TARC Rocket Simulation Difficulty

[diyaerospace]

Hello,

I am on a TARC team and working to hopefully make it to qualifications this year. I have built the rocket and flown it 8 times so far. However, I am having difficulty getting within 15 ft of the target altitude.

For context here are all of the flights.

Flight 1:
Weight 511g
Altitude 700ft
Backtracked Cd 1
F30-8FJ

Flight 2:
Weight 494g
Altitude 943ft
Backtracked Cd 0.815
F42T

Flight 3:
Weight 460g
Altitude 832ft
Backtracked Cd 0.855
F30-8FJ

Flight 4:
Weight 460g
Altitude 817ft
Backtracked Cd 0.888
F30-8FJ

Flight 5:
Weight 425g
Altitude 828ft
Backtracked Cd 0.89
F30-8FJ

Flight 6:
Weight 405g
Altitude 889ft
Backtracked Cd 0.83
F30-8FJ

Flight 7:
Weight 419g
Altitude 872ft
Backtracked Cd 0.83
F30-8FJ

Flight 8:
Weight 428g
Altitude 817ft
Backtracked Cd 0.925
F30-8FJ

All of these flights accounted for wind and other atmospheric conditions. The backtracked Cd was done with rocksim. Right now I don't know how to get closer to the target altitude, anyone have any suggestions? I will backtrack the Cd in Openrocket to see if there is any noticeable difference. I also averaged all the Cd's to get an average Cd of 0.88.

Thanks for your help,
Walter

 

[mikec]

I think this is just within the errors of simulation and motor variation, but you have to post files for anyone to express an informed opinion.

I doubt you have an exact measurement of the wind profile over the whole flight, exact launch angles, etc.

 

[kjohnson]

If you plot out your altitude and mass for your F30 flights, you've got a fairly standard trend started. As long as you aren't changing other things between flights (launch angle, weather and wind conditions) you should be able to hone in on a mass that gets you closer to the target. There are going to be things outside your control like motor variance (see the 2 flights at 460g mass), but you should be getting closer just with mass adjustment alone.

 

[sriegel]

Remember, too, that the sim is just a starting point. Any variation from the assumptions of the sim in your build or real-world weather and launch conditions will give different results.

I might suggest flying at least two - three flights at any given mass to give you a better average value for that mass. Any single data point is less reliable. It could be spot on or it could be a fluke.

Your team is off to a good start on data collection. Continue to build that trend line to interpolate or extrapolate to the ideal mass.

 

[diyaerospace]

I think this is just within the errors of simulation and motor variation, but you have to post files for anyone to express an informed opinion.

I doubt you have an exact measurement of the wind profile over the whole flight, exact launch angles, etc.

Rocksim Simulation

 

[mikec]

FWIW. I only have OR 15.03, which may or may not translate this file correctly. A few potential issues:

You have many component mass overrides and then a single stage override for mass and CG, so the component mass overrides do nothing. This rocket weighs 350g pounds dry in the sim file and sims to 987 feet on an F30, which is not very close to what you are getting in reality. How are you entering the weight into the sim? When I override to 428g I get 844 feet compared to 817 actual which is only 3.3% low.

Fin cross-section is square. OR overestimates the drag of such fins.

OR says the Cd is 0.65. You can't change the Cd directly in OR, you can only influence it with surface finish and fin cross-section.

You had a diameter discontinuity between the nose cone and the body tube. Not sure how much that affects the sim result.

sriegel's comments are dead on, you may have gone as far as you can with sims.

 

[diyaerospace]

This rocket weighs 350g pounds dry in the sim file and sims to 987 feet on an F30, which is not very close to what you are getting in reality. How are you entering the weight into the sim? When I override to 428g I get 844 feet compared to 817 actual which is only 3.3% low.

The rocket's weight listed above is the wet mass. Your Cd of 0.65 is to low when compared to the backtracked Cd of 0.88. If you increase the Cd and decrease the weight that should get your sim to match mine.

-Walter

 

[Andrew Brown]

Did you factor in the temperature variation of the altimeter? The reported altitude varies quite a bit over even small temperature differences.

 

[mikec]

Your Cd of 0.65 is to low when compared to the backtracked Cd of 0.88. If you increase the Cd and decrease the weight that should get your sim to match mine.

Even with square cross-section fins and all surface finishes set to "rough" I can only get to a Cd of 0.81 at mach 0.3 and apogee of 900 feet (starting at sea level, I assume you have entered the altitude of your launch site but above sea level will only increase the apogee by a little.)

Whether the altimeter reports different altitudes as a function of temperature or not, my understanding is that you are stuck with whatever it says: section 2.6 of https://rocketcontest.org/wp-conten...merican-Rocketry-Challenge-Rules-v5.16.22.pdf "The peak altitude of the rocket as recorded by this altimeter... post-flight will be the sole basis for judging the altitude score "

Again, at this point my advice is to set the sim aside and build up more flight data with the same motor.

