067
PCT Application No. PCT/IN2022/050125
Technology Readiness Level
TRL- 3/4
Proof of Concept Ready & tested and validated in Laboratory stage.