Learning Forum

[Juwelmojumder1]

Hi, 

I'm trying to simulate the solar radiation for the solar collector using the DO radiation model available in ANSYS-Fluent. I'm using the solar ray tracing model as a solar load, and thus solar ray is converted to heat, and absorbed by the working fluid. I'm using the incident solar radiation value, 10,000 W/m^2 (, which is equivalent to 10 suns) as an input to the collector's top surface (Jusy a Face, not body), which is expected to be uniform on the face. I considered the non-gray surface and solar wavelength was in a range between 0.5um to 1.5 um. 

I've checked with angular discretization, mesh study. Then, I've done the set up of the DO model following the above criteria, like as follows: 

After that, I followed the necessary fluid properties, and set up for the solar inlet face as follows:

I found a good convergence, but I'm not sure about the outcome (Wall flux--->surface incident radiation, /m^2), as it shows not exactly the same value that I input as incoming solar radiation. Moreover, wall flux--->sun heat flux is zero on the top face. 

and 

I suspect the reason for such a curve rather than a uniform manner is the absorption coefficient value of the working fluid. In fact, there is no solid body considered for glass (solar window) but only face. 

Could you please guide me, whether my DO model set up is ok or not?

Is the direction value ok in terms of Sun direction vector, beam direction?

And, which solar radiation value (surface incident radiation/incident radiation) represents my actual incoming radiation? 

 

Thanks in advance. 

 

 

[Amine Ben Hadj Ali]

Do you want to account for radiation or solar load only? Actually you do not need to set anything apart "Solar Ray Tracing" settings under "Radiation Tab".

[Juwelmojumder1]

Hello Amine, 

Thanks for your reply. 

Do you want to account for radiation or solar load only? 

Response:

Non-gray DO model has been applied in order to solve the radiative transport equation (RTE) for the solar collector system. Meanwhile, I've to use wavelength-dependent optical properties of working fluid in simulation. The full details of this method can be found in many references.

Again, the sun’s rays that enter the computational domain can be modeled by the ray tracing algorithm. By applying the solar load in ANSYS-Fluent, the calculated heat by the ray tracing algorithm is coupled to the energy equation via a heat source term.

What happens in simulation in terms of energy equation If I (1) check only the DO and uncheck the solar ray tracing (2) uncheck of the DO and check with only solar ray tracing, and (3) checking both DO and solar ray tracing in the radiation model?

Is radiation in DO model, and solar heat flux in solar load represent the same incoming daylight solar radiation?

Then, how to set up the radiation model for my case? 

Actually you do not need to set anything apart "Solar Ray Tracing" settings under "Radiation Tab".

Response:

Regarding the radiation tab for the "solar inlet face/solar window", then, how I will define the beam direction as it appears in both cases (DO/solar ray tracing) when checked?

 

Thanks

Juwel

 

[Amine Ben Hadj Ali]

I wont define anything ad everything is provided via solar load model.

[Juwelmojumder1]

Could you please explain in detail for this particular problem?

Kind regards

Juwel

 

[Amine Ben Hadj Ali]

Solar load model calculates for you the thermal load either automatically by providing where you are or explictely by providing the fluxes as you did in the solar load panel.

[Juwelmojumder1]

What happens if I only check with DO model and nothing in solar load?

If I do so, I get radiative heat flux on the window surface too. But there is no surface incident radiation under 'wall flux'.

 

Juwel

[JavadAhmadi]

Hej and I hope everything is well with you.

Actually this is the same problem I am facing now in the modeling of double skin facades which in terms of the structure of the problem is totally the same as collectors.

At first, I think, solar load coming from the sun is just importable from the " SOLAR LOAD" tab in radiation sections and in the part of the boundary condition radiation tab it is not necessary to add the same data as you added accordingly. My main problem is starting exactly in this part what is differences between radiosity, reradiation, and irradiation in FLUENT and in which sections of DO or SOLAR LOAD we can import these parameters, especially in 2D problems? Has anybody had this problem already?  

[lieyna]

Hi, Juwel,

 

May I know, how do ypu know your solar wavelength was in a range between 0.5um to 1.5 um and how can we know the value of Angular discretization as what you set? Does it vary base on our case?

