Dynamic stability analysis info

A thread on the arocket mailing list got me wondering about this topic since I have some strange (but still safe) behavior from one of my own rocket designs. I found these resources about it:

Basics of Dynamic Flight Analysis (Parts 1 through 6) - Apogee newsletters 192 - 198 (minus 194)

Experiments with Dynamic Stability Analysis - Apogee newsletter 253

https://www.apogeerockets.com/newsletter/newsletter_archive

Define strange behavior. Rocket details also useful as would be a sim file.


Mark Koelsch
Sent from my iPhone using Rocketry Forum

markkoelsch said:

Define strange behavior. Rocket details also useful as would be a sim file.

Click to expand...

Minor tail wag just prior to motor burnout in very light winds even though stability margin is more than adequate at all velocities according to a Rocksim static stability margin plot for flights. This test vehicle was involved:

https://www.rocketryforum.com/showt...R-1-2-altitude-of-Rocksim&p=709474#post709474

Used to research this problem, identical to the problem raised by the creator of the thread:

https://www.rocketryforum.com/showt...R-1-2-altitude-of-Rocksim&p=701621#post701621

The ratio of the width of the flat leading fin edges of the test vehicle to the airframe diameter is even greater in the sub-scale test vehicle than in the HPR version. I suspect that turbulent flow around the fin due to the flat leading edge is causing variable forces that are extremely sensitive to angle of attack, resulting in highly variable forward Cp movement not predicted by either Rocksim or Openrocket. However, my experiment did at least show that for rockets with significant flat leading edge fin area, Openrocket's altitude predictions are much more trustworthy than Rocksim's.

What everyone tends to forget is that if you do not have 3-D GC data, as built 3-D aerodynamic data, and 3-D thrust axis data on your rocket as built, it really doesn't matter how good the sim is, you can't predict how your rocket will behave under all but most modest profile flights.

Most sims will assume a 3-D axisymmetric distribution of weight, a defined aerodynamic profile and a perfectly aligned thrust axis. The as built and put on the pad rocket may not conform to those assumptions and would not behave exactly as predicted even if the sim had the capability to solve for it.

A significant deviation would be asymmetric fins not properly described and an accidental off-axis thrust alignment due for example to a slightly off-axis nozzle. This would cause the rocket to roll and pitch aka coning, but this behavior would not be modeled, because the data was not input into the model, and the model may not have the proper equations to provide the simulation even if you had input the data.

Really guys, don't expect more out of any hobby sim than you can get from the data you input, or can input, to it.

Bob

bobkrech said:

What everyone tends to forget is that if you do not have 3-D GC data, as built 3-D aerodynamic data, and 3-D thrust axis data on your rocket as built, it really doesn't matter how good the sim is, you can't predict how your rocket will behave under all but most modest profile flights.

Most sims will assume a 3-D axisymmetric distribution of weight, a defined aerodynamic profile and a perfectly aligned thrust axis. The as built and put on the pad rocket may not conform to those assumptions and would not behave exactly as predicted even if the sim had the capability to solve for it.

A significant deviation would be asymmetric fins not properly described and an accidental off-axis thrust alignment due for example to a slightly off-axis nozzle. This would cause the rocket to roll and pitch aka coning, but this behavior would not be modeled, because the data was not input into the model, and the model may not have the proper equations to provide the simulation even if you had input the data.

Really guys, don't expect more out of any hobby sim than you can get from the data you input, or can input, to it.
Bob

Click to expand...

I'm fully aware that sims aren't very sophisticated and make a lot of assumptions that aren't necessarily true. However, what I've experienced with two different motors at the same point in flight, near max v, along with the large flat leading edge area of the fins of the test vehicle leads me to suspect that there is a reduction in fin effectiveness caused by the turbulence caused by the flat leading edges, turbulence effects that are definitely not simulated by the simulation software. I'm investigating dynamic stability topics to follow Sherlock Holme's advice - "Eliminate all other factors, and the one which remains must be the truth."

Without doubt the flat leading edge is the worse possible shape for a fin as it most likely generates the most drag and the least lift with is exactly counter to what you want from a small fin.

I would be using RASAero to investigate this effect since there is a lot of high speed aerodynamics embedded into that simulator. You will find it interesting to read the technical papers on the Rogers Aerospace website.

Bob

I've been playing around with RASAreo lately. Expect around 30% or more increase in altitude for even a very basic bevel.

bobkrech said:

Without doubt the flat leading edge is the worse possible shape for a fin as it most likely generates the most drag and the least lift with is exactly counter to what you want from a small fin.

I would be using RASAero to investigate this effect since there is a lot of high speed aerodynamics embedded into that simulator. You will find it interesting to read the technical papers on the Rogers Aerospace website.

Bob

Click to expand...

It's been an interesting path to this point. I intentionally created a high-drag HPR design to reduce altitude with a given motor, a design shown in this post:

https://www.rocketryforum.com/showt...R-1-2-altitude-of-Rocksim&p=701621#post701621

I suspected that a flat leading edge, besides providing the desired high drag, would have a negative effect on fin effectiveness, but didn't sim the design in both Rocksim and Openrocket until after the rocket was built, initially simulating it only in Rocksim. When I finally did and saw the huge, apparently max-velocity-dependent difference in altitude between the two simulators, I had to experiment to determine which produced a more accurate simulation and that turned out to be Openrocket as implied by this fight test of a sub-scale version of the HPR design:

https://www.rocketryforum.com/showt...R-1-2-altitude-of-Rocksim&p=709474#post709474

Not a definitive test considering the limited number of flights, but enough to give me much more confidence in Openrocket than Rocksim predictions for that particular rocket design with its six, square leading edge fins. It was during those two flights that the extent of what I assume to be the flat plate leading edge turbulence effect on fin effectiveness was seen near max velocity.

BTW, I didn't design the high-drag rocket to use tube fins because I've already built custom 24mm and 29mm-powered designs and because as shown in the excellent multi-part series in Sport Rocketry on tube fin rockets by Larry Brand, the bizarre Cd curve of tube fin rockets is highly velocity and specific design dependent and even though I've had great results with accurate predictions from Rocksim with 24mm and 29mm tube fin flights, I wasn't confident that I'd get the same with HPR level 1 or, especially, level 2 flights. I suspect that at high velocities the Cp of tube fin rockets may vary in as strange a fashion as their Cd, but that's just a hunch.

I've already tried using RASAero but, unfortunately, it only allows the simulation of three or four fin rockets while my design has six.

Because the ratio of flat leading edge area to total fin area is significantly lower in the HPR version of the design (I wanted the sub-scale version to exaggerate the flat leading edge effect to see which simulation software was best at predicting altitude with flat leading edge fins), the fin effectiveness reduction will probably be significantly less with the full-scale HPR version. Now that I know what delay length to use from the results of the sub-scale simulator prediction accuracy test, the real point of that test, the HPR version will be flown on an low-H motor and I'll be watching closely for any tail wag.