I am trying to build a subroutine for the contact force calculation in a gearbox model. The reason I am trying to use a subroutine instead of the default contact formulation is to take into account the time varying mesh stiffness of the contact in terms of the load angle between the mating gear teeth.
From what I understood throught test models and documentation, the default way that Adams calculates the contacts in a model is to identify them seperately (in different TRACKS) and calculate the forces individually, summing them to estimate the total contact force between two bodies at each step. So in my case, this would mean that if two sets of teeth come into contact, adams calculates the two forces and sums them to estimate the total contact force. This however does not take into account the individual stiffness change, only the change in the contacting areas.
Is there a way to extract information regarding the orientation of the contact point without adding markers to each tooth?
If I have not understood the way Adams estimates contacts, please do correct me as the rest of this formulation is based on the spaculation of Adams estimating the forces for each track seperately.
The reason that I didn't use the Gear moduel is because I would also like to include the axial compressive, fillet foundation and shear stiffnesses. Also, I am trying to simulate faults (such as broken gear teeth) which doesn't seem to be possible usong the Gear module.
Well nobody said that you can't use NxM contacts for a gear pair with N/M teeth.
Only issue is that with a contact it'd be hard to accomplish what you try.
The contact algorithm determines the penetrated volume and calcluates "the middle" - the contact force is applied there. What you can control in the CNF/CFFSUBs is the forces but not the point of application.
So If I were about to do this, I'd use Gear/AT to save the work and maybe just add some force formulation to include my special effects.
If you need to apply forcdes to individual locations, you somehow need markers or nodes.
In the past I've been using a GFOSUB to distribute the load of a hydrodynamic bearing to the nodes of the bearing surface.
Other than that the most promising approach will be using mnf-based gears/teeth and some usersub-type that can apply foces to nodes like a MFOSUB.
The problem with the NxM contacts is that the number of active contact at each point is large. I have a model with helical gears having a gear ratio of about 4, which means that for every 1 contact, I'd need 4. The model is already fairly large and thus it is rendered useless if the simulation time ramps up from 15 minutes (current), which is already fairly large, to something even higher.
Just to clarify, the Gear/AT is the Gear Machinery module in Adams right? It is not a seperate plugin/software.
A/Machinery has several gear modules ranging from simple contact methods to Advanced 3D contact method which is old version of Gear AT (v2015), hence missing some very interesting features.
Gear AT is has been standalone product (plugin) for Adams up to 2021 release. It is going to become a part of Adams core product soon. Installer of Gear AT 2021 can be access from Solution Download center (SDC), however it can be used with Adams versions ranging from 2018.1 to 2021.2.2 .
Adams Gear AT considers the contact over the complete flank at the same time.
All changes of the meshing during contact according to load changes or misalignments, etc… are considered.
The underlying flexible tooth is completely 3D, which means that all 6 DOF of displacements and strains are connected at every location.
We don’t use slice theory as the analytical or semi-analytical approaches do for simplification!
We use a full 3D-FE-based approach (without the need of co-simulation…, that’s our “trick”)
In Gear AT we calculate the contact pattern over width and involute at the same time for every location on the flank.
If you are interest in a highly accurate dynamic representation we strongly recommend to use Gear AT, as it is the highest standard for gear representation within Adams.
According to your description, it seems we already offer all needed capabilities.
I suppose that an effect of broken tooth could be modelled by imposing radial tooth deviation.
We will be happy to talk to you in an online-session about your use case and requirements. If you are interested, please feel free to contact Mr. Christof Rachor (at hexagon.com).
Thanks for clarifying. Gears AT might work in my case, but I am not sure our University License covers it. I believe that Gears AT is not listed in our available MSC ONE tokens (unless it is implicitly listed along with Adams).
An online-session would be great but I am having trouble finding Mr. Rachor's contact information.