Thanks,

Regards,

LINA

 

 

[Juwelmojumder1]

Hi Javad, 

If you need to simulate solar radiation model especially with DO, then you need to strongly follow up the optical properties of the working fluid, such as the extinction coefficient of the solar participating mediums for the radiative transport equation. In 2D solar load doesn't suitable, and optical properties are not considered in solar ray tracing, while DO irradiation reflects the same thing with DO solar model. irradiation or normal radiation is the same meaning as an amount of incoming radiation (W/m2) incident on any surface. 

 

[Juwelmojumder1]

Hi LINA, 

Incoming radiation wavelength is divided into UV, Visible, and Infrared, where the range is not equal with each other. However, the visible range is considered up to 1150nm. Afterward, infrared is dominated. So it hard to omit UV and visible range, while you can limit the infrared area the range you want. 

[lieyna]

Thanks Juwel,

 

However, I am not clear about radiation model usage. If I need to account radiative heat transfer for cooling effect in my model, as my wire heated by joule heat that i assign in cell zone, do i still need to use one of radiation model to do radiative heat transfer for cooling my wire to ambient? 

As I'm read in ANSYS help and heat transfer book, in mu understanding, by using radiation model it will account the radiative heat transfer. If not, the radiative heat transfer is not account in our case. Please correct me if I'm wrong.

 

I hope you or anyone else can share the opinion regarding this. 

Thanks.

Regards,

LINa

[Juwelmojumder1]

hI LINa,

It is always nice to share my own understanding with you! I studied to solve my own problem as You might have noticed that I haven't received any detailed solution from any experts in this group. Experts are running out of time!!!! 

No worry, I'll try to answer your question!

In your numerical solution, where you need to solve the Navier stroke equation, especially when you focus on the energy equation, you have to consider the heat source. Now, such a heat source can be added either as constant heat flux or any solar radiation model available in Fluent. Both options allow you to solve your numerical solution but the desired accuracy in heat transfer can't be obtained always that comes with a trade-off. 

In your case, you can add the radiative heat transfer coefficient value as a boundary condition (on the thermal tab in Fluent), which is only available for the radiation model. There are many resources of formula to calculate this value that you should cover already. Remember that this coefficient is a heat loss coefficient with surrounding (for your case). You also can add convection heat transfer coefficient value as it conjugates with radiative heat transfer but will depend on your problem.  

Thanks

[EngineerMech]

Hello Juwelmojumder1, and others

 

I would like to ask you what the optical properties (absorption coefficient, refractive index, scattering coefficient) you have used for the materials that will participate in radiation. I am modeling a thermal PV cell with all its components(glass, eva, tedlar, cell), so I hope you or others can help in this. Also, according to your knowledge is that necessary to model as non-gray or just not add any wavelength? thanks in advance.

[Juwelmojumder1]

Hi,

The non-gray radiation model offers your model more practical. Hereby, the higher the number of bands higher the accuracy. Typically should not go higher than 6-10, which makes compatible in terms of computational cost. Again, it will depend on geometry types. The optical properties of your solar cell significantly matter in your DO modeling. You have to calculate it based on what types of cell material you are using. 

https://refractiveindex.info/?shelf=glass&book=HIKARI-BK&page=J-BK7

You may follow the above link to extract some material properties, but you might need to calculate further to get the absorption coefficient of the cell materials. 

the absorption coefficient of solid is analytically calculated, and you will need a complex refractive index of the material. 

you may find interesting in...

https://www.sciencedirect.com/science/article/pii/S0196890418300062

Thanks 

Juwel

[EngineerMech]

Hello Juwel,

 

Thanks for your response. My model is PV/T just like your model I just add all PV layers(glass, EVA,CELL+EVA, EVA, Tedlar, AIR, iNSULATION). I have activated the DO model but used it just for air and deactivated for all solid material and that gave me +3oC higher than experimental results WITHOUT any wavelength and it is very good for me. If I include the glass to participate in radiation the cell temperature will be higher than the experimental by about 25oC which is not correct as I consider the absorption coefficient for the glass is 26 1/m (for wavelength <4.25(from paper), even though I didn't activate the non-gray section in DO model) and that is what confused me, what absorption coefficient I should take for the glass if I don't activate the non-gray model? and what it should be if I have more than one wavelength because in fluent I can specify one value for the absorption coefficient in material properties? thanks again for your response and for any suggestions from others.