I assume your reported masses are measured just before you go to the pad?

If you are not using a rail then you can get a fair amount of variation from rod whip. I always recommend using a long stiff rail for TARC.

I'm not sure what a good strategy to account for/adapt to wind is. Ideally your rocket design would minimize weathercocking, and it doesn't look too bad in that regard though a little overstable (2.2 cal).

 

[jmasterj]

You can override the Cd in the new version of OR.

 

[Andrew Brown]

Whether the altimeter reports different altitudes as a function of temperature or not, my understanding is that you are stuck with whatever it says: section 2.6 of https://rocketcontest.org/wp-conten...merican-Rocketry-Challenge-Rules-v5.16.22.pdf "The peak altitude of the rocket as recorded by this altimeter... post-flight will be the sole basis for judging the altitude score "

For scoring, this is absolutely true. However, the simulator models true altitude. The altimeter makes zero temp corrections and needs to be manual converted if you want to correlate to true altitude. If your goal it to correlate to simulation, just use the formula to correct the altimeter output. Again, not applicable for scoring. Trying to tweak Cd when your altimeter can vary as much as it does across temp is looking at the wrong cause. Further, trying to compare multiple launches that may have been at different temperatures, and trying to get within 15 ft (less than 2%), is a challenge if you don't adjust for temp.

 

[diyaerospace]

The altimeter makes zero temp corrections and needs to be manual converted if you want to correlate to true altitude.

Do you know how to manually temperature correct the altimeter data. It is worth noting that I put a Jolly Logic Altimeter 2 and a Firefly altimeter in the payload bay. Both fell within feet of each other.

I assume your reported masses are measured just before you go to the pad?

Yep!

If you are not using a rail then you can get a fair amount of variation from rod whip. I always recommend using a long stiff rail for TARC.

All of these flights used a rail.


Thanks for everyone's help,

Walter

 

[Andrew Brown]

Do you know how to manually temperature correct the altimeter data. It is worth noting that I put a Jolly Logic Altimeter 2 and a Firefly altimeter in the payload bay. Both fell within feet of each other.

Yep. In the TARC team handbook, there is a section on using the altimeter.

https://rocketcontest.org/wp-content/uploads/Handbook-for-ARC-2023-V23.2_compressed.pdf

These guys also had a good discussion in their presentation competion from a couple years ago:

https://www.nar.org/wp-content/uploads/2021/08/Northville-HS-1st-in-Presentation-2021.pdf

Neither of those altimeters temperature compensates, so it makes sense that they are consistent with each other. The NAR and FAI competitions require that you either use a temperature compensating altimeter, like the Adrel Altimeter, or you correct manually when reporting scores. If you work with any competition fliers, they can help you. TARC requires that you go off of the reading with no compensation. It basically means you have a moving target for actual altitude based on the temperature of the day.

Our kids employ the method of calculating how high the rocket needs to actually go to make the altimeter read the target value based on the temperature, then adjust mass to hit that actual altitude (not the altimeter target value). My kids are in Michigan, where the temp can vary from 10 degrees to 50 degrees this time of year. That basically means we need to be able to get it 770 ft at 10 degrees and 835 ft at 50 degrees, and a linear interpolation in between to make the altimeter read 850 ft.

 

[cerving]

Actually, the altimeters do temperature compensate (look up the data read algorithm for the MS5607 baro chip), however ALL of the digital baro chips have very slow temperature slew rates so it's pretty worthless for rocketry flights. But, everybody is on an even playing field (they specify that everybody has to use a PF altimeter), so it really doesn't matter... they're all going to be equally "wrong".

 

[mikec]

These guys also had a good discussion in their presentation competion from a couple years ago:

https://www.nar.org/wp-content/uploads/2021/08/Northville-HS-1st-in-Presentation-2021.pdf

Wow, that's an amazing report! I would not have guessed the PF altimeters had so much variation with temperature.

Also note that they found a 2-cal stable rocket had too much variation with wind and went with an intentionally less stable design.

 

[mikec]

Actually, the altimeters do temperature compensate (look up the data read algorithm for the MS5607 baro chip)

There are two flavors of temperature compensation: one is the effect of temperature on the pressure reading itself, which the sensor is trying to compensate for. But the other is the "standard atmosphere" equation, which is used to convert pressure deltas to altitude deltas. That's the thing that assumes 15C temperature. Altimeters could correct for that if they could get a reliable reading of the ambient temperature, but they are usually way off since they are internally mounted in a typically hot rocket. So most of them don't try.

