cancel
Showing results for 
Search instead for 
Did you mean: 

CBUSH Behavior Quirk

Experimenter
Experimenter
The CBUSH seems much more sensitive to non coincident nodes than the CELAS2 element. I have seen it give 0 stiffness unexpectedly, but it seems to be related to the nodes being some distance apart. Does anyone else have insight on this problem? Are the rotations at a cbush coupled with the translations like a beam element? This could cause this problem and I suspect the reason for this behavior.

Additional thoughts:
There is NO SUCH THING as a coincident nodes. In general a user can not guarantee their pre processor will write out exactly the same coordinates for two nodes.
It is a lot of extra work to move nodes to be very close to each other. Also how close is close enough?
Non-coincident nodes joined by linear springs introduce small numerical errors (non positive definite matrix). Usually 1 mm distance in a vehicle model is acceptable as shown from experience using CELAS2 elements.
Craig Birkett
Daimler Trucks North America
Chassis Analysis & NVH
Portland, Oregon
3 REPLIES

Re: CBUSH Behavior Quirk

Creator
Creator
The CBUSH is a different element than CELAS and I'm not surprised that you're seeing this behavior. The CELAS simply connects DOF to each other and since it doesn't know anything about relative location it isn't sensitive to that. A CBUSH depends on the relative location of the GRIDs. If they're coincident it acts like a CELAS (except that it has its own coordinate system). If they're not coincident it works more like a beam element. It's equivalent to placing a pair of coincident GRIDs at the center of the element. The coincident GRIDs are connected to each other with spring elements and attached to the outer GRIDs with rigid elements. That means that if you don't have rotational stiffness at the two ends, and you don't define rotational stiffness for the CBUSH you'll get a mechanism, which will look like zero stiffness. You either need to make the GRIDs coincident, or give the element rotational stiffness.
Paul Blelloch, Ph.D.
Director, Aerospace Analysis
ATA Engineering, Inc.
11995 El Camino Real
San Diego, CA 92130
(858) 480-2065

Re: CBUSH Behavior Quirk

Creator
Creator
P.S., You shouldn't get this behavior if both GRIDs that the CBUSH are connected to have rotational stiffness. Lots of elements in Nastran, however, don't have rotational stiffness. These include solid elements, ROD elements, BEAM elements with end releases, SHELL elements with no K6ROT, etc.. You can control the location on the CBUSH where the coincident GRIDs and springs are applied. If you have rotational stiffness at one end and not the other you can move the stiffness to the end without rotational stiffness and eliminate the mechanism. I happen to really like CBUSH elements because they eliminate the common problem of grounding with CELAS elements, but you do have to watch out for this behavior.
Paul Blelloch, Ph.D.
Director, Aerospace Analysis
ATA Engineering, Inc.
11995 El Camino Real
San Diego, CA 92130
(858) 480-2065

Re: CBUSH Behavior Quirk

Experimenter
Experimenter
Thanks for this very useful information! I suspected as much but could not find this information in the documentation. Also a number of support people at NX are unaware of this subtlety.

I much prefer the coupling of rotation to displacement to be removed from the CBUSH as it leads to easily missed errors mentioned in my first post. I don’t think most users are aware of this effect either. The idea of a using what is essentially a beam element to model a bushing seems wrong to me also. Other error checking can be done to check for hidden constraints due to non-coincident springs.

I'd be interested to hear other user's opinions on this matter also.
Craig Birkett
Daimler Trucks North America
Chassis Analysis & NVH
Portland, Oregon