[abstract] PARAMETER ESTIMATION OF THE COPERNICUS DECOMPRESSION MODEL.

Rubicon Research Repository/Manakin Repository

[abstract] PARAMETER ESTIMATION OF THE COPERNICUS DECOMPRESSION MODEL.

Show full item record


Title: [abstract] PARAMETER ESTIMATION OF THE COPERNICUS DECOMPRESSION MODEL.
Author: Gutvik, CR; Zeljko, D; Dunford, R; Brubakk, AO
Abstract: INTRODUCTION: Clinical manifestation of decompression sickness (DCS) may occur as a consequence of decompression stress. The majority of current decompression models have been validated using clinical symptoms as an endpoint criterion while the underlying mechanisms have been ignored. DCS is a rather scarce phenomenon and we have a poor understanding of why individuals respond so different to decompression stress, so from a modelling perspective this is not a very good approach for validation. Our working hypothesis is that there is a statistical correlation between vascular bubble formation and risk of severe DCS, and the Copernicus project was started to develop a decompression model where vascular bubble formation should be used as a stress predictor and thus validated through ultrasonic bubble measurements. METHOD: 3 different data sets of man dives with recorded bubble scores were combined to fit the model. A total of 185 bounce dives collected by R. D. Nishi (DCIEM, Canada), 188 multilevel dives collected by R. Dunford (DAN USA) and 84 experimentally designed decompression dives collected by Z. Dujic (Split, Croatia) were used, giving a total of 457 dives. The DCIEM data were used to fit an explicit, polynomial bounce dive model. The Croatia data and DAN data were pooled with the smooth bounce dive model and an optimization problem was formalized to minimize the sum of all errors between the dynamic bubble model, Copernicus, and the bubble scores. The error terms were given customized “punishment” to avoid underestimation of the evolved gas phase. RESULTS: The smooth bounce dive model gives a strong fixation plane around the operating domain of the model, while the Croatia data affect the parameters that give sensitivity in the decompression shape. The DAN data are a bit weaker but help normalizing the sensitivity on the model. DISCUSSION: Although bubble scores can be an additional tool for evaluating the stress/risk of a dive, it’s a common problem that the variance of the measurements is relatively high. Even though we believe that using bubble measurements require less data than using clinical symptoms, it's apparent that the present data pool also contains so much variability that it would be advantageous with even more experiments. Yet, the correlation between the data sets and the model predictions seems to be strong enough to trust the model to produce adequate simulations.
Description: Abstract of the Undersea & Hyperbaric Medical Society 2008 Annual Scientific Meeting June 26-28, 2008 Salt Lake City Marriott Downtown, Salt Lake City, Utah.
URI: http://archive.rubicon-foundation.org/7925
Date: 2008

Files in this item

Files Size Format View
abstract.txt 193bytes Text file View/Open

This item appears in the following Collection(s)

  • UHMS Meeting Abstracts
    This is a collection of the published abstracts from the Undersea and Hyperbaric Medical Society (UHMS) annual meetings.

Show full item record

Browse

My Account