Optimizing Fin Shapes in Rocksim

[Anonymous]

I'm currently delving into optimizing fin planform shapes and their geometric dimensions for a water rocket project highschool project. to obtain maximum stability (which I would determine by the static margin) and I've been using Rocksim for this purpose, and I've come across a methodology question that I'm hoping someone here might be able to shed some light on.

In my approach, I've been contemplating the idea of altering one variable at a time to find its optimal value before moving on to the next. For instance, I'd first optimize the root chord length while keeping other parameters constant, and then, once that's optimized, I would proceed to optimize the next variable, like semi span, and so on.

However, I'm uncertain about how effective this method is in Rocksim, especially considering the complex interactions between various fin parameters. My main questions are:

1. How is Rocksim programmed to handle the optimization of fin shapes? Can it effectively account for the nuanced interactions between different fin dimensions when only one variable is altered at a time?
2. Is the method of optimizing one variable at a time, then moving onto the next, a viable approach in Rocksim? Specifically, I'm interested in understanding how this approach might impact the overall performance and stability of the rocket, given the interdependencies of fin dimensions. Might a certain fin span only apply to be optimal when adjoint to a certain sweep angle? And the many instances of contradictions inside the rocketry literature (elliptical is optimal due to its induced drag working, but for low reynolds number rectangular/parallelograms are better).

I have been reading many different articles for my research purposes (such as articles posted on the apogee website, Stine's Handbook of model rocketry, Gregorek's TR11 report, countless rocketry forum threads and many other articles and presentations such as this. I am obviously interested in the subsonic regime, but there are still so many factors to take into account. It's almost like the more information I read, the less clear my head gets about an approach to my optimization problem (I guess that is why they say ignorance is a bliss).

P.S I found an article with a similar methodology to what I am aiming to accomplish (I will not be doing CFD, but their OpenRocket methodology aligns somewhat), yet I still don't understand how the interdependency of factors was adressed, nor how they reached some of their conclusions such as the optimal number of fins being three without testing, I am just taking a guess and assuming the rocket models they made ALWAYS had 3 fins, and therefore it was a controlled variable).

On another note, does anyone dispose of a copy/method of contact with Mr. Cipolla, who created FinSim. I was looking into using it for a possible section on flutter analysis in my research, but all copies have been taken down from the website and I have not found any way to contact the man himself.

 

[SolarYellow]

The "3 fins is best" is old lore. If you do a good job with your investigations, you'll likely find that an optimized 4 fin design yields a higher apogee than a 3 fin design of the same proportional planform and stability factor. It's only by a very small percentage, though. The topic has been hashed over at length on this forum previously. There are many other factors involved that would lead one to choose 3 or 4 fins in various situations.

Your basic objective is to minimize drag for whatever stability factor your team chooses to target. A reasonable general approach is to hang as much fin area as far back on the rocket as you can. The Black Brant V provides a useful example of how one might do that and is obviously well proven. In service, there may be a variety of other considerations that could lead you to choose a different planform or to alter that concept in various ways for your rocket.

Without being able to sit in a room and have a conversation with you, my gut says you are trying to rely excessively on "mechanical" analysis, just setting up the equations and working through them to get an optimized result, and not enough on logical analysis, understanding intuitively what's happening and using that to direct your quantitative investigation.

You don't want "maximum" stability. That will be too much drag. Figure out what stability factor you want your rocket to have, and iterate around a baseline that achieves that minimum required stability factor, while seeking out the highest apogee.

All the factors are related and interdependent. I don't think you'll achieve a good result trying to sweep through one at a time and find the best answer for that single element, then combine all those "best" answers.

The best thing you can do is spend some time just trying changes in Rocksim and seeing what happens. You don't need to take a formal approach at first, just dive in and look around, see what you can learn. Figure out what kinds of changes do what. Play with it until you have an intuitive feel for it, then go back and set up your formal analysis matrices to sweep through and collect data.

 

[Anonymous]

The "3 fins is best" is old lore. If you do a good job with your investigations, you'll likely find that an optimized 4 fin design yields a higher apogee than a 3 fin design of the same proportional planform and stability factor. It's only by a very small percentage, though. The topic has been hashed over at length on this forum previously. There are many other factors involved that would lead one to choose 3 or 4 fins in various situations.

