Draft and hull specs/attributes about vesseljs HOT 5 CLOSED

Double check calculations and push master when done.

from vesseljs.

I took the liberty to push to master even before everything is OK. That is because I have fixed countless bugs, and everything seems to be working at least as well as before.

I have introduced a small debug tool under examples: https://shiplab.github.io/vesseljs/examples/PX121_calculations.html

Ixwp and Iywp are still a long way off. But as far as I can see, all other values look quite reasonable. I will of course continue debugging, and it will be interesting to have @MrEranwe 's work on #11 as another debug tool.

from vesseljs.

Status for calculations run on the block case (#11):
The calculation of wetted area ("As") does not yet account for end caps. But even when adding two times the calculated projected area (which should give an upper bound, at least approximately), the number is still much lower than the manually calculated wetted area. This suggests that there is at least one other error too, either in the calculation of wetted area or in that of projected area.
The calculation of center of bouyancy ("Cv") is broken, and returns (0,0,0).
Waterplane Ix and Iy (Ixwp, Iywp) are off by a couple orders of magnitude.
Cb is way off. I expected exactly 0.5 (not accounting for trim), and it returns 0.99725.
Submerged volume ("Vs") does not align with the assumed displacement. A bug here is probably causing the Cb problem too.
The calculated weights and drafts are very different. I would like to see how the manual weight calculation is done, as I know the automatic calculation for the hull right now is very wrong.

The rest of the calculated values are within the margin expected from the difference in accounting or not accounting for trim. (I have not checked the stability values yet.)

I will look into the problems, and try to solve them during this week.

@MrEranwe Maybe you can add manually calculated values for Ixwp, Iywp, yz coordinates for centers of gravity and buoyancy, projected area (along x axis, that is)?

from vesseljs.

from vesseljs.

Now, in dev, it almost passes the blockCase_compare_json.html test. Only the z
component of the center of submerged volume deviates signfificantly from
the hand-calculated value. It looks like more zero-offset areas are sent
to the patchColumnCalculation function than I would expect. (Potential
bug.)

I have (temporarily) introduced simpler volume and area calculations
during debugging, and will try to reintroduce the bilinear patches, as
they give higher accuracy when they work (at least for the volume, where there is a siple analytical solution to the integral).

from vesseljs.