Simultion result differences between OpenRocket and RockSim

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mpitfield

Well-Known Member
TRF Supporter
Joined
Jul 6, 2013
Messages
4,902
Reaction score
460
Location
Toronto, Ontario
I just installed the latest version of OR, imported and converted a rocket from RockSim and ran the same simulation.

In OR I get 24,800 @ Mach 1.91 and in RS I get 20,300 @ Mach 1.81. Anyone have an opinion as to which sim is more likely to be closer to accurate, and why the big difference?
 
Not a worthwhile question. The motor you use will be subject to variations that exceed the difference.
 
I've had similar problems except with more variation. I've had a rocket with RockSim sim to 14000 @ mach 2.4 and in OR to 24000 @ mach 2.5
 
Not a worthwhile question. The motor you use will be subject to variations that exceed the difference.

Hi Peter, I am trying to understand your answer. Are you suggesting that due to motor the variations, it is not a worthwhile exercise in simulating? Is this just in general or this particular motor?

I realize these simulators are more of a best guess for dummies and don't not take everything into account and or every variation, so I generally take them with a grain of salt. However my personal experience has been that my RockSimulations are generally pretty accurate, in both max velocity and AGL. Now to put this into perspective I have only been at it for just under two years, so my experience and sample data is limited, possibly giving me a false sense of confidence.
 
Hi Peter, I am trying to understand your answer. Are you suggesting that due to motor the variations, it is not a worthwhile exercise in simulating? Is this just in general or this particular motor?

I realize these simulators are more of a best guess for dummies and don't not take everything into account and or every variation, so I generally take them with a grain of salt. However my personal experience has been that my RockSimulations are generally pretty accurate, in both max velocity and AGL. Now to put this into perspective I have only been at it for just under two years, so my experience and sample data is limited, possibly giving me a false sense of confidence.

Peter is right when comparing your sim to the real flight - the errors in the motor thrust data or the performance variation of the motor itself will dominate any other parameters.

As for this particular OR vs. RS discrepancy, plot drag coefficient vs. Mach in each software, and that will reveal the answer.
 
Does importing the model also import the simulation parameters? Things like the altitude of the launch field, simulation time step, etc. could be affecting the sim as well.
 
RockSim is not suitable for flights over Mach 2.

Thanks for the tip. This is not the first time I have heard that. However first I need to break Mach 1, so if and when I ever attempt a launch over Mach 2 I will keep that in mind

Peter is right when comparing your sim to the real flight - the errors in the motor thrust data or the performance variation of the motor itself will dominate any other parameters.

As for this particular OR vs. RS discrepancy, plot drag coefficient vs. Mach in each software, and that will reveal the answer.

Thank you Buckeye this is feedback I can understand, or at least act on. I ran both scenarios and I can visually see differences, however only from a high level as I do not understand the implications at a technical level.

Output from RS

View attachment 260892

Output from OR

View attachment 260891


Does importing the model also import the simulation parameters? Things like the altitude of the launch field, simulation time step, etc. could be affecting the sim as well.

No it seems to strip off the simulation data and just import the basic rocket. Interestingly none of these, within a reasonable range, has that much of an effect on the simulations. At least not on the scale of what each sim is producing.
 
Last edited:
Hi Robert, out of curiosity and not that it will reveal much, however did you launch based on the sim and if so which one was closer?

The sim is of a 38MD I plan on building in the future, but I haven't gotten the time to test and see the difference.
 
Thank you Buckeye this is feedback I can understand, or at least act on. I ran both scenarios and I can visually see differences, however only from a high level as I do not understand the implications at a technical level.

There you go. See the huge Cd jump after Mach=1 in Rocksim? It is adding about 50% more drag to the rocket than OR. That is causing the bulk of your 4000' discrepancy. I see this all the time. The longer your rocket is supersonic, the lower the altitude from Rocksim. Which one is correct? I don't know. Somebody more versed in rocket aerodynamics can answer that.

Also, you can export both sets of data to a text or csv file and plot them together in Excel for easier viewing.
 