Your basic objective is to minimize drag for whatever stability factor your team chooses to target. A reasonable general approach is to hang as much fin area as far back on the rocket as you can. The Black Brant V provides a useful example of how one might do that and is obviously well proven. In service, there may be a variety of other considerations that could lead you to choose a different planform or to alter that concept in various ways for your rocket.

Without being able to sit in a room and have a conversation with you, my gut says you are trying to rely excessively on "mechanical" analysis, just setting up the equations and working through them to get an optimized result, and not enough on logical analysis, understanding intuitively what's happening and using that to direct your quantitative investigation.

You don't want "maximum" stability. That will be too much drag. Figure out what stability factor you want your rocket to have, and iterate around a baseline that achieves that minimum required stability factor, while seeking out the highest apogee.

All the factors are related and interdependent. I don't think you'll achieve a good result trying to sweep through one at a time and find the best answer for that single element, then combine all those "best" answers.

The best thing you can do is spend some time just trying changes in Rocksim and seeing what happens. You don't need to take a formal approach at first, just dive in and look around, see what you can learn. Figure out what kinds of changes do what. Play with it until you have an intuitive feel for it, then go back and set up your formal analysis matrices to sweep through and collect data.

You have exactly nailed the problem I have with mechanical analysis. I have been doing some random testing, but it is complicated to design a realistic rocket and take into account so many parameters at once. The logical analysis of results I find complicated to really understand, since, yes, I can read many different papers and research articles, but the fundamentals of the inner workings of RockSim are something I can't end up analyzing.

I did realize, as you pointed out, that using the term maximum stability would be incorrect, what I meant was a minimized optimal value for stability. Rockets which are too stable (at least in the static stability definition) often do not have the best apogee because of the weathercocking, so I actually meant optimum stability which would probably be a static stability with the range between 1-2 cal.

Thanks for the pointers and clarification, it is much appreciated!

 

[tfish]

@Anonymous

If you post your simulation here..guys can see what you've done and give some input.

Tony

 

[Scott_650]

Tim Van Milligan, owner/proprietor of Apogee and the “man behind the curtain” for RockSim has written many articles for his Peak of Flight newsletter on using RockSim to optimize rockets along with several general topic type articles on fin shape and the aerodynamics of rockets. Well worth the time to look through the PoF topic index on the Apogee site.

 

[MarsLander]

PoF topic index on the Apogee site.

https://apogeerockets.com/Peak-of-Flight?pof_list=archives&m=education
^^^Index here scroll down the page.

 

[Anonymous]

Tim Van Milligan, owner/proprietor of Apogee and the “man behind the curtain” for RockSim has written many articles for his Peak of Flight newsletter on using RockSim to optimize rockets along with several general topic type articles on fin shape and the aerodynamics of rockets. Well worth the time to look through the PoF topic index on the Apogee site.

I have read both his "Technical publication 16 - What Type of Fin Shape is Best", which concludes that rectangular/parallelogram shaped fins are the best due to their efficiency at lower Reynolds numbers and also his "PoF Newsletter 442 - What is the best fin shape for a model rocket?", which indicates that the clipped delta or elliptical are the best due to their lower drag values, with small fin span (just enough to get the margin of stability to 1). I have not mentioned the aspect of airfoils (though it is present in both of the articles mentioned), seeing as there is no way to simulate changes in airfoil incrementally in Rocksim and my investigation is looking to not do a simple literature review, rather taking a more hands-on approach (I will comment on the importance of airfoils, but not simulate their effects due to my lack of knowledge of CFD simulations)

 

[Anonymous]

Another doubt I have in relation to the simulations is, if optimizing for stability, would it be better to look at the static stability margin inside Rocksim without the added weight of the motor (+500 grams worth of water), since in the real world the loss of water happens so fast that the relationship between CG and CP during flight would be that of the empty mass of the rocket?

 

[Anonymous]

This is just a test example, but I don't understand how Rocksim reaches such information.

For example, Rocksim predicts the rocket below to get a max altitude of 47.44 m and to have a stability margin of 1.20, since I am guessing the further back the fins reach, the further back Rocksim decides to move the CP, where realistically a rocket like this would be inviable.


While having a more realistic rocket, where the only parameter varied is the tip chord length to a smaller and realistic one, we obtain a max altitude of only 14.49m and a stability caliber of -0.68, which therefore indicates that the simulation might be using an approach more guided by stability caliber than the drag that such area of the fins would produce.

