Catapult Optimization Tutorial

[sspoldi]

Hi Everyone,
Last year a lot of people inquired about the design and optimization of the electric catapult we used in Stronghold. Since the basic architecture shows up a lot (like in our hammer in 2014), I thought I would attempt to develop a description/tutorial that other teams could use to help their design process.

I've attached a tutorial PDF and an Excel spreadsheet that walks through the preliminary sizing and design we used to validate the concept. Please let me know if it helps, I will try to be more diligent about posting our optimization programs before the end of build season, so other teams can benefit (or at least share our misery!).

Full Disclosure: I wrote the spreadsheet using LibreOffice on my Linux laptop and exported to Excel format. Laura had no problems opening it up on her machine (Windows 10?), but if anyone has a problem, I'll convert it to real excel and post.

Thanks, see you all in about 3 weeks!

Cheers,
Steve.

[GeeTwo]

Thanks for sharing! I've already posted a link to this on our team's slack site.

[sspoldi]

Originally Posted by GeeTwo

Thanks for sharing! I've already posted a link to this on our team's slack site.

Thanks a lot. FYI, I've swapped out the excel file in the first post, my conversion from open office wasn't quite fully functional, sorry about that

The new one (edited in post 1) works properly in excel.

Thanks, Steve.

[Knufire]

Originally Posted by catapult_2016.pdf

I cannot stress enough how important it is to use
appropriate units. Everyone at the shop makes fun of slugs, but that’s the only way equations like F = MA work.

Everyone makes fun of slugs until they take a dynamics class...

[dardeshna]

This is really awesome. I'm hoping we will do some more numerical driven design this year and a great example to use.

One question though (I may be stupid and have missed something), but this method doesn't seem to account for the torque on the arm due to gravity. Was it negligible in the grand scheme of things?

[GeeTwo]

Originally Posted by dardeshna

One question though (I may be stupid and have missed something), but this method doesn't seem to account for the torque on the arm due to gravity. Was it negligible in the grand scheme of things?

The force needed on the ball (for 45 degrees) was 10#, which is about 15x the weight of the boulder. Including the weight (or more properly moment of inertia) of the throwing arm would actually be the next most important feature to add, with gravity being well behind that. Both of these are taken care of in the 2x target for force.

[sspoldi]

Originally Posted by dardeshna

This is really awesome. I'm hoping we will do some more numerical driven design this year and a great example to use.

One question though (I may be stupid and have missed something), but this method doesn't seem to account for the torque on the arm due to gravity. Was it negligible in the grand scheme of things?

Hi Devin,
Yes there are a few things missing. The short list is:

1. Effect of gravity
2. Arm starting angle (more like -10 degrees)
3. Actual ball release angle (I don't really what this is)
4. Actual height the boulder leaves the catapult (at least 1.5 feet up)
5. Aerodynamics

I intentionally kept it simple to illustrate that basic principles are all you need to effectively optimize. The Octave script we used last year actually covered 1 thru 4, and pretty much came up with the same results and gear ratio.

I spent a few minutes yesterday installing Octave on the Windows 7 side of my laptop, If you're a glutton for punishment install it and run the attached script (ignore the warnings!), that's what we ran. You'll notice a lack of useful comments, but it does have gravity (and a cool 3D plot).

Cheers,
Steve.

P.S. Remove the .txt extension, CD won't upload a .m file, not sure why.
Oh yea, gravity wasn't a big factor!

[Ether]

Very nicely done Steve. Thanks for posting this!

I wanted to point out to any students who may be reading this thread some issues with Euler integration.

In order not to derail this thread, I started a new thread here. Comments welcome.