Redistribution of decompression stops to optimise integral supersaturation

[harvalen]

There is no meaningful on-gassing at 6 msw. The difference between 3 and 6 msw is extra .3 atm driving the off-gassing -- and presumably adding to supersaturation, so the question is what's more "better": the faster off-gassing driven by higher delta-P or the lower integral supersaturation from the lower delta-P.

I think we're saying the same thing. In the end, looking at integral supersaturation without taking maximum instantaneous supersaturation into account is probably not going to give a useful answer.

 

[David Carron]

All of the preceding is no-doubt correct.

However, ZHL will just as happily give me an ascent profile with the last stop at either 3m or 6m but cannot tell me which is preferable. Optimising for integral supersaturation seems to suggest that 6m might be preferable and I can certainly apply my own reasoning about tissues and loading and what-not but what I was really hoping for was some hard research results that would have something to say about the optimal last stop depth.

 

[inquis]

However, ZHL will just as happily give me an ascent profile with the last stop at either 3m or 6m but cannot tell me which is preferable.

I suspect a 9m last stop would be even better based on the ISS metric alone, for the same reasons 6m is better than 3m.

I'd still like to understand whether you are computing the area under the super saturation vs time curve or under the super saturation vs sqrt(time) curve.

Also, how are you "redistributing" the fixed deco time across stops? Are you lowering GFlow then adjusting GFhigh until the specified time is matched, repeating as GFlow is further decreased? Or something else?

 

[Sascha314]

Whats exactly the mathematical definition.

Without knowing it, my educated guess is: if you give a GF, both profiles will bring you up with the same saturation of the most saturated tissue.
I believe the other tissue will not differ too much from it, if you change the last stop. So the off gasing after surfacing is not vastly different. However at 3m you off gas quicker, so releasing more gas at this stop. Is this also a component of your final result? I believe then its the driving factor.

One should visualize all 16 tissue as a function of time. Like with a color coding fornthe amount of on or off gasing. Reducing to one number is oversimplifying.

 

[dmaziuk]

I suspect a 9m last stop would be even better based on the ISS metric alone, for the same reasons 6m is better than 3m.

Except at 9 m on air you may actually be on-gassing something: the no-limit line is somewhere around 7.8 m. So you'll probably be "even better" stopping at just above the no-limit depth; you should get longer TTS on the flip side, but the increase may not be meaningful in practice (I'd say one half-time of the fastest TC is probably within the margin of error anyway).

 

[David Carron]

I suspect a 9m last stop would be even better based on the ISS metric alone, for the same reasons 6m is better than 3m.

I'd still like to understand whether you are computing the area under the super saturation vs time curve or under the super saturation vs sqrt(time) curve.

Also, how are you "redistributing" the fixed deco time across stops? Are you lowering GFlow then adjusting GFhigh until the specified time is matched, repeating as GFlow is further decreased? Or something else?

The integral supersaturation is very much what it says on the tin: the integral from t = 0 to t = +inf of the sum over all tissues of max(0, tissue inert gas pressure - pamb). Sorry for the absence of any math formatting on that formula. Numerically, this is calculated by sampling the instantaneous super saturation at 5 second intervals during the ascent and for the 48 hours after the surfacing and using the trapezoid rule to approximate the integral.

For any given dive, I first calculate a complete ZHL ascent profile then throw away everything except the total stop time. I then do an exhaustive search over all possible combinations of stops and stop durations which sum up to the original stop time (I limit this search to integer stop durations in minutes, and use the start of off-gassing to define the deepest potentially useful stop to limit the search space to a few million combinations per dive), simulating the dive and ascent with the stop combination, retaining the one with the lowest IS.

As a slight update to all of this, I tried comparing the raw ZHL16 80/80 ascent to the same ascent but calculated for the last stop at 6m. Although, as far as ZHL16 is concerned, these two ascents are equivalent in terms of safety, the slightly longer 6m profile systematically results in a lower integral supersaturation for the dive.