GPS DriftCast (GPS Drift 2.0) Vastly improved landing location prediction based on winds aloft forecasts

[dvdsnyd]

OK, question...

The help tab shows the entering of weathercock data as positive numbers. My Rocksim sims show the range as negative numbers. Do I enter the number as a negative (per the Rocksim plot) or a positive number, per the help example? The help topic is silent regarding the impact of negative vs. positive range numbers from the simulation.

From the looks of it, Rocksim always provides the range as a negative number, as in the rocket got pushed downwind (as it would whether it weathercocked significantly or not) from the launch location, which is the opposite direction of where the rocket will float upon chute deployment. That would indicate the rocket will float back the way it came past the launch point in the direction of the wind.

If I DO enter the numbers as negative, the resulting landing location is FARTHER from the launch location than if I enter positive numbers. So, is the GPD DriftCast algorithm, in effect, accounting for the weathercocking as a positive number INTO the wind?

Enter as positive numbers.

 

[kjhambrick]

OK then.

The TNT Texas Shootout ( TXSO ) is this weekend ( Sat May 25 - Mon May 27, 2024 ).

I've got three potential flights in mind.

1 - T'Pring's P'Toy on I211W to 6239 ft
2 - Spock's Johnson on I435T to 6319 ft
3 - Nocturnal Missions on K110T to 8856 ft

I ran GPS Driftcast v2.2 for Sat and Sun on T'Pring's P'Toy.

As of this morning, Saturday looks pretty iffy, wind-wise.

This is a screenshot of my GPS DriftCast WorkBook:


And this is the kml file overlaid on googleearth with a ruler on the 10:00 am ground track:


Eek !

I am not intimately familiar with the site but I believe that might be the Brazos River up there by all those landing pins ...

EDIT: Nope it is just a creek ... the Brazos runs N-S about 3-4 miles east of the Rocket Ranch Launch Site ...

Sunday looks MUCH better using this morning's forecast:


And the rocket should land in the "ideal" recovery area:


The downside is that I need to leave Seymour on Sunday morning or early afternoon at the latest.

Maybe I should finish Hedley Labar and send him on F67W EconoJet motors to 1150 ft to test HED with Tinder Piranha Line Cutters



I'll be checking the winds later today and tomorrow eve ... maybe things will improve.

Any which way, I will be sure to report after the weekend.

-- kjh

EDIT: added a gratuitous pic of Hedley

 

[Buckeye]

No tailcone, CP at 57.006, CNa = 14.479, apogee = 11,926 ft:


Fake tailcone, CP at 55.789, CNa = 13.353, apogee = 13,025 ft:


Edit: here is the plot ... not much 'upwind' change ...

So ... would this count as an "inverse Base Drag Hack" ?

Yes, not much upwind change. The tailcone volume of the Firestick is pretty big. Anyway, my lack of sim weathercocking must be due to the overall small stability margin in OR.

Haha, yeah inverse BDH, but at least a tailcone is a real thing!

 

[JoePfeiffer]

This really shows how much we could use a winds aloft model in OR.

 

[Charles_McG]

I use the multi-level wind add-in for OR all the time. But I have to hand input the data from windy to the extension. I use it for drift analysis - though without the knifty map plotting. Just the north-east 2D graph.

 

[kjhambrick]

@Charles_McG --

Does the "Multi Level Winds Aloft" add-in work with the latest version of Open Rocket ( i.e. version 23.09 ) ?

If so, I definitely want that !

Thanks !

-- kjh

 

[Charles_McG]

Yes.
https://www.rocketryforum.com/threa...ta-period-is-now-finished.171295/post-2293133

 

[MarsLander]

I

Yes

I tried adding it into OR, following the github instructions.

This doesn't exist...Simulation options -> Add extension -> Flight -> MultiLevelWind

Ideas?

 

[dvdsnyd]

This really shows how much we could use a winds aloft model in OR.

More than just the winds aloft model, the ability for OR to FETCH the forecast data and use it, instead of relying on users to copy and paste lines of wind direction and speed data into a table.

 

[Charles_McG]

I

I tried adding it into OR, following the github instructions.

This doesn't exist...Simulation options -> Add extension -> Flight -> MultiLevelWind

Ideas?

I'm running 22.02

 

[kjhambrick]

I'm running 22.02

<<snip>>

Ahh Haa !

I am running OR version 23.09 ...

That 'splains it

-- kjh( shucks )

 

[Charles_McG]

Ahh Haa !

