Making a Wheel Rotate in Ansys Rigid Body Dynamcs

  • 78 Views
  • Last Post 1 weeks ago
mina1 posted this 2 weeks ago

Hi Peter,

 

Why there is no RBD simulation for wheels rolling on a frictional surface. Could let me know when we have an axle and wheel rolls on a surface except for frictional force contact between wheel and rail surfaces, what contact or joint should we add to see the wheel realistically rolling? I tried revolute but it did not work. Also tried general and let the freedom in rolling and translational still did not worl.

 

Please help me in this regard.

 

Thanks,

 

Mina

 

Order By: Standard | Newest | Votes
peteroznewman posted this 2 weeks ago

If you put 3 revolute joints between 3 wheels and a body, maybe you can roll that tricycle down an inclined plane.

mina1 posted this 2 weeks ago

Hi Peter,

Thanks for your response. My case is the same as this which was mentioned with Dale. I tried to use frictionless contact between axle and wheel, frictional contact between wheel and surface, but it gives me no link error. I tried with revolute joint and general joint in combination with frictional contact on the wheel surface but still, I could not be able to roll the wheel. I used two-step gravity analysis as you mentioned. When I monitor the wheel angular velocity visually and probe it it is zero rolling.

 

I would appreciate your constructive comments.

 

Thanks

 

Mina

Geometry

mina1 posted this 2 weeks ago

Hi Peter,

 

Could you kindly let us know what is the problem in Rigid Body Dynamics that does not let us roll the wheel? As mentioned I tried everything but it did not work.

I would be able to send my WB file. Please let us know how can I send the file.

 

Thanks

 

Mina

mina1 posted this 2 weeks ago

Peter

I could not attach the WB file here. Please let me know how can I send it confidentially to you.

 

Thanks

peteroznewman posted this 2 weeks ago

Mina,

Unfortunately, the Rigid Dynamics solver is not very robust at computing contact. Sometimes it works adequately as it did for the escapement mechanism.

It would be better for you to define your analysis goals. What is it that you don't know that a model can compute for you? Please reply with a specific question that you want to answer and document that with figures and numbers.

mina1 posted this 2 weeks ago

Hi Peter,

 

Thanks for your guidance. I am trying to capture the reaction torque. I am trying to send the file. My last effort the wheels only slide on the rail. Please kindly let us know which simplification I can apply to remove contact issue. My concern is not contact forces, I just wanted to simulate the system realistically and evaluate the required torque.

 

Thanks

Attached Files

peteroznewman posted this 2 weeks ago

Hi Mina,

Please fully define "reaction torque" and "required torque".  Do you know how to draw a Free Body Diagram?  Make one for the system that contains the components you want to evaluate. Show the inputs and output and constraints.

mina1 posted this 2 weeks ago

Hi Peter,

Thanks. I want to use a ball screw actuator to move this cart. I used grounded revolute joint on the shaft that I defined the screw joint and I put the input value load (rotational acceleration) there and it some times it moves in a sliding pattern. My input is rotational acceleration and the output would be required torque. The coefficient of friction is 0.13 between the surfaces. Screw joint has 25 mm/rev lead. The ball screw is connected to the axle of the wheels. How can I find al alternative way of 

not using rolling friction.

Thanks

mina1 posted this 2 weeks ago

Hi Peter,

 

Could you kindly let us know did you review the Archive attachment? I also replied you now but it shows moderation awaiting.

I used screw joint with 25 mm/rev lead. It is a ball screw system. My input is rotational acceleration and output is required torque considering mu 0.13 between the wheels and the rails. I got confused with the joint and contact combo. Both joint and contact have friction coefficient but it does not work. The analysis type is rigid body dynamics. How can I replace the friction effect with the alternative force or displacement.

 

Thanks

 

Thank

Attached Files

peteroznewman posted this 2 weeks ago

 Hi Mina,

Thank you for the post.

I see now that you have a ball screw that rotates. The nut or carriage on the ball screw is on the wheel axle. You say the input is rotational acceleration of the ball screw and the output is the torque required to achieve that acceleration. The specification is still incomplete. What is the mass of the cart? That will have a direct effect on the required torque.

