BRW
Contributor
List,
Interesting discussions on related Threads, so maybe
some will find this interesting too.
Modern deco science and technology modeling,
validation, and testing are very global in approach,
not just a "plunk 'em and dunk'em" and pulse'em
and record'em" process which is too differential,
cannot be successfully employed across the gamut
of real diving, and suffers from extrapolation problems.
At least when not coupled across a global platform.
Instead, statistical approaches today permit meaningful
extrapolation and correlation of models.
Nobody, given unlimited bank accounts and everlasting
life, could ever perform all the wet and dry tests necessary
to "validate" a table, algorithm, or schedule. Much more is
needed these days, and can be done with models and real
diving data. And wet tests often differ from lab tests,
so real diving data is more valuable.
So, in the last 15 yrs or so, a paradigm shift has occurred.
Had to occur too.
Modern deco technology relies on large data bases, and correlations
across the full set using statistical methods, maximum likelihood,
and risk of DCS analysis. This could not be done in the past
because of computing power, data, and probably inertia.
And the methodolgy holds for rec and tec diving. This really
is what DAN Project Dive Exploration is all about
This is the approach used by the USN, DCIEM, C & C, DAN PDE,
and military, commercial, and scientific sectors. Plus RGBM.
It's recounted in RGBM In Depth (pp 29 +), Basic Deco Theory
And Apps (Chap 7), Technical Diving In Depth (Chaps 7 - 8)
Plus the published papers listed at end (and Refs therein).
It goes like this.
You can't really test all dives and profiles in controlled
experiments, because of the numbers of possibilities,
costs, and related factors. Because DCS is binomially
distributed, you need to do 6 - 9 repeats of a
chamber experiment, Doppler, or wet test, to get
good stats on DCS incidence rates in the 1% range.
Even rec diving is impossible. You have to really
think globally while maximizing statistical reliability.
So today we employ a model-statistics approach to testing
algorithms, using broad base data and models, followed
by field testing (real diving) and closely follow field
stats provided by vendors, insurance companies,
training agencies, etc. Any spikes or minispikes
in DCS rates then raise a flag.
Also recall that staging algorithms like the USN,
Buhlmann, etc tested relatively few profiles in restricted
ranges, and then released tables for general consumption.
Stats from real diving, later, suggested DCS rates were
in the 1/10,000 range. Minispikes in DCS were
treated with ad hoc diving protocol changes, lessening
of NDLs, etc. But these early approaches did not
really start at the beginning with a global statistical
look at available data. At least, nothing reported
until the past 15 yrs or so suggest it from my best
recollections.
Here's how things go these days, and is an approach
taken by USN, DAN, DCIEM, etc, but I will confine
specifics here to RGBM (and what we have done).
1) -- modern deco science requires a data bank of profiles,
a coarse grain model with the best dynamics, and
a computer to fit the model to the data bank with
risk analysis in maximum likelihood. RGBM Data
Bank has some 2300 profiles with 2 - 3% incidence
of DCS (the higher the DCS the better as anybody
doing this will tell you) and that is used for analysis.
and folding of the model over data to get risk estimates
for model parameters and overall DCS rate. So,
for rec RGBM, this rate is 1/200,000,
and for tec RGBM it's 1/23,000. For rec (no deco to
light deco) diving, this can be done in 10 - 15 minutes
on a desktop PC with megahertz speed. For tec diving,
it takes 1 - 2 hrs. For rec RGBM, only certain segments
of RGBM Data Bank are apropos. For tec RGBM, all of RGBM
Data Bank is needed;
2) -- then you cut the rec and tec RGBM algorithms loose and
watch the reported DCS rate in the field (real diving
world) and check trends (minispikes in DCS rate);
3) -- such field testing of RGBM has been online with
NAUI, C & C, and selected individuals for the past 8 yrs;
4) -- so far, RGBM has virtually zero DCS rate across
rec diving (100,000s of meter, and rec Table
dives) and tec diving too (1000s of meter dives
and tec Table dives). But, of course, somebody
WILL get hit. Has to happen. Will happen.
