Subscribe to rss feed

Some preliminary flight simulations | Project planning and misc | Forum

 
You must be logged in to post user permissions login Login register Register


Register? | Lost Your Password?

Search Forums:


searchicon 






Minimum search word length is 3 characters – Maximum search word length is 84 characters
Wildcard Usage:
*  matches any number of characters    %  matches exactly one character

topic

Some preliminary flight simulations

print
small tagNo Tags
UserPost

2:03 am
December 15, 2009


Luke Maurits

Adelaide, Australia

Admin

posts 1483

offline
link
print
1
0
ratedowngrey
rateupgrey

Post edited 9:18 am – December 15, 2009 by Luke Maurits


In order to be able to get some approximate hard numbers up on the Wiki for OHKLA in order to make it look more like we know what we are doing, I have done some basic flight simulations today.  There is ample work for improved accuracy, but at the very least these figures should be not terrible.

Here's the gist of how these work:

  • A rocket is completely specified by its engine burn time, engine thrust, empty body mass, full fuel mass, payload mass, radius and drag coefficient.
  • Simplifying assumptions: engine thrust is constant for duration of burn, fuel mass decreases at a uniform rate throughout burn.
  • Air density (for drag calculations) is computed from altitude according to the formula here.
  • Because I know very little about aeronautical engineering, my way of handling supersonic drag is as follows:  At all times during flight, the rocket's coefficient of drag is D*f(M), where D is a fixed "base drag coefficient" and f(M) is a "drag scaling function" which is a function of Mach number M (yes, Mach number calculations take into account change in air density with altitude).  The shape of f(M) is such that f(M) = 1 if M < 0.6 or M > 1.8.  In between these values, the drag is increased.  Between Mach 1.0 and Mach 1.2, f(M) = 5.  Between Mach 0.6 and Mach 1.0, it increases linearly from 1 to 5 and between Mach 1.2 and Mach 1.8 it decreases linearly from 5 to 1.  These numbers were "inspired" by this article.  With drag coefficients calculated thusly, drag forces are computed using the standard drag equation.  This is probably the biggest fudge in the entire simulation, if a real aeronautical engineer can advise me on how to improve this, by all means do so.
  • Equations of motion are integrated using RK4.

My main aim here was to get a feeling for what the engine thrust and burn time need to be.  For the purpose of some initial investigation, I have used an empty rocket mass of 30kg, a fuel mass of 100kg, a payload mass of 5kg and a radius of 15cm.  These values I got by looking at the values for RocketLab's Atea-1 rocket and adding "pessimism" factors.

I had no idea what a realistic base drag coefficient for a rocket should be, so I experimented with D = 0.1, 0.2, 0.3, 0.4.

The graph below shows (for these various drag coefficients) the required burn time (s) for a given thrust amount (N) in order for the apogee of the flight to be 100 km (or, reading it sideways, the required thrust amount for a given burn time).

Ohkla thrust v timemouse

Now, it's important to realise that these are probably underestimates.  This assumes that all of the rocket thrust is directed perfectly upward.  In real life there will be wind, which will cause the rocket to tilt somewhat, causing some of the thrust to be wasted on horizontal translation.  This is a little tricky to model because, as the rocket tilts, the angle of attack for the stabilising fins will change, causing asymmetric generation of lift to counter the tilting.  We're gonna have to model this eventually, but since this a first stab I have left it out.

Note that if my crappy supersonic drag approximation results in more drag than we can actually expect, this will counteract the understiamtion due to the above point about wind, and these estimates may be okay or even overestimates.  For now I will assume that they are underestimates.

Now, taking a pessimistic drag coefficient estimate of 0.3, a thrust of about 5000 N for 50 seconds is adequate according to the above graph.  Since these are underestimates for the rest of this thread I will use 5000 N for 60 seconds (a rather unprincipled correction).  For these values:

  • 100km of altitude is reached after about 106 seconds
  • Maximum altitude is 175 km after 230 seconds
  • Maximum velocity is about 1715 m/s or 6174 km/h
  • Maximum acceleration experienced is about 120 m/s/s, or a bit more than 12 g
  • Max Q is about 46000 Pa

These figures seem relatively reasonable:  RocketLabs claim a max launch acceleration of 15 g for their rocket (which is lighter than ours, so the slighter higher result makes sense) and claim a maximum velocity of "> mach 5″.  Now, this is kind of technically incorrect – the speed "mach 5″ is air pressure, and hence altitude, dependent.  When a rocket is at its fastest, the atmosphere is very thin, so the speed of sound is very fast and the mach number is actually quite low.  Assuming that by "mach 5″ they mean 5 times the speed of sound at sea level, that's about 6180 km/h which is extremely close to the maximum speed from my simulation.  CSXT claim that one of their rockets reached space in one minute and thirty seconds, only 16 seconds faster than my simulation.  So we are certainly seeing figures which match those for real world OHKLA-esque rockets.