I am running OR version 23.09 ...

That 'splains it

-- kjh( shucks )

I just updated to 23.09 and I see exactly what I posted above.

 

[kjhambrick]

Thank you !

It is worth a shot!

-- kjh

 

[JoePfeiffer]

More than just the winds aloft model, the ability for OR to FETCH the forecast data and use it, instead of relying on users to copy and paste lines of wind direction and speed data into a table.

When I think of what I'd want in the model, that's part of it.

 

[kjhambrick]

Another day, a new forecast.

Once again, this is "T'Pring's P'Toy" on an AT RMS 38/480 I211W to about 6200 ft or so.

GPS DriftCast Screen Shot for Saturday, May 25:


Lot's of red boxes there

This is the google earth landing scatterplot:


Further out than the previous forecast but the landing scatter plot looks a little less hazardous ... ???

This is Sunday, May 26.

GPS DriftCast Screen Shot:


Not so RED

This is the google earth landing scatter plot:


Interesting ... while the Saturday flight lands farther out, it might actually be a shorter walk if I can drive up to CR 442 then notth on the dirt road to the landing area.

I don't know the site that well ( errr ... at all ), and it's been a bit rainy lately so I'll have to see what's what when I get there ...

-- kjh

EDIT: Maybe Saturday isn't so bad ?

My 6200 ft flight should land in the same area that Matt's AT DMS H13W, 9200 ft flight landed as reported by @manixFan here on TRF > For Sale Custom Comspec Receiver Compatible Beacons

 

[dvdsnyd]

When I think of what I'd want in the model, that's part of it.

I did open Feature Request #2477 on the OR Github for this feature.
It references this thread and GPS DriftCast

 

[dvdsnyd]

I updated the spreadsheet...even though I said I wasn't going to...
This time to *hopefully* add the functionality to recall a saved forecast.
For now the program saves just one forecast - the last one that was ran.
This is saved in the Saved Forecast Tab.

The full instructions are located in the help tab of the workbook.

Quick run down:

User selects the Saved Forecast option in the launch date
This sets the saved forecast flag in the program.
The cell under the launch date fills to green as an input.
User selects this cell and selects the saved date via the drop down.
***Most important****
The user MUST then update the launch window and begin launch time inputs.
This is done by going to the Saved Forecast tab and grabbing that information there.

Now when running the sim, the saved forecast will be used instead of a new one from the internet.
This is a bit of a beta/experimental release.
Everything should be stable, minus the functionality of the recall forecast, but please let me know if you run into issues!

It's great seeing experiences from users. I'm still not totally sure what to think on some of it. I think the weathercock portion needs some massaging.
Just not sure what to do yet.

Thanks!!!!!
Dave

 

[kjhambrick]

Nive, @dvdsnyd !

I need to finish some odds -n- ends for work so I can leave for the weekend but I'll take it for a spin.

-- kjh

 

[kjhambrick]

The bad news:

I just noticed Windy's latest forecast for the TNT Seymour Rocket Ranch:


Those low clouds early Saturday look ominous.

The good news is that my wife has decided to stay home so she will be here Sunday when 'the girls' come home from Lake Palistine ( don't tell her I said that )

That means I can stay later on Sunday and maybe get a flight or two in.

I loaded all the TX sites I know into GPS DriftCast 2.25 Beta and attached here in case anyone wants a copy with the TX launch sites:


Same Rocket, same I211W but ran a new scatter plot with GPS DC v2.25 for Sunday:


It looks very good !

Maybe I can fly that 28 year-old K1100T to 8800 ft after all

-- kjh

p.s. I'll play with the new Saved Forecast feature but first I need to finish up for work and then pack for the weekend ...

 

[Buckeye]

Well, here is part of the problem. OR and RS predict very different weathercocking. I set up both simulators the same as possible and to my flight in post #31. I would have added RAII also, but it does not permit fins on a boattail like the MAC Firestick.

Here are the inputs the DriftCast weathercock table for 3 simulations. I also plotted them for a more graphical view:






RockSim predicts way more weathercocking than OpenRocket, up to 2.5 times. Rocksim CP method, with more stability margin, naturally gives more upwind weathercock. Check.

What is going on with OR at 20 mph? In fact, at 30 mph, OR predicts no weathercocking at all.

Actual flight upwind distance was about 1800 ft with 10 mph surface winds. I don't have the winds aloft recorded, but 25 mph or less from what I remember when I was playing with the forecasts. I think I will be using Rocksim in the future.

