Auto Sizing of CANDU Nuclear Reactor Fuel Channel Flaws from UT Scans

Area of Science:

• Nuclear Engineering
• Materials Science
• Non-Destructive Testing

Background:

• Nuclear power plants require regular inspections during outages to ensure safety and reliability.
• Canada Deuterium Uranium (CANDU®) reactors utilize fuel channels containing pressure tubes, which are critical components housing fuel bundles.
• Current inspection methods for CANDU® pressure tubes rely on manual analysis of Ultrasonic Testing (UT) data.

Purpose of the Study:

• To develop and evaluate automated algorithms for detecting and sizing flaws in CANDU® reactor pressure tubes.
• To compare the performance of proposed automated methods against traditional manual analysis of UT scans.

Main Methods:

• Implementation of two deterministic algorithms: segmented linear regression and average time of flight (ToF) within ±σ of µ.
• Analysis of UT scan data to locate, measure, and characterize pressure tube flaws automatically.
• Comparison of automated results with manually analyzed data from experienced analysts.

Main Results:

• The segmented linear regression algorithm achieved an average depth difference of 0.0180 mm when compared to manual analysis.
• The average ToF algorithm yielded an average depth difference of 0.0206 mm against manual analysis.
• Both automated methods demonstrated results comparable to the 0.0156 mm depth difference observed between two independent manual analyses.

Conclusions:

• The proposed automated algorithms for pressure tube flaw detection and sizing are effective and comparable to manual analysis.
• Adoption of these algorithms in production can lead to substantial cost savings through reduced time and labor.
• The developed methods enhance the efficiency and reliability of essential nuclear power plant inspections.