There you go. See the huge Cd jump after Mach=1 in Rocksim? It is adding about 50% more drag to the rocket than OR. That is causing the bulk of your 4000' discrepancy. I see this all the time. The longer your rocket is supersonic, the lower the altitude from Rocksim. Which one is correct? I don't know. Somebody more versed in rocket aerodynamics can answer that.

Also, you can export both sets of data to a text or csv file and plot them together in Excel for easier viewing.

I just input my rocket into RASAero, a very no-nonsense application, I like the simplicity. Whether it is any better is apparently a matter of opinion but I have run the sim now in OR, RS and RASAero.

I tired to upload the CSV from RASAero however TRF does not like the CSV format so I converted it to an XLSX TRF didn't like the .8MB size for that format, so here is a screenshot of the simulation output with both the max AGL and high Mach number highlighted in the raw data @ 3.7 sec

View attachment 260935

Comparative numbers from all three simulators

OpenRocket 24,800 @ Mach 1.91
RockSim 20,300 @ Mach 1.81
RASAero 26,000 @ Mach 1.82

So final conclusion, there is no consensus among the simulators, is it reasonable to average out the results and say 23,700 @ Mach 1.6?

One more pic, this is the CD vs Mach output for RASAero.

View attachment 260936
 
Last edited:
Does importing the model also import the simulation parameters? Things like the altitude of the launch field, simulation time step, etc. could be affecting the sim as well.

OK, we are assuming you input everything identically into each of the 3 software - mass, weather, launch site, finish, etc. Right?

You now have 30% discrepancy with RASAero in the mix. Cd alone isn't the culprit, I now think. There is probably something else amiss. Check the motor file used by each software. Are the thrust vs. time points the same? The are multiple versions of every motor file floating around the inter-webs. You may have 3 different versions of the same motor.
 
OK, we are assuming you input everything identically into each of the 3 software - mass, weather, launch site, finish, etc. Right?

You now have 30% discrepancy with RASAero in the mix. Cd alone isn't the culprit, I now think. There is probably something else amiss. Check the motor file used by each software. Are the thrust vs. time points the same? The are multiple versions of every motor file floating around the inter-webs. You may have 3 different versions of the same motor.

I used the most recent (April 4th 2015) motor files from ThurstCurve https://www.thrustcurve.org/motorsearch.jsp?id=1085. With RockSim I used the .eng file and OpenRocket and RASAero I used the RASP .eng file. I will look at the data from each file and see, with my untrained eye, if there are any major discrepancies.

As far as conditions, yes my conditions are all very similar or exact where they can be. With RockSim you have the most parameters for launch conditions which I have at higher than expected values, except temperature as temp seems to make my sim AGL higher. Within the range of conditions that I would launch, from no wind to breezy 8-15Mph winds, consistent to very turbulent wind, as well as low strength to high strength multiple thermals, the sims vary but nowhere near as great than between the simulators.

I need to check the weight values I input into RASAero as I may be off by 990gr due to my interpretation of the lift off weight and motor weight, this will no doubt have an effect on the outcome, which I will post.
 
I will say this. Apogee has admitted RockSim's drag algorithms are inaccurate in reference to nose cones. It imposes less drag on stubby nose cones and more drag on long conical or ogive nose cones.

I'm actually having the reverse issue. RockSim is giving me better numbers than OpenRocket.
 
I will say this. Apogee has admitted RockSim's drag algorithms are inaccurate in reference to nose cones. It imposes less drag on stubby nose cones and more drag on long conical or ogive nose cones.

I'm actually having the reverse issue. RockSim is giving me better numbers than OpenRocket.

That would account for some of the difference. The 54mm carbon fiber Tomach uses a 6:1 Sears-Haack (Von Karman) nosecone.
 
I used the most recent (April 4th 2015) motor files from ThurstCurve https://www.thrustcurve.org/motorsearch.jsp?id=1085. With RockSim I used the .eng file and OpenRocket and RASAero I used the RASP .eng file. I will look at the data from each file and see, with my untrained eye, if there are any major discrepancies.

