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

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
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."
The approach that I use is just dirt-simple (or maybe brute force is a better term). From Windy, I take the predictions for 5 models and just average them, perhaps throwing out an outlier. From there, I just determine the wind speed for each thousand-foot altitude range during the flight, or perhaps larger intervals depending on the data. For the descent, I just take drogue and main rates of 80 and 25 ft/s to determine the time for each altitude range, and assume that the drift is equal to the wind speed. I'm normally starting with a pretty good estimate of the apogee position if it's a vertical flight. Getting the apogee position on the current flight was a little more challenging. The ascent, with and without drift, is important because the wind is the highest at the point where the rocket is slowing down. The process is the same, though, where the time spent in each altitude range on ascent is calculated from the simulation and then the windspeed is applied. The biggest problem is that Windy only gives 3-hour time increments and it is not obvious what time is associated with the data.

With the above approach on a 10K flight, I can typically get the apogee position within a few hundred feet and the landing position within perhaps 500 feet. That's good enough to avoid all obsticles except for the runways at Hearne.

Jim
 
The approach that I use is just dirt-simple (or maybe brute force is a better term). From Windy, I take the predictions for 5 models and just average them, perhaps throwing out an outlier. From there, I just determine the wind speed for each thousand-foot altitude range during the flight, or perhaps larger intervals depending on the data. For the descent, I just take drogue and main rates of 80 and 25 ft/s to determine the time for each altitude range, and assume that the drift is equal to the wind speed. I'm normally starting with a pretty good estimate of the apogee position if it's a vertical flight. Getting the apogee position on the current flight was a little more challenging. The ascent, with and without drift, is important because the wind is the highest at the point where the rocket is slowing down. The process is the same, though, where the time spent in each altitude range on ascent is calculated from the simulation and then the windspeed is applied. The biggest problem is that Windy only gives 3-hour time increments and it is not obvious what time is associated with the data.

With the above approach on a 10K flight, I can typically get the apogee position within a few hundred feet and the landing position within perhaps 500 feet. That's good enough to avoid all obsticles except for the runways at Hearne.

Jim

Thanks, Jim.

I am confused on your ascent analysis. Do you have wind in the apogee simulation? If so, there should be some weathercocking into the wind, right? However, it sounds like you are using the winds aloft to push the rocket with the wind. ????

For weathercocking into the wind, the sim results are very inconsistent. See this thread.
 
I should let Jim explain how he shoots for an apogee position for his flights, but he was flying his VOS guidance system for his flight track.

This means that Jim's final apogee position over the ground wouldn't follow the normal vector sum of ( upwind ascent path ) + ( downwind drift ) + ( random YEET and YOINK ) that you and I and most other fliers have to try to calculate.

Jim's final apogee position over the ground will be a function of ( programmed VOS flight profile ) + ( downwind drift ) + ( random YEET and YOINK ) only.

-- kjh
 
Thanks, Jim.

I am confused on your ascent analysis. Do you have wind in the apogee simulation? If so, there should be some weathercocking into the wind, right? However, it sounds like you are using the winds aloft to push the rocket with the wind. ????
He is using his VOS. No weather cocking...lol
 
Thanks, Jim.

I am confused on your ascent analysis. Do you have wind in the apogee simulation? If so, there should be some weathercocking into the wind, right? However, it sounds like you are using the winds aloft to push the rocket with the wind. ????

For weathercocking into the wind, the sim results are very inconsistent. See this thread.
I have a vertical stabilization system on the rocket, so aside from an initial tilt and recovery when up to speed, I don't get weathercocking. Once vertical, however, the rocket will still drift with the wind on ascent. Actually, the rocket will drift with the wind even if not vertical.

Jim
 
Dave --

I believe I've got HEDLEY LABAR ready for his maiden voyage ( :) ??? does a male rocket make a maiden voyage ??? :) )

HEDLEY is a 29mm BMS School Rocket I've been working on to learn Head End Deployment.

I've signed up for the Tripoli Houston Launch at Hearne, TX but my attendance is still a little up in the air because our granddaughters are coming home sometime Saturday and I am the designated picker-upper.

Anyhow, I decided to prepare for the launch, just in case.

Attached is the latest version 2.25 of GPS DC with the flight info for HEDLEY on an AT F67W Econojet Motor.

I also fixed the GPS coordinates of the Hearne, TX launch pad so it is in the proper spot instead of in the airport terminal :)

This is a screenshot of the Excel Program:
hedley-C40608-f67-gps-dc-screen.png

The Data was there under the [Saved Forecast] Tab this time !