See https://www.nar.org/wp-content/uploads/2016/07/NARCON-2017-Altimeter-Design-Wolf.pdf

 

[Steve Shannon]

Don’t get hung up on a simulation. If you’re within 100 feet in your sims, build the rocket and fly it a bunch of times. That is the only place tweaking means something.

 

[diyaerospace]

Thanks for the temperature compensation tip!

I adjusted all the previous flight data to account for temperature.

Then I plotted all the points and a line of best fit in DESMOS. From this model it looks like a mass of 411 grams will hit an altitude of 860ft (with temperature correction). Given the temperature of our next planned launch the altimeter should read 10ft low for an altitude of 850ft.

Desmos | Graphing Calculator

www.desmos.com

 

[diyaerospace]

Hello,
Here are all the flights I did after my last post.


Flight 9:
407 grams
844 ft
Payload Section Weight: 183
1:10 duration


* larger spill hole cut in payload section's parachute

Flight 10:
405 grams
869 ft
Payload section weight 185
0:48 duration

* larger spill hole cut in payload section's parachute

Flight 11:
412 grams
814ft
Payload section weight 196
0:36 duration

*new launch day

Flight 12:
408 grams
888 ft
payload section weight; 155
0:50


The first 3 flights seem to be headed in the right direction, after we got so close with flight 9 I shouldn't have adjusted the weight.

Flight 12 is an outlier as the altitude was far higher than expected. At this point I have a lot of flights to refine my desmos model so I don't know what caused it. Anyone have any suggestions?

Here is my process for finding optimum mass,
I first use my desmos model with a line of best fit to find the correct mass to get to 850 ft. In this case it was 407 grams.

I input the weather conditions of the flights in the desmos model and the predicted weather conditions at the launch site to find the difference. Rocksim predicts an 8ft increase in altitude so I bring the total weight to 811 grams. (1 grams has a 2ft altitude change)

At the launch site I measure the temperature and find if the altimeter will read high or low. In this case the altimeter would read 6ft low (55F) so I removed 3 grams.



Thanks for everyone's help,
Walter

 

[diyaerospace]

Anyone have any ideas?

 

[mikec]

Are you measuring the wind speed for each flight? Some of this variation could be weathercocking-related. Your rocket is more stable than it probably should be based on the presentation posted upthread.

 

[diyaerospace]

Are you measuring the wind speed for each flight?

not directly, I use data from close weather stations.
Flight 12 (which is the flight I am most concerned about) had very low wind speed.

 

[mikec]

not directly, I use data from close weather stations.
Flight 12 (which is the flight I am most concerned about) had very low wind speed.

If it was the highest and all the others had higher wind speed, then this could all be weathercocking. Or it could be motor variation, delay time variation, etc.

A cheap anemometer might be of use.

I'm not sure what your breakout of payload section mass and total rocket mass means. If you are moving the CG around depending on where you are putting ballast mass, that may also affect weathercocking.

 

[diyaerospace]

If you are moving the CG around depending on where you are putting ballast mass, that may also affect weathercocking.

The stability margin for flight 12 was 3 which is about 1/2 a margin lower than the previous flights. I tested the altitude difference in rocksim and found it to be less than one foot.

 

[mikec]

The stability margin for flight 12 was 3 which is about 1/2 a margin lower than the previous flights. I tested the altitude difference in rocksim and found it to be less than one foot.

You need to let go of the idea that what simulation tells you has any resemblance to reality at this level of detail.

Read all of https://www.nar.org/wp-content/uploads/2021/08/Northville-HS-1st-in-Presentation-2021.pdf

Original rocket designs were based on suggestions from TARC guides and
discussions with mentors suggesting a stability of 2 calibre or higher. These
rockets resulted in excessive altitude variations due to weathercocking.
● Some research and simulation models in OpenRocket made us realized that
over-stability was the primary cause.
● The more stable the rocket was, the more variation we simulated based on
wind variation.
● We decided to test newer designs in the 1.5-1.25 caliber range, due to the
decrease in wind variation exhibited.

 

[Steve Shannon]

Anyone have any ideas?

Stop simulating and continue building and flying.
The tolerances for motor certification are great enough that if you do get a perfect flight you had some luck.

 

[Richardx]

Time is running out but see how many of these you can do

Biggest problem to tackle is the consistency of your flight. If the rocket can not fly consistently, it will be almost impossible to hit the target. So focus on that first. Watch the lift off. Is it waggling or launching straight? Is the boost phase straight or wiggly? What is the velocity at liftoff on the simulation?
To correct inconsistent flights, you will need to look at stability margin, weight distribution, fin shape, thrust to weight ratio.
I have seen too many teams trying to tweak the rocket for altitude before having the necessary consistency.

