Message Boards Message Boards

GROUPS:

Correct way to model/approximate force disturbance for force controller

Posted 16 days ago
240 Views
|
2 Replies
|
0 Total Likes
|

I'm currently working on simulating (in a very basic way) and designing a force controller for a linear drive.

There is a circular saw blade mounted on the linear drive cart. The linear drive will be set at a specific feedrate, given through a single between 1 and 10. (Unimportant atm for the question I think anyways...). Eventually the sawblade will come into contact with an object that will produce dampening feedforce of something around 40N that will jump +/- 10N...(learned through experiments). I would like to design the controller (PI, or LQI) with mathematica to reduce the feedrate in order to reduce the process forces, or rather control them to a specific referenced feedforce (25N for example)....Eventually a function of feedforce will be given and tracked during the entire cutting process. In this case, when reference force rises, the controller should increase the feedrate, and respectively drop it when the referenced force drops.

The feedforces can be directly measured with dynometer, and the feedrate adjusted programmatically. There is currently no position control for this system. Only feedrate can be given...hence the want to design this controller.

At the moment I'm completely unsure how to even model the collision of the sawblade with an object...Though for my purposes at the moment, I believe it's not required, as the saw motor dynamics are unimportant for the current form of simulation. I would like to model the linear movement of an object, that encounters a breaking/dampening/load force to approximate this collision/cutting process..If this is at all possible.

So far, as pictured, I have designed a simple rotational to linear rail, with a mass and force, plus position sensors.

model

This is agreeably very basic. However, if this load or dampening force or breaking force can be added to the system the way I imagine it.... I believe it will be accurate enough for my current purposes (Though I will gladly be taught how to do it better! :) )

My first thoughts were to add the breaking as pictured with a simple pulse signal...however I am unsure if this is ideal. How would one go about this?

Thanks for the help!

Attachments:
POSTED BY: Mor Bo
Answer
2 Replies

Mor,

In my experience, the brake is a good way to model collisions or systems that suddenly change.

Caveat: I did not spend any time studying the problem or looking at any literature so this may be half-baked!

I would model the blade hitting the material as a combination of extra mass, springiness, and damping. You can then adjust all three to get a behavior that makes sense:

enter image description here

I also changed the brake behavior a bit to make it act more like a switch and to make the brake supply only the force demanded by the material.

I hope this helps.

Regards

Neil

Attachments:
Posted 15 days ago

Hello Niel, Thanks for the help! I will try this verison and see if I can get it to work...

Though, I had a realisation moment, maybe you have an idea....The proposed controller will reduce the feedrate of the saw, as the breaking force gets too high... which should also reduce the breaking force, or increase the feedrate, if the force is too low....Which means the breaking force is also coupled with the feedrate...Or is this already compensated for in your fixed spring/mass/dampening design?

I was thinking of directly adding the the speedsensor values as a modifying 'source' for the break somehow. Any ideas?

Thank you again, I will give your model version a test soon and let you know my results!

POSTED BY: Mor Bo
Answer
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract