Route Planning with GRIB - Work in Progress

[PjotrC]

Originally Posted by dacust

they also have a straight table for wind vs. speed.

Here's one I copied.

Thanks, included in the test set...

Unless I am misinterpreting the gybe angle value, the spline-approximation has some problems here.

I also added a very low (0.1) speed value for angle 0. Always better to avoid explicit 0 as speed.

P.

[jonasaberg]

Maybe someone will know in more detail, but as far as I understand the model used for VPP calculations is based on experimental data and these are given at apparent wind angles (AWA) of 27, 50, 80, 100 and 180 degrees.

Hence, it might not make sense to let the spline go from 27 (or the corresponding true wind angle) down to AWA=0. At least not all they way down to 0.

Also, somehow the fluctuation in the spline need to be handled.

I think the best would be to base the polar on a spline made on the data before the VPP calculation, and not on the datapoints that are the output of the VPP.
Might be wrong here...

I think the Run VMG is the dead run, at TWA=AWA=180.
It doesnt make sense otherwise....

[jonasaberg]

I managed to correct myself two times in my last post.

I think that the ORC certificate should be read like this:

Run VMG = speed of boat at AWA=TWA=180 degrees

Gybe angle = The TWA where the VMG is highest, that is, the TWA when the boat makes the highest speed component in the direction the wind is blowing.

[PjotrC]

Originally Posted by jonasaberg

Run VMG = speed of boat at AWA=TWA=180 degrees

Right, looks much better now...

[jonasaberg]

Better, but then I found this page
Create Polar Diagram - Oppedijk.com

That makes sense....so I have to correct myself again. Getting tired of being wrong... ;-)

Gybe Angle= Best TWA to maximize the downwind component of the speed of the boat.
Run VMG= the component of the boats speed downwind (always smaller than the boat speed)

Beat Angle=The TWA that makes for the highest speed component upwind
Beat VMG the component of the boat speed toward the wind. (always smaller than the boat speed)

Sorry for this mess....

I then have an additional comment: You will see after correcting the curves that the derivative of the VMG with respect TWA is higher for TWA lower than the Beat angle, and that the derivative is lower for TWA higher than the Beat angle.

This explains why it is so dangerous to go to high when sailing against the wind, the boat loose speed very rapidly if you go to high, it is safer to go a little to low instead, you then loose just a little VMG.

Hence, it doesn't make sense to draw the spline all the way down to zero. Better with a straight line as they do here:
Create Polar Diagram - Oppedijk.com

And, sorry if it sounds like I am complaining. I am not, I really appreciate your work.

/Jonas

[jonasaberg]

Pjotr,
Have a look at this. They have a lot of good ideas. It is good but not opensourse.
Sailplanner

Also they have a free polar program, I am pretty sure it is based on the same equations I discussed earlier.
http://www.polardesigner.org/

/Jonas

[dacust]

Originally Posted by champ8242

This is the sheet you are referring to.

Bart

Originally Posted by jonasaberg

Yepp. I am pretty sure it is.
I also have some documentation and the accompanying paper somewhere.

It is a bit tedious to use...

/J

Looks like it. And complicated is right. 15 tabs, macros all over the place.

It opens just fine in OpenOffice. I haven't tried doing anything in it, so can't say if OO runs it all OK. And, I don't know enough about the parameters, acronyms, etc. to even try.

-dan

[jonasaberg]

Dan-
Theres a tab called #Runit# or something.
In it there is a box saying "Write a zero to reset"
do that. Klick inanother box and check that the polar plots are empty (the tabs to the far right)

Then go back to "Runit" again and write a "1" in the box to prepare it for a new run.
Then just to the left of the box is a button saying "Run one parameter" or something like that.

Try if it works. Maybe OpenOffice use the "old" macro language. I tried it in Excel2007 and it did of course not work. (I have had to rewrite my Excel2003 macros for 2007...and they changed from 200 to 2003 as well...
Writing macros in Office is begging for eternal code maintenance.)

