Rocksim and Rasaero were extremely...conservative?

[Buckeye]

Here is a first for me. The simulations were well under the altimeter measurement. 3FNC Loc Vulcanite on an Aerotech H128 with Perfectflight MAWD:

Altimeter = 1939'
Rocksim (best scenario) = Rasaero (best scenario) ~ 1775'

Best scenario settings include "polished" finish (hardly), no turbulent flow in Rasaero, and some tweaking of rod angle (eyeballed at first) to get max altitude for the given wind speed (measured at ground with my anomometer). Mass, components, and other launch site conditions were faithfully recorded and simulated. Well below Mach. The latest eng and rse files from the Aerotech website were used.

Even under best scenario simulations, I can't achieve what the altimeter reports. Using a fixed Cd in Rocksim requires Cd ~0.4 to achieve 1939'. Not likely for a slender rocket with my finishing skills.

I then dug into the motor files. I went to Thrustcurve.org and noticed that the H128 has two versions of simulator files, one with more impluse than the other. I plugged in the hotter eng file into Rasaero, and voila, the simulation now says ~2050', which seems like a more realistic correlation.

My conclusion: Rocketeers fret way too much over their simulation settings while taking the accuracy of thrust curves for granted. Motor files and variation are all over the place. Some published thrust curves have T=0 at t=0, while some start impulsively at t=0. This inconsistency also makes a difference in the simulation. Put standard deviation bounds on the motors and hope for the best in your sims.

 

[Handeman]

I agree, then though in the published variability of the motors from the thrust curves and you would be hard pressed to get a sim very close to an actual flight.

 

[Buckeye]

Exactly right. It would be nice to see the motor thrust variation published in the eng files and then have the simulation software work with it in some manner. Using mean or nominal motor data in the flight simulations gets me in the ballpark for altitude, time to apogee, mach delay, and ejection delay. That's about all you can ask for.

 

[Chuck Rogers]

Here is a first for me. The simulations were well under the altimeter measurement. 3FNC Loc Vulcanite on an Aerotech H128 with Perfectflight MAWD:

Altimeter = 1939'
Rocksim (best scenario) = Rasaero (best scenario) ~ 1775'

Best scenario settings include "polished" finish (hardly), no turbulent flow in Rasaero, and some tweaking of rod angle (eyeballed at first) to get max altitude for the given wind speed (measured at ground with my anomometer). Mass, components, and other launch site conditions were faithfully recorded and simulated. Well below Mach. The latest eng and rse files from the Aerotech website were used.

Even under best scenario simulations, I can't achieve what the altimeter reports. Using a fixed Cd in Rocksim requires Cd ~0.4 to achieve 1939'. Not likely for a slender rocket with my finishing skills.

I then dug into the motor files. I went to Thrustcurve.org and noticed that the H128 has two versions of simulator files, one with more impluse than the other. I plugged in the hotter eng file into Rasaero, and voila, the simulation now says ~2050', which seems like a more realistic correlation.

My conclusion: Rocketeers fret way too much over their simulation settings while taking the accuracy of thrust curves for granted. Motor files and variation are all over the place. Some published thrust curves have T=0 at t=0, while some start impulsively at t=0. This inconsistency also makes a difference in the simulation. Put standard deviation bounds on the motors and hope for the best in your sims.

Actually, the mistake was that there are two sets of Aerotech H128W motor data in the RASAero rasp motor file. When you run RASAero, pull up H128W on the motor input screen and you'll see two H128W (AT). We had inadvertently left in old data from the original development version of RASAero (the 1.5 sec burn 155.8 NS version) which should have been replaced with the updated/more accurate 1.3 sec burn 177.8 NS version, which closely matches the 172.9 NS from the TMT testing. The "hotter" motor you added from the thrustcurve.org web site was already in RASAero, it was the second H128W.

So the old data should have been deleted, instead they were both included. The one that should have been left in was the "hotter" engine data, the 1.3 sec burn 177.8 NS version, that produced an altitude prediction of approximately 2050 ft, a 5.7% error.