You say the friction coeff between the wheels and rails is 0.13 so you could compute two torques on the ball screw. The first torque is calculated assuming an infinite friction between the wheel and the rails (like a rack and pinion, no slip). The second torque is the value at which slipping of the wheel begins. Again, more detail in the specification is required. What is the normal force on each wheel? From that you can calculate the limiting tangential force, and from that the wheel torque. You can calculate back to the ball screw torque, but you must also include the efficiency of the ball screw, which is another missing piece of the specification.

mina1 posted this 2 weeks ago

Thanks Peter for your guidance.

 

Could you elaborate more? I did not understand what is the formula your were suggesting. The required torque desired only on the ball screw, not the wheels. I assigned steel for all the masses. I could monitor the mass from the properties bar.

How can I include the efficiency on the ball screw? By having user-defined results and multiplying the efficiency on it? The efficiency is 90%. The normal force on each wheel would be sharing the moving mass symmetrically.

The screw joint only works on RBD not Transient. Also, it has RZ and Z free which means the wheels cannot move in the Y perpendicular direction.

I wanted to verify motor sizing required torque for this ball screw.

Please kindly explain more.

If I consider infinite friction the torque will go up and does not match. Would you think sliding the wheel instead of rolling it still gives us good results. Would you know if there is any differences?

 

Another thing is how can I equally distribute the weight between two-wheel axles. Do you agree this is 1 DOF problem?

I saw either in Solidworks and Ansys at the end they suggest bushing parameters. But there is no way how much values should I include

and it is too sketchy. I would like to know your idea in this regard as well.

 

Thanks

peteroznewman posted this 2 weeks ago

Forget about ANSYS for a minute and do some basic system calculations.

Convert the rotational acceleration to a linear acceleration (a) using the screw constant (c) of 25 mm/rev.

Compute the force (F) required to move the cart mass (m) as F = ma.

Convert the force to a torque (T) using the screw constant. Divide that torque by 0.90 to account for the efficiency of the ball screw. That is the required torque.

I see now that the wheels are free to rotate on the axle and it is the carriage that pulls the axle forward. In my previous post, I wrongly thought there was a transmission from the ball screw to the axle, like there is in a car. Forget about the wheel slipping.  What is more important is the bending of the axle due to the force (F) applied at the center, and the mass of each wheel at the end and the linear acceleration (a).

Another concern is what keeps the two wheels on the rails? There has to be at least another wheel or other constraint to keep those wheels on the rails and the axle perpendicular to the ball screw.

mina1 posted this 2 weeks ago

Dear Peter,

Thank you very much for your thorough response. I already have the basic system calculations and my problem is which joint-contact combo addresses the simulation issue with rolling the wheel on the rail. Once I tried revolute, on the axle-wheels+frictional contact did not work+general joint did not work. Another time, I tried no-separation contact between axle-wheels and frictional contact between wheels-rails still did not work. It actually moves it but actuated motion by ball screw does not roll the wheels. I added two grounded general joints on each wheel free to rotate and free in axial motion still I could not roll it. Also I just realized there are two contact in Ansys frictional and forced frictional which the latter one is only active in RBD. I would not know the differences.

I tried frictionless contact between the wheels-axle but it detaches from the axle.

 Also, there is no concern about the bending of the axle since the axle is rated way more than the load capacity. That is why I made everything rigid. Only actual required torque is desired.

My main concern is which combo correctly makes the wheel roll? If you still recommend me to stay with those simplified formula please let us know.

 

Thanks,

 

Mina

 

 

peteroznewman posted this 2 weeks ago

Dear Mina,

What is your interest in seeing the wheel roll?  Why is it important? If what you want is a cartoon that shows the wheel rotating as the cart moves, then you can make a cartoon and use a revolute joint load to set the angle to match the motion of the cart.

Below is one reason that could be important to assess wheel rolling.
If the acceleration of the axle is so high that the inertia of the wheel causes it to slip instead of roll on the track, that could be a concern for wear of the wheel and tracks. In that case, you can do a calculation on the force (Fb) required at the wheel radius to balance the linear acceleration at the wheel axle given the wheel moment of inertia (Iw). You can compare the force (Fb) with the limiting frictional force (Ff) calculated from mu*Fn.  If Fb<Ff then the wheel will roll. If Fb>Ff, then the wheel will slip.

Kind regards,

Peter

mina1 posted this 2 weeks ago

Dear Peter,

 

Thanks for your advice. No, the cartoon is not our interest. If you would confirm I could resume the analysis with sliding friction and forget the rolling situation and interpret the data with hand calculation for the wheel angular speed and acceleration. Is it a good assumption replacing sliding friction to rolling friction? Back to your kind comments having the wheel rolling could show the attainable linear acceleration as well.

