Weather Routing

[seandepagnier]

There were some recent changes submitted for landfall detection. These were mostly to prevent traveling too close to land. I did not test them so I cannot be sure your result is not from this. Maybe an older version works better.. Are you building from source or using which version?

For this reason I have delayed releasing the plugin, there are still issues.

[rgleason]

Originally Posted by rgleason

As I said before, have you tried to check inverted?

bump?

[SCarns]

What the most recent release for OSX? I lost my hard drive and the one online is still 1.10. I'm sure I had a newer version than that... Anyone have a link?

[ptulp]

Originally Posted by Nereus32

What the most recent release for OSX? I lost my hard drive and the one online is still 1.10. I'm sure I had a newer version than that... Anyone have a link?

I have a version 1.12 from May here: https://github.com/ptulp/weather_rou...gin_1.12.0.pkg


I still have to build the latest version.


Peter

[SCarns]

Originally Posted by ptulp

I have a version 1.12 from May here: https://github.com/ptulp/weather_rou...gin_1.12.0.pkg


I still have to build the latest version.


Peter

Excellent... Thanks Peter!

[rgleason]

Small Update from Sean, compiled for Windows and O4.8.4 and O4.8.6
https://github.com/rgleason/weather_...s/tag/v1.12.1e

[rgleason]

New update from Sean, with improvements from Theeko and Did-g..
There is also a very nice informational, Pull Request that has not been merged yet by Sean. I've run 4-5 routes successfully from this plugin.

Try checking View > Settings > "Apparent Wind on Barbs for Route (not true wind)"
Give this version some exercise please and report.

Sean's update for Windows Opencpn v4.8.4 & v4.8.5 Weather_routing_pi here

[Gilletarom]

Hello,


But, now, not compiled.

[rgleason]

Sorry Gille, now attached.

[Gilletarom]

Hello,

Originally Posted by rgleason

Sorry Gille, now attached.

RGleason : Thank you.

Sean :

I try your last WR bêta version avec O486. The changes made work.

I had already tested Didg's work. It's good.
I saw the change in true wind and apparent wind: Bravo.

I take this post to report items that remain in English even if OpenCPN is in French. They are visible on the attached screenshots. There are:
- 3 on the first screenshot
- 15 on the second screenshot.

Gilletarom

[rgleason]

Gille, there is a new weather_routing_pi update of trivial fixes by Theeko and did-g
It works fine. Don't know if the terms have been entered in internationalization.

[Gilletarom]

hello Sean, Hello Didg, Hello Theeko,

Congratulations for the improvements made to WR.

Today, I tested version 1.121.g with a grib file containing both wind data and current data for the Channel area, between France and England.
I use boundary built so that WR does not go on the land. As a result, I do not check the box "Detect the earth".

Two tests

- Find routes from Cherbourg (France, North Cotentin) to Brixham (England, south).
- Search a route from Cherbourg to Dielette (west of the Cotentin, south of Cape Hague.)

The first test was successful. Slightly slow, but we get 15-minute isochronous roads in 15 minutes between the French port and the English port. See the screenshot with the green road.

The second test failed. WR is unable to determine a weather-route between these two French ports.

I had already realized, with previous versions of WR, several attempts to bypass the Cotentin passing north of Cherbourg. For example, from Saint-Vaast-la-Hougue to Cherbourg or Dielette.

In all these cases, WR is unable to pass "over the Cotentin". I tested other Grib files, or with other start times. This does not change anything.

See the attached screenshots. (Attention, the black road visible on the screenshots, is a route traced "by hand", without using WR, between Cherbourg and Dielette)

Gilletarom.

[rgleason]

Gille, can you upload the grib and boundary?

[Gilletarom]

Hello Sean, and All,


Here are several links to download:

- A Grib file: DL.FREE.FR
- My boundary: DL.FREE.FR
- My configuration Cherbourg-Dielette: DL.FREE.FR
- My configuration Cherbourg-Brixham: DL.FREE.FR
- The polar of my boat: DL.FREE.FR

I think with all these files, you can try to redo my own tests.


If you look at the result of the Cherbourg-Brixham configuration, we observe that the isochrones are calculated inside an area delimited by a curve that looks like an ellipse.

But, if I look for what are the centers of the ellypse, I observe that this ellipse does not have for axis of symmetry, the line "Cherbourg-Brixham".

This phenomenon is present for any starting point and any point of arrival.

So there is, there, a first BUG. (???)

Moreover, if you look at the result of the Cherbourg-Dielette configuration ", we observe that the isochrones are calculated within an area delimited by the boundary on the one hand, and probably by an ellipse on the other hand , in the context of Cherbourg and Diellette.
Obviously, in this context, WR can not determine a weather-route and announces an error.
Unfortunately WR is apparently not able to modify the ellypse if it does not determine a weather-route.

This is the second BUG.

It would probably be necessary for the user to define, or modify, the characteristics of the ellipse so that WR is able to determine a weather-route. This would solve this second bug.

[seandepagnier]

The reason for the ellipse is the settings you are using in the configuration. You have

MaxDivertedCourse="90"


You should increase this value to widen the search area.