Generally RASAero predicts approximately 40% of rockets to within +/-5%, and approximately 80% of rockets to within +/-10%. (Detailed comparison are on the Comparisons with Altitude Data page on the RASAero web site.) Variations in motor data from the motor data used by RASAero undoubtedly is a contributing factor in some of the altitude prediction errors drifting up from +/-5% to 5-10%. A bigger factor I believe is getting an accurate lift-off weight, measured just prior to taking the rocket out to the launcher. The consistent +/-5% altitude comparisons we get are typically when an accurate lift-off weight has really been nailed down.

Accuracy of motor data used for flight simulations is certainly an important issue. But this particular example was just a basic mistake on our part, we left in old data which should have been deleted. If we hadn't made this mistake (left in both sets of motor data), or if you'd picked the second H128W motor then the initial altitude comparison would have had a RASAero predicted altitude of approximately 2050 ft, a 5.7% error.


Chuck Rogers
Rogers Aeroscience

 

[Buckeye]

Thanks for the response, Chuck. I did see the two H128s in RASAero at first. However, I replaced the entire motor database with this file from the Aerotech website, assuming it was more current and "official:"

Aerotech 3-24-10A.eng

I did the same with the rocksim database, Aerotech 3-24-10.rse

Both of these files contain only one H128, the lower one with 155.8 N-s.

I then found the 177.8 N-s version on Thrustcurve:

https://www.thrustcurve.org/motorsearch.jsp?id=131

There are 4 simulator formats, but they are not the same. The rocksim format is the lower impulse motor, while the ALT4, rasp, and compuroc files are the higher impulse motor. Confusion abounds.

I agree on the mass input. Just a few grams here and there (wadding, BP, ematches, motor adapter tubes, etc) can make an appreciable difference in the simulation.

 

[Chuck Rogers]

Thanks for the response, Chuck. I did see the two H128s in RASAero at first. However, I replaced the entire motor database with this file from the Aerotech website, assuming it was more current and "official:"

Aerotech 3-24-10A.eng

I did the same with the rocksim database, Aerotech 3-24-10.rse

Both of these files contain only one H128, the lower one with 155.8 N-s.

I then found the 177.8 N-s version on Thrustcurve:

https://www.thrustcurve.org/motorsearch.jsp?id=131

There are 4 simulator formats, but they are not the same. The rocksim format is the lower impulse motor, while the ALT4, rasp, and compuroc files are the higher impulse motor. Confusion abounds.

I agree on the mass input. Just a few grams here and there (wadding, BP, ematches, motor adapter tubes, etc) can make an appreciable difference in the simulation.

We have really tried to distill all of the of the various motor data files down to what we consider the "best", most current motor data for each motor. The RASAero user doesn't have to sort through all this, when he chooses a motor from our motor menu, he gets (although we made an error in this case) what we consider to be the "best", most current motor data for that particular motor.

Yes, mass (lift-off weight) does matter. I consistently ask in requests for altitude comparison data for the rocket to have been weighed just prior to launch, to get the actual measured (not estimated) lift-off weight.


Chuck Rogers
Rogers Aeroscience

 

[Buckeye]

Well, here we go again. Same LOC Vulcanite, same MAWD, now with CTI I216

altimeter = 7677'
RASAero = 6400' to 6900', depending on how much I fiddle with finish and turbulence model.

Even when I tried using RockSim's much maligned Cd calculation with the best possible finish, the max altitude is still only 7400 ft.

I am probbably the only guy on this forum complaining that the simulations are too low!

For those of you playing at home, here are the details:

mass = 966 g
dia = 57.4 mm
rod angle = 0 deg
temp = 80 F
elevation = 1033 ft
pressure = 29.5 in Hg

So, either there is significant motor thrust variation (again), nobody has a good subsonic/transonic Cd model, or my altimeter is wacked.

 

[Chuck Rogers]

mass = 966 g
dia = 57.4 mm
rod angle = 0 deg
temp = 80 F
elevation = 1033 ft
pressure = 29.5 in Hg

In addition to the data above, for threads like this I think it would be a good idea to also note the max Mach number of the flight, so the readers can get an idea of what Mach number range we're looking at for the rocket.