Once you have consistent flights, you can tackle the altitude adjustment. A few common mistakes. Is the ejection delay long enough? I have seen many team's rocket ejection nosecone up. This means the altimeter inside the payload section is ejecting up to higher altitude at random height. Is the vent hole size for the altimeter too big. Large vent holes will produce inconsistent data. Lastly, launch minimum of 3 times with exact same configuration before making any changes. This will avoid reacting to bad weather or abnormal launch conditions.

I teach my teams that there are 3 phases
phase 1 - straight, stable, and consistent flight. do not worry about the altitude or time
phase 2 - target altitude and parachute time
phase 3 - consistency across different launch condition (we had to launch in rain last year at the National fly off)

Good luck!

 

[Andrew Brown]

The stability margin for flight 12 was 3 which is about 1/2 a margin lower than the previous flights. I tested the altitude difference in rocksim and found it to be less than one foot.

A stability margin of 2.5 to 3 is a lot for a competition rocket. Even for a sport rocket, it's a lot. It's fine in simulation as conditions are always perfect in sim. It's just solving the same equations and getting the same result. I don't use RockSim, but I believe it's pretty similar to OpenRocket. In OR, you could sweep wind conditions and see how it affects altitude. OR even has a feature where you can change turbulence and wind standard deviation to give more realistic variations, but as others have posted, sim can only guide you to trends and help identify issues. It's not absolute. TARC looks deceiving easy on the surface, but it's a tough competition with near equal parts engineering, practical knowledge, data analysis, minimization of variables, practice, and luck.

The 2021 winning presentation from Northville has been quoted a few times already, but I actually think their presentation from 2022 was better despite some lower flight scores. They really struggled that year and the presentation provides a lot of clues as to what can go wrong and still attempt to mitigate.

https://rocketcontest.org/wp-content/uploads/Northville-HS-2nd-Place-2022-Presentation-2.pdf_compressed.pdf

 

[Ez2cDave]

The first 3 flights seem to be headed in the right direction, after we got so close with flight 9 I shouldn't have adjusted the weight.

Flight 12 is an outlier as the altitude was far higher than expected. At this point I have a lot of flights to refine my desmos model so I don't know what caused it. Anyone have any suggestions?

Here is my process for finding optimum mass,
I first use my desmos model with a line of best fit to find the correct mass to get to 850 ft. In this case it was 407 grams.

I input the weather conditions of the flights in the desmos model and the predicted weather conditions at the launch site to find the difference. Rocksim predicts an 8ft increase in altitude so I bring the total weight to 811 grams. (1 grams has a 2ft altitude change)

At the launch site I measure the temperature and find if the altimeter will read high or low. In this case the altimeter would read 6ft low (55F) so I removed 3 grams.



Thanks for everyone's help,
Walter

Walter,

You need some "Real world" data, not "Simulated data" . . .

WHY ?

(1) The "sim" always uses a perfectly vertical flight path which, almost never, occurs in Reality.

(2) The "sim" uses an ideally-performing rocket motor, every time, which never occurs in Reality.

(3) The "sim" uses a "standard descent rate" ( no "oscillation" that causes parachutes to "spill" & increase the descent rate ), which never occurs in Reality.

(4) The "sim" uses a constant wind velocity ( or zero ) which, almost never, occurs in Reality.

"Real World"

( 1 ) Keep the MASS constant for EVERY flight, until you get a meaningful "Real World" database, for reference.

( 2 ) Pack the Recovery System the SAME way, EVERY time, for consistent deployment.

( 3 ) Always maintain and verify the Launch Angle, in degrees from vertical . . . Use a stiff rail or a tower, to eliminate "rod whip".

( 4 ) Keep a log of the "Production Numbers", found on each motor, record that and correlate it to flight performance, for every flight.

( 5 ) Once you have good baseline data, adjust mass to achieve the desired altitude "range" ( include any changes in Recovery System, too ). To decrease Altitude, INCREASE Mass or INCREASE Drag, without increasing mass ( "Drag Plates" are a good example of increasing Drag, without much effect on Mass ). Remember that changes in Mass will ALWAYS affect the Decent Rate !

( 6 ) VERY IMPORTANT - Only change ONE VARIABLE, at a time . . . Failure to do so will get you "lost in the data", causing you to "chase ghosts" and destroy consistency !

( 7 ) FINALLY - You will NEVER be able to overcome "variations" in Motor Performance, so watch your data closely and you will see "anomalies" caused by motor inconsistency, given that all other variables were unchanged.

Dave F.

 

[diyaerospace]

Thanks for the informative responses. I learned a lot.

One question I have is how to quantify the effects of temperature and elevation on the final altitude achieved. The Northville team minimized drag on their design to limit these variables but did not detail how to model them in their mass-altitude plot. Are temperature, elevation, and humidity effects low enough to ignore?

Thanks,
Walter