[Juwelmojumder1]

hI, 

It's all about calculating the RTE equation in the solution. This equation has been using to solve solar participating mediums. You can pick any option whether gray (where, a constant property of mediums are taken on entire wavelength) or non-gray (where wavelength-dependent properties of participating mediums). In your gray cases, the absorption coefficient of glass is actually an average value of the entire wavelength, that's all). For the non-gray case, I suspect your optical properties values under individual wavelength ranges are not taken accurately, that made difference in solution. For this, just dig a more on % of incoming energy in UV, visible, and infrared that incident on the surface. and calculate properties accordingly. I hope you will find it. Again, semitransparent surfaces including air in your model are ought to consider participating as a cell zone condition, except the opaque surface. 

Thanks

Juwel

[EngineerMech]

Hi,

thanks again for your help. Actually I m interested just in the gray model, therefore, I use the law in the attached image, which is suggested by someone, to get the average absorption coefficient. For example for the glass with 3mm thickness and transmissivity=0.95, the absorption coefficient will be around 17 1/m. This value gives the wrong results. The absorption coefficient needs to be around 1000 1/m to get reasonable results.

Thanks

[lieyna]

Dear Juwel, may i know your absorption coefficient air for wavelength  between 0.5um to q,5um that you set in material air?

 

Hi, 

I'm trying to simulate the solar radiation for the solar collector using the DO radiation model available in ANSYS-Fluent. I'm using the solar ray tracing model as a solar load, and thus solar ray is converted to heat, and absorbed by the working fluid. I'm using the incident solar radiation value, 10,000 W/m^2 (, which is equivalent to 10 suns) as an input to the collector's top surface (Jusy a Face, not body), which is expected to be uniform on the face. I considered the non-gray surface and solar wavelength was in a range between 0.5um to 1.5 um. 

I've checked with angular discretization, mesh study. Then, I've done the set up of the DO model following the above criteria, like as follows: 

After that, I followed the necessary fluid properties, and set up for the solar inlet face as follows:

I found a good convergence, but I'm not sure about the outcome (Wall flux--->surface incident radiation, /m^2), as it shows not exactly the same value that I input as incoming solar radiation. Moreover, wall flux--->sun heat flux is zero on the top face. 

and 

I suspect the reason for such a curve rather than a uniform manner is the absorption coefficient value of the working fluid. In fact, there is no solid body considered for glass (solar window) but only face. 

Could you please guide me, whether my DO model set up is ok or not?

Is the direction value ok in terms of Sun direction vector, beam direction?

And, which solar radiation value (surface incident radiation/incident radiation) represents my actual incoming radiation? 

 

Thanks in advance. 

 

 

[Juwelmojumder1]

Hi EngineerMech,

The absorption coefficient for any semitransparent glass has a lower value compared to the opaque medium. For the glass, it's value 1000 1/m  seems unrealistic to me. The value 26 1/m is sound good in your case. You must check the other set up like the boundary condition if you are not close to the validation. 

Thanks

Juwel

[Juwelmojumder1]

Hi Lieyna,

The absorption coefficient for air is negligible. Many previous studies assumed it as zero. 

 

Thanks

Juwel

[EngineerMech]

Thanks, Juwel, It looks like I will just include the air in my simulation as it is very good in the validation. Including the glass or any other material gives incorrect results.  My boundary condition: all sides walls are insulated and not participate in solar ray tracing, the upper wall of the glass participate as the semitransparent, bottom wall (opaque) doesn't participate in solar ray tracing, the interface(coupled wall) between the air and the Tedlar (opaque) participates in solar ray tracing with 0 absorption value, the interface(coupled wall) between the air and the insulation(opaque) doesn't participate in solar ray tracing. My last question and I hope you have the answer: why only the interface(coupled wall) between the air and the Tedlar that participates in solar ray tracing as a BC works and if I use the other interface wall the ray-tracing model doesnt work? thanks in advance for your help and wide knowledge.

