Weather Routing Plug In - Does it Actually Work

[seandepagnier]

Originally Posted by duckpaint

Bingo! it worked! wow, I can't believe it. It is really difficult! It would be very usefull to have a Manual or some help in order to get it going! (If you use the help menu, it redirects to a URL that does not exist)

I added more text to the error message for the next version.

I will try now to use the polars for my own boat (I disagree that you may

You disagree?!?
[quote[
have different polars for the same boat, the polars from the ORC certificate are the maximum speed for each wind/angle combination, assuming that you made the necessary changes in your sails)

Regards.

Duck.-[/QUOTE]

Assuming you do this, then ok, but it takes time to change sails, and also, for cruising purposes, you don't just change sails exactly at certain times, so there is overlap in the cross over chart (which I support) which means you don't bother to shake the reef until later, because the wind might pick up again, and you end up doing a lot less work, while sailing a little slower.

Otherwise, you can use an overlap factor of 0%, and the multiple polars is still useful so you can see on the plot when and where sail changes do occur.

Finally, if you want to use a single polar, that is also perfectly acceptable, and should work as well.

[seandepagnier]

Gilletrarom,

Sorry for the delay... when you posted this I was unable to respond.

Originally Posted by Gilletarom

Hello Sean d'Epagnier,

Here joined, screen copy with the same test that in post #36, when WR failed to create a weather-route.
Just, I change the level of zoom. We can see that WR try to create a route, in the direction of the south. And so, not in the direction of the north.

This is strange. Why it failed ?

Here join the gpx file of the boundary and the grib file that I use (You can try à 18.00 H the 29/12).

Best regards.

Gilletarom.

The screenshot looks very strange, but I really need to see your weather routing configuration xml file to determine what is going on.

[Gilletarom]

[QUOTE=boat_alexandra;2323807]Gilletrarom,
Sorry for the delay... when you posted this I was unable to respond.[/QUOTE

No problem.

The screenshot looks very strange, but I really need to see your weather routing configuration xml file to determine what is going on.

Sorry, but I destroyed all data.

B.R. Gilletarom

[Gilletarom]

Hello Sean d'Epagnier,

Originally Posted by boat_alexandra

Gilletrarom,
....
The screenshot looks very strange, but I really need to see your weather routing configuration xml file to determine what is going on.

First, see post #43. I download the data that are in this post.

Then, this morning I try to reproduce that occured, with the same grib file and same boat. I do not success.

But made other test. See screen copy.

The wind is not a problem, not too big, not enough. The polar contain very numerous data .

But, WR cannot go from D7 to A3. See D7 and A3 on screen copy.

Here joined the opencpn/plugin/WR folder, the grib file and the gpx file of the boundary that i use.

So, I think that you have all the data that you need to reproduce my test.

B.R. Gilletarom.

[seandepagnier]

I think my first reply got lost.

Anyway, you should increase the diverted course (basic constraints) to allow the route to complete. If the calculated course is very near the edge of the map, then the diverted course was not enough.

This may take longer to compute, but you can speed it up a little bit by reducing the max search angle (under additional) to say 90 or 60 or...

[stelian]

Hello everybody,

Can someone please explain why I get this route when going from Perros-Guirec (north of Brittany, France) to the Scilly Islands ? (latest OpenCpn git version, latest WR plugin from git, testing on Linux)



I have a feeling this has something to do with my polar. When I choose another polar (tried with 60ftmono.pol and Alberg35.pol) the calculated route seems fine, but when I choose DUFOUR_34.txt or my own boat_dufour31.csv I get bad results...

http://popies.net/tmp/WR/bad-routing.png
http://popies.net/tmp/WR/WeatherRout...figuration.xml
http://popies.net/tmp/WR/dufour31.xml
http://popies.net/tmp/WR/boat_dufour31.csv
http://popies.net/tmp/WR/gfs-latest.grb

[hasse_A]

I also have got such crazy results. Unfortunately I do not have a plot of it.
Next time I will document the plot.
Anyhow you are not alone of getting crazy results.