Here's a published bibliography in addition to books.
1) -- "New Looks And Decompression Models",
B.R. Wienke, USJNP Proceedings 131, 2002
(Hawaii)
2) -- "Effects Of In Water Decompression Profile On
Bubble Formation After Dives With Surface
Decompression", A.O. Brubakk, A.J. Arntzen,
B.R. Wienke, UnderSea Hyp Med
3) -- "Modern Decompression Theory And Methods"
B.R. Wienke, Norwegian Underwater Intervention
International Symposium Proc 13, 2001 (Bergen)
4) -- "Computational Reverse Dive Profiles: Contrasts
And Comparisons", B.R. Wienke, Proc Smithsonian
Wkshp 247, 1999
5) -- "Project Dive Safety And Computers In Diving:
Outlook, Availability, And Application, B.R. Wienke
et al, UHMS/DAN Dive Computer Wkshp Proc 6,
1999 (Cozumel)
6) -- "Understanding Dive Tables And Testing Procedures",
B.R. Wienke, S Pac UnderSea Med Soc J
24, 209 (1995)
7) -- "On Nonstop Time Limits, Safety Stops,
And Ascent Rates", B.R. Wienke
S Pac UnderSea Med Soc J 22, 192
(1993; Proc AAUS Wkshp Ascent Rates,
36, 1991
8) -- "Acclimatization Reduces Incidence Of
Decompression Sickness", C.E. Lehner,
B.R. Wienke, E.H. Lanphier, UnderSea
Hyper Med 21, 22 (1994)
9) -- "Bubble Number Saturation Curve And
Asymptotics Of Hypobaric and Hyperbaric
Exposures", B.R. Wienke, Int J. BioMed
Comp 29, 215 (1991)
10) -- "Numerical Phase Algorithm For Decompression
Computers And Applications", B.R. Wienke
et al, Comp Biol Med 22, 389 (1992)
11) -- "Critical Tension Envelope For Multilevel
Diving Within The US Navy Tables",
B.R. Wienke, UnderSea Hyper Med 18,
63 (1992)
12) -- "Bubble Model For Repetitive Diving",
B.R. Wienke, S Pac UnderSea Med
Soc J, 21, 197 (1991)
13) -- "Modeling Phase Volume Constraints
Under Repetitive Decompression",
B.R. Wienke, Math Comp Model
16, 109 (1992)
14) -- "Reduced Gradient Bubble Model",
B.R. Wienke, Int J BioMed Comp
26, 237 (1990)
15) -- "Modeling Dissolved And Free Gas
Phases Under Decompression", B.R.
Wienke, Int J BioMed Comp 25,
193 (1990)
16) -- "Altitude Excursions And The 24
Hour Rule", B.R. Wienke et al,
S Pac UnderSea Med Soc J 22,
114 (1990)
17) -- "Phase Dynamics In Diving", B.R.
Wienke et al, Proc AAUS Wkshp
On Safe Ascents 13, 1990 (Woods
Hole)
18) -- "N2 Transfer And Critical Pressures
In Tissue Compartments", B.R. Wienke,
Math Comp Model 12, 1 (1989)
19) -- "Tissue Gas Exchange Models And
Decompression Computations: Review"
B.R. Wienke, UnderSea Biomed Res
16, 53 (1989)
20) -- "Computational Decompression And Operational
Models", B.R. Wienke et all, Proc Third Int
Conf Supercomputing 233, 1990 (New York)
21) -- "Phenomenological Models For Nitrogen
Transport In Tissues", B.R. Wienke,
Nuovo Cimento 8D, 417 (1988)
Regards to all, and Merry Xmas,
Bruce Wienke
Program Manager Computational Physics
C & C Dive Team Ldr
NAUI BOD Vice Chairman Technical Diving
Interesting discussions on related Threads, so maybe
some will find this interesting too.
