Uncertainties in NDE Reliability and Assessing the Impact on RI-ISI

Abstract

A major thrust in the past 20 years has been to upgrade nondestructive examinations (NDE) for use in inservice inspection (ISI) programs to more effectively manage degradation at operating nuclear power plants. Risk-informed ISI (RI-ISI) is one of the outcomes of this work, and this approach relies heavily on the reliability of NDE, when properly applied, to detect sources of expected degradation. There have been a number of improvements in the reliability of NDE, specifically in ultrasonic testing (UT), through training of examiners, and improved equipment and procedure development. However, the most significant improvements in UT were derived by moving from prescriptive requirements to performance based requirements. Even with these substantial improvements, NDE contains significant uncertainties and RI-ISI programs need to address and accommodate this factor. As part of the work that PNNL is conducting for the U. S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research, we are examining the impact of these uncertainties on the effectiveness of RI-ISI programs. One of the primary objectives of in-service inspection, including a RI-ISI program, is to manage potential degradation that may occur, but that had not been foreseen through previous operating experience. However, RI-ISI programs in the U.S are primarily based on history, looking back at past failures in the operating fleet. Therefore, RI-ISI may not adequately manage degradation events that are yet to occur, such as those that may have a long incubation (initiation) time, but a potentially fast growth rate. For this reason, RI-ISI will always be reactive to such failure events. Successful ISI needs to determine what NDE is required, when and how frequently it needs to be applied, how effective the NDE must be and where the NDE needs to be applied. Both flaw detection and accurate characterization need to be addressed. This paper will examine the reliability and uncertainties of NDE, and how these may impact RI-ISI.