/Hans

[seandepagnier]

Originally Posted by stelian

Hello everybody,

Can someone please explain why I get this route when going from Perros-Guirec (north of Brittany, France) to the Scilly Islands ? (latest OpenCpn git version, latest WR plugin from git, testing on Linux)



I have a feeling this has something to do with my polar. When I choose another polar (tried with 60ftmono.pol and Alberg35.pol) the calculated route seems fine, but when I choose DUFOUR_34.txt or my own boat_dufour31.csv I get bad results...

http://popies.net/tmp/WR/bad-routing.png
http://popies.net/tmp/WR/WeatherRout...figuration.xml
http://popies.net/tmp/WR/dufour31.xml
http://popies.net/tmp/WR/boat_dufour31.csv
http://popies.net/tmp/WR/gfs-latest.grb

This is definately a bug. I thought I fixed it, but it is a problem in how it computes the final route segment (from the last isochron to the actual destination)

I will try to debug it soon. If your provided information can repeat the route everytime I can fix it.

For now, you can use the cursor route (in yellow) which should be much closer to the correct route when you move the mouse over the destination.

[stelian]

Originally Posted by boat_alexandra

This is definately a bug. I thought I fixed it, but it is a problem in how it computes the final route segment (from the last isochron to the actual destination)

I will try to debug it soon. If your provided information can repeat the route everytime I can fix it.

For now, you can use the cursor route (in yellow) which should be much closer to the correct route when you move the mouse over the destination.

Yes, the yellow route seems correct.

You should be able to reproduce easily the issue with my files, just load the grib and set the route date to "31/03, 12h". Choosing other time/date combinations give other "interesting" routes...

Thanks !

[stelian]

Hi boat_alexandra,

Did you have the chance to look into this issue ?

Thanks,

[seandepagnier]

Yes.

I am attempting to download the grib. 30mb is big, I can transfer about 150mb in 24 hours.

[stelian]

Let me know if you cannot download the grib, I can try to reproduce the issue with a smaller one.

[stelian]

While trying to reproduce this particular issue with a smaller grib, I came upon something else: how does WR deal with winds stronger than the ones listed in the polar file ?

The polar for my boat only goes up to 20 kts, and the grib I downloaded right now lists winds going up to 35 kts. Trying to route stops with "Polar: no data", and I suppose this is because the polar does not say how to deal with this stronger winds. Am I right ?

If this is correct, how can I still force the routing ? Is there some parameter I can set so that winds stronger than 20 kts should be dealt in the same way as 20 kts ? Or should I edit the polar, and in this case what should I put inside (just one new column for, let's say 100 kts, with the same data as the 20 kts column ?)

Thanks,

Stelian.

[seandepagnier]

Originally Posted by stelian

While trying to reproduce this particular issue with a smaller grib, I came upon something else: how does WR deal with winds stronger than the ones listed in the polar file ?

I finally got the grib.

The polar for my boat only goes up to 20 kts, and the grib I downloaded right now lists winds going up to 35 kts. Trying to route stops with "Polar: no data", and I suppose this is because the polar does not say how to deal with this stronger winds. Am I right ?

yes

If this is correct, how can I still force the routing ? Is there some parameter I can set so that winds stronger than 20 kts should be dealt in the same way as 20 kts ? Or should I edit the polar, and in this case what should I put inside (just one new column for, let's say 100 kts, with the same data as the 20 kts column ?)

This is a common complaint, and a lot of people think it should just use the highest wind speed. Other routing programs silently do this, and sometimes the complaint is that the plugin dosn't work when the other program does.

You could add a column for 100 knots, but keep in mind, if you do, it is completely unrealistic. Boats do not sail the same speed all wind speeds above 20 knots that they do at 20 knots on every wind angle. In fact, even if they could you would never want to.


So maybe it's nice to "complete" a routing, I don't really see what the point of it is if the results are surely wrong. You might do better to actually define the speed you sail in 30 knots, in many cases, you might not point very close to the wind, or go very fast upwind, but it also depends on seastate which I have yet to deal with.

[seandepagnier]

stelian

The reason for the funny routes is that your destination forced the router to pass through land. The algorithm isn't so smart, and it tries to find the best final route to the destination, but all of them passed through land so were invalid.


The reason it picked the one it did, is there was an exception for more than 60 miles away since the landfall test becomes very expensive and slows it down too much as it has to load lots of cells. So it "found" a round that did pass through land that was more than 60 miles away for the final propagation, this was wrong.


For now I disabled landfall detection for the final propagation, I think the result is better...