Building a new generation of fission reactors in the United States presents many technical and regulatory challenges. One important challenge is the need to share and present results from new high-fidelity, high-performance simulations in an easily usable way. Since modern multiscale, multi-physics simulations can generate petabytes of data, they will require the development of new techniques and methods to reduce the data to familiar quantities of interest (e.g., pin powers, temperatures) with a more reasonable resolution and size. Furthermore, some of the results from these simulations may be new quantities for which visualization and analysis techniques are not immediately available in the community and need to be developed.
This paper describes a new system for managing high-performance simulation results in a domain-specific way that naturally exposes quantities of interest for light water and sodium-cooled fast reactors. It describes requirements to build such a system and the technical challenges faced in its development at all levels (simulation, user interface, etc.). An example comparing results from two different simulation suites for a single assembly in a light-water reactor is presented, along with a detailed discussion of the system's requirements and design.
[1]
Bucholz,et al.
SCALE: a modular code system for performing standardized computer analyses for licensing evaluation
,
1981
.
[2]
Jeff McAffer,et al.
Eclipse Rich Client Platform
,
2010
.
[3]
Neeti Pokhriyal,et al.
Anomaly Detection for High Fidelity Core simulators
,
2013
.
[4]
Andrew Siegel,et al.
Software design of SHARP
,
2007
.
[5]
Derek Gaston,et al.
MOOSE: A parallel computational framework for coupled systems of nonlinear equations
,
2009
.
[6]
Timothy J. Tautges,et al.
Scalable parallel solution coupling for multiphysics reactor simulation
,
2009
.
[7]
David Andrs,et al.
Multidimensional multiphysics simulation of nuclear fuel behavior
,
2012
.
[8]
David A. Bader,et al.
Hierarchical Data Format
,
2011,
Encyclopedia of Parallel Computing.