[Juwelmojumder1]

Hi EngineerMech,

Firstly, As you are solving the RTE, then you don't need to use the solar ray tracing in the solar load. DO irradiation is there to provide you the solar load. So first look at the basic setup. Not necessary to follow up on the problem I posted here as it was my very first attempt. I'll prefer you to visit.

https://www.afs.enea.it/project/neptunius/docs/fluent/html/ug/node481.htm#:~:text=ANSYS%20FLUENT%20provides%20a%20solar,that%20enter%20a%20computational%20domain.&text=The%20ray%20tracing%20approach%20is,sources%20in%20the%20energy%20equations.

You will find, DO-solar ray tracing doesn't match!

Secondly, try to use symmetric on the sidewalls, and keep lower the boundaries as much you can. 

[lieyna]

Dear Juwel,

As you mentioned the air absorption coefficient assumed 0, could you please give me any source from previous study (if you have) that mention about that? My email : noorlinazainuddin@gmail.com. I tried to find but still cannot get it. I really appreciate with your help. And, I tried using solar load in DO model. I found that, my result seems like ok using solar ray tracing than DO irradiation. What difference between them?  Thank you an advanced

[Juwelmojumder1]

The value of the refractive index of air  (dry) has been considered as 1 due to the negligible change of properties of air over the wavelength. Indirectly, when the transmittance of the air is assumed as 1 (https://www.shimadzu.com/an/uv/support/uv/ap/solar.html) then the extinction coefficient of air can be taken as 0, analytically. For an accurate prediction of it, one can follow this. http://www.cnofs.org/Handbook_of_Geophysics_1985/Chptr18.pdf. 

 

Thanks

Juwel

 

[adlyna]

Dear Juwel, did you find the answer for question you asking? Because what I found, there is nothing effect of solar by using the solar irradiation/Do irradiation. Instead using solar tracing is have a good result and can see the effect of solar irradiation of the temperature result by changing the different value. I think this is because using solar tracing, we can set the absorptivity of material in each particular boundary condition and also can sign which wall [participate to solar tracing. But, using solar irradiation and enter the any value, the temperature result is same means it have no effect by solar amount. I did not sure why, but I read in manual fluent it was stated that, nThe solar load model's discrete ordinates (DO) irradiation option provides you with an easy means of applying a solar load directly to the DO model. Unlike the ray tracing solar load option, the DO irradiation method does not compute heat fluxes and apply them as heat sources to the energy equation. Instead, the irradiation flux is applied directly to semi-transparent walls (which you specify) as a boundary condition, and the radiative heat transfer is derived from the solution of the DO radiative transfer equation.nSo base on statement, does I need to do as your method which is at the top wall change from opaque to semi transparent and give the value of solar irradiation at the box like you do in your beginning question in this discussion? But, I am using opaque as I am model a solid wire surrounding with air and do a heat transfer analysis. The sun is located on top of wire. Here, I am using solar calculator for solar location. So, I was play with different value of solar to see the effect to wire temperature. So, what function of solar irradiation (enter value) part a for DO irradiation at the solar load option, as it cannot see the effect of solar irradiation?.Did you have an idea?.Thanks,nLINAn

[Juwelmojumder1]

Hi Lina, nI answered my understanding in your another post. nI can assure you that DO model is best fitted for the semitransparent medium, but your model is opaque. So don't need to worry about absorption/scattering coefficient of the participating mediums. Apply your Irradiation flux directly to the opaque surface in the model section, while OFF the radiation model, but active only the ray tracing with other relevant solar load set up if requires. nThanksnJuweln

[adlyna]

Thank Juwel for your suggestion. The reasons I used radiation model, because I have include radiative heat transfer effect to my system. The wire have an experience of radiative heat transfer to the air. I read in manual, if we want the radiative heat transfer, we need to on the radiation model. Correct me if I am wrong. Thannks

[adlyna]

Dear Juwel,nnCould you please explain what meaning of this word: 'I've checked with angular discretization, mesh study. Then, I've done the set up of the DO model following the above criteria' and how you find the angular discretization theta and phi division (5) and theta and phi pixel (3). nBecause I found(but playing in random value), the angular discretization is influence in my result. but by increasing tat value give more computational cost time. I hope you can share your knowledge regarding this. Thanks in advanced.n