Roadcase stacker [defunct]

This project came about of an afternoon of roadcase stacking. Or, rather, attempted roadcase stacking. Roadcases come in many sizes (though usually only the one shape), so I started looking at whether there was a way to automatically calculate which roadcases could go on top of each other.

Part 1 was planning out the process of checking a single pair’s stackability, in one permutation.

The simple calculator took 9 measurements on each case, measuring the interior and exterior wheelwell distance for both cases, plus the height. Height seems like the least important factor so far, because structurally neither our pallet stacker nor our ceilings get tall enough for it to matter much, so it works out without many hitches. The rest of the measurements could be calculated from those given.

Given imperfect measurements and a propensity to say “eh, it mostly fits, it’s okay,” it made sense to have the calculator do some margin-of-error calculations. The simple calculator took one roadcase’s wheelbase dimensions and checked whether it slotted into the other case’s wheelwells. If it didn’t, it computed a difference variable, which it used in some more lenient calculations later on. Calculators of various tolerance checked for fit if the wheels faced in certain directions - one checked for precise fit, one checked with a 1/4” (6.35mm) tolerance. That felt like enough to account for measurement inconsistencies while not having a silly amount of leeway. That first file also included some lines for checking certain criteria like whether the wheels’ pivot points lined up with the wheelwells - things to help narrow down where the problems were, if need be.

A test set of data showed the simple calculator worked.

Part 2 is setting that simple calculator up to check both permutations (case A on case B or case B on case A).

I worked off of the following basic flow:

I wanted to have it output individual spreadsheets and not a matrix both for ease of programming and because our roadcase inventory changes fairly frequently, so being able to recalculate these spreadsheets in a more piecemeal fashion was actually relatively important to me. And also matrices are hard.

As of 12 August 2020, I have a rough sketch of it written up, but upon testing it looks like there’s something wonky somewhere since it’s not giving the expected results. Tracking that down and updating things will be a process. For now, if you’re interested, here is the test road case measurements csv (these are highly unrealistic measurements), the long-form code for over and the the code for over as a function. Just please know that it’s known broken.

To add: revised arithmetic, a picture for which measurements are which, code for each road case under the others.

Marked defunct 12 nov. 2022