As far as conditions, yes my conditions are all very similar or exact where they can be. With RockSim you have the most parameters for launch conditions which I have at higher than expected values, except temperature as temp seems to make my sim AGL higher. Within the range of conditions that I would launch, from no wind to breezy 8-15Mph winds, consistent to very turbulent wind, as well as low strength to high strength multiple thermals, the sims vary but nowhere near as great than between the simulators.

I need to check the weight values I input into RASAero as I may be off by 990gr due to my interpretation of the lift off weight and motor weight, this will no doubt have an effect on the outcome, which I will post.

Yep, both motor files are the same. Impulse, ave thrust, and profile are identical. No worries, there.

Temperature, barometric pressure, and launch site altitude are the most critical site parameters. Launch angle more so than wind. Humidity and latitude don't have much impact in the simulation in my experience.

990g seems like a lot. That's an entire HPR rocket for me! You only need to input "dry weight" of the rocket (no motor). The simulator will add the mass of the motor (and casing) from the motor file when you choose it.

Motor, mass, drag coefficient are the big hitters, in that order.
 
The problem with these 3 software is that the drag models are hard coded by the developers. The user can't override them (other than Rocksim's constant Cd). So, the user is relegated to futzing with surface finish and square edges to generate variation.

Since you seem willing to dig into the data, here is another suggestion, which introduces yet another software. Download the oldie but goodie wRASP32. It is also a simple interface, that I like. It has the nice feature of allowing you to input your own Cd curve. So, export Cd vs. Mach from RS, OR, and RA. Then input these curves one at a time into wRASP32, keeping all else constant. This way, you can isolate the impact of the 3 different aerodynamic drag models.
 
I saying the difference between the two programs calculations is less than the variations to be expected from one motor to the next of the same batch. Sim programs only get you in the general vicinity. Using a recording altimeter will enable you to find the differences between the details of your rocket and the details in the simulation. Then go back to the simulation and make the tweaks that give observed results if you really need to. You will find, however, that the next flight with the same motor will produce different results and you will end up chasing your tail. Never forget: "The perfect is the enemy of the good."

Hi Peter, I am trying to understand your answer. Are you suggesting that due to motor the variations, it is not a worthwhile exercise in simulating? Is this just in general or this particular motor?

I realize these simulators are more of a best guess for dummies and don't not take everything into account and or every variation, so I generally take them with a grain of salt. However my personal experience has been that my RockSimulations are generally pretty accurate, in both max velocity and AGL. Now to put this into perspective I have only been at it for just under two years, so my experience and sample data is limited, possibly giving me a false sense of confidence.
 
The comment about Rocksim's handling of nosecones is correct. What works fairly well as a workaround (tested in an accurate sim file versus flight data) for a Von Karman is to change the sims nosecone to an actual cone of the same mass and length.

Apogee should fix their nosecone issue. It has been an issue forever.

That said, both OR and RS are based on Barrowman. Barrowman has several assumptions, and one that is recall is that you need to be subsonic.
 
Apogee should fix their nosecone issue. It has been an issue forever.

Agreed. Look at the goofy spike in Cd in post #9. Any supersonic flight is suspect is RS. This has been an issue forever. You would think that a commercial (not free) software would finally fix this fundamental problem that is identified time after time by its users.
 
Agreed. Look at the goofy spike in Cd in post #9. Any supersonic flight is suspect is RS. This has been an issue forever. You would think that a commercial (not free) software would finally fix this fundamental problem that is identified time after time by its users.

It is also equally suspect in any sim based on Barrowman. This would include OR as well. You have to understand the limitations and go with it. I have good confidence in RS to about Mach 1.3-1.4. This is only because I have been able to quantify simulation to real world data.

Now, what is better is Rocksim Pro. I used it as a demo for a while, and it is far more capable. What is cool is that you can input your own aero info. So, say you have information for Cd ( or other aero info)from Rasaero or a CFD program you can input that into RS Pro. I wish it did not cost what it does- if it were in the 150-200 dollar range I would snag it in a heart beat.
 
Back
Top