AB2 model

[Juha]

Hi,


Are there some public references for the AB-2 model?
Is the interpretation of the AB-2 model used by Decocheck available publicly?
Looking at the tissue saturations window Decocheck appears to use quite a large number of compartments. I counted 48.

Juha

[Ben Daglish]

Hi Juha,

JP Imbert has only published one paper on the AB model, available all over the place - for instance, www.bsac.org/techserv/ndc/doc2004/doc2004jpimbert.htm .

As the model is a 'single equation' (rather than being a set of M-Value-like approximations to a curve), in theory you can use as many compartments as you like - as you'll see in the paper above, COMEX found (and I concur) that 15-20 values per decade (based on a logarithmic scale) seems about right, so indeed Decochek uses around 50 compartments. (Ideally of course we'd just integrate the curve...but it's a bit hard... )

Cheers,
   Ben

[Juha]

Hi,

Well, that link didn't work for me, but google found this
http://jacquet.stephan.free.fr/MT92_plongee.pdf

english version is after the french paper.

BTW, the tissue saturation window of DecoChek doesn't scale the window very nicely at the surface.... The saturation levels that are higher (by a large enough margin) than ambient pressure don't fit in the window.

Juha

[Ben Daglish]

Hi all,

The link to that JP Imbert paper should have been http://www.bsac.com/legacyorg/techserv/ndc/doc2004/doc2004jpimbert.htm . If anybody's interested, I've got some notes on it (and decompression theory in general) that I prepared for Mark's talk to the BSAC - it's a sort of 'intermediate' approach for those who haven't done maths for a while

The basic crack though is that it's a perfusion/diffusion-based model with some basic bubble mechanics thrown in. JP take the case of a single arterial bubble 'stuck' at the interface of artery and tissue, and then models the inert gas flow in/out on both sides ( much like others, such as the Hemplemann/Hennessey Bulk Diffusion model that's the basis for the BSAC tables). The theory is then simply that safe ascent depends upon keeping the bubble at the same size or smaller.

What's marvellous though, and the reason that we chose it for Decochek (other than the fact that in the form of the MT-92 tables, it's been tested with many thousands of dives with, as far as we can gather, a lower DCS incidence rate than any other model) is that the maths then reduces down to just two parameters. Neo-Haldanian models generally have two parameters per compartment - from 10 to 40 parameters, depending on the model. VPM / RGBM use a good 10 or 20 parameters - Hill's thermodynamic models even more.

With all models, what happens is that the theory gets developed and the equations produced. *Then*, the equations (which without exception produce exponential-type curves) get fitted to existing data - the thousands of dive-hours actually logged - by tweaking the constants. Even those who know nothing about the maths of curve-fitting can, I'm sure, appreciate that it's a lot easier to tweak an equation with two parameters that one with 20. The AB model basically manages to juggle the various parameters that exist in other models (tissue & gas diffusion rates, surface tension, skin elasticity etc.) all into a form where they either cancel each other out, or simply get added together to become one of the two constants. It's beautiful and elegant, and as soon as I saw it (having previously implemented all the other models mentioned), I knew it was a good 'un.

msg me if you want any more info. I may well prepare a set of pages soon on my own take on deco theory - if anybody's desperate though, I've got a number of notes / resources available.

Ben

[landercarlos]

Hi Ben,

Did you finish your model theory paper?
Could you send me some info, notes or resources about the model?
Is the equation a binomial-probabilistic?

Thank you,

[Pato]

Greetings all,

I am also very curious about more on this AB2 model and decompression theory. Any info is welcomed.

Cheers,
Frederico.

[Mark Ellyatt]

Hi Pato,

Apart from Ben's posts above about AB2 model details there is only the general web 'info' that applies to all deco models freely available via the internet. The 2 links above contain days of reading material pertinent to this model but for a more general read about decompression either download the helpfile in Decochek and have a look at the last section or obtain the services of an tech instructor who can explain the topic in detail.

The TDI decompression procedures power point has a good entry level chunk of decompression theory but this is really where the easy to understand deco-science lesson finishes unfortunately.

I would be very careful about using any web forum suggestions 'practically'- they are generally plentiful in supply but have proved painful on more than 1 occasion 

Regards

[landercarlos]

Hi Mark and Ben,

So after these years; did you finished your theory papers?

It will be nice if you can share it.

To my knowledge  is that Mr. JP is working with his model but I think he is incline on VPM... dunno really

After all I think that for the extremes dives  (as all you guys make) the probabilistic models has advantages.

My best,