CFD has been utilized to predict flow behavior of fluid in many applications, such as wind and water turbine and heat transfer. In the case of the natural ventilation, there is air flow driven by buoyancy (temperature difference) or wind effect. For the buoyancy effect, can CFD predict the air flow (pattern and quantity) inside the room when the temperature differences are known?
You can use Boussinesq approximation. If you have your own algorithm, you have to add an additional term in the z-momentum equation corresponding to the direction of gravity. For simulations of HVAC system, convergence works fine. Benchmark case can be Reyleigh-Benard convection.
Dear Mr.Rachman,
Yes, you can develop your own code or use commercial CFD codes. Many commercial CFD codes can predict natural ventilation.
Also, you can find many publications about this subject. Check "Building and Environment" and also "Energy and Buildings" journals and you can find many researches focused on this subject.
yes you can use any commercial CFD codes, but mostly and this is our many years experience, the natural convection is quite unstable situation (vortices and plumes change shape and position) so you may be forced to solve the case as transient, so you will need more CPU time.
yes, you can use CFD. one of my students used Fluent to study the natural convection in Fluent. you need to use gravity option and expansion coefficient to consider the buoyancy effect in simulation.
You can solve the Boussinesq equations. I think that many commercial codes have that option.
There is gravity option in commercial software and a part of it you define your own code.
You can use several CFD code to solve this problem (commercial or open-source or your own code). It depends on what you want to analyze but you need to take care about turbulence (DNS, LES, RANS, vortices methods...). In fact, you haven't a strong driven force (contribution of gravity, thermal and velocity) so your flow will be highly effect by the turbulence. I know CFX has a tutorial to simulate this kind of flow if you want a practical example.
yes, You can use both commercial CFD (like commercial codes) and even Lattice Boltzmann method to solve this problem for both Boussinesq and non-Boussinesq approximations of density-temperature relationship.
Dear Rachman,
Besides gravity you must select the Boussinesq aproximation. You can also define a UDF for the gravity. Software Fluent its a great option to simulate natural convection (as previous cited by Faik Hamad).
You may check OoenFOAM solvers e.g.
http://openfoamwiki.net/index.php/BuoyantBoussinesqPisoFoam
and since open source you can customize it to your needs.. (but i think standard solvers should do in your case and some tutorials available should help)
As mentioned by others, you cab use any OpenSource Or Commercial Software for the simulation. You can use Boussinesq aproximation or Ideal Gas Model. However, for building simulation natural convection is Turbulent is most cases and hence very diffuclt to converge.
You can use Boussinesq approximation. If you have your own algorithm, you have to add an additional term in the z-momentum equation corresponding to the direction of gravity. For simulations of HVAC system, convergence works fine. Benchmark case can be Reyleigh-Benard convection.
You can try also open source free software like openFOAM. openFOAM provides
standard solver that address Boussinsesq and no Boussinesq, laminar and turbulent and other buoyant flows.
you can use any software like FLUENT or written by yourself by adding the gravity effect in the y-momentum equation.actual air flow is turbulent but usually for human comfort inlet velocity is very low. so you can check the buoyancy effect taking flow as laminar.
I used Fluent to model the natural heat transfer flow in enclosed annuls by applying Boussinesq approximation. The results was compared well with experimental data.
You can use a comecial code like Fluent (or other) or you can develop you our own code.
You can find many literature about this in Fluent tutorials.
All the above answers are correct. But it may be that you will have to run the job as transient as the natural convection is often unstable and as a steady you may not reach a good convergence.
Yes you can using ANSYS-fluent. In fluent you need to activate the gravity (body force) in the direction of interest.
Yes sure. The canonical example is the "Differentially Heated Cavity" but also the Taylor-Rayleigh instability if hot gas is topped by cold gas and Rayleigh Benard convection (cold top wall and hot bottom wall).
All the mentioned answers are valid. However, it is mentioned that Rayleigh Benard convection can be used as a benchmark. R-B is very unstable problem, can not be used as benchmark, because the solution is sensitive to the boundary and initial conditions and error in the numerical method. Differentially heated cavity is the best candidate for testing natural convection codes.
With my students we are running nice 2D examples using CCM+ (Cd adapco) : An underwindow radiator with rising plume and above the radiator a cold window in a room. You obtain two large vortices, and depending on the driven force (temperature difference on radiator and/or window), the vortices move up and down. Try it.
Yes it can easily do it by fluent and i already deal with a problem in which i was predicted the same.
See "A study of wind and buoyancy driven flows through commercial wind towers"
http://www.sciencedirect.com/science/article/pii/S0378778811001009
http://eprints.whiterose.ac.uk/79898/
- Badges
- Science topic