CNS calculation headaches

[Riccardo Brama]

Robert,

  I totally agree with you on the use of a two functions interpolation. One to be used up to 1.5 the other one (that seems to me to be well approximated by a linear interpolation) for higher values.

   That would lead to less "underestimation" and to a behavior that seems to be more realistic.
   
   Cheers,
   R.

Inviato da iPhone di 
Eng. Riccardo Brama, Ph.D.
Chief of Engineering @Dive Industries

> Il giorno 10 ago 2019, alle ore 17:04, Robert Helling <helling at atdotde.de> ha scritto:
> 
> Willem,
> 
>> On 10. Aug 2019, at 16:10, Willem Ferguson <willemferguson at zoology.up.ac.za> wrote:
>> 
>> An interesting alternative, Robert. I am not happy with the deviation at 1.5 and 1.6. One would have to check what the effect of these two points are on the power curve. What is the effect on the overall fit of the power curve if one omits those two points? What of a 3rd order polynomial that could in principle accommodate the inflection at 1.4? I am not averse to a mathematical solution because the linear interpolation also causes some inaccuracy.
>> 
>> 
> 
> here is the same on a log scale:
> 
> <PastedGraphic-2.pdf>
> 
> I would not be happy to fit this with a line for all points including the last two. Rather, I would use a new line for the last three points (and extrapolate that) for values above pO2=1.5bar.
> 
> 
> Robert
> _______________________________________________
> subsurface mailing list
> subsurface at subsurface-divelog.org
> http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.subsurface-divelog.org/pipermail/subsurface/attachments/20190810/5c2a93a6/attachment.html>