Modern deco science and technology modeling,
validation, and testing are very global in approach,
not just a "plunk 'em and dunk'em" and pulse'em
and record'em" process which is too differential,
cannot be successfully employed across the gamut
of real diving, and suffers from extrapolation problems.
At least when not coupled across a global platform.
Instead, statistical approaches today permit meaningful
extrapolation and correlation of models.
Nobody, given unlimited bank accounts and everlasting
life, could ever perform all the wet and dry tests necessary
to "validate" a table, algorithm, or schedule. Much more is
needed these days, and can be done with models and real
diving data. And wet tests often differ from lab tests,
so real diving data is more valuable.
So, in the last 15 yrs or so, a paradigm shift has occurred.
Had to occur too.
Modern deco technology relies on large data bases, and correlations
across the full set using statistical methods, maximum likelihood,
and risk of DCS analysis. This could not be done in the past
because of computing power, data, and probably inertia.
And the methodolgy holds for rec and tec diving. This really
is what DAN Project Dive Exploration is all about
This is the approach used by the USN, DCIEM, C & C, DAN PDE,
and military, commercial, and scientific sectors. Plus RGBM.
It's recounted in RGBM In Depth (pp 29 +), Basic Deco Theory
And Apps (Chap 7), Technical Diving In Depth (Chaps 7 - 8)
Plus the published papers listed at end (and Refs therein).
It goes like this.
You can't really test all dives and profiles in controlled
experiments, because of the numbers of possibilities,
costs, and related factors. Because DCS is binomially
distributed, you need to do 6 - 9 repeats of a
chamber experiment, Doppler, or wet test, to get
good stats on DCS incidence rates in the 1% range.
Even rec diving is impossible. You have to really
think globally while maximizing statistical reliability.
So today we employ a model-statistics approach to testing
algorithms, using broad base data and models, followed
by field testing (real diving) and closely follow field
stats provided by vendors, insurance companies,
training agencies, etc. Any spikes or minispikes
in DCS rates then raise a flag.
Also recall that staging algorithms like the USN,
Buhlmann, etc tested relatively few profiles in restricted
ranges, and then released tables for general consumption.
Stats from real diving, later, suggested DCS rates were
in the 1/10,000 range. Minispikes in DCS were
treated with ad hoc diving protocol changes, lessening
of NDLs, etc. But these early approaches did not
really start at the beginning with a global statistical
look at available data. At least, nothing reported
until the past 15 yrs or so suggest it from my best
recollections.
Here's how things go these days, and is an approach
taken by USN, DAN, DCIEM, etc, but I will confine
specifics here to RGBM (and what we have done).
1) -- modern deco science requires a data bank of profiles,
a coarse grain model with the best dynamics, and
a computer to fit the model to the data bank with
risk analysis in maximum likelihood. RGBM Data
Bank has some 2300 profiles with 2 - 3% incidence
of DCS (the higher the DCS the better as anybody
doing this will tell you) and that is used for analysis.
and folding of the model over data to get risk estimates
for model parameters and overall DCS rate. So,
for rec RGBM, this rate is 1/200,000,
and for tec RGBM it's 1/23,000. For rec (no deco to
light deco) diving, this can be done in 10 - 15 minutes
on a desktop PC with megahertz speed. For tec diving,
it takes 1 - 2 hrs. For rec RGBM, only certain segments
of RGBM Data Bank are apropos. For tec RGBM, all of RGBM
Data Bank is needed;
2) -- then you cut the rec and tec RGBM algorithms loose and
watch the reported DCS rate in the field (real diving
world) and check trends (minispikes in DCS rate);
3) -- such field testing of RGBM has been online with
NAUI, C & C, and selected individuals for the past 8 yrs;
4) -- so far, RGBM has virtually zero DCS rate across
rec diving (100,000s of meter, and rec Table
dives) and tec diving too (1000s of meter dives
and tec Table dives). But, of course, somebody
WILL get hit. Has to happen. Will happen.
