On Mon, 06 Oct 2003 09:06:44 -0400, Glenn Ashmore
tempted fate with:

While you guys are working on programs, I want to see one that produces
polars from actual data. Say you would set the sea state and then when
you get the boat really cooking you hit a button and it records the wind
data and boat speed. Then it would sort it all out and plot polars for
each sea state.

Another option would be to just constantly record the data and then use
peak boat speeds maintained for more than say 30 seconds to select the
points to plot.

It would be really interesting to invite a hotshot crew of racing
sailors to sail a boat and produce a set of polars then do the same with
your regular crew and try to match the results. It would be a great way
to refine sail trim and helm techniques.

Glenn

Oddly enough, my cpRepeater program might be good start on this. The
current version captures all the data you want, including sea state.
It can export it as a CSV file easily imported into just about
anything. Based on your specs, I'd probably need to up the sample
frequency or else add the manual feature you mentioned. Unless you
can live with a 1 minute sample rate. The part I don't have is the
actual generation of polars. That's on the list to research as an
add-on someday. My gut feel is that filtering the data is tougher
than generating the polars. The manual approach you mention would
take that burden off me and put it on you, where it belongs. :-)
Adding the manual sample bit would be easy enough. Hmmm..



Glen "Wiley" Wilson usenet1 SPAMNIX at worldwidewiley dot com
To reply, lose the capitals and do the obvious.

Take a look at cpRepeater, an NMEA data integrator, repeater, and
logger at http://www.worldwidewiley.com/

Glen Wilson wrote:

Oddly enough, my cpRepeater program might be good start on this. The
current version captures all the data you want, including sea state.
It can export it as a CSV file easily imported into just about
anything. Based on your specs, I'd probably need to up the sample
frequency or else add the manual feature you mentioned. Unless you
can live with a 1 minute sample rate. The part I don't have is the
actual generation of polars. That's on the list to research as an
add-on someday. My gut feel is that filtering the data is tougher
than generating the polars. The manual approach you mention would
take that burden off me and put it on you, where it belongs. :-)
Adding the manual sample bit would be easy enough. Hmmm..

The 30 seconds was just an arbitrary period to filter out surges due to
surfing etc. Manually picking the record point would be a problem
because if you are like me, when you really get the boat in the sweet
spot the last thing you think about is pushing a button! :-)

Analysis would be the real problem. First you would have to group by
sea state and wind speed to get a data set and sort by true wind angle
to get the polar coordinate then pick the highest normal boat speed to
get the vector length.

That last step would be the most complicated. Maybe look at everything
within 2.5º+ - of the true wind angle in 5º increments, throw out
everything below the median to eliminate acceleration and use the mode
or the upper standard deviation of the remaining values.

What I am after is not a theoretical value for boat performance from a
VPP but more the relative performance of the boat and crew together. I
would like to be able to gather data for several weeks and plot it out.
That could be used as a base line for tuning the boat and crew. Then
after a few months of working with the boat repeat the process and see
how much we have improved.

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com