Otherwise the link I posted above is easier to use...it also have some documentation to it...

/J

[PjotrC]

Originally Posted by jonasaberg

Hence, it doesn't make sense to draw the spline all the way down to zero. Better with a straight line ...

I've uploaded a new kit to opencpn_2.5.1010_setup.exe - 4shared.com - online file sharing and storage - download .

I've added linear interpolation for computing polar-derived speeds.

Jonas, you are right that smoothing the curves forward of the beat angle is harmful. So, I am interpolating only after the 2nd (twa, kts) pair. This, accidentally, leads to similar polar specification as Bluewater does use. I attach appropriate OCPN polar for the ORC sample.

Looking at the recent posts I have the impression that I am solving a slightly different problem... I am not pursuing mathematical perfection in computational sailing... I try to build a practical (i.e. easy to use) tool that helps a cruising sailor in mid-term (3-5 days) passage planning, based on GRIB data. Wind is only one of important factors, and I do not want to overemphasize it. Basically I would probably be happy with the fixed PlanSpeed, but of course there IS a difference between upwind and downwind....

Thanks in advance for help in the testing.

Piotr

[sinbad7]

Piotr..

Downloaded your latest version of OCPN plus the .pol file.

The attached result does not make sense. 8 kn speed with the wind W6??

Tore

[PjotrC]

What is wrong with this?

Seems that the ORC polar specifies just that kind of performance.

Try some cruising profile, or make your own, perhaps will be more realistic.

P.

[jonasaberg]

Originally Posted by PjotrC

Looking at the recent posts I have the impression that I am solving a slightly different problem... I am not pursuing mathematical perfection in computational sailing... I try to build a practical (i.e. easy to use) tool that helps a cruising sailor in mid-term (3-5 days) passage planning, based on GRIB data. Wind is only one of important factors, and I do not want to overemphasize it. Basically I would probably be happy with the fixed PlanSpeed, but of course there IS a difference between upwind and downwind....

I started out using a homebrew tool that eventually became a commercial product. Initially it:
-Plotted the grib on a rudimentary map with only coastlines.
-I could plot the grib at different times in the future.
-I then could place the boat at a position, and view the corresponding polar-diagram at that position at a predicted wind in a future time.
By playing with this it greatly helped in deciding how to sail a track.

It was good during racing, but I found it to be a terrific tool for +24h cruise with only me and my wife on hour 42 feet yacht. It has made it possible to sail around weather and to adjust the start of the sail to get very comfortable and fast cruising instead of beating agains the wind or getting a boring dead reach.

The next iteration was with routing, and that has been a perfect tool to use but I don't like the current version of it.
I would love to have it integrated into OpenCPN.

I think a good model, weather prediction and polars are all needed.
With only one of the three you will have a worse result. And the prediction is always going to be the weak link.

The Polar can always be adjusted, but the VPP diagram calculated by a VPP program is a very good start, and the speed values can easily be adjusted with a percentage after comparing a little with your own measured results.

[PjotrC]

Originally Posted by jonasaberg

-Plotted the grib on a rudimentary map with only coastlines.
-I could plot the grib at different times in the future.
-I then could place the boat at a position, and view the corresponding polar-diagram at that position at a predicted wind in a future time.
By playing with this it greatly helped in deciding how to sail a track.

I think this is just what 2.5.1010 does... , maybe a bit better on the charting side...

[jonasaberg]

yea, sorry, need to get to a computer i can install and test on...
;-)

[PjotrC]

Originally Posted by jonasaberg

yea, sorry, need to get to a computer i can install and test on...
;-)

Jonas,

If you prefer a source patch, I will post this as well. However, I had to extend the plugin interface, so I am not sure it will run unmodified on other platforms. Some hand combat might be needed in this area.

Would be great to use your experience...

Piotr