General Mechanical

Meshed gears under pretension problem

• Klearchos Terpos
Subscriber

Greetings everyone,

I am trying to model the following problem I encountered in a gear test rig:

As shown in the first picture, we have two gears under pretension in a still state.

As shown in the second picture, we now apply a known external torque and the system marginally starts moving. What happens is that the compressed tooth flanks create a static friction moment which inhibits motion. As soon as the static friction becomes kinetic the system starts moving.

What I am aiming for is to find the value of the pretension torque.

As far as ANSYS is concerned, I have the gear models ready as I used them in setting up a static structural simulation. Hence, the mesh and frictional contact between the meshing tooth flanks is set and ready.

However, for this study I will need to use transient structural. I have added my boundary conditions (a very small rotational speed for each gear and the known external torque) but I am unsure of how to proceed with the supports in order to get the desired results.

What supports or boundary conditions would you use?

You can see my setup so far in the following picture:

Thank you for helping out!

• peteroznewman
Subscriber

Why do you say "for this study I will need to use transient structural"?

Other than a brief period when the frictional force transitions from sticking to slipping, it is essentially a statics problem with no significant inertial forces. You are using very low velocities.

If the static friction and dynamic friction are the same value, then there is no release of strain energy when the slipping begins so no significant inertial forces develop.

I suggest you use a Static Structural model. Both gears have a Revolute Joint to Ground on the hole ID. One gear has a Joint Load of Rotation, the other gear has a Joint Load of Moment.

A multistep analysis is used. In step 1, the Rotation is set to 0, holding that gear fixed, and the Moment is ramped up to the 40 Nm torque. At some point in step 1, the contact goes from sticking to slipping.

In step 2, a small rotation is ramped on. Sliding continues and the contact point moves up or down the tooth, depending on which tooth you are looking at.

• Klearchos Terpos
Subscriber

I think I have followed your instructions precisely. The simulation has run, but I how exactly can I probe/calculate the pretension torque I am looking for and which causes the gears to require such a high moment to move that little?

• peteroznewman
Subscriber

One gear has a moment load, the other gear rotates to move that load. You can Probe the Reaction Moment on the Joint that has the rotational input. If you put 40 Nm as a torque load on one gear, the driving gear with the rotational input will probe with a higher value. The size of the torque depends on the shape of the gear tooth and the coefficient of friction you assigned to the contact.

• Klearchos Terpos
Subscriber

That makes sense, but apparently I am missing something since the result is 40 Nm, identical to the input load. Apparently, the theoritical value is about 175 Nm. Here is a picture:

Any ideas on what I did wrong?

• peteroznewman
Subscriber

What direction is the rotation of the driven gear?  Is it advancing against the moment on the other gear or is it retreating from the moment on the other gear?  Those two directions of rotation will have a different reaction moment if there is a significant friction on the tooth contact.  If the contact is frictionless, then the moment will be the same in each direction.

• Klearchos Terpos
Subscriber

Thank you for your time. It is advancing against the moment. I 've experimented with many configurations, but I haven't managed to get the pretension moment to show up. All I am getting is 40 Nm and I reckon it is quite impossible for the pretension to coincide with the external moment.

Here are there settings of the frictional contact:

• peteroznewman
Subscriber

I just remembered something about involute gear teeth profiles. Ideally, they don't slide, they roll on each other. That means the friction would not change the answer and the correct output torque is identical to the load of 40 Nm.

Now if you have a non-involute tooth face (a bad idea) then you will get sliding at the contact point and then the direction of rotation will make a difference.