So, for rough planning and design purposes, we can work with the rough figures:

  • Engine thrust = 5000 N
  • Engine burn time = 60 s
  • Time to zenith = 4 mins
  • Max acceleration = 12 g
  • Max velocity = 6175 km/h ("Mach 5″)

I probably won't get much more done today, but in the near future I'll make a Wiki page dedicated to the results of simulations (which we can refine over time), which will include links to the GPLed code I wrote for this and more graphs of various things (altitude and velocity vs time, etc).  EDIT: The simulation results will go here, the code and instructions will go here.  Keep an eye out!

Input on everything here is extremely welcome, particularly from people who know more about supersonic aerodynamics than I do.

Main CLLARE workgroups: Mission Planning, Navigation and Guidance. I do maths, physics, C, Python and Java.

7:53 am
December 15, 2009


Rocket-To-The-Moon

Altus, Oklahoma, USA

Member

posts 685

offline
link
print
2
0
ratedowngrey
rateupgrey

Post edited 1:59 pm – December 15, 2009 by Rocket-To-The-Moon


Excellent work! These rough numbers at least give us something to design towards. It seems like these simulations plan on having a relatively low thrust motor burn for a substantial amount of time. CSXT's solid rocket only burned for 14 seconds and then it coasted upward for over a minute (sorry, dial-up is too slow to provide the source). I have no problem with a long burn motor for several reasons. For one it reduces the maximum acceleration which will make off the shelf electronics more feasible and secondly it makes it possible that the Selene Modular Booster can simply be a cluster of OHKLA hybrid rockets. For Selene staging it might be best to have around a 120 second burn time so that the three stages burn for a total of six minutes.

This assumes that all of the rocket thrust is directed perfectly upward.  In real life there will be wind, which will cause the rocket to tilt somewhat, causing some of the thrust to be wasted on horizontal translation.

For OHKLA I don't think this will be a problem. As long as we keep the rocket pointing directly upward the wind will only result in drift. We only waste fuel if we try to correct the drift (so that we don't have to recover the rocket downwind). If the wind is blowing at an average of 5m/s from 0-15,000m then the rocket will only drift 5 meters for every second that it is in this altitude range. This is negligible during ascent, but quite significant during parachute descent.

One option would be to measure the wind direction/velocity during the ascent (inertial…may be difficult to get this accurate) and then over-correct "upwind" after passing 15,000m so that it will land close to the launch facility under parachute. It is best to make this correction after the rocket is a high velocity so that we have the thrust vector not perpendicular to Earth for as short of a time as possible.

Alternatively we could just look at a weather map and skip the thousands of lines of code required to measure and correct the drift.

Main Workgroups: Propulsion & Spacecraft Engineering

8:06 am
December 15, 2009


Luke Maurits

Adelaide, Australia

Admin

posts 1483

offline
link
print
3
+1
ratedowngrey
rateupgrey

Rocket-To-The-Moon said:

For OHKLA I don't think this will be a problem. As long as we keep the rocket pointing directly upward the wind will only result in drift. .


Of course this is true, but I don't think we will be able to keep the rocket pointing directly upward.  Let's restrict our attention to the two dimensional plane which is perpendicular to the Earth's sufrace and parallel to the wind, and suppose the wind is blowing from left to right.

The wind will exert a drag force on the part of the rocket which is above its centre of mass, causing a clockwise torque.  At the same time, the wind will exert a drag force on the part of the rocket which is below its centre of mass, causing a counter-clockwise torque.  There will be some rotaton about the centre of mass if those two torques do not perfectly balance.  Whether or not this happens depends on the location of the centre of mass and the difference in the drag coefficient of the part of the rocket above it (including the slender nose cone) and the drag coefficient of the part of the rocket below it (including the stabilising fins).  Note that here I'm talking about the drag experienced by these shapes when travelling "sideways".  I think it would be extremely hard to design the rocket so that these torques balanced.  Actually, since the centre of mass will be moving as the fuel burns, I think it would be literally impossible.

Or am I completely confused?  I haven't actually read anything about this issue from an authoritative source, I am merely trying to apply what I remember of rotational dynamics.  I haven't thought about this kind of thing for a while.

Main CLLARE workgroups: Mission Planning, Navigation and Guidance. I do maths, physics, C, Python and Java.

