Nuclear power plants (NPPs) are composed of various types of pipes, and these piping systems are designed, manufactured, and monitored according to strict standards because various types of loads can affect the effective functioning of an NPP. Under the existing design criteria for seismic loading conditions, equipment and facilities of NPPs are designed to exhibit an elastic behavior, and the structural integrity is evaluated by applying linear elastic analysis and stress-based acceptance criteria. However, when an earthquake exceeds the design criteria, a large-amplitude load is repeatedly applied; consequently, the nonlinear behavior of the equipment should be considered in flaw evaluations. In the current procedure for flaw evaluation, the stress intensity factor range is generally used as the main fatigue parameter. However, as this evaluation is an analysis based on linear elastic fracture mechanics, simulating an elastic–plastic behavior is difficult.
Therefore, this study aims to develop a fatigue crack growth model that can be used under conditions of large and variable loads such as seismic loads and to analyze the characteristics of crack growth in pipes. In this study, as a factor of crack growth, the concept of change in net-section strain energy, which can be applied to both elastic and plastic regions and has the advantage of simplifying the relationship, was applied. First, the results of finite element analysis on plate specimens were compared with the experimental results of other studies, and we confirmed that these results can be applied regardless of plastic deformation. Moreover, a crack growth model is proposed to predict the crack growth rate. The performance of this model was verified through comparison with experimental results and existing crack equations under variable loads. The validated methodology was extended to semi-elliptical cracks in the piping system. Unlike cracks in plate specimens, a geometrical correction factor was considered owing to the complex shape of cracks in pipes. The crack growth rate could be predicted through a neural net fitting, and the accuracy was confirmed by comparison with actual experiments results. In contrast to the crack evaluation method based on the American Society of Mechanical Engineers standard, the proposed crack growth prediction methodology obtained the results by considering plastic deformation. Thus, the results showed that the proposed method can be used to evaluate the behavior of cracks under various seismic loading conditions.
원자력 발전소 배관들은 다양한 종류의 하중이 발생하면 발전소에 심각한 손상을 일으킬 수 있기 때문에 엄격한 기준에 따라 설계된다. 설계기준을 초과하는 지진이 발생하면 큰 진폭의 하중이 반복적으로 작용하기 때문에 배관 균열 평가 시에 비선형거동이 고려되어야 한다. 그러나 기존의 평가 절차는 선형탄성 파괴역학을 기반으로 하기 때문에 탄소성 거동을 모사할 수 없다. 본 연구에서는 균열 성장의 인자로서 탄소성 거동을 모사할 수 있는 순면적 변형률 에너지 변화량을 사용하여, 하중비, 균열 속도와의 관계를 이용하여 변동하중 조건에서의 균열 예측 모델을 개발하였다. 또한 배관에 존재하는 반타원형 균열에 형상에 맞도록 기존의 인자들을 확장하고 기계학습을 이용한 인자들의 관계 생성을 통하여 균열 예측방법론을 제시하고 검증하였다. 마지막으로 지진하중이 배관에 부하 되는 상황에서 현재 사용되는 균열 평가 방법과 본 연구에서 제시된 방법론을 통해 균열 진전을 예측하고 결과를 비교 평가하였다.