High-Performance Computing Probabilistic Fracture Mechanics Implementation for Gas Turbine Rotor Disks On Distributed Architectures Including Graphics Processing Units (GPUS)

We present an efficient Monte Carlo (MC) based probabilistic fracture mechanics simulation implementation on heterogeneous high-performance (HPC) architectures including CPUs and GPUs for large heavy-duty gas turbine rotor components for the energy sector. A reliable probabilistic risk quantification requires simulating millions to billions of MC samples. We apply a modified Runge-Kutta algorithm to solve numerically the fatigue crack growth for this large number of cracks for varying initial crack sizes, locations, material and service conditions. This compute intensive simulation was demonstrated to perform efficiently and scalable on parallel and distributed architectures with hundreds of CPUs utilizing Message Passing Interface (MPI). In this work, we include GPUs in parallelization strategy. We develop a load distribution scheme to share one or more GPUs on compute nodes distributed over network. We detail technical challenges and strategies in performing the simulations on GPUs efficiently. We show that the key computation of the modified Runge-Kutta integration step speeds up over two orders of magnitude on a typical GPU compared to a single threaded CPU supported by use of GPU textures for efficient interpolation of multi-dimensional tables. We demonstrate weak and strong scaling of our GPU implementation, i.e., that we can efficiently utilize large number of GPUs/CPUs to solve for more MC samples, or reduce the computational turnaround time, respectively. On seven different GPUs spanning four generations, our probabilistic fracture mechanics simulation tool ProbFM achieves speedups ranging from 16.4x to 47.4x compared to single threaded CPU implementation.

Probabilistic Fracture Mechanics Evaluation of a BWR Feedwater Nozzle

A probabilistic fracture mechanics (PFM) evaluation was performed for the nozzle blend radius and nozzle-to-shell weld of a boiling water reactor (BWR) feedwater nozzle using the PFM methodology in Electric Power Research Institute (EPRI) Boiling Water Reactor Vessel and Internals Program (BWRVIP) BWRVIP-108-A and BWRVIP-241-A, which are the technical basis for inspection relief in ASME Code Case N-702. Using a finite element model of the feedwater nozzle, stress analysis was performed for plant-specific piping loads and bounding transients, which were grouped by severity and projected cycle counts. Monte Carlo simulations were performed using the VIPER-NOZ (Vessel Inspection Program Evaluation for Reliability, including Nozzle) PFM software to determine probabilities of failure for the reactor pressure vessel (RPV) with an inspection population of 25% of the feedwater nozzles every ten years for sixty years of plant operation. The results show that the probabilities of failure for normal operation and low temperature over pressure (LTOP) event meet the acceptance criteria for RPV failure in NUREG-1806 by the U.S. Nuclear Regulatory Commission (NRC). Thus, there is potential to seek regulatory relief to reduce the inspection population of BWR feedwater nozzles from 100% to 25% every ten years using the technical basis of ASME Code Case N-702.

Benchmarking Study on Probabilistic Fracture Mechanics Analysis Codes xLPR and PASCAL-SP Considering Primary Water Stress Corrosion Cracking

Probabilistic fracture mechanics (PFM) is expected as a more rational methodology for the structural integrity assessments of nuclear power components because it can consider the inherent probabilistic distributions of various influencing factors and quantitatively evaluate the failure probabilities of the components. The Japan Atomic Energy Agency (JAEA) has developed a PFM analysis code, PASCAL-SP, to evaluate the failure probabilities of piping caused by aging degradation mechanisms, such as fatigue and stress corrosion cracking in the environments of both pressurized water and boiling water reactors. To improve confidence in the analysis results obtained from PASCAL-SP, a benchmarking study was conducted together with the PFM analysis code, xLPR, which was developed jointly by the U.S. Nuclear Regulatory Commission (NRC) and the Electric Power Research Institute. The benchmarking study was composed of deterministic and probabilistic analyses related to primary water stress corrosion cracking in a dissimilar metal weld joint in a pressurized water reactor surge line. The analyses were conducted independently by NRC staff and JAEA using their own codes and under common analysis conditions. In the present paper, the analysis conditions for the deterministic and probabilistic analyses are described in detail, and the analysis results obtained from the xLPR and PASCAL-SP codes are presented. It was confirmed that the analysis results obtained from the two codes were in good agreement.

Development of Guideline on Seismic Fragility Evaluation for Aged Piping

The seismic probabilistic risk assessment is an important methodology to evaluate the seismic safety of nuclear power plants. In this assessment, the core damage frequency is evaluated from the seismic hazard, seismic fragilities, and accident sequence. Regarding the seismic fragility evaluation, the probabilistic fracture mechanics can be applied as a useful evaluation technique for aged piping systems with crack or wall thinning due to the age-related degradation mechanisms. In this study, to advance seismic probabilistic risk assessment methodology of nuclear power plants that have been in operation for a long time, a guideline on the seismic fragility evaluation of the typical aged piping systems of nuclear power plants has been developed considering the age-related degradation mechanisms. This paper provides an outline of the guideline and several examples of seismic fragility evaluation based on the guideline and utilizing the probabilistic fracture mechanics analysis code.

High-Performance Computing Probabilistic Fracture Mechanics Implementation for Gas Turbine Rotor Disks on Distributed Architectures Including Graphics Processing Units (GPUs)

We present an efficient Monte Carlo based probabilistic fracture mechanics simulation implementation for heterogeneous high-performance (HPC) architectures including CPUs and GPUs. The specific application focuses on large heavy-duty gas turbine rotor components for the energy sector. A reliable probabilistic risk quantification requires the simulation of millions to billions of Monte Carlo (MC) samples. We apply a modified Runge-Kutta algorithm in order to solve numerically the fatigue crack growth for this large number of cracks for varying initial crack sizes, locations, material and service conditions. This compute intensive simulation has already been demonstrated to perform efficiently and scalable on parallel and distributed HPC architectures including hundreds of CPUs utilizing the Message Passing Interface (MPI) paradigm. In this work, we go a step further and include GPUs in our parallelization strategy. We develop a load distribution scheme to share one or more GPUs on compute nodes distributed over a network. We detail technical challenges and solution strategies in performing the simulations on GPUs efficiently. We show that the key computation of the modified Runge-Kutta integration step speeds up over two orders of magnitude on a typical GPU compared to a single threaded CPU. This is supported by our use of GPU textures for efficient interpolation of multi-dimensional tables utilized in the implementation. We demonstrate weak and strong scaling of our GPU implementation, i.e., that we can efficiently utilize a large number of GPUs/CPUs in order to solve for more MC samples, or reduce the computational turn-around time, respectively. On seven different GPUs spanning four generations, the presented probabilistic fracture mechanics simulation tool ProbFM achieves a speed-up ranging from 16.4x to 47.4x compared to single threaded CPU implementation.