Examples of DX initiatives Phased Array UT Inspection System for Tube to Tubesheet Weld Joints:
kantanPAUT®
- ・Winner of the Grand Prize at the 2nd Deep Learning Business Awards
- ・Our case studies featured in the "Deep Learning Textbook: Official Text for the Deep Learning Generalist Certification (JDLA Deep Learning for GENERAL) 3rd Edition“
Background
- ・Due to fluid leakage at the tube to tubesheet weld joints of the shell and tube heat exchanger during operation, a precise inspection becomes necessary.
Issues before the initiative
- ・The precise inspection of thousands of tube to tubesheet weld joints required an enormous amount of time and cost.
Response
- ・Developed an automated defect inspection system based on images of flaw detection data using AI.
Necessary data
- ・Utilized phased array technology to image ultrasonic waveform data.
- ・Automatically acquired flaw detection data at 1° increments over the full 360°.
- ・Leveraged PaaS to automatically determine flaw detection data in the cloud.
Outcome
-
・Realized substantial savings in both inspection time and cost;
- 1.Visually re-evaluated only images that AI identifies as defective.
- 2.Accurately differentiated noise similar to defects to reduce false positives.
- 3.Reduced inspection time by approximately 80% compared to the traditional method.
