Abstract
Fatigue is identified as a significant degradation mode that affects nuclear power plants world-wide. Recent research on the interaction between fatigue degradation and the influence of PWR environment has caused international concern and triggered numerous research programs [1]. In this context, several codes & standards, including the RCC-M code, have included some technical mandatory or non-mandatory sections to address the issue. In RCC-M, this is compiled in the Rules in Probation Phase 2 and 3 [2].
Due to the novelty of these rules, there is room for improvement for the specific and practical implementation of these rules. AFCEN has hence launched a benchmark exercise at the end of 2019 to help increase the quality of these rules.
Part 1 of this paper [3] states that EDF and CNNC/NPIC have launched an effort to benchmark their respective codes on fatigue calculation including the EAF algorithm. In the second part of the benchmark, the two companies started the code comparison based on a benchmark case provided by AFCEN.
As stated previously, the 2016 edition of RCC-M code integrates the modifications made to the Code in Probation Phase 2 and 3(RPP)[2], which respectively modify the fatigue design curve for austenitic stainless steels and Nickel base alloys, as well as integrate environmental effects in the fatigue evaluation for austenitic stainless steel components. In this paper, a comparison between RCC-M RPP and NUREG/CR-6909 rev.1 [3] is proposed. The comparison focuses on the technical details of the strain rate calculation and transient combination method.
The cumulative fatigue usage factor with or without considering EAF according to RCC-M RPP – 2 and RPP – 3 is given by EDF, using code_aster and its POST_RCCM operator. CNNC/NPIC will provide multiple sets of results including cumulative fatigue usage factors according to RCC-M RPP and NUREG CR/6909 rev. 1 respectively using its own software. Comparison of selection for peak and valleys points, Sn and Fen values are also presented. Differences of the algorithms of the two codes are also discussed.
