Rocksim success

[RodRocket]

How successful are the actual flights compared to the sim files? If I build a rocket in Rocksim, how close will the actual flight be? My weather features seem to be shut off for the trial version. And that may change my projects completely when I start messing with wind and such.

I will be purchasing the full version soon. Just got to wondering how close the sim really is.

See ya,
Rod

 

[rdbones]

How successful are the actual flights compared to the sim files? If I build a rocket in Rocksim, how close will the actual flight be? My weather features seem to be shut off for the trial version. And that may change my projects completely when I start messing with wind and such
See ya,
Rod

I have seen quite a wide range of accuracy with Roc-Sim. It mainly deals with the performance of the rocket. I did a MD54 scratch built and the estimated altitude was off by about 23%, while a friend of mine did a scratch built Phoenix and the estimated altitude was off by about 10 feet.

I use it mainly for design, I look at the simulations to get a SWAG and then I fly the rocket to get an accurate altitude and then I adjust the CD to accurate estimations.

 

[troj]

RockSim, in my experience, is rather optmistic -- most of the time, your rocket will under-perform, relative to RockSim.

After you get a couple flights, you can then start tweaking parameters such as the Cd (drag coefficient) based on actual flights, and get the numbers more accurate.

-Kevin

 

[Buckeye]

The problem with tweaking the Cd after the fact is that now the constant Cd becomes the catch-all for every variation that took place during flight - weather, winds, motor thrust, etc. Also, backtracked Cd on one motor doesn't always translate to a different motor, especially if one is Mach and the other is not.

My recent RockSim simulations are low compared to my altimeter. Rare.

https://www.rocketryforum.com/showthread.php?t=24730

I used to fret about matching sims and altimeter altitude data. I gave up. I think thrust variation is the biggest culprit, both in the eng files as well as in real life.

I now use the flight sims to get a feel for Mach delay, ejection delay, flight time, and approximate altitude. That's about all you can ask for.

 

[MarkII]

Everything that has been posted so far about RockSim's estimates are true in my experience. But much of the discrepancy may be due to inexact estimates of mass, inaccurate finish settings (which are themselves rather general), imprecise or inaccurate launch site and conditions settings (this is also difficult to match with the exact conditions at the time of your launch, even if you knew all of them) and the presence of some design elements that may be difficult to model accurately (not as big of a problem as the others, but still a nagging one in a small number of designs).

With all of that being said, the estimates can often be surprisingly close when the program is given appropriate and accurate data to work with. My pre-flight simulations seemed to agree almost to the inch with my Javelin XL's flight for my Level 1 certification. I didn't have an altimeter on board to verify this, but it sure looked right and the time intervals and the landing data were all absolutely dead on. It was the maiden flight for that rocket, which I had only completed the night before and had only attached the rail buttons an hour beforehand, but it went exactly as I had expected. Not all of the sims nail the real thing this precisely, but time and time again they have been close enough for me to regard the program as an incredibly useful tool.

 

[plano-doug]

RockSim, in my experience, is rather optmistic -- most of the time, your rocket will under-perform, relative to RockSim.

I'd say the number one source of error is the users failing to enter the actual, measured rocket weight into the program.

(They should also move the weight around to get the sim's CG properly located, too.)

Doug

.

 

[FastCargo]

I'd say the number one source of error is the users failing to enter the actual, measured rocket weight into the program.

(They should also move the weight around to get the sim's CG properly located, too.)

Doug

.

I would say Openrocket has this same limitation. I've noticed it gets quite accurate when I put in the actual mass of the rocket..

FC

 

[RodRocket]

Thanks for that tip Doug. Guess to do it absolutely right one would enter each part weight as they built the rocket in the sim. Then adjust the total weight after building the actual rocket, to account for glue, paint and difference in materials.

 

[gpoehlein]

What the rest of them said.

