Route Planning with GRIB - Work in Progress

[sinbad7]

Piotr..

Would it be too much work to include date/time and wind forecast for each leg in the Track/Route mouse over? Would be much quicker than having to display the Route Manager.

Tore

[PjotrC]

Tore,

The main reason for having the Route Properties Dialog open is the need for Departure Date/Time specification. Without it, it is not known when any WP will be reached, unless the route happens to be an Active one...

The other reason is not to overload the mouseover or right-click menus. I assume that if someone is doing some real planning, it is worth opening the dialog.

So, I try to keep the original style of OCPN by adding as little as needed to obtain useful new functionality. Of course the Planning Dialog could be made completely differently.

When testing, please keep in mind one thing, that did not surface yet - the rows in RouteProp Dialog describe Route Legs rather that Route Waypoints, so the Speed is given for the _previous_ hop, while Wind Forecast - for the _next_ hop...

Currently I am thinking on if/how to display the polar while editing/selecting a waypoint _inside_ a route.

Please keep digging...

P.

[sinbad7]

NOTE .. Re my #71

Please note new link to the polar file library: 4shared folder - POLAR FILES

Tore

[jonasaberg]

Originally Posted by sinbad7

NOTE .. Re my #71

Please note new link to the polar file library: .....

Tore,
I think it is poor use of resources to use a filesharing site to share files that are less than 1 kB in size. It has its benefits for multi-MB files but it is silly to wait to download a file that contains a internet link or a few lines of text.

A better idea is to start a new thread and for users to add their polars, one .pol file per post as a text file. It is easy, just rename it to boatname.pol.txt and of you go...

/Jonas

[sinbad7]

Jonas..

Can not see your point re poor use of resources storing ANY size file in a filesharing site. Your concern is of course the waiting time imposed for free users of the resource which is easily remedied by paying their support fee.

Setting up a polar file sharing library as a mere thread is not very satisfactory as it easily gets lost in the hundreds of threads accumulating here.

Better then to set up the library under the opencpn.org URL if this is desireable,but I will maintain the '4SHARE' library for the benefit of all other navigation programs.

Tore

[sailorF54]

A large number of polars for (mostly) european boats in downloadable ORC certificates
Federation Francaise de Voile
Select 2009, 2010 or 2011 and hit 'Rechercher' to refresh
Data at the bottom of the second page

[jonasaberg]

Originally Posted by sailorF54

A large number of polars for (mostly) european boats in downloadable ORC certificates
Federation Francaise de Voile

Very good.
The Swedish site is similar with partly different types of yachts, and of course, since Sweden is much smaller than France, not as many certificates. ORC are not the most common handicap system in Sweden/Scandinavia.

A common problem is that the since the certificates are valid for only one year they tend to be erased from the webpages whenever they overhaul the design of them. They don't serve a purpose for the sake of racing after they have expired.
If someone could be bothered to download the files and store them elsewhere it would be great. Preferably in a zip-file, and if so on the site Tore suggests or a similar site.

/J

[PjotrC]

I think it might be too early for collecting polars for OCPN... perhaps it is still better just to note the existing repositories serving well their well understood purposes.

As I mentioned in an earlier post, I am trying now to build a tool helping in cruising passage planning:

- estimating time to complete a given route under a given forecast, including comparison of different routes
- viewing the GRIB from the route progress perspective rather than from time perspective only
- providing a wind forecast summary along the route

Essentially this is ready in 2.5.1010. Still some small improvements to come.

A polar diagram of boat performance vs. wind is nice to have, but does not result in dramatic improvement - on a longer passage a fixed cruising speed estimate gives surprisingly good approximation, and average upwind VMG is just about equal to

beat_speed*cosine(beat_angle)

But yes, the polar has to be there.

However:

1. The syntax of polar specifications is just one aspect.

2. The method of approximation used is another aspect.

3. The interpretation of chosen items (beat angle, run VMG, etc.) is yet another aspect.

4. The requirements for data (specific items need/need not be present) is yet another one.

5. Depending on the application, there might be a need for specifying only real sailing speed, or a passage average, perhaps combined with motoring speed.

