Friday, July 9, 2021

How to... choose radiation model in Ansys Fluent

 Many times in numerical modeling we have a dilemma whether we have to take into account the radiation model, and if so, what exactly to choose.  

Types of radiation models ansys fluent
Which model of radiation to pick in Fluent


There are many theories behind the use of radiation solvers.  First of all, the main theory is the maximum value of the temperature that we take into account in our model.  It is usually assumed that the radiation model should be used at temperatures above 500 C.  Another point of view is that quasi-vacuum conditions or very high domain pressures should also assume the phenomenon of radiation.

In my opinion, the easiest way to verify the validity of taking into account the radiation model is to perform an additional analysis on a very simplified model.  The model should be modified geometrically so that its size (number of finite elements) is on a scale of 1:10 compared to the original.  It is important to model all physical and material conditions as closely as possible to the main simulation.

Then, two analyzes should be performed, which will differ from each other by the absence or presence of the radiation model.  Successively, we read the average HEAT FLUX from several areas of the main solid domains and compare these two variants.  If the mean values ​​differ less than 5% from each other then you can simplify the model by not defining an additional solver in the form of radiation.

It is important not to consider the wall heat transfer coefficient at all.  This default coefficient in CFD analyzes is so sensitive (depending on the boundary conditions, geometry, number of finite elements or the size of the time step) that its fluctuations for the same physical case may differ even by several dozen percent.  It is worth writing the User Defined Function yourself based on the literature equations of the HTC coefficient and heat flux from the simulation.

Where to pick model radiation in fluent ansys
Window where U pick radiation model in Fluent

If it comes out from our additional analyzes, that it is justified to use radiation in our model, we must consider which solver will best fit for our case.  The fluent program offers a choice of 6 typical radiation models.  Three main types are used in about 90% of analyzes created by CFD users.  Below I will try to explain them a little bit more.

Surface-To-Surface model. 

 It is a model abbreviated as S2S.  In this model, radiation is only resolved between the surfaces of solids that "see each other".  Regardless of the gas used in the S2S domain, it does not take into account the issues of absorption or scattering of radiation in the gas.  It is very good to use it in case of thermal problems for pressure 1Bar and gas type air.  This solver is also very suitable for problems modeling a quasi-vacuum or a thermal process with argon.

DO Model 

The DO model (Discrete Ordinates) is characterized by a completely different type of approach to determining heat transfer by radiation.  The characteristic basis of this model is solving the radiation energy transport equation.  It is the DO that takes into account the radiation loss through absorption and dispersion in the gas.  The model is widely used in problems where high pressures and the proportion of gases very susceptible to radiation are taken into account.

Monte-Carlo Model.  

The radiation model with the highest computational accuracy of all offered by Fluent.  It takes into account all radiation phenomena, but it is a model that greatly extends the computation time.  The Monte-Carlo method itself is not limited to counting radiation models.  It is widely used in many fields of engineering and science.

I hope that the above entry has explained at least a little what are the differences between the radiation models in the fluent program.        I also hope that I have explained in which cases radiation should be used and how to effectively check the model in terms of the influence of certain phenomena on the obtained results.


No comments:

Post a Comment