 

[mikec]

If the rocket is unstable then it just won't fly straight and the max altitude will be greatly reduced.

I don't think Rocksim's simulation is realistic enough to optimize fins in any particular way beyond predicting stability at low AOA, which is about all a Barrowman-based approach can do.

In my very limited experience water rockets are more stable than strict Barrowman would predict. Some have attributed this to jet stability effects. Rocksim really won't model this.

 

[Anonymous]

How would you suggest I direct my research then? If you don't mind me asking, since it seems I am very limited in my optimization. Should I just establish a range of only 2 parameters or something similar?

 

[mikec]

I suggest a focus on actual flight testing rather than simulation. At least use the simulation to predict something and see if that really happens or not.

 

[dhbarr]

water rocket project

You're deeply subsonic -- it's not that aerodynamics don't matter, it's just that things like nosecone heating, fin flutter, etc. simply cannot affect you. Your primary focus is mass, and in particular where the water is sitting and how it's moving.

 

[cls]

How would you suggest I direct my research then? If you don't mind me asking, since it seems I am very limited in my optimization. Should I just establish a range of only 2 parameters or something similar?

You have "analysis paralysis". Go build several different models and collect lots of flight data

 

[BABAR]

You have "analysis paralysis". Go build several different models and collect lots of flight data

Theoretically valid but I imagine would be challenging in practice.

there is always a certain amount of simple random variation even under truly identical conditions. Sometimes this is referred to as background “noise.” So even if you could assure the winds were constant, temperature constant, you can fly the same rocket with exactly the same setup and you are going to get a range (hopefully narrow) on numbers, from which you can calculate an average and a standard deviation.

add To that the uncontrolled variables.

winds

temperature

motor itself. The official Estes motor chart says “Delays have a tolerance of plus or minus 10% or 1 second, whichever is greater.” They don’t say what the variability is in the PROPELLANT load, but I’d be surprised if it wasn’t similar to that of the delay charge. Really, please, somebody with inside poop correct me if I’m wrong.

something difficult to account for is wear and tear on the rocket itself. Many of my rockets are not exactly pristine after 10 or more flights, fins get dinged, etc. many problems are easily fixable, but not sure it is EXACTLY the same rocket after the fix.

I don’t do water rockets (yet!), so not sure of their durability.

I am curious how many amateur rocket scientists have really gone this route? In my personal experience most of my “experiments” have ideally been “binary” as mainly I try somewhat radically different designs, so mainly either

it worked!

or

it didn’t work

but even then sometimes its

It sorta worked

not meant to discourage experimentation, but I would think distinguishing subtle difference in fin construction may be challenging to detect with actual flights. On the good side, water is cheap in most places, so that’s a plus compared to burning a lot of motors!

any chance you’d have access to a wind tunnel at a local college? I think Tim Van Milligan’s daughter actually did some cool stuff with the assistance of the Air Force Academy. (One of my friends there really hoped to put a Magpie in the hypersonic wind tunnel, never happened of course.)

anyway, doesn‘t hurt to ask.

I commend you on your aspirations and wish you the best.

 

[BABAR]

You have "analysis paralysis". Go build several different models and collect lots of flight data

Some interesting info in this post

https://www.rocketryforum.com/threads/thrust-and-impulse-variance-in-commercial-apcp-motors.172341/

 

[Rschub]

I have a method to calculate the perfect fin size exactly, but it only works on spherical rockets in a vaccum.

 

[Anonymous]

I have finally decided to just control some variables in the simulation (such as span and number of fins). The only pair of questions I have left is whether in Rocksim stability always correlates to height, from my tests it seems to be so, but in the other similar research project (linked in my first instance of the thread), it doesn't seem to be so, though they are using OR instead of RS. Since if it is so, I don't have to create two different data tables, one for stability and one for apogee, rather I just have to do 1 for either of the 2 results and since they are correlated it saves me time.

Also, does anyone know if the weird attachment that fins do sometimes happens in RS due to the transitions has any impact on the calculations or is it a simple visual bug?

 

[Rschub]

In general, you should only have one variable at a time so you can quantify the results without any other changes contaminating the data.

 

[kjhambrick]

And when you do fly and if you care about answering such design Q's with real data, fly a 6-DoF flight computer.

One can actually measure what's what that-a-way ...

