SimScale CAE Forum

Fan simulation error


#1

Hello All,

I am trying to run an axial fan simulation enclosed within a pipe. Unfortunately I am unable to get the case to run and the solver log is not of much help to me. I would appreciate it if someone could give me some direction.
Cheers!

https://www.simscale.com/projects/jwick/test1/


#2

Hi @jwick,

Other than illegal cells reaching around 44. I don’t see that many issues. You might wana to re-mesh and check your geometry to keep it as error free and remove all small features as much as you can at the moment. Will check again later.

Cheers.

Regards,
Barry


#3

hey barry,

Thanks for the input. I tried remeshing my model but the mesher fails to capture my rotating zone (solid) completely. What should I do ?

P.S

Look under FanTestNewMesh


#4

Hi @jwick,

The previous mesh had no issues capturing that right? What did you change between this mesh and the last one?

Cheers.

Regards,
Barry


#5

Hey @Get_Barried,

Not really. Apparently I overlooked that problem. None of the meshes seem to capture the solid completely. The top face is always missing.

Thanks and regards


#6

Hi @jwick,

Yep for the rotating zone only the front and back faces are showing. Looks like its not meshing the circular area of the rot zone.

I have some suspicions on what is going on, but nothing concrete yet. I suggest for now, try uploading a FanTestNew geometry in STEP format and ensure in CAD that your rotation zone is 100% watertight. I’ll copy the project over and mesh on my end to see what I can do. Get back to you on this tomorrow or so.

Cheers.

Regards,
Barry


#7

hey @Get_Barried .

Appreciate the help mate. I did what you said, uploaded a new step model and created up a new project as well. Running into the same problem though.

https://www.simscale.com/projects/jwick/fantestairflow/


#8

Hi @jwick,

I think I figured out the problem. You need to ensure that there is a gap (tiny one if need be) between the fan and the casing itself on all sides, thus it only can encompass the rotating geometries and there should be no intersections. The MRF zone then needs to occupy that space between fan and casing (aka the gap) and has to enclose the entire fan as well. You will need to edit this in the CAD.

I do think there is something strange going on however, probably to do with the settings somewhere but no idea what it could be at the moment. What I’ve mentioned above is a quick fix that would hopefully enable you to start your sims at least.

Do try it out and let me know how it goes.

Cheers.

Regards,
Barry


#9

Hey @Get_Barried

Thanks ! I will try again after making changes to the CAD model. In the meantime I tried running a simulation on another simpler geometry. I have trouble getting the simulation to the run. Can you check my solver log. Is there something wrong with my BC ?

thanks and regards,

John wick


#10

Hi @jwick,

Have you managed to resolve the issue?

Sorry, I didn’t see this earlier. From what I can recall, your BCs are fine.

Cheers.

Regards,
Barry


#11

Hi @jwick, I just took a look at your setup, the simulation before last looked ok, but the latest one has an inlet with o velocity, if you dont know the velocity you could assign a pressure inlet and let the simulation take care of it. I also noticed a bit of instability, the same instabilityies I see if my mesh isnt of sufficient quality or right on the edge of being acceptable, you could overcome this in the numerics however I noticed that the side of the cells at the base mesh are longer in the x directon than they are in the y and z direction, ammending this might give you the mesh to want (but I would recomend refining the blade one level further).

Best,
Darren


#12

Hi @1318980

Thanks for the feedback mate. I have moved on from that case study to another. I am designing a micro air turbine prototype and my design goal is to make sure the turbine gets sufficient air speed to spin at approx 5000 rpm, I found this post to be quite useful Torque and speed from rotating turbine? as it closely resembles the outputs I am interested in. What I understood from that post is that I need to simulate different cases with fixed inlet velocity and different rotational speeds. So when the moments are zero I get my rpm value for that particular speed ? Does that mean the torque is going to be zero as well. Or is there a method of calculating the inlet air velocity required to spin the turbine at a specified rpm.
Cheers

John


#13

So power, torgue times angular velocity, needs to equal the mechanical resistance of the turbine, more power means the turbine will speed up and less means is slows down, so to put it simpley the sum of power will be zero. To ensure the fan maintains the desired RPM it needs to match the power lost, but if you do not know that at this stage, a better idea might be to create a curve that shows wind speed and power so that you can easily know where you design might sit. If you want to vary fan spped also, you add another variable, you will then need to know the relation ship between power, wind speed and angular velocity, I suggest tackiling the first idea first inlet velocity vs power.

Let me know if you need any further explaination.
Best,
Darren


#14

Hi @1318980 Darren,

https://www.simscale.com/projects/jwick/turbinetest/

So I have run a simulation successfully. The inlet velocity is 10m/s and have set my turbine to rotate at 5000 rpm which is what I want. When I look at the force and moments , they are quite low. Is my set up correct ? I have a couple of questions though, I am still new to cfd and I am on a steep learning curve :stuck_out_tongue:

  1. Does the pressure moment in the y axis give the output torque in Nm ?
  2. How do I calculate the mechanical resistance of the turbine ?

Thanks,

John


#15

Hi @jwick,

Pressure moments should be in Nm for SI unit projects.

Thats something you will either have to know or make an estimate, if you know the components you might be able to do that calulation if you dont you could try taking typical efficiencies, and the power output to make the estimations of power lost.

The setup seems logical, CFD is a long process to understand and we are more than happy to help. The results are likely to be somewhat innacuarate at the moment, I would look at firstly adding layers to your mesh, observing the Y+ on the surfaces (look up boundary layer modelling in CFD) and also have a look cat some of the posts on the forum about mesh independence, which is essentially ensuring the results are not effected sufficiently by discretisation error (error caused by lack mesh fineness). This is alot to take in but start small and keep asking questions, the community is a great place to get this knoledge!

Best,
Darren

p.s. the turbine seems really small, 10cm wide is that correct?


#16

Hi @1318980

Thanks for your valuable feedback. I’ll definitely do a grid independence study with boundary layers. Right now i wanted to get a general idea of the calculations that are involved to get the necessary parameters i’m interested in. Except for the torque most of the calculations seems to be analytical. Any useful references you could throw my way? Yes. The turbine is pretty small and that’s the intended design. The idea is using a micro turbine which uses high speed air as an input power source ( presumably a blower of sorts). Asking an engineers opinion, do you think this idea is feasible? I intend to 3d print prototype test it but wanted to run some cfd calculations to get an idea of the rpm to change my design accordingly.

Thanks and regards,
John


#17

@jwick, when I first started CFD I did a bit about turbines, maybe the resources at the bottom of the write up might be usefull:

Looking back it was pretty basic but might get you off the ground.

Whats the purpose?

Best,
Darren


#18

Thanks darren. The purpose is to power a cutter tool. Check out this video. You’ll get the idea. https://youtu.be/Fj1gir4tJbU


#19

Ah I understand the purpose now! Yes, this makes sense. Another thing you might be interested in in power curves, where you could see how much torque the device has at different RPM’s so, at each stage you could see the power and how much RPM is likely to be lost when x amount of power is drawn. This will help in calculating how much the device would loose speed as it works, where it travels towards peak power rather than away, to make this logical, it becomes more powerful as it slows down.

But just a few ideas, good luck with the project!

Best,
Darren


#20

Good insight. Will keep it in mind.

Thanks Darren !