Page 1 of 1

International Date Line and Full polars (ver 1.83)

PostPosted: Tue Feb 23, 2010 11:56 am
by omar
I found a problem with full resolution polar files,
like clipper_pro-full.csv. These files have much
more detail both in wind speed and wind angles.

The extra detail solved an imprecision in vrtool sail
suggestion (sail changes in angles between polar vertices).
Problem is that it broke some of the WR code. My fault.
I have corrected the issue in version 1.82.

If you use "full" polar files and Weather Routing,
you are advised to update to the new version.

Version 1.82 (beta) is available in zip format:
:arrow: http://www.tecepe.com.br/nav/vrtool/vrtool.zip
> Download and unzip in \{program files}\vrtool directory\

Note: Even with the new version, the WR module
is not working very well with the full polar file.
This is because it was designed to work with polar files
with vertices some 5 degrees apart. If polar vertices are
only 1 degree apart, the same number of branches
in the WR track tree will cover a smaller angle.
Raising the number of branches causes the tree to grow fast,
resulting slow calculations.

I found it difficult to run WR with full polars.
I have no solution for that yet.

Re: Full polar bug fix (ver 1.82)

PostPosted: Tue Feb 23, 2010 3:04 pm
by BobNegraes
So, that's why it doesn't work in my computer.... I use pro-full polars.

Let's try again with this new version... AND IT WORKS!!! GREAT JOB, Omar!

Fair Winds!

Re: Full polar bug fix (ver 1.82)

PostPosted: Tue Feb 23, 2010 3:39 pm
by Neptune
GREEEAAAT :lol: :P

This version works even on my Vsta PC - Something that was not possible till now..

Congratulation

Re: Full polar bug fix (ver 1.82)

PostPosted: Tue Feb 23, 2010 6:03 pm
by omar
Regarding Weather Routing, I'm using it configured
with 11 to 15 branches in each tree node.
This means that it tests 5 to 7 different angles to each side of
the direct course (left and right). Using the polar chart with
vertices 5 degrees apart we get 25 to 35 degrees of coverage
each side. That normally includes the best track.

To get the same kind of coverage with full polars would require
using 50 or more branches (not sure). I don't know if this will work.
If using the Weather Routing module, my advise for now is to
use the old - smaller - polar file ( with [x]Interpolation checked )
For other program tools, the full polar seems to work fine.
And is certainly much better in sail suggestion :)

Other corrections in version 1.82:
:arrow: Changed the minimum wind angle used by the track
designer to engage in auto close hauled mode (from 40 to 30)
This improves the tool accuracy for designing windward tracks.
:arrow: Changed color indication of wind angle "too small" in the
boat attitude panel (40 to 30)

Re: Full polar bug fix (ver 1.82)

PostPosted: Wed Feb 24, 2010 2:15 pm
by Ricoo
That's great news. Thanks omar. ;)

Weather Routing and the International Date Line (ver 1.83)

PostPosted: Sun Feb 28, 2010 12:28 pm
by omar
I just fixed a problem in Weather Routing module
with the International Date Line

Since next clippper leg crosses it, the update is
in order. Same address. Again :x

http://www.tecepe.com.br/nav/vrtool/vrtool.zip

Re: International Date Line and Full polars (ver 1.83)

PostPosted: Sun Feb 28, 2010 7:40 pm
by Ricoo
Omar, you're the best. I have to try this right away !
...
Ok, I tried it. Isochrons are always very flat with full polars, so I guess we may still use the classical polars for weather-routing.
Thanks anyway, this is on its way to excellence. ;)

By the way, my test passed successfully over 180°.

Re: International Date Line and Full polars (ver 1.83)

PostPosted: Mon Mar 01, 2010 8:48 pm
by michithesailor
Hi Omar!

Great Job - Weatherrouting works fine for normal routing.

But if I try a to generate a route from Africa to Florida (and using UGRIB data with 12h importfilter) I get "To many points >2M"...?

What I have to do?

Thanks & kind regards
Michi

Re: International Date Line and Full polars (ver 1.83)

PostPosted: Tue Mar 02, 2010 6:54 pm
by omar
michithesailor wrote:if I try a route from Africa to Florida...I get "To many points >2M"...

If start and destination are too far apart, the route tree grows wide fast,
resulting in more than 2 million points. The calculation aborts at this
point, before your computer runs out of memory.