Core Design: Questions on Codes Used

[Uranium]

I have a few questions regarding codes typically used in core design:

1) Which codes need to be coupled with a TH module and which do not? From what I have gathered, steady state codes can be coupled with a code like RELAP to model transients.

2) Are independent codes that are coupled to one another than codes with a TH module internally? SIMULATE can run transients, but is the TH side of it as good with whatever it is using the those calculations?

3) PARCS seems to be a fairly new code being developed (not very new, but new enough). How does this code differ from SIMULATE and why was it developed?

4) Are Monte-Carlo codes truly the best...once the computing power is there? It seems that probabilistic would be a most accurate, but how much worse are deterministic codes and how much better can they become?

Thanks for your help. I hope my questions make sense.

Regards,

Uranium

 

[Astronuc]

PARCS was developed by a group at Purdue for the NRC. The NRC uses independent codes to audit the works of others. AFAIK, traditionally core simulators has simplistic T/H and fuel rod models in order to obtain the fuel temperatures for cross-section calculations (and Doppler). Some models have been upgraded, and the move now is to couple nuclear with T/H codes.

Codes like RELAP/RETRAN are (primary) system codes as opposed to subchannel codes like VIPRE/COBRA.

SIMULATE is a licensed production code from Studsvik/Scandpower. The code is used by utilities and is an independent method from the vendor (fuel supplier) codes. The NRC uses their own methods, which are independent of those methods used by vendors or utilities.

http://www.nrccodes.com/

The NRC currently develops and maintains the following codes:

TRACE: The TRAC/RELAP Advanced Computational Engine. A modernized thermal-hydraulics code designed to consolidate the capabilities of NRC's 3 legacy safety codes - TRAC-P, TRAC-B and RELAP. It is able to analyze large/small break LOCAs and system transients in both PWRs and BWRs. The capability also exists to model thermal hydraulic phenomena in both 1-D and 3-D space. This is NRC's flagship thermal-hydraulics analysis tool.

RELAP5: Small break LOCA and system transient analysis tool for PWRs or BWRs. It has the capability to model thermal hydraulic phenomena in 1-D components.

PARCS: Purdue Advanced Reactor Core Simulator for reactor kinetics simulation. The Purdue Advanced Reactor Core Simulator (PARCS) is a computer code that solves the time-dependent two-group neutron diffusion equation in three-dimensional Cartesian geometry using nodal methods to obtain the transient neutron flux distribution. The code may be used in the analysis of reactivity-initiated accidents in light water reactors where spatial effects may be important. It may be run in the stand-alone mode or coupled to other NRC thermal-hydraulic codes such as RELAP5 and TRACE. Through coupling with TRACE, the capabilities for RAMONA, NRC's legacy kinetics code are recovered.

SNAP: Symbolic Nuclear Analysis Package, SNAP, is a graphical user environment designed to assist the NRC code user in all aspects of input model development. This involves such tasks as the 1) the actual development of code input decks, 2) executing the models that have been developed, 3) visualizing the code output, and 4) managing the history of new and old legacy models alike.

FRAPCON/FRAPTRAN: FRAPCON is a computer code used for steady-state and mild transient analysis of the behavior of a single fuel rod under near-normal reactor operating conditions. FRAPTRAN is a computer code used for transient and design basis accident analysis of the behavior of a single fuel rod under off-normal reactor operation conditions.

MELCOR: Integral Severe Accident Analysis Code: Fast Running, parametric models

ACGrace - formerly called Xmgr5, this program is used for plotting graphics for all NRC's major safety codes.

SAPHIRE - Systems Analysis Programs for Hands-on Integrated Reliability (SAPHIRE) is used for performing probabilistic risk assessments.


I believe that Monte Carlo are used given the substantial uncertainties in so many variables. Deterministic methods require good resolution (and lots of computing power), as well as intimate knowledge of the local physics. We aren't there yet.