Difference between system subchannel and CFD codes

  • Thread starter Meb15aa
  • Start date
Hi everyone, I am researching how numerical simulations have evolved over the years in nuclear reactors for assessing the thermal hydraulics inside the reactor core.

I have found vague information in regards to the three different main numerical techniques but want to learn more.

So far, system codes are very basic and simulate the entire reactor core as a single node, where macroscopic parameters like total pressure can be extracted. Sub channel codes take this one step further by simulating codes per each sub channel providing better information about the physics inside the reactor. Finally CFD provides detailed flow physics in 3D as opposed to the latter two simulating in 1D but at the expense of computational time.
I would like to learn more about these three methods particularly the reason for evolving the precision of numerical technique. would anyone be interested in sharing their experience, or providing literature around this subject.
Thank you,


Staff Emeritus
Science Advisor
The simplest 1D codes were developed when computational resources were limited - limitations on processing and memory. As computers evolved, so did the numerical methods.

One should look into the difference between Lagrangian and Eulerian methods. Basically has to maintain mass, momentum and energy in either single phase (liquid or gas) or two phase (liquid with vapor).

One should look into thermal hydraulic codes, RELAP, COBRA, and RETRAN

History of RELAP - www.ans.org/pubs/journals/download/a_38253
Thank you for the reply Astronuc, very useful information.


Staff Emeritus
Science Advisor
I should have also mentioned VIPRE, which was derived from COBRA, hence the reference to a snake, and TRAC. All three codes have been adapted by different organizations.

Two major CFD codes in the industry are CFX and STAR-CD or STAR-CCM+

This is where we are now - https://www.osti.gov/pages/servlets/purl/1332948
okay cool, thanks once again.

There is also work being done is coarse grid CFD as as a means to fill the gap between the existing but limited sub channel codes and highly detailed but time consuming 3D CFD models. It generally works by extracting a representative part of the mesh model from a highly resolved CFD model and overlapping that information with a coarse grid model to re calibrate the data an basically correct it.


I have also read some papers declaring current methodology (subchannel codes) cannot meet the safety requirements of future generation reactors, therefore explaining one of the reasons why the industry is looking towards other means like the link you sent and coarse grid methodology to get relatively quick but informative data.

Moreover, going back in time, the reason why simulations were introduced were due to experiments being very time consuming, expensive and dangerous especially when considering accident scenarios. Chernobyl, 86 was an experiments that had gone disastrously wrong. Simulations need to be validated through experiments but due to the vast stored data over the decades, simulations are becoming increasingly dominate in the industry (Also due to rise in computational capabilities).

Very interesting area of work I believe :)

Want to reply to this thread?

"Difference between system subchannel and CFD codes" You must log in or register to reply here.

Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving