Learning Forum

[nicolo piergiovanni bagnoli]

Hello, my name is Nico and together with a few fellow student we are enrolled in a class called data-driven design. We wish to use ansys to collect trajectory data of several axe designs being thrown at the same speed and same rotational speed (initial conditions). I was hoping someone would be so kind to help me get started! I have mocked something up on ansys Motion but have a few questions:

-First of all, is this the correct toolbox for the task?

-Secondly, I am afraid the speed and rotation I have given the axe are fixed and that the actual inertia of the axe is not doing anything to change the initial speed and rotation. How can I be sure that those are only initial conditions and do not persist for the whole simulation?

-Third: How can I map the trajectory of the axe to have some data to download (like the center of mass position) and how can I make sure each axe design I load has the same starting point in space in respect to a given target? Should I model the target in the same file? Or is there a way to create an ambient setting of sorts...

Thank you in advance an I am sorry if some of my requests seem a little simplistic!

[peteroznewman]

Hello Nico, this sounds like a fun project!

I am more familiar with Rigid Dynamics analysis so I would start there.

One idea is to design an axe throwing machine. Imagine something like a closed-end tube that the axe handle sits in. The tube is angled back -80 degrees from vertical with the axe handle in the tube. The tube accelerates forward and rotates just like your hand does when throwing an axe. The tube decelerates and stops angled forward at 80 degrees from vertical while the axe handle slides out of the tube and launches into space.

By designing an axe throwing machine, you will have a standard “test fixture” to throw different axe shapes in exactly the same way to study how axe shape and mass distribution affect the trajectory.

[nicolo piergiovanni bagnoli]

Well that is a great idea, i though there would be a way to d what the machine does without actually simulating the machine but that is also a good option, any good tutorials for learning the rigid dynamics toolbox so I have a good understanding of it and not just winging it without a solid foundation?

Thanks a lot again and have a splendid weekend!

[nicolo piergiovanni bagnoli]

Good morning, so I tryed rigid dynamic and it seems to be the best option, Im having trouble with boundary conditions however. As you can see in the image below when giving the arm of the catapult a constant rotation velocity i cannot seem to find a way for it to hit the base and stop. I tried giving the connection a maximum rotation but then the simulation stops when that condition is reached so I would not be able to see the path of the ball when i launch it. (yes i am using a ball for now:)) Could you help me better understand what boundary condition is needed?

[peteroznewman]

You don’t need much for the base, just a rod across the path of the arm that represents the stopping angle.  The arm will be at the starting angle in the CAD.  I expect you used a revolute joint at the axle of the arm.  That can be a joint to ground, you don’t need a base. 

Don’t use a constant velocity boundary condition for the reason you discovered. It just keeps going. Use a rotational (torsion) spring on the revolute joint with a preload and a spring rate. The preload torque will accelerate the arm from rest and the frictionless contact between the rod and the arm will stop the arm when contact is made, and the ball will keep going forward.

You can soften the deceleration of the arm and make the solution more likely to continue if you put a translational joint between the rod and ground, perpendicular to the arm at the point of contact.  Then you can put a spring between ground and the rod.  Adjust the mass of the rod and the spring rate to get the arm to decelerate at the rate that launches the ball nicely.

[nicolo piergiovanni bagnoli]

Thank you! Im just having trouble fingind the "torsion spring" in the rigid dynamic interface, could you give me some indication on where I could find such?

 

Also just for interest: I was having trouble with ansys when defining the boundary region between two spheres, it would stay stuck at 1% and I would have to restart my system. I then just made the object to launch a cube and the launching "basket" also a cube and the problem disappeared? Do you know why this happens? Thank you infinitely for your time!!!

 

[peteroznewman]

The problem with a sphere in what should have been a larger radius sphere is that the ball is not stable and a smaller radius ball rolls around in the larger radius pocket causing lots of computation to keep track of those small motions.  If the two spheres were of equal radius, that may also be a problem.

Try creating an arm with a conical cavity to take the ball. Then there is a well defined circle of contact. If you have a Rough contact, the ball will be stable in the cone during the launch.  

Use a regular Spring to Ground and add the Preload to that.

On the revolute joint, you can define a Stop so you don’t need the rod for the arm to crash into.

[nicolo piergiovanni bagnoli]

Thanks, everything works perfectly!

 

Lets say now I had to repeat the simulation with 100 balls of different sizes and store the results (maybe distance travelled), is there  way in ansys to do this kind of analysis? Like maybe have a file for the base and the add all different size ball files to analyze? Is there a way to create a script for such a task? Thank you for your time and patience:)

[peteroznewman]

Glad to hear that Nicolo!

Workbench, DesignModeler, SpaceClaim and Mechanical all support Parametric Design capabilities. If you build the throwing arm in DesignModeler with a Parameter for the size of the ball so that as the ball diameter changes, the ball remains tangent to the cone, then you can access that Parameter in Workbench. There will be a Parameter set in the Project Schematic with a Table of Design Points. In there you can paste in 100 different rows for the ball size.  In Mechanical, you can create an output parameter for horizontal distance at a specific time. That would be a Directional Deformation and would include the distance travelled while in the cone.

[nicolo piergiovanni bagnoli]

This information was very useful, I thank you for your time!