Thermal Hydraulic Calculations using NS

[koab1mjr]

Hi all

I am very much a layman when it comes to Thermal Hydraulics and this may be a naive questions but here it goes...

On the job, i use various codes to examine two phase flow in the core loop. Despite being vetted and validated numerous times over the course of many years (TRAC, RELAP etc...), these codes cannot resolve some of the more complex higher order effects of flow, two phase behavior in particular. Many of the codes I use have origins in the 60s-70s and things have come a long way since then, and with CPU power being more accessible I had the following questions:

1) Why is implementing a multi-phase code that uses the NS instead of a momentum balance so much more difficult?

2) What could be gained by switching schemes? I know CFD has its own issues, but I feel it could help with some problems if enough man power is behind it.

3) Both equations represent a momentum balance (NS and conservation of momentum) but what are the key differences?

Thanks in advance.

 

[jlefevre76]

Hmm, can't say I've ever had to deal with two phase flow, but I know commercial packages like Fluent contain built in codes to handle that kind of analysis. I'm guessing you either work for the government or a large contractor, so in either case I can't see a problem with you getting a license (and a few parallel licenses for using multiple processors) to run the simulations you want.

As far as being able to describe why they are so much harder...I can't help there too much. They just are. Having two phases makes things more complicated in terms of how the flow behaves because you have...two phases, effectively two different fluids (one gas and one liquid, as least I'm assuming from context) interacting and mixing. That also means two sets of thermal properties (conductivity, viscosity, etc).

Check out some existing commercial software packages that can handle two-phase flows and see if they would meet your needs. Then hit up your department/boss/supervisor for the money and computing resources to solve whatever you're hoping to solve.

 

[Thermalne]

1) Both use Navier-stokes, but its how system codes solve the equations vs. CFD codes. Also, the problems fundamentally look at different things. System codes look at a much larger picture (full primary system) in only 1 or 2-D. Whereas CFD codes solve for much smaller picture (a single pump or T-junction) and capture 3-D effects not able to be modeled using system codes. I'd highly recommend you study fluids (both single and multiphase) more in-depth if you continue to use system or CFD codes. There is a lot of basic information you seem to be missing.

2) CFD is inherently more computationally intensive and can not be solved for large systems such as a PWR core without the use of a super computer. Also, two phase flows are extremely complex and poorly posed mathematically. There is no easy analytical solution (or no analytical solution) and are almost always transient in nature.

3) Again both of these are solved using Navier Stokes equations, but it is how they are solved and what they are solved over. I'll try and get a more complete explanation soon.

Hi all

I am very much a layman when it comes to Thermal Hydraulics and this may be a naive questions but here it goes...

On the job, i use various codes to examine two phase flow in the core loop. Despite being vetted and validated numerous times over the course of many years (TRAC, RELAP etc...), these codes cannot resolve some of the more complex higher order effects of flow, two phase behavior in particular. Many of the codes I use have origins in the 60s-70s and things have come a long way since then, and with CPU power being more accessible I had the following questions:

1) Why is implementing a multi-phase code that uses the NS instead of a momentum balance so much more difficult?

2) What could be gained by switching schemes? I know CFD has its own issues, but I feel it could help with some problems if enough man power is behind it.

3) Both equations represent a momentum balance (NS and conservation of momentum) but what are the key differences?

Thanks in advance.

 

[koab1mjr]

I know for a fact in that my main TH code does not use the NS equation in its solution. It is a momentum balance and rather than differentials we have volume averages. This leads me to my point, in some instances such an approach is not sufficient to resolve flow features that could material impact performance.

In BWRs, annular flow this is of critical importance. One would want to know about the thickness in liquid film something on the order of 50 microns. You have no shot at calculating this with CVs. I would figure you would have to employ a CFD formulation to have a shot at it.

 

[Astronuc]

Here is some introductory material on CFD. I want to elaborate on this topic when I have more time.

"When one is using FLUENT or another finite-volume code, it’s useful to remind oneself that the code is finding a solution such that mass, momentum, energy and other relevant quantities are being conserved for each cell."

http://infoscience.epfl.ch/record/169604/files/EPFL_TH5227.pdf

http://pbadupws.nrc.gov/docs/ML0708/ML070860521.pdf

http://pbadupws.nrc.gov/docs/ML1125/ML112570022.pdf

AREVA's Topical Report mentions NS equations and how they used in the COBRA-FLX T/H code.
http://pbadupws.nrc.gov/docs/ML1015/ML101550179.pdf

The subchannel-flow model equations used in the COBRA-FLX code are derived from the local-instantaneous formulation of the three-dimensional Navier-Stokes equations.
. . .
Two forms of the basic fluid flow equations are used in the COBRA-FLX code - the
conservative and transportive forms. The conservative form of the mass conservation
equation is the form used above. The transportive form is obtained by expanding the
spatial derivative in Equation 2-23 [continuity equation] to get . . . .

This is probably a paper worth getting:
Abstract - http://www.jsme.or.jp/monograph/pes/1999/ICONE7/PAPERS/TRACK07/FP7491.PDF

See also - Hydra TH: A Thermal-Hydraulics Code for Nuclear Reactor Applications
http://www.casl.gov/docs/CASL-U-2013-0100-000.pdf

And http://www.latp.univ-mrs.fr/IJFV/IMG/pdf/saturne.pdfThe three conservation equations involve mass (continuity), momentum and energy, for each phase. Two phase flow, particularly beyond nuclear boiling, is difficult to solve on a fine scale. Having details of the liquid droplets, or locally rapidly changing phase mass fractions, is challenging.

 

[jlefevre76]

The three conservation equations involve mass (continuity), momentum and energy, for each phase. Two phase flow, particularly beyond nuclear boiling, is difficult to solve on a fine scale. Having details of the liquid droplets, or locally rapidly changing phase mass fractions, is challenging.

Right, there are codes made to characterize droplet size and behavior. I have no experience with them, but I THINK (but I'm not sure) that Fluent can do stuff like that. I seem to remember someone in my CFD class from BYU showing an animation of droplets pouring out of a hose or nozzle or something. We were supposed to use either Fluent or Adina for that course. I have seen that Fluent has options for dealing with multiphase flows, but exactly what capabilities it has, I'm not certain.

On a separate note, as was mentioned, CFD codes are only used to model a single component (or even part of a component when the said component is geometrically complex). If you are trying to model every component of a reactor with CFD, you could spend a career doing that and never finish. In fact, many researchers spend their careers working out the finer details of individual components using CFD (usually using custom codes), and will successfully model a few components over the course of their entire career ( looking at all the different variables that could change, including Re, Pr, droplet size, dimensions, materials, etc ).

 

[jlefevre76]

Hey, I just stumbled across this today and thought maybe it would be worth taking a look at. I don't know what sorts of things it supports, but it's free, and compared to
spending money on a license for CFD software, this might save you a bit of money.

NPARC
http://www.grc.nasa.gov/WWW/wind/release.html

NASA has a ton of other freeware that is carefully controlled, but freely given to military personnel (and private companies so long as they're not working on a foreign contract). Anyway, hope this helps a bit.
http://www.innovative-cfd.com/cfd-solver.html