The Core of Non-Destructive Testing: ECT
ECT (Eddy Current Testing) is a non-destructive testing method that utilizes electromagnetic induction to accurately diagnose corrosion, cracks, and wear in metallic equipment. It is particularly effective in precisely analyzing the condition of critical equipment such as heat exchanger and steam generator tubes, maximizing their stability and value. This method extends the lifespan of equipment, prevents unexpected failures, and enables efficient facility management.
Vision
Shaping the Future of ECT Inspection with AI Technology
DEEP-AI's goal is to maximize equipment safety and revolutionize inspection efficiency through ECT technology. By integrating AI and data analytics, DEEP-AI overcomes the limitations of traditional inspection methods, providing customers with more reliable and trustworthy results. Through continuous research and development, DEEP-AI enhances the precision and application scope of ECT technology, leading the future of non-destructive testing.
R&D
High-Speed Data Processing and Analysis
Processes and analyzes large inspection datasets in real time, maximizing inspection efficiency.
Key Research Areas
1. Development of Real-Time Data Processing Systems
• Implementation of High-Speed Data Collection and Processing Algorithms
• Enhancing Analysis Speed Through Parallel Processing
• Cloud-Based Data Storage and Management Systems
2. Establishment of an Integrated Inspection Data Analysis Platform
• History Data Management and Tracking
• Generation of Customized Analysis Reports by Equipment
• Statistical Analysis and Visualization of Inspection Results
AI-Based Automatic Detection of Structures and Defects
By utilizing AI, reliable automatic defect detection and classification from ECT signals is achieved.
Key Research Areas
1. Development of Deep Learning-Based Structural Classification Models
• Implementation of an Optimal Calibration Signal Algorithm for Structural Detection
• Accurate Classification of Structural Types Through Extensive Learning
2. Development of a Deep Learning-Based Defect Classification Model
• Accurate Defect Type Classification Through Learning of Various Defect Patterns
• Improved Accuracy in Measuring Defect Size and Depth
3. Development of an Automated Inspection Report Generation System
Achievements
Obtained EPRI Certification from the Electric Power Research Institute
It is the first AI-based non-destructive solution to obtain certification from the Automated Analysis Performance Demonstration Dataset (AAPDD) program, organized by the Electric Power Research Institute (EPRI) in the United States.
PoC with Refining and Petrochemical Companies
We have conducted Proof of Concept (PoC) projects with Saudi Aramco, LG Chem, Hanwha TotalEnergies Petrochemical, SK Innovation, S-OIL, HD Hyundai Oilbank, and AI-KWANG TECH, successfully demonstrating the effectiveness and potential of DEEP-AI's technology.
A Total of 5 Patent Applications Filed
We have filed patents related to defect signal interpretation, including a signal acquisition system and method for eddy current testing of heat exchanger tubes, and an eddy current testing data labeling system.