Could kindly let us know which joints and contacts are closer to what we are looking for?

 

Thanks

peteroznewman posted this 2 weeks ago

Dear Mina,

I changed the ballscrew General joint to a revolute to ground. The nut/carriage has a screw joint to the ballscrew, the axle is fixed to the nut/carriage and the wheels were changed be on revolute joints to the axle. Now this arrangement has no ability for gravity (or any force) to apply the weight of the cart to the tracks. I replaced the Fixed Joints that hold the rails to the ground with Translation Joints that allow the rails to have vertical travel. Then I applied an upward force on the rails to create a contact force on the wheels. The upward force should equal the weight on each wheel.

Now the wheels roll along the tracks.

Kind regards,

Peter

ANSYS 2020 R1 archive attached.

Attached Files

mina1 posted this 2 weeks ago

Dear Peter,

 

Thank you very much for your kind support. I was wondering if there is a possibility to open the file in our older version (R3). I could not open the file, unfortunately. Please kindly give us advice.

 

That is a really smart move. Appreciate that. I am wondering what will happen for the friction? In this new config. did you kindly consider that

or we should stay with the calculations?

 

Respectfully Yours,

 

Mina

peteroznewman posted this 2 weeks ago

Dear Mina,

It's helpful to mention what version of ANSYS was used when attaching an archive to a post (I updated my post above). I have eight versions of ANSYS installed on this laptop, so I could have done this in 2019 R3 if I had known. Unfortunately, there is no way to simply go back to older versions.

Are you working on the Student license?  If so, you can download ANSYS Student 2020 R1 and install it to open my file to examine it closely. Then you can open your original file in 2019 R3 and make it look like my file. The only risk of having two versions installed, is when you double click on a .wbpj file, it will open the newer version even if the file was created in the older version. If you don't notice this and save the file, you will not be able to open that file in the old version again, so be careful and backup important files. You could uninstall ANSYS Student 2020 R1 after you have examined my file.

If you have a Research license, there is a risk to install a Student license of a newer version as it might "break" the Research license in the older version.

I haven't studied what happens to the wheels as the friction is reduced. There are more important design details such has how the axle stays perpendicular to the rails (rack and pinion is one solution), and how the carriage/nut is connected to the axle and the ballscrew. You need a design that can tolerate angular misalignment between the ballscrew and the axle and rails.

Kind regards,

Peter

 

mina1 posted this 2 weeks ago

Dear Peter,

 Thank you very much for your kind support. It was a student version and as you suggested I am going to update it to the newer version.

This is actually a simplified version of the designed system which the angular misalignment took into account. We were looking for the motor sizing verification with the calculated results. 

I was wondering when I check the redundancy if the actual system has 1DOF joints redundancy in total, should we stick to the 1DOF in the model to get the legit results?

That would be awesome if you could kindly let us know how could we incorporate the friction effect alternative on the system. What would be your smart advice to deal with the friction.

 

Another thing I was wondering in case we put the normal force on the rail should we delete the gravity?

 

Thank you very much,

 

Mina 

peteroznewman posted this 2 weeks ago

Dear Mina,

Yes, delete gravity as it is canceling some of the applied force on the rail.

Attached is a simplified version of your model with just one wheel on a rail with 100 N of contact force and an axle on a translation joint being accelerated at 1 m/s^2.

When the Coeff of Friction is 0.2, the force to accelerate the system is 4.26 N.

When the Coeff of Friction is 0.0, the force to accelerate the system is 3.65 N.

The force to accelerate the mass is less when there is no friction on the wheel because the wheel is not experiencing any rotational acceleration.

You can calculate the 3.65 N simply from the mass in the moving parts, excluding the rail = 3.65 kg *  1 m/s^2 = 3.65 N.

You can calculate the difference between 4.26 N and 3.65 N, which is the force at the wheel radius needed to create the torque on the wheel to rotationally accelerate the polar moment of inertia of the wheel to maintain the rolling relationship.

Between 0.2 and 0, there is a limiting COF when the wheel starts slipping. The normal force and the wheel radius also affect this limiting COF.

ANSYS 2020 R1 archive attached.

Attached Files

mina1 posted this 1 weeks ago

Thanks, Peter,

 

Please let me go through it and get back to you.

 

Thank you very much,

 

Mina

Close