Learning Forum

[Ella Chapman]

Hi everyone, 

I've been developing a model that includes a fan boundary condition, however I am unsure how to find the pressure jump so that this fan runs at the correct speed. There's not much help online regarding the internal fan BC so I'm a little stuck. The fan should be running at a speed of 6m/s. 

Thank you!

[Rob]

You want to fix the fan speed? The reason the option is there is because we may not know the pressure profile so want to let the fan speed vary. 

[Ella Chapman]

Hi Rob,
How am I supposed to set up the fan BC? I was under the impression that I needed to know the pressure jump..

[Rob]

Are you setting the cell zone fan or surface zone? They're different. The solver will calculate flow based on the inputs, so you'd define the various boundaries and let the solver figure out what speed or pressure you get. 

[Ella Chapman]

I've been setting the surface zone as the fan BC. The fans are supposed to emulate the fans in a server, so if I'm assuming it on the inlet then I'm not getting a fast flow for the fans. 

[Rob]

OK, so it's inside the domain somewhere?  You need all the information to fill in the below. 

If the model starts at the fan surface (ie you're just interested in flow inside the server) have a look at velocity boundary options. 

[Ella Chapman]

Hi Rob, I've shown exactly how I want the air flow to work, as well as how I've set up the boundary and the velocity contour. The fan boundary is set on one of the interfaces as discussed prior. 

 

 

Is it possible to make the velocity more constant along the boundary. Thanks for your help so far!

[Ella Chapman]

Oops, the other two did not appear. My bad!

 

[Rob]

That may be mesh related - there are only 5-6 cells between the bottom fan & the wall. I'm also guessing you've not got a thin wall at the sides: the mesh is too fine for a wall & shadow pair. 

[Ella Chapman]

So should I decrease the amount of mesh I've used?

[Rob]

Between the fan & wall, no, you need more mesh. At the sides of the zone you don't need to model the wall with a physical thickness, a coupled "thin" wall will be sufficient. 

[Ella Chapman]

Okay, I've tried changing the mesh - The thermal analysis seems to be doing something I would kind of expect but the velocity is still odd to me. 

Also, thank you for the help so far!! You are a life saver haha

[Rob]

There isn't any flow - look at the velocity scale. Assuming you did set the fan (you won't be the first to have missed a step in a model), how did you initialise the flow?

[Ella Chapman]

Okay, I have an inlet at the top and then the outlet at the bottom. The problem I'm still getting is that there's no velocity where the fans should be. I just initialised it using hybrid initialization and it's coupled. 

[Rob]

Odd. Hybrid should avoid the trap of zero flow giving zero flow for cases without external boundaries. If you replot with a velocity scale of around 0 to 0.1 m/s how does it look? 

[Ella Chapman]

This is what I'm getting if I set off the autorange and put it as 0 to 0.1m/s.

[Rob]

Can you check for stray walls? And that the fan is on! 

[Ella Chapman]

That's what I'm confused about - I don't know how to model the fan! I'm not sure exactly sure how I'm supposed to properly model it, I'm currently just using a surface zone boundary condition.. Do I used the polynomial pressure jump and if so, how? 

[Rob]

You'll need a pressure jump to get some flow moving. Otherwise there's nothing for the solver to force the movement with. 

[Ella Chapman]

Which I asked about prior, how am I supposed to do that?

[Rob]

Sorry, as you posted the panel yesterday I assumed you'd found the option, click on the Edit button next to polynomial and put in the fan curve. You'll find additional information if you click on Help on the panel.