Validation of Valve Leak Quantification With Non-Intrusive Acoustic Emission Technology

Abstract

In the late 1980s and early 90s, several companies tested a range of acoustic devices for monitoring valve leakage during the check-valve diagnostic system research performed at the Utah State Water Research Laboratory as part of two separate nuclear-industry-sponsored initiatives. The acoustic sensor technology and analysis techniques evaluated were found helpful but no progress was made in non-intrusively quantifying the leak rate through the valves tested during these programs. Around that same time, oil & gas companies in the UK were experimenting with detection and quantification of valve leakage using acoustic emission (AE) technology. The AE sensors and signal-processing technology selected for the UK oil & gas effort responded to much higher frequencies compared to the sensors and systems used during the nuclear-utility initiative in the U.S. This research led to new products for detection and quantification of valve leakage in oil & gas applications. Because of minimum leak threshold and accuracy concerns, non-intrusive acoustic valve leak measurement has remained an elusive goal for commercial nuclear power. Various general-purpose acoustic tools have been trialed to detect leakage with mixed results because of complications caused by plant and system acoustic characteristics. Several of today’s moderately successful check-valve diagnostic systems employ acoustic sensors and can detect the most likely event representing flow cutoff when a check-valve disc fully closes, but leak-rate quantification with any of these systems is not possible. Correlation methods and other AE analysis techniques that have been developed to quantify leakage in steam systems have been generalized as small, medium, and large leakage classifications with no clear criteria for these levels. During the last couple of years, nuclear-plant engineers responsible for programs for compliance with Appendix J, “Primary Reactor Containment Leakage Testing for Water-Cooled Power Reactors,” to Part 50, “Domestic Licensing of Production and Utilization Facilities,” of Title 10, “Energy,” of the Code of Federal Regulations (Appendix J to 10 CFR 50) have made extensive use of a new acoustic valve leak-detection system known as MIDAS Meter®. Appendix J valve testing (also known as Type C testing) requires that sections of nuclear-plant piping be isolated by closing a number of valves, thereby creating a confined pressure boundary. The isolated piping within the boundary is pressurized with approximately 344.7 kilopascals (kPa) [50 pounds per square inch (psi)] of air and the leak-tightness of the boundary is evaluated. When the isolated piping exhibits excess leakage or cannot maintain the test pressure, the valves creating the boundary are evaluated one by one to find the culprit leaker. The process of finding and correcting the problem valve can take from hours to several days and may become an outage critical-path activity. Appendix J engineers have enjoyed considerable success with their newfound ability to quickly and confidently identify the leaking valves with MIDAS Meter® and remove their test programs from the critical path. MIDAS Meter® is a high-frequency acoustic-emission-based system which includes algorithms that convert the acoustic emission signal to leak rate. The basic algorithms were first developed from the field results obtained during the early development work for UK oil & gas operators and refined over the next 20 years. Though not originally validated under a quality-assurance (QA) program of the 10 CFR 50 type, nuclear plants that own MIDAS Meter® have been eager to go beyond simple troubleshooting and use the leak quantification results for nuclear applications, including safety-related decisionmaking. In order to support owners and avoid improper application of this very successful new tool, Score Atlanta embarked on an extensive validation program consistent with 10 CFR Part 50 requirements. A purpose-built leak-test flow loop and valve simulator apparatus were constructed in the Atlanta facility and testing began in early 2013. To support Appendix J users, the air testing was performed first and completed in July 2013. The water testing followed and should be completed in early 2014. Numerous combinations of leak path, leak-path geometry, and differential pressure were created and evaluated during the air phase of the program. Pressure was limited to 1034 kPa [150 psi] for air testing. The water testing includes pressures up to 8,618 kPa [1,250 psi] and a similar number of varying leak paths and pressure test points. This paper discusses the preliminary results of the test program, including any special limitations required for use of AE-derived valve leak results in nuclear safety-related applications. The full results of the test program and guidance for nuclear safety-related use of the technology are expected to be available ahead of the 2014 ASME-NRC Valve Symposium. Paper published with permission.