6. The convenient method, by which the polar is constructed, may be different for many classes of users.


ad. 1. The syntax : Easy enough, the format accepted here is syntactically the same as for Bluewater Racing and Expedition, and similar enough to ORC Certificate contents and (I think) MaxSea. The other formats (e.g. Euronav, VRTool) are meant rather for use just with the respective programs. Rich as they are, I do not believe adopting any one of them would be a good move now.

ad. 2. Method of approximation: There seems to be no standard on this, yet it makes a big difference on end result. If splines are used, the curve looks nice end-to-end, but the speed forward of the beat angle falls down faster or slower, depending on what is the first point provided. Bluewater suggests that it should start ideally at 0 degrees, but at the same time requires the 2nd value to be the optimal beat angle. This results in rather smooth transition from beat angle down to 0, inviting sailing much closer to the wind, than is possible in reality. I have chosen the linear interpolation, applied only after the 2nd point of the polar, i.e. after the beat angle. This results in a sharp line distinguishing between the course which is "possible" and "impossible" to sail. To acheive the same kind of sharp line with spline approximation, it is necessary to put the first data point much closer to the beat angle, than to 0. So many programs will give very different results with the same polar...

ad. 3. Interpretation of chosen items : These values are of greatest use to the optimizing algorithms, not implemented here (yet). Still, it would be nice to use the existing polar files in unmodified form as much as possible.

ad. 4. Requirements for data : For weather-routing algorithms the polar needs not be complete (through 360degs), it may leave some angles unspecified (never to be used), for hand-made real planning this is not acceptable. A polar must give some reasonable speed value for any angle in any wind speed. The no-go values may be represented by just very low speed (I use now 0.11 kt, which means - "we will not really sail this way"). This difference means that the polars prepared for racing simulation and automatic route optimization may need to be adapted, before being used in manual passage planning. There is also the issue of "zero speed" which will wreak havoc in any manual plan. The ready-made, existing polars may differ widely as to what is assumed about their "unspecified" sectors, and they not always cover all the 0-360 deg range as well as higher wind speeds.

ad. 5. Application : I see broadly three kinds of polars, all fitting in the same syntax:
- pure "racing" or "sailing" polars
- tactical passage planning polars - requiring the user to explicitely plan the beating legs and tacking points
- strategic passage planning polars - just showing the average passage speed (speed with wind-on-the-nose is 0.4-0.5 of the speed on a closest reach, so what?)
(both planning polars may, to a different degree, express the motoring capability or preference)

Within the current simple syntax there is no way to specify configurations (spinnaker or jib, motoring or sailing, etc.), yet the motoring option is so important in passage planning, that it is necessary to include it somehow. It is a bit of a hack, but currently I use the decimal fraction of 0.11 to visibly mark a motoring speed. 0.11 means "dead in the water" and 1.11, 2.11 3.11 4.11... all mean "motoring at XX knots". These values are usually not lost in interpolation, and are well visible in Route Properties passage plan.

ad. 6. Polar construction method : I believe that for the cruising sailor the polar is built entirely by owner's experience with his specific boat and his specific sailing style. It could be built automatically by analyzing the Voyage Data Recorder logs, but more often is just drawn up by the owner, asking himself: well, with such wind and such point-of-sail I usually make xx knots on a passage. (S)he will not be choosing interpolation nodes to suit best the spline drawing. (S)he will just record a number of points well remembered from past experience. This means that the polar should be built without additional tools - it should be just a list of points easy to record and manipulate.

So, perhaps too early yet to label any given file "OpenCPN Polar File".

Piotr

[PjotrC]

I posted a new version to opencpn_2.5.1018_setup.exe - 4shared.com - online file sharing and storage - download. No known bugs and I think all functionality there.

The isochrone (polar) is displayed with Append Waypoint To Route, and with Drag Routepoint (when GRIB and departure time available).

The GRIB display is synchronized to waypoint ETA on

- selecting the row on Route Properties Waypoint List
- start (menu selection) and finish (left click) of Append Waypoint
- start and finish of Waypoint Dragging

Enjoy,

Piotr

[Viking Sailor]

Piotr,

As a cruiser, I don't find polars to be all that useful. I try to sail downwind as much a possible. When I do sail upwind it is almost always a short term or tactical maneuver. Thus, the COG is naturally predetermined, and speed tends not to be a critical factor.