The rocket flight had some unique features which may be at play - boat tail affecting CP, and a dual thrust motor. The flight also started downwind 200 ft before turning upwind.

 

[MarsLander]

OR and RS predict very different weathercocking.

Would like to see these compared to RASAero...

I was the one that first modified the GPS spreadsheet to include the weathercocking. I had the same concern at the time I built it, just never got around to comparing RS and OR and RasAero...thanks for doing this @Buckeye

The two biggest variables in this process/approach, are exactly what you are describing, the weathercocking...and the other, that I discussed with @dvdsnyd is the source of the winds. The original used Windy. The new one uses another site API. When I run one vs. the other (old vs. new) I get different results (the wind forecasts are different).

My intent, at some point, is to do comparisons of all six combinations vs actual launch. No idea when I'll get to do it.

 

[Buckeye]

Would like to see these compared to RASAero...

I was the one that first modified the GPS spreadsheet to include the weathercocking. I had the same concern at the time I built it, just never got around to comparing RS and OR and RasAero...thanks for doing this @Buckeye

The two biggest variables in this process/approach, are exactly what you are describing, the weathercocking...and the other, that I discussed with @dvdsnyd is the source of the winds. The original used Windy. The new one uses another site API. When I run one vs. the other (old vs. new) I get different results (the wind forecasts are different).

My intent, at some point, is to do comparisons of all six combinations vs actual launch. No idea when I'll get to do it.

That would be interesting. I will use another simple rocket design suitable for RASAero and compare the 3 software again for weathercocking.

Keep in mind Windy also has about 5 different wind models, one of which is the same as the API call from DriftCast. Some days they are similar, some days different. I can live with that, since it is weather prediction after all. The ease of fetching one of the data as DriftCast does far outweighs any discrepancies with other models, in my mind. Based on a sample of one, the forecast works well!

What bothers me is nearly identical software, solving the same equations of motion, with the same algorithm and giving very different results. OR is 6 DOF, RS is 3 DOF, but I am not sure if that should matter with a constant wind vector.

 

[sr205347d]

Don't expect wind forecasts to be super accurate, especially at lower altitudes. Wind can be highly variable at lower altitudes due to thermals. Even where and when there are no thermals, the wind forecast is a bit of a guess. I owned a little four seater airplane for 18 years, and noticed quite a difference between forecast and actual wind where I flew, below 10,000 feet. Maybe it is more accurate these days. It should also be more accurate at higher altitudes.

You can see different values for the forecast wind in Windy depending on the forecast model you select.

 

[dvdsnyd]

Sorry, I rushed this out before heading out of town for the weekend.
In version 2.25 Beta, the recall forecast function doesn't work.
I've tested version 2.26 more, it should be working now.
Sorry for the fumble.
Thanks,
Dave

 

[kjhambrick]

Sorry, I rushed this out before heading out of town for the weekend.
In version 2.25 Beta, the recall forecast function doesn't work.
I've tested version 2.26 more, it should be working now.
Sorry for the fumble.
Thanks,
Dave

NP Dave.

Being a SW developer, I understand how things happen :

Speaking of which ...

I flew "T'Pring's P'Toy" on that AT RMS 38/480 I211W with a Blue Raven and a Comspec Tracker onboard on Sunday.

Liftoff was at 09:19 am -- you can kind-sorta see the rocket start to wiggle here.


OR sim'd to 6200 ft or so but apogee was actually about 7000 ft even though the rocket wiggled violently about a second into the flight ( still looking at Blue Raven Data as to why ).

The rocket was definitely transonic when it wiggled but I can't find any obvious reason in the Blue Raven data ...

The actual GPS coordinates of the Pad and the mesquite tree I landed in are marked ( I somehow duplicated each forecast pin but the map is still readable )

So here we go ...

I found a bug in the GPS Map Camera App -- There was no internet so the GPS coordinates remained at the Hearne, TX site !

This is tha Pad with bad coordinates ( note Hearne, TX ) and then a picture of Pad C2 later in the day when I had a 5G connection ...


This is where the rocket landed in a mesquite tree -- lucky me -- TP would have landed in a bog if not for the tree


This is a google earth pic showing the ( duplicate ) pins as well as the distance from pad to the landing tree ( 4873 ft ):


Note that the landing track is fairly parallel to the predicted drift path !

EDIT: these are altitudes ( PAlt - raw pressure altitude, DAlt - pressure altitude adjusted for temperature and site altitude, IAlt - inertial altitude )

I will call this a pretty good GPS DriftCast with questionable sim data.