Generally we recommend that for the most accurate altitude predictions from RASAero the rockets should be run with an Equivalent Sand Roughness (Surface Finish) of Smooth (Zero Roughness), and that the All Turbulent Flow Box Not Be Checked (the default laminar, transition to turbulent, turbulent flow). I'm assuming that with the above selections the RASAero predicted altitude is the 6900 ft you referenced.

Compared to the 7677 ft from the altimeter, the RASAero prediction of 6900 ft has an error of -10.12%.

Again, from the Altitude Prediction Comparison Table on the Comparisons with Altitude Data Page on the RASAero web site, RASAero predicts approximately 40% of rockets to within +/-5%, and approximately 80% of rockets to within +/-10%. This error is a little high, but it's just slightly over 10%.


Chuck Rogers
Rogers Aeroscience

 

[Buckeye]

Chuck:

With smooth surface and turbulence unchecked, the altitude from RASAero is 6867 ft. Max Mach number = 0.93. I did a quick curve fit and derivative of the altimeter altitude data and got a similar max velocity of 1110 ft/s.

So, transonic regime?

 

[JDcluster]

How much does this Vulcanite weigh anyhow?
It's hard to get good numbers without good data to start with.
Garbage in= Garbage out routine applies.


Way back when I first started flying H's and above I flew a stretched Callisto that is very similar to the Vulcanite and only got 1400 ft ( if I remember correctly).


JD

 

[Buckeye]

Well, here we go again. Same LOC Vulcanite, same MAWD, now with CTI I216

mass = 966 g
dia = 57.4 mm
rod angle = 0 deg
temp = 80 F
elevation = 1033 ft
pressure = 29.5 in Hg

JD: See above.

 

[Chuck Rogers]

With smooth surface and turbulence unchecked, the altitude from RASAero is 6867 ft. Max Mach number = 0.93. I did a quick curve fit and derivative of the altimeter altitude data and got a similar max velocity of 1110 ft/s.

So, transonic regime?

Barely. In RASAero the subsonic aerodynamic prediction models are run up to Mach 0.90. The Transonic models start at just over Mach 0.90 and end at just under Mach 1.05. The supersonic aerodynamic prediction models start at Mach 1.05. Although the vertical cylinders of the rail guides begin to get a transonic drag rise at Mach 0.80, you can see the drag rise on the CD plot for the Caliber ISP Example Run on the RASAero web site.


Chuck Rogers
Rogers Aeroscience

 

[Rocketjunkie]

I would like to know how to add motors to the database such as my EX motors.

 

[Chuck Rogers]

I would like to know how to add motors to the database such as my EX motors.

RASAero uses the commonly used rasp motor data file format. The motor data file can be found in the /My Documents/RASAero directory. The file name is rasp.eng.

You create the motor data in the rasp file format, and then in WordPad or NotePad append the data to the rasp.eng file. It will then show up in the motor list in RASAero.

Before you edit the rasp.eng file, be sure to save a copy in case you've messed up the file. Just archive the first version and subsequent versions as you modify the file.

The ThrustCurve.org web site;
https://www.thrustcurve.org/

has information on the rasp motor file format at:
https://www.thrustcurve.org/raspformat.shtml

and has extensive resources and applets on creating and manipulating rasp format motor files. The RockSim software has a motor editor, and you can save the motor data in the rasp file format (NOT the RockSim format), and then append the data to the RASAero rasp.eng file.

We skipped adding a motor editor (being focused on developing very accurate aerodynamic prediction models) because most of our users use commercial motors (we have an extensive motor database built-in), and we figured the EX/Tripoli Research users already have worked with and are familiar with the rasp motor file format.


Chuck Rogers
Rogers Aeroscience

 

[Rocketjunkie]

RASAero uses the commonly used rasp motor data file format. The RockSim software has a motor editor, and you can save the motor data in the rasp file format (NOT the RockSim format), and then append the data to the RASAero rasp.eng file.

I have them in a RockSim .eng file. I'll resave the file in rasp format. Thanks for the info.