2026-01-14 アルゴンヌ国立研究所（ANL）
Argonne’s framework could improve maintenance schedules for critical components in nuclear power plants, saving millions of dollars on operating costs. (Image by Shutterstock.)
＜関連情報＞
- https://www.anl.gov/article/cutting-nuclear-power-plant-costs-argonne-develops-framework-for-smarter-maintenance
- https://www.sciencedirect.com/science/article/abs/pii/S1350630725007319
給水加熱器における管の疲労破壊の数値解析と実験検証 Numerical analysis with experimental validation of tube fatigue failure in feedwater heaters
Yeni Li, Akshay J. Dave, Thomas W. Elmer, Matthew Retzer, Richard B. Vilim
Engineering Failure Analysis Available online: 8 August 2025
DOI:https://doi.org/10.1016/j.engfailanal.2025.109990
Highlights
- Develops a multiphysics simulation for fatigue analysis in feedwater heater tubes.
- Builds surrogate model using Gaussian Process regression to predict stress based on operating conditions.
- Quantifies fatigue damage using cumulative damage theory and infers remaining useful life.
- Validates model with inspection and replacement data from operating nuclear power plants.
- Supports predictive maintenance and novel heat exchanger design.
Abstract
This study introduces a comprehensive methodology for analyzing fatigue failures in feedwater heater tubes. It employs a 3D simulation approach that integrates condensation processes, conjugate heat transfer, and one-way fluid–structure interaction. This multi-physics simulation provides detailed insights into the coupled thermal and mechanical stresses affecting feedwater heater tubes during operation. To validate the simulation results, the model’s stress profiles are compared with eddy current inspection data and heater replacement records from two operating nuclear power plant units. Gaussian Process regression is used to model von Mises stress as a function of operating conditions, enabling uncertainty quantification and stress prediction. Cumulative damage theory is then applied to estimate the progression of fatigue-induced damage in the tubes, and their remaining useful life (RUL). The RUL predicted by this methodology accurately envelopes actual plant data for multiple heaters, including the time at which initial tube maintenance was performed (23–25 years) and the time of heater replacement (27–29 years). Additionally, each startup is projected to reduce useful life by approximately 2.8–4.5 % (the limits would be determined by safety factor in range of 1.25–2.0).
The results provide guidance for prioritizing maintenance actions, such as identifying high-risk locations for inspection or early intervention, especially at tube support plate joints where degradation is most common. By aligning model predictions with actual inspection and replacement data, this research boosts confidence in predictive maintenance planning. The significance of this research lies in its potential to connect high-fidelity simulations with practical decision-making processes. It represents a crucial step toward predictive maintenance strategies, equipping operators and engineers with tools to anticipate and address fatigue issues while enhancing system reliability and efficiency.
The analysis methodology can also serve as an important tool for designing and evaluating novel heat exchanger concepts. For example, the growing interest in utilizing nuclear reactor heat to power endothermic industrial processes presents new challenges for heat exchanger design. Higher temperatures, new heat transfer surface geometries, and advanced materials push the design space into a realm with limited operational data. The methodology proposed in this work can expedite the design process by predicting the mechanical stress a novel heat exchanger design will encounter throughout its intended service life and confirm that damage levels for duty cycle events are within allowable limits before committing the design to qualification tests.