There are many more conditions that affect how fast I drive the boat then just the wind speed and direction. These conditions include, but are not limited to, the sea state, how tired the crew is, visibility, shipping, and last but not least how interesting the book I'm reading is. Further, the actual weather often does not agree with the predictions.

I'm not suggesting polars should not be included -- just noting that including them may not be worth putting a lot of effort into. Simple is sometimes better.

Best regards,

Paul

[P_Dub]

[QUOTE=PjotrC;798927]I posted a new version to opencpn_2.5.1018_setup.exe - 4shared.com - online file sharing and storage - download. No known bugs and I think all functionality there.



Your version 2.5.1018 does not work with Windows 7.
Bye
Pete

[PjotrC]

Originally Posted by P_Dub

Your version 2.5.1018 does not work with Windows 7.

2.5.1010 was OK on a Win 7 Starter... I see no reason for 1018 not to work. Will check at next opportunity...

Thanks for trying.

Piotr

PS. I restored the 1010 for download, perhaps you might check... Parallel installation essential and no other plugins than grib and dashboard installed.

[jonasaberg]

Originally Posted by Viking Sailor

Piotr,
As a cruiser, I don't find polars to be all that useful. I try to sail downwind as much a possible. When I do sail upwind it is almost always a short term or tactical maneuver. Thus, the COG is naturally predetermined, and speed tends not to be a critical factor.

Paul and others,

I agree with most of what Paul is saying. When cruising we do the same, when we are just two in crew we don't sail with spinnaker in a strong winds, we sail conservative,take in a reef much earlier, typically use a smaller sail than with a full crew, don't change sail every 15 minutes if the wind is varying etc...

This said, when sailing in the Baltic sea in summer the wind will vary in direction and strength. We have found the use of weather prediction and polars a very strong tool in choosing when to sail and how to sail to avoid sailing into the wind with a small crew.
The polar will tell you the maximum you can perform with your yacht, which you will not use when cruising, but it is nevertheless a very good starting point. An example is to reduce all the numbers with 10-20% and it will give you a much better approximation than just using a fixed speed.

I don't say it is the holy grail but my experience is that it can indeed be valuable if you sail in an area of the world when the wind is not a steady breeze that just change with the seasons but steadily vary during the days.
This is also the reason I would like a tool that take the weather changes in the future into account in the weather routing.

My dream scheme would be:
1. Download gribs that predict the weather every 3 hours in the future for the next couple of days (say, 5).
2. Use the polar and the predictions to calculate an optimized route for the length of the prediction, 5 days.
3. When the next prediction arrives, this is used to make a new optimization from the ships position to the target for the next 5 days.
4. an so on cyclically...

This of course, is based on the availability of Internet connection, but this is typically not a problem in northern Europe, there is 3G access almost everywhere along the coastlines, and when you leave the coastline you typically are without coverage for less than 24 hours.

As I understand this is a bit iut of Pjotr's scope, but I just wanted to vent my opinion! ;-)

Keep up the good work!

/Jonas

[PjotrC]

Fair winds and following seas to all of you, whatever your sailing style and weather prediction method...

a bit off topic, maybe...

The connectivity landscape in Scandinavia is very different even from the rest of Europe.

Attached is a plot of communications pattern I made last year during one of the trips. We had journalists on board and the charter agreement required us to stay connected as much as possible, because the Sponsors expected regular reporting. After looking into roaming and airtime tariffs, I had drawn up a Communications Policy with the following priorities

- free WiFi
- local 3G
- Inmarsat C
- roaming 3G (in extremis)

I fitted a formidable external 3G antenna and a modest WiFi outfit, a stack of routers down below, drawer filled with 3G modems and prepaid SIMS for Poland, Scandinavia, Germany, Netherlands, Belgium, France and UK.

I really tried hard to stay connected all the way.

Today I clicked a bit to convert my mailbox contents into GPX.

The Squares show messages sent by Internet email, the Circles - by Inmarsat C. I used Inmarsat only if I could not get 3G. I dropped harbours from the plot, leaving only positions really underway...

I remeber only once getting a GRIB underway, when I needed it.

Just an experience,

Piotr

[nohal]

Piotr,
tested your latest build on Win7 - works as advertised. Really interesting work you are doing. I'm fighting a bit with the UI (constantly loosing the route start timestamp etc. - this definitely needs some touches), but definitely it's a nice feature.

Pavel