Please keep it going !

-- kjh

 

[dvdsnyd]

NP Dave.

Being a SW developer, I understand how things happen :

Speaking of which ...

I flew "T'Pring's P'Toy" on that AT RMS 38/480 I211W with a Blue Raven and a Comspec Tracker onboard on Sunday.

Liftoff was at 09:19 am -- you can kind-sorta see the rocket start to wiggle here.

Hey KJH,
Yea, it's been harder than I thought. Amazing how little bugs get through.

Nice flight. I was hoping the landing prediction would be closer.
Kind of like Buckeye's flight, the overall "form" is there, but the magnitude is off.

I know that the 2.25_Beta doesn't work for the saved forecast, but if you still have a sheet with the last forecast run, all should not be lost.
You should be able to copy the Saved Forecast Tab output from 2.25 and bring it into 2.26, and reclaim the sim. Make sure all the inputs are the same with regards to the launch window and beginning launch window.
That way you may be able to play with descent rates or anything else you thought may have needed adjustment.

Just a thought.

Thanks for the feedback and testing! I genuinely appreciate it!

Dave

 

[kjhambrick]

Thanks Dave.

I've got a deadline at work today but I should be able to get to it tonite.

Fortunately, I've got data from my Blue Raven so I might be able to tweak the inputs to match the actual flight.

-- kjh

 

[kjhambrick]

Dave --

Jim Jarvis posted his latest VOS Flight Video here: TRF > Forums Advanced Rocketry Topics [Unrestricted] > Staging, Airstarts & Clusters > I could use just a little guidance > Post #947

And then he posted his pre-flight "Recovery Prediction" here: TRF > Forums Advanced Rocketry Topics [Unrestricted] > Staging, Airstarts & Clusters > I could use just a little guidance > Post #951

Note that the shape of Jim's actual recovery phase GPS track was astoundingly parallel to his predicted track.

Jim usually posts a similar Recovery Prediction a day or two before each AARG launch on the Groups.IO > AARG > Messages List for each Launch announcement Message.

His predictions have been uncannily precise for my flights and most of the other member's flights at each launch land in "Jim's Ballpark".

I've always wondered how he calculates the direction changes with changes in altitude under the chute.

Maybe we can coax him over here or you can ask your Q's over there so @JimJarvis50 can share his secrets ?

HTH !

-- kjh

p.s. the I211W flight I referenced above in Post #85 was embarrasingly ugly. "T'Pring's P'Toy" wiggled and coned and tilted and rolled like never before thru the thrust phase of the flight. I lost sight of the rocket at burnout so I can't say that TP was able to "Straighten Up and Fly Right" after burnout but the onboard Blue Raven data suggests that is the case. I've got some Q's but I don't know what all they are yet ... more to come in a separate thread where I might have Qs and As ...

 

[Buckeye]

And then he posted his pre-flight "Recovery Prediction" here: TRF > Forums Advanced Rocketry Topics [Unrestricted] > Staging, Airstarts & Clusters > I could use just a little guidance > Post #951

Note that the shape of Jim's actual recovery phase GPS track was astoundingly parallel to his predicted track.

Jim usually posts a similar Recovery Prediction a day or two before each AARG launch on the Groups.IO > AARG > Messages List for each Launch announcement Message.

His predictions have been uncannily precise for my flights and most of the other member's flights at each launch land in "Jim's Ballpark".

I've always wondered how he calculates the direction changes with changes in altitude under the chute.

Maybe we can coax him over here or you can ask your Q's over there so @JimJarvis50 can share his secrets ?

The wind forecasts are pretty good, so getting a nice parallel track should be pretty common. I achieved that in post #31. I am sure Jim is using the same speed and bearing approach as DriftCast.

The challenge seems to be getting the apogee location from the trajectory simulation. I'd like to see how Jim did that "with and without drift."

 

[kjhambrick]

The wind forecasts are pretty good, so getting a nice parallel track should be pretty common. I achieved that in post #31. I am sure Jim is using the same speed and bearing approach as DriftCast.

The challenge seems to be getting the apogee location from the trajectory simulation. I'd like to see how Jim did that "with and without drift."

Yes, and then when one has to factor in seemingly random YEET and YOINK events (*) during the thrust phase

-- kjh

Apologies to @Rschub ( see TRF > Forums > Rocketry General Discussion Forum > High Power Rocketry (HPR) > Post #40 )