Here's a published bibliography in addition to books.
1) -- "New Looks And Decompression Models",
B.R. Wienke, USJNP Proceedings 131, 2002
(Hawaii)
2) -- "Effects Of In Water Decompression Profile On
Bubble Formation After Dives With Surface
Decompression", A.O. Brubakk, A.J. Arntzen,
B.R. Wienke, UnderSea Hyp Med
3) -- "Modern Decompression Theory And Methods"
B.R. Wienke, Norwegian Underwater Intervention
International Symposium Proc 13, 2001 (Bergen)
4) -- "Computational Reverse Dive Profiles: Contrasts
And Comparisons", B.R. Wienke, Proc Smithsonian
Wkshp 247, 1999
5) -- "Project Dive Safety And Computers In Diving:
Outlook, Availability, And Application, B.R. Wienke
et al, UHMS/DAN Dive Computer Wkshp Proc 6,
1999 (Cozumel)
6) -- "Understanding Dive Tables And Testing Procedures",
B.R. Wienke, S Pac UnderSea Med Soc J
24, 209 (1995)
7) -- "On Nonstop Time Limits, Safety Stops,
And Ascent Rates", B.R. Wienke
S Pac UnderSea Med Soc J 22, 192
(1993; Proc AAUS Wkshp Ascent Rates,
36, 1991
8) -- "Acclimatization Reduces Incidence Of
Decompression Sickness", C.E. Lehner,
B.R. Wienke, E.H. Lanphier, UnderSea
Hyper Med 21, 22 (1994)
9) -- "Bubble Number Saturation Curve And
Asymptotics Of Hypobaric and Hyperbaric
Exposures", B.R. Wienke, Int J. BioMed
Comp 29, 215 (1991)
10) -- "Numerical Phase Algorithm For Decompression
Computers And Applications", B.R. Wienke
et al, Comp Biol Med 22, 389 (1992)
11) -- "Critical Tension Envelope For Multilevel
Diving Within The US Navy Tables",
B.R. Wienke, UnderSea Hyper Med 18,
63 (1992)
12) -- "Bubble Model For Repetitive Diving",
B.R. Wienke, S Pac UnderSea Med
Soc J, 21, 197 (1991)
13) -- "Modeling Phase Volume Constraints
Under Repetitive Decompression",
B.R. Wienke, Math Comp Model
16, 109 (1992)
14) -- "Reduced Gradient Bubble Model",
B.R. Wienke, Int J BioMed Comp
26, 237 (1990)
15) -- "Modeling Dissolved And Free Gas
Phases Under Decompression", B.R.
Wienke, Int J BioMed Comp 25,
193 (1990)
16) -- "Altitude Excursions And The 24
Hour Rule", B.R. Wienke et al,
S Pac UnderSea Med Soc J 22,
114 (1990)
17) -- "Phase Dynamics In Diving", B.R.
Wienke et al, Proc AAUS Wkshp
On Safe Ascents 13, 1990 (Woods
Hole)
18) -- "N2 Transfer And Critical Pressures
In Tissue Compartments", B.R. Wienke,
Math Comp Model 12, 1 (1989)
19) -- "Tissue Gas Exchange Models And
Decompression Computations: Review"
B.R. Wienke, UnderSea Biomed Res
16, 53 (1989)
20) -- "Computational Decompression And Operational
Models", B.R. Wienke et all, Proc Third Int
Conf Supercomputing 233, 1990 (New York)
21) -- "Phenomenological Models For Nitrogen
Transport In Tissues", B.R. Wienke,
Nuovo Cimento 8D, 417 (1988)
Regards to all, and Merry Xmas,
Bruce Wienke
Program Manager Computational Physics
C & C Dive Team Ldr
NAUI BOD Vice Chairman Technical Diving
