Component Testing Proposal to Quantify Margins in Existing Environmentally Assisted Fatigue (EAF) Requirements

Abstract

Research has led to an update of the various existing fatigue rules to account for aggravating effects, chief among them being the effect of PWR and BWR primary system environments on fatigue life. Among the many documents and evaluation rules published around the globe, NUREG/CR-6909 [1] has established itself, since its issue in 2007, as a reference document on the international scene. It is also the technical basis document of NRC Regulatory Guide 1.207. Regulatory Guide 1.207 requires application of NURREG/CR-6909 to new US plants and plants seeking a renewed license. After the issue of this first version of NUREG/CR-6909, international experts in fatigue pointed out some potential areas of improvement of the document and compiled the recommendations into a unified EPRI roadmap documents [2,3]. These documents identified areas where knowledge gaps exist and has been used to guide recent experimental efforts led by EPRI [4]. Additional international testing campaigns [5,6] have contributed also to industry’s understanding of the physical phenomena of fatigue and the interaction between various aggravating effects and these efforts are continuing today. [7] These new results, ongoing research, and industry concerns related to the requirements of NUREG/CR-6909 have led to the issue of Revision 1 of NUREG/CR-6909 as well as alternative rules to account for EAF in crack initiation phase [7] and crack growth phase [8]. Nevertheless, these requirements are built along the same philosophy as previous ones, namely the idea of adding conservatism on top of conservatism, thus not recognizing the accurate probabilistic margin that exists. Operating plan experience has been overall very positive and demonstrates the high degree of conservatism afforded by the existing requirements. With the extension of reactor lives beyond 40 years or even 60 years being considered, margins are growing thinner. An assessment of how well the new requirements predict fatigue failure when compared to a real component’s operational history provides invaluable insight in support of the safe operation of nuclear power plants. Such a comparison makes up the process of checks and balances that any new requirement should be judged against and modified accordingly where necessary. Testing on small scale fatigue specimens enters in the same logic of studying one or an interaction of limited number of effects to support requirements covering only the specific effect(s) studied. Such an approach masks margin evaluation and does not reveal the inconsistency between fatigue failure prediction and operating plant experience. To fulfill the objective of assessing the margins in these test methods and their associated results, a solution is to carry out a full scale component test under prototypical test conditions. This paper proposes views on what such a test, including preparation for such a test, should be provided in a proposal of conditions to be tested and ideas for test geometries.