8:16 am
December 15, 2009


Rocket-To-The-Moon

Altus, Oklahoma, USA

Member

posts 685

offline
link
print
4
0
ratedowngrey
rateupgrey

Luke Maurits said:

Or am I completely confused?  I haven't actually read anything about this issue from an authoritative source, I am merely trying to apply what I remember of rotational dynamics.  I haven't thought about this kind of thing for a while.


This sounds completely logical.

In order to keep it pointing directly upward we would need a vertical gyro that stays parallel to the surface of the Earth. Any deviation would be sent to the fin actuators to compensate. This should cancel out any torques that the wind puts on the rocket and give use a perpendicular flight profile. As the atmosphere thins with altitude the fins will become ineffective, but so will the wind.

CSXT used the entire rocket as a gyro by spinning it (canting the fins). This is definitely an option but it does pretty much make it impossible to record ascent video.

Main Workgroups: Propulsion & Spacecraft Engineering

8:39 am
December 15, 2009


Luke Maurits

Adelaide, Australia

Admin

posts 1483

offline
link
print
5
0
ratedowngrey
rateupgrey

Do we necessarily have to move the fins in response to tilt?  I think that the natural tendency of fixed fins, parallel to the rocket body, would be to act against the tilting, i.e. to correct for it.  Whether or not this was enough to keep the rocket upright would depend on the precise size of the fins and their shape, and also the speed of the winds.  If we could get away with fixed fins, I think our design philosophy would dictate that we do so.

Spinning the rocket is a possibility to consider.  We could still record video and use our inertial measurements of the spin to figure out how much to rotate the individual frames by to correct for this motion, resulting in roughly steady video.  I don't know how well this would work, though.

Regarding your earlier comment on these simulations using low power, long life burns – this wasn't something I consciously decided to do, I just decided to let the program range from 1000 to 10000N because they were nice numbers that seemed sensible at the time.  If higher thrusts are realistic it would be a simple matter to extend the plot shown here further to the right and look at burn times for higher powered engines.  I agree with your arguments for favouring longer burn times, by the way, but it can't hurt to look at the numbers for the other approach to fully inform our choice.  I'll try to produce this extended graph tomorrow.

Main CLLARE workgroups: Mission Planning, Navigation and Guidance. I do maths, physics, C, Python and Java.

11:46 am
December 15, 2009


Rocket-To-The-Moon

Altus, Oklahoma, USA

Member

posts 685

offline
link
print
6
0
ratedowngrey
rateupgrey

Luke Maurits said:

Do we necessarily have to move the fins in response to tilt?  I think that the natural tendency of fixed fins, parallel to the rocket body, would be to act against the tilting, i.e. to correct for it.


I'm not 100% sure, but I think that the wind will still be able to torque the rocket about its CG. The fins serve to align the thrust vector with the CG…I think. This is total intuitive guesswork, please correct me if this is off target.

Main Workgroups: Propulsion & Spacecraft Engineering

12:22 pm
December 15, 2009


rpulkrabek

Member

posts 349

offline
link
print
7
0
ratedowngrey
rateupgrey

Post edited 6:28 pm – December 15, 2009 by rpulkrabek


I was blocked to this post at work because you have "simulations" in the title and the proxy at work thought I was trying to play a game, which was a bit frustrating.

Anyway, I wanted to remind that we once had a member that also did some similar simulations. sirhcdobo from reddit posted a Matlab code that solves ODE atmospheric trajectories (http://www.reddit.com/r/tothem…..uation_i/). I thought we can continue this as well for OHKLA. His comments about drag coefficients he used, which he said were about standard for any space going material shapes, was, "in terms of the drag coefficient ill state some numbers for the hyshot mission that i have just completed.at 0 angle of attack the CD is approx .15 for mach numbers 0-.9 it then jumps up to approx .41 for mach number about 1 by mach 2 CD is down to .2 before decaying away to .15 again. this is pretty standard for space going materials and shapes."

The Matlab code is in Dropbox. I looked at it, and he was right, it could use a bit of work, mostly to comment the code, since I often found myself wonder what each variable was for and such. There are three plots that come from the code. They are here:

i1Baamouse

i05XHmouse

TN8Hvmouse

Image 1: http://i.imgur.com/i1Baa.jpg

Image 2: http://i.imgur.com/i05XH.jpg

Image 3: http://i.imgur.com/TN8Hv.jpg

As a side note, I am currently looking into seeing if I can perform such an FEA/CFD that will help decide a nozzle shape and perhaps the resulting thrust for a known combustion. This will come later and depend on how much time I have. Hopefully it'll come out right.

