Digital twin model generation for crack propagation in steam generator tubes: Methods and implementation

Crack propagation in the steam generator tubes are critical factor influencing the safe operation of nuclear power plants. Accurate prediction of crack growth behavior is vital for effective equipment health monitoring and maintenance. A fundamental framework and key technologies for the implementation of digital twin technology in nuclear power plants were presented in this study. And a novel crack propagation digital twin model generation technique for steam generator tubes was introduced, which integrates experimental data with extended finite element simulation results. Particle Filter method and reinforcement learning were conducted to evaluate respective strengths and limitations in crack propagation prediction. Both experimental data and extended finite element simulation data were incorporated into the model training process. The results demonstrated that both digital twin model generation techniques have distinct applicability conditions and high accuracy. This study could provide new insights and technological approaches for the health monitoring of steam generator tubes.