-- kjh

 

[Anonymous]

It took so time, but I got a first batch of iterations done for the parameters to change. Note that none of the values get to 1 caliber of stability, but to fix it we have decided to add some nose cone weight to the simulation, which allows for the CG to move forward and have fin shapes with better stability. Thanks for all the help!

In case anyone is interested in seeing some of the results, even if they are not exact, they are adjoined in a file.

 

[BABAR]

I have a method to calculate the perfect fin size exactly, but it only works on spherical rockets in a vaccum.

Is that the IJDM* rule?



*It Just Doesn’t Matter

 

[Rschub]

 

[Anonymous]

I am pretty much finished with my stability analysis for the new dataset, but was wondering if including airfoil in my research would be interesting considering the effect it has on the drag of the rocket. Can anyone vouch for another program to do this analysis (as Rocksim isn't accurate to the degree needed as stated by the man, Van Milligan himself)?

I was looking at QBlade, "a cross-platform simulation software for wind turbine blade design and aerodynamic simulation which uses XFOIL, allows for the user to rapidly design custom airfoils and compute their performance curves, extrapolating the performance data to a range of 360°Angle of attack, and directly integrate them into a wind turbine rotor simulation".

I wouldn't be that interested in the wind turbine functionality, but it seems easier to learn and maneuver than a program such as Openfoam, which would require coding skills and meshing skills (of which I lack) and an understanding of how exactly to set up a CFD to obtain correct results. Having never worked before with such a program that seems like a daunting task.

If it is paid, but offers a free trial or is not overly expensive I would consider getting it, given that it can speed up the workflow and allow for setup and analysis. Thanks!

 

[SolarYellow]

I expect that to distinguish between various low-drag symmetrical airfoils, you would need to have a high-fidelity flight sim program, and by that, I mean higher fidelity than Rocksim. If you get that level of technological sophistication, you should replow all the work you did in Rocksim at the higher fidelity, too.

If you have the capability to manufacture an airfoil fin accurately, I'd pick one of the standard symmetrical subsonic low drag airfoils and get on with designing your rocket. If you can't manufacture it accurately, does it really matter?

It's definitely possible to burn waaay too much time and effort attempting to prove that your design is as close as reasonably possible to perfect before you start building. My advice is to get to something good enough and then start building. There is wisdom in my sig, but there's also the reality that the literature around model rockets is limited. It's possible to get to the end of the "library" without having all the information you need to demonstrate that your design is the best of all possible designs. In fact, it's certain. Once you've digested the majority of the information available in the literature, you reach a point of diminishing returns in continuing to analyze. You'll learn sooo much more actually building and testing than you can learn by continuing to read and analyze the literature.

 

[mikec]

AFAIK the classic Barrowman stability equations don't say anything about fin airfoil. The only use of airfoil information in simulations is for drag, not stability.

 

[dhbarr]

Airfoils are not likely to make a measurable difference in this flight regime, unless your fins are quite thick. If you'd like to be able to tell the difference, try a set that are completely squared off vs. a set that have any reasonable airfoil whatsoever. Still might be within the measurement errors, though.

 

[Anonymous]

AFAIK the classic Barrowman stability equations don't say anything about fin airfoil. The only use of airfoil information in simulations is for drag, not stability.

I am well aware, but my research is centered on looking for the best fin design in terms of stability and apogee, and seeing as drag has an effect on apogee, and the airfoil has more of an effect on drag than planform shape, it would be an interesting aspect to consider and investigate.

 

[kjhambrick]

@Anonymous --

This is a fun tool: Airfoil database search (NACA 4 digit)

One of my favorites for rockets fins is the NACA 16-006 (naca16006-il) ( not that I can actually make one with my stone knife and bearskin tools )

Lots of info to absorb over there ...

One important take-away is the relatively low values of alpha ( Angle of Attack ) before a symmetric fin stalls ...

HTH

-- kjh

EDIT: grammar

 

[Alan15578]

I am well aware, but my research is centered on looking for the best fin design in terms of stability and apogee, and seeing as drag has an effect on apogee, and the airfoil has more of an effect on drag than planform shape, it would be an interesting aspect to consider and investigate.

Often true, but aspect ratio and sweep can often have more significant effect by reducing wetted area. Then again, you might just as well be researching if red rockets fly higher than blue rockets.