This is a screenshot of googleearth with the landing scatter data:
hedley-C40608-f67-ge-screenshot.png

Hmmm ... those landing spots are pretty close to the asphalt ... maybe I need to scoot my 2020 mini-button rail a little east :)

My fingers are X'd that I can actually make it to the launch but I'll let you know how it goes either way !

Thanks Dave.

-- kjh
 

Attachments

  • gps-dc-2.25-updated-hedley-C40608-f67-hearne.xlsm
    1.1 MB · Views: 0
Last edited:
Dave --

I believe I've got HEDLEY LABAR ready for his maiden voyage ( :) ??? does a male rocket make a maiden voyage ??? :) )

HEDLEY is a 29mm BMS School Rocket I've been working on to learn Head End Deployment.
.....

-- kjh

kjh -
Looks great! Hoping you get to fly as well.
We are supposed to have a launch on Sat, but we've gotten close to 3 inches of rain this week alone, and had gotten a bunch last week as well.
Fields/roads will be iffy.

One thing I would recommend is going to the latest 2.26 version of DC.
I fixed the forecast recall function so that it actually works in that version.

Thanks,
Dave
 
Thanks Dave.

I downloaded and saved release 2.26 last week but I inadvertantly ran release 2.25 beta this morning ( ! DOOH ! )

Good luck with your weather and your site conditions this weekend and thanks again for the CLUE :)

-- kjh
 
Did forecast last night for LDRS and the Viper, about 1pm yesterday afternoon.

Included weathercocking based on Rocksim. Launched about 2pm, landed on the 2pm pin. Was maybe 2-400' due east of the 2pm simulation.

Close enough for me to use it again in the future. I did not have GPS onboard. Just going from the location of the trees. Certainly within the error of all the simulations, wind perturbations, rocket gyrations, yada yada yada.

1717720198735.png
 
Did forecast last night for LDRS and the Viper, about 1pm yesterday afternoon.

Included weathercocking based on Rocksim. Launched about 2pm, landed on the 2pm pin. Was maybe 2-400' due east of the 2pm simulation.

Close enough for me to use it again in the future. I did not have GPS onboard. Just going from the location of the trees. Certainly within the error of all the simulations, wind perturbations, rocket gyrations, yada yada yada.

View attachment 649259

Nice! What was the max altitude?
 
Did forecast last night for LDRS and the Viper, about 1pm yesterday afternoon.

Included weathercocking based on Rocksim. Launched about 2pm, landed on the 2pm pin. Was maybe 2-400' due east of the 2pm simulation.

Close enough for me to use it again in the future. I did not have GPS onboard. Just going from the location of the trees. Certainly within the error of all the simulations, wind perturbations, rocket gyrations, yada yada yada.

View attachment 649259

That's really Awesome!!
This was your L1 flight, correct?
 
The approach that I use is just dirt-simple (or maybe brute force is a better term). From Windy, I take the predictions for 5 models and just average them, perhaps throwing out an outlier. From there, I just determine the wind speed for each thousand-foot altitude range during the flight, or perhaps larger intervals depending on the data. For the descent, I just take drogue and main rates of 80 and 25 ft/s to determine the time for each altitude range, and assume that the drift is equal to the wind speed. I'm normally starting with a pretty good estimate of the apogee position if it's a vertical flight. Getting the apogee position on the current flight was a little more challenging. The ascent, with and without drift, is important because the wind is the highest at the point where the rocket is slowing down. The process is the same, though, where the time spent in each altitude range on ascent is calculated from the simulation and then the windspeed is applied. The biggest problem is that Windy only gives 3-hour time increments and it is not obvious what time is associated with the data.

With the above approach on a 10K flight, I can typically get the apogee position within a few hundred feet and the landing position within perhaps 500 feet. That's good enough to avoid all obsticles except for the runways at Hearne.

Jim

@JimJarvis50
I would be curious to see if you employed GPS DriftCast to your methods in one of your next flights. Look for version 2.26 at time of this writing.
Turn the weathercock table off, and set your launch latitude/longitude to your predicted apogee and run it.
I think the approach I am using for drift is probably pretty similar to you - mostly just brute force.
Only difference is the forecast. Within 45 hours of the launch, the forecast is RAP, which is one of the forecasts Windy uses.
Further out, Open-Meteo is used, which is similar to GFS I believe.

Thanks,
Dave
 
Back
Top