2D CFD Simulation torque problem

[alexandre75]

Hello guys,

I am a french student working on a 2D model of a savonius rotor on autodesk simulation CFD 2015. I would like to compute the torque applied on the savonius rotor. When I try to compute the torque I have an error message. Do you know what goes wrong ? My protocol is the following:

- I apply the material and the rotating region

- I apply the boundary conditions (input speed on the inlet and pressure on the outlet and no air flow on the 2 other sides is assumed if I'm correct)

- I then solve using the finer mesh (given that the computing time is really short)

When the coomputing is done, I try to use the wall computing tool but the software stops. 

I need to know the torque in order to know the power for a given wind celerity.

My questions are : is my protocol anyhow wrong ? is there any other way to compute those ?

Attached : the support file for my simulation, the computed file, and a preview of the problem.

[alexandre75]

Hey again, 

I did a screenpic of my simulation in order for you guy to see an other problem I have. Is it normal to see the shape of the rotating region in the simulation as if the rotating region was a solid cylinder ? Maybe this is why I have not the right results ! 

Thank you,

Alex

[pmenzies10]

HI Alex,

In your first post, you attached this file:  2DParametrable_25_support.cfz ‏71 KB

This will open in the UI.  In your most recent post, you posted a file that will not open in the UI.  Please try again.

Thanks,

Peggy

[alexandre75]

Here is my support file.

[pmenzies10]

Success!!  This support share file will open in the UI and solve.   Can you share with me exactly what you are trying to accomplish?

My understanding from earlier posts is that you want to study an external free spinning device to determine the torque generated for a given wind speed.  In your support share file, you have specified the rotation of the device.  Please help me understand more specifically what you intend to accomplish.

[pmenzies10]

1)  Perhaps this section from the documentation will be helpful:

Analysis Types

The parameters that define a Rotating Region are based on the type of analysis to be run. There are three different scenarios: Known Rotational Speed, Known Driving Torque, and Free Spinning. The type is selected from the drop menu as described in Step 3, above.

Analysis Type--Known Rotational Speed

Enter the rotational speed of the rotor in either radians per second or RPM.

A variable rotational speed can be entered by changing the Variation Method to Table, and entering data points for rotational speed vs. time.

Analysis Type--Known Driving Torque

This method is useful for modeling a device that is rotated by a known driving torque (such as from a motor). Torque can be entered as a constant value or as varying with time or RPM using a piece-wise linear data table.

(The direction of applied torque is set as the rotational direction on the main Material Task dialog.)

If there is a resistive torque acting on the device, subtract that from the Known Torque value. For example, if the known motor torque is 100 N-m, and the resistive torque is 5 N-m, then apply a value of 95 N-m.

In addition to torque, enter the inertia of the rotating device. This is commonly the rotational inertia of the rotor and shaft and anything that is connected to the shaft (such as a motor or flywheel if the rotating device is a turbine). An easy way to determine an approximate inertia is to multiply the combined mass of the rotor, shaft, and shafted accessories by the average radius squared. This approach is reasonable if the intent of the analysis is to run the device to a steady state condition.

If the intent of the analysis is to obtain a detailed time history of the rotational speed, then a more precise value of inertia is necessary.

Analysis Type--Free Spinning

In this case, the rotor starts with no rotational speed, and will “spin up” based on the applied fluid loading. Specify the inertia of the mechanical components and the rotor. The steady rotational speed will occur when the net hydraulic torque is zero.

If the device is free spinning, but a known resistive torque exists:

1. Set the Analysis Type to Known Driving Torque,
2. Apply the resistive torque as a negative value.

This will cause the device to spin up due to the surrounding flow, and will find a steady rotational speed when the net hydraulic torque is zero.

2)  Some other tips:

• Your mesh is still very coarse - consider a mesh refinement region centered around the rotating region (cylindrical or spherical)
• You also may want to consider running a Frozen Rotor analysis

[alexandre75]

Haha yes success and thank you ! So, I am trying to compute the avarage hydraulic torque applied on the Savonius.

I allready read all the possible documentation about the rotating region.

The thing is when I plot the torque on the last revolution, I see that it goes from -0,2 N.m to 0,5 N.m and according to a paper I am trying to reproduce, there should not be any negative torque value on the last revolution.

Moreover, when I do my simulation, we can clearly see the shape of the rotational region as if it was a "solid" cylinder... I were able to get rid off (more or less) the shape of the region of rotation by taking a time step of 0,01 rad/s which I think is very low and takes a lot of time.

So my questions are:

- What number of cells should I take ?

- What time step would you reccomend ?

- Is there a relation between the time step, the number of cells and the rotational speed ?

- How can I get rid off the shape of the rotating region that messes up my simulation ?

- How come I get negative torque values ?

Thank you because I really need some help !

Alex

[pmenzies10]

Hi Alex,

If you have read and understood all the possible documentation, then I can only recommend seeking a consulting firm.  I'd be happy to make a recommendation.

Otherwise, I think it is time for other members of the community to chime in.

Good luck,

Peggy

[alexandre75]

Yes.. I tried a lot of things. What consulting firm would you recommend ? Anyway, I hope other members of the community will help me on that one ! Thank you for everything !

Alex

[pmenzies10]

Try http://www.simspecs.com/  for a consultant firm.  They are very good.

[alexandre75]

Hi and thank you for this advice. But I feel I am really near of the solution.. I have one more question:

When I compute the torque with CFD Simulation, I see that my hydraulic torque is two times smaller than the torque I am supposed to get. As I have two solid in the rotational region (see support file posted before), is it maybe because CFD simulation computes the torque on only ONE of the solid. And in order to get the TOTAL torque I have to multiply by two the torque that gives me CFD Simulation ?

If not, where could this factor two come from ?

Thank you for your answers,

Alex

[pmenzies10]

I believe the factor is just conincidence.  The torque reported is from the rotating region - implies both blades.