I stopped worrying about the altitude Rocsim says a rocket will achieve. Mostly, I'm looking more for stability margin and ejection velocity (have I gotten the delay right?). The only time I've been concerned with altitude in Rocsim was to tweak a design for LESS altitude! (To explain, I've been workin on my design for my Level 1 cert rocket, and I want the rocket to fly low and slow to maximize my chances for a successful cert flight.) my experience has been like most - most of the time Rocksim is very optimistic on its altitude prediction.

 

[plano-doug]

Thanks for that tip Doug. Guess to do it absolutely right one would enter each part weight as they built the rocket in the sim. Then adjust the total weight after building the actual rocket, to account for glue, paint and difference in materials.

I've weighed the individual parts before, but I usually just try to get the final weight adjusted using a single mass object, then move that weight around until the CG matches the finger balance point. It's a lot less tedious that way

What I'm usually looking for is the optimal delay number, trying to minimize the stress on the recovery system. Still, if you get one out in a cluster or some weather cocking on your stager, there can be quite a jerk on the shock cord, but at least I started with the best delay...

Doug

.

 

[John Beans]

Rocsim is a super-useful simulator, and I find it an essential tool for gauging motor delays and stability in custom airframes.

That having been said, real life has gusting winds and chaotic movements and unexpected rotations. And while the total impulse of a most engines is probably accurate to <5%, the mix of thrust and burn time that it produces is not nearly so (20%?). So a particular engine which "burns quickly" and produces a lot of initial thrust may not go as high as another "identical" one that produces a lower peak speed and doesn't waste as much energy on air resistance.

Take a set of three motors in a pack, fly the same rocket three times, and observe different flight paths, speeds, and apogees each time. Which of the the three is the "right" answer?

 

[RodRocket]

I guess it is like the weather, you can predict all you want but no one will know what actually happens until it happens.

Thanks for the tips guys, got a lot of learning to do.

 

[bradycros]

I guess it is like the weather, you can predict all you want but no one will know what actually happens until it happens.

Thanks for the tips guys, got a lot of learning to do.

Maybe someone knowledgeable in RocSim and Open Rocket could make a sticky post(s) here in Rocketry Electronics and Software to help others learn how things are done, eh?

 

[UPscaler]

My darkstar simulated to 1501 feet on an H123 actual altitude was 1498.

My little John simulated to 6,450 feet, actual altitude was 4,000 something.



It varies

Braden

 

[Zephhyr]

My 3 fin stretched Nike Smoke scratch build on a J420 sims to 4399, at the launch yesterday it went 4591 integrated altitude, (showed 4966 max barometric altitude) according to my G-Wiz HCX.

 

[MarkII]

My darkstar simulated to 1501 feet on an H123 actual altitude was 1498.

My little John simulated to 6,450 feet, actual altitude was 4,000 something.



It varies

Braden

One day it's off by 0.002%.
Another day it's off by 38%.

Sounds like one rocket weathercocked a bit. My own overall experience is that it is a bit closer to the first number than to the second one. Being off by more than a third is rare. Once you factor in the inability to precisely predict the launch conditions, it gets reasonably close, and every once in awhile, it actually nails the flight so well that it's scary. But as with any algorithm, garbage inputs lead to garbage outputs. It can't fix bad data.

 

[RodRocket]

Next pay day, I am going to find a small scale, smaller than a typical postal type scale to start weighing parts as I build a rocksim rocket.

And maybe to see what happens, build the same rocket using the rocsim numbers and see what the difference will be. You think the programmers went through this process when they added the parts and pieces into rocsim? And maybe came up with an average weight for a particular part.

Been working so much with rocsim that I am getting behind with building. Oh well, it's too darn cold to paint anyways.

See ya,
Rod

 

[Fe Dude]

I always wind up swiping my buddy's scale before a launch to double check the weight and re-run all of my sims (mostly to get a general idea of what delay to use in an ideal situation).

 

[GDJ]

On my serious builds I end up weighing every piece and enter that weight into OR (Rocksim would probably be similar) and I find that gets the sims fairly close to reality. There will always be a variance, but if one tries to eliminate the variables as much as humanly possible, then the results will be better.

As always, YMMV.....