I also once described how we measured thrust in a lab I took part in. We can use this method in the future too, I hope.

7:49 pm
December 15, 2009


Luke Maurits

Adelaide, Australia

Admin

posts 1483

offline
link
print
8
0
ratedowngrey
rateupgrey

Thanks for reminding me of those earlier posts, rpulkrabek, I had forgotten about them.  The information drag coefficient ranges for various Mach numbers is especially useful.

I think I will update my code to use those drag numbers before I do what I said I'd do earlier, extend the thrust vs burn time graph further into the realm of more powerful engines.

I will also have a look at the Matlab code to see if it is taking into account effects which I have overlooked.

Things are looking good for our simulations!

Main CLLARE workgroups: Mission Planning, Navigation and Guidance. I do maths, physics, C, Python and Java.

1:12 am
December 18, 2009


Luke Maurits

Adelaide, Australia

Admin

posts 1483

offline
link
print
9
0
ratedowngrey
rateupgrey

I have done a second version of the thrust vs required burn time graph, with two changes:

  • Thrust values range from 1,000 N to 50,000 N instead of 1,000 N to 10,000 N
  • Drag coefficients now match the description given in rpulkrabek's post earlier.

Here it is:

800px Ohkla thrust v time2mouse

As you can see, once the thrust gets to around 25,000 N we start to see seriously diminishing returns.  Extra thrust does very little to decrease the burn time – this is because if the rocket goes fast enough while low enough (in the thick atmosphere), drag forces can get large enough to balance out the thrust resulting in no acceleration – the same sort of general situation that leads to a terminal velocity for falling objects.

Because the previous drag model was extremely conservative compared to more realistic models, this graph shows lower required burn times for any given thrust.

Main CLLARE workgroups: Mission Planning, Navigation and Guidance. I do maths, physics, C, Python and Java.

9:37 pm
April 15, 2011


spacelaunch

Member

posts 6

offline
link
print
10
0
ratedowngrey
rateupgrey

The simulations and assumptions above are generally correct.

One issue not addressed however is at which stage the rocket becomes aerodynamically
stable. For a fin stabilised rocket you really want to leave your launch rail
at above 4G of acceleration, less than this & you will most likely need
active guidance to assure near vertical ascent. A high thrust rocket at low
altitude will incur increased drag loss, a low thrust rocket will incur greater
G losses. This often mitgated by most unguided sounding rockets by using either a two stage configuration or a dual thrust phase rocket motor (boost/ sustain mode).

If you’re building a single stage hybrid rocket to achieve the goal of the
program (above 100km), you could look at throttling the motor. An ideal scenario would be to
have a dual phase hybrid which gives a short duration high thrust initial
phase, then throttles back to a coasting low thrust long burn phase. You could
achieve this by using a high regression rate (wax) fuel boost grain, housed in
a larger low regression rate grain & two propellant feed valves.

One feed valve would control a peripheral high flow injector for the boost
phase & would actuate for say 4 to 5 seconds, then would be throttled down
as the sustainers valve is brought open. A single drive (electric or pneumatic)
& gears could be used to actuate both valves. I would suggest ¼ turn plug
valves for the application due to their low actuation torque requirements, and
ability to throttle.

Guys at this stage I would steer you toward two excellent & free
software packages for determining rocket performance. These are Launch 2.2A
& Aerolab, which have both been developed by Hans Olaf. These programs have
been accepted as accurate for purpose of gaining launch permission at the Woomera
rocket range in Australia,
by the defence department controlling body. Which is a good testament to the
accuracy of their results; I have used them for a number of years & can attest
to the accuracy also.  Launch is for the
flight simulations & Aerolab is for drag & stability all the way to
supersonic velocities. It will save you valuable time and effort that would
expended on trying to develop & validate your own code for the job..

I’m going to run some numbers & come back to you shortly with some
figures on the above dual phase thrust scenario.

small tagNo Tags

About the CSTART – Collaborative Space Travel and Research Team Forum

Forum Timezone: UTC -6

Most Users Ever Online: 59

Currently Online:
7 Guests

Currently Browsing this Topic:
1 Guest

Forum Stats:

Groups: 4
Forums: 36
Topics: 514
Posts: 3815

Membership:

There are 1139 Members

There are 2 Admins

Top Posters:

Rocket-To-The-Moon – 685
brmj – 402
rpulkrabek – 349
DenisG – 69
antinode – 64
J. Simmons – 46

Recent New Members: peterpaul008, Sandra, melody, devinbutler30, Manoj Kumar, Andrew

Administrators: Luke Maurits (1483 Posts), Rizwan (170 Posts)



 
share save 120 16