Noise reduction 2

[laxman]

Hi,
I have a small dc motor, which fits into a enclosure. As soon as the running dc motor is fitted into the enclosure (press fit), the motor noise gets vastly amplified. This needs to eliminated. Can anyone suggest me some tips to handle this problem in the FEA point of view!.

Thanks
Laxman

 

[brad]

You may want to simultaneously post this on the FEA area of the website.

My two cents--note that although I've got a good deal of experience in FEA NVH, I don't consider NVH my area of expertise. However, this is a good start.

It sounds like you've got a motor operating at or near the resonant frequency of the casing.

I presume that you can determine the amplified frequency and the motor's operating frequency (possibly one and the same). The quickest, dirtiest approach is to run a normal modes analysis and see where the natural frequencies of the system line up, and which frequencies are near the motor's driving frequency. You will probably see some resonant frequencies of the structure in this area. You can try and engineer these normal modes away from the driven frequency of the motor. This analysis is fairly quick and painless.

The more detailed analysis is to do a forced frequency response of the structure. This will not only give the resonant frequencies, but also give the physical response to a prescribed varying load at those frequencies. If you have not previously done forced frequency response, first deal with a normal modes approach, and see if you can increase or decrease the problem modes.
Brad

 

[laxman]

Hi Bradh,
Thanks for your input, it was very helpful, Regarding the forced frequency response analysis, could you please put some more light on it. I have done frequency response analysis in nastran before but I don't know for this problem, what would be the load value I need to apply (should I use the torque value as the load input?)

 

[laxman]

Hi Bradh,
Thanks for your input, it was very helpful, Regarding the forced frequency response analysis, could you please put some more light on it. I have done frequency response analysis in nastran before but I don't know for this problem, what would be the load value I need to apply (should I use the torque value as the load input?)

Thanks

 

[GregLocock]

You should probably use the measured vibration on the motor casing to drive your acoustic model of the enclosure. I don't know if you could succesfully model the casing vibration starting from the electrical characteristics of the motor - best guess is not.

The torque on the motor is probably NOT a good load case toi use to excite your model since the chances are that the noise is at some harmonic of the operating speed, whereas torque is a DC load case.

Cheers

Greg Locock

 

[brad]

In Nastranspeak, this would still be SOL 111, so you are
on the right path.
After reading Greg's response, I need to add a bit.

I would still recommend as a starting point dealing with SOL 111, and using either the moment from the motor torque, or else separately applying two point loads, both orthogonal to the moment axis (i.e., if you torque is about x, then apply a load in y, and in another subcase a load in z).

The moment about x would simulate excitation due to the torque, the loads in y and z would simulate the possible effects of torqueshaft imbalance on your system.
Just apply a unit load/moment in all cases, and measure the drive point mobility response of the system. See if you do indeed have excitation in the area of concern.

Just to clarify (as it appears I have up to this point completely contradicted Greg) . . .

I was initially approaching this as a structure-borne vibration issue. This approach which I've described would address that.

I think Greg may be approaching this as an airborne acoustic problem, in which case my approach isn't relevant for that. I'm not as familiar with airborne acoustic modeling . . .

 

[GregLocock]

Hi bradh.

". I'm not as familiar with airborne acoustic modeling . . ."

Lucky you!

Yes, I was assuming that at some point the nasty issues of structure/air interaction were important.

I am still a bit dubious that the excitation of the system's acoustic problem will bear any particular relationship to the drive torque, but then I know very little about the vibration characteristics of electric motors. By analogy with i.c. engines I would have expected a lot of excitation at no. of poles*running speed, and odd harmonics of that.


Cheers

Greg Locock

 

[brad]

Greg,

I certainly wasn't lamenting my lack of structure-airborne acoustic background. I am ignorant in that area by intent (nasty stuff!).

I am also not sure that the torque will directly impact it; however I have seen that effect in automotive powertrain. It comes out of any inconsistencies in torque output of the motor. I woudn't initially expect those to be as significant for an electric motor, but I will also acknowledge that I don't have a lot of experience in dealing with electric motors.

Brad

 

[Rob45]

Here's as much simulation as you need: your enclosure is "making" the noise, while the motor vibration (not its' torque, but more likely its' imbalance) is driving that noise.
You need to a) isolate the motor from the structure, or b) damp the motion of the surfaces of the structure, with e.g. viscous damping sheet.
As a last resort, you build another enclosure around the first one.