VPM-B oddity

[Rick Walsh]

Hi Robert,

I read and responded to your email when I woke up at 4am on Friday morning
thanks to a car alarm that wouldn't stop, and my phone was blinking to say
I had notifications.  This time I'm actually awake, looked at your
screenshot, and have tested myself.

On 11 September 2015 at 04:26, Rick Walsh <rickmwalsh at gmail.com> wrote:

> I think we are following the model, and I agree it's silly.  I believe the
> reason is:
> 1. Crushing pressure is calculated at the greatest depth.  The crushing
> pressure helps get rid of bubbles according to the model.
> 2. The Boyle's law compensation takes the baseline value as the ceiling at
> the start of deco, taken as the end of the bottom phase (last user input
> point) of the dive.  A deeper ceiling increases the calculated deco time.
>

I was half correct.  The greater crushing pressure from the 73 m bounce
shortens the deco time.  But the ceiling *is* greatest at the end of the
bottom phase, so my second point doesn't apply.

> The model assumes that deco starts at the last input point.  This is ok
> for a square profile dive.  But for a multi-level dive, especially with a
> deep bouncing, it doesn't make much sense.  I think a better assumption
> would be that the Boyle's ceiling be taken is the deepest ceiling at any
> point in the dive.  This gives very nearly the same result for a square
> profile dive, but in a multilevel dive such as your example, the
> decompression time would increase with the deep bounce, as one might expect.
>

What I am referring to here is just another oddity in VPM-B, at least how
we have implemented it, which could be addressed by taking the Boyle's
ceiling as the deepest ceiling.  Consider this 100 m dive as an example
(for those playing at home, I've used 18 m/min descent, 9 m/min ascent,
12/70 back gas, 50% and 100% deco gas).

Subsurface dive plan
based on VPM-B at nominal conservatism

depth    duration    runtime    gas
100m    6min    6min    (12/70)
100m    14min    20min
66m    4min    24min
66m    1min    25min
60m    1min    26min
60m    1min    27min
---snip---
6m    17min    125min    oxygen
3m    0min    125min
3m    29min    154min
0m    0min    154min

Say you remember that there's something interesting at 66 m, so you decide
you'll pad out the first deco stop so you get to spend 5 min at that
depth.  Should be acceptable; you're not breaking the ceiling, but your
deco time could be extended as a result.  You manually add in waypoints at
66 m @ 24min and 66 m @ 29min.  Now the plan is:

Subsurface dive plan
based on VPM-B at nominal conservatism

depth    duration    runtime    gas
100m    6min    6min    (12/70)
100m    14min    20min
66m    4min    24min
66m    5min    29min
51m    2min    31min
51m    1min    32min
---snip---
6m    16min    118min    oxygen
3m    0min    118min
3m    24min    142min
0m    0min    142min

The effect of the user deliberately extending the first deco stop is that
they brush of 12 minutes of deco.  In my opinion, that's silly and
dangerous.

For comparison, I ran the same plan in the VPM-B Fortran program, and got
total run times of 157 min (3 minutes longer than Subsurface) for 100 m
constant depth, and 168 minutes with a manually entered 5 minute stop at 66
m.  The Fortran program does increase decompression time for spending extra
time at 66 m, which is what it and we should do.

I will send a patch or two to use the deepest ceiling as the Boyle's law
compensation baseline.  And maybe another to warn about the crushing
pressure issue if I can think of a way to do it.

Cheers,

Rick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.subsurface-divelog.org/pipermail/subsurface/attachments/20150912/8046ac11/attachment.html>