# New type of rocket propulsion design, and why it might not work?

• escape75

#### escape75

TL;DR Summary
propulsion design
Hi,

*Excuse the crude visualization and possibly imprecise description - I'm not a rocket scientist :)

So, I have a friend that has come up with a "new" concept for a rocket propulsion design, and although I have a hunch that it might not work better than existing designs I cannot really explain it properly or find the issue with his design, so here it is ...

Instead of having 2 rockets firing outwards into space (left side), we will have 2 identical rockets firing into a curved chamber that will then in turn dispose of the gasses outwards (right side). His design is basically what you see on the right side part of the diagram.

I came up with this simplification of how we could test it. (The fuel tanks are in blue). There would be a platform, mounted on 2 rails, in a way that it could only move forward/backward, and when we start the burn on both sides of this "system" which way would be platform move ?

According to him because of the better design and how mass hits the curved walls the platform would move left, but to me it seems the exhaust would be hindered on the right side so thrust would be lowered and the platform would move to the right ?

It would be interesting to see what everyone's thoughts are on this, and if there's any type of calculations that can be performed instead of just saying that it would move this or that way ...

## Answers and Replies

Summary:: propulsion design

According to him because of the better design and how mass hits the curved walls the platform would move left,
Welcome to PF.

It seems like there would be more friction/heating losses in the longer curved ducting sections for the right side design, so it would generate less thrust. Where exactly are the combustion chambers in his design?

russ_watters
Summary:: propulsion design

have a hunch that it might not work better than existing designs
You didn't say in what way the new design is expected to be better than a conventional rocket. And if better, why?

Thanks :)

The combustion chambers would be just next to the fuel tanks in both cases, and yes there would be for certain heating loss in the curved ducting, as well as more friction, but there would also be additional forces from the exhaust gasses acting on the curved wall adding to the thrust, however since the speed of the ejected gasses would most certainly be slower how would that translate to total thrust,- one gained by forces acting on the curved ducting minus the losses that occur while doing so,... I wonder.

Thanks :)

The combustion chambers would be just next to the fuel tanks in both cases, and yes there would be for certain heating loss in the curved ducting, as well as more friction, but there would also be additional forces from the exhaust gasses acting on the curved wall adding to the thrust, however since the speed of the ejected gasses would most certainly be slower how would that translate to total thrust,- one gained by forces acting on the curved ducting minus the losses that occur while doing so,... I wonder.
All that matters is the momentum of the ejected material. If it's slower, that's bad.

dlgoff, Benjies, russ_watters and 3 others
Those "additional forces" already happen in the expansion bell of a normal rocket engine. This will add extra losses and would substantially reduce efficiency (and make it very difficult to cool - the heating rate along the outside of that curve would be extremely high).

berkeman
Exactly, that makes sense to me.

However he would still like to test things out, and instead of rockets use compressed CO2 cartridges.

However I'm not sure this would even simulate things properly as we would most likely see thrust achieved by having the compressed air pushing against the atmosphere and not by having momentum because of the ejected mass times speed ...

However he would still like to test things out
Does he not understand how rockets work, and the law of Conservation of Momentum?

I mean, it's good that he wants to try some experiments, but which part of our explanations does he want to test out? If it's just the path the propellant follows before being ejected, there are other ways to test that out. Maybe just a couple marbles on two sides of a doubly-inclined plane resting on an air-hockey table (very low friction), where the right marble is forced to follow a more circuitous path...?

russ_watters
However he would still like to test things out, and instead of rockets use compressed CO2 cartridges.
How about a single tank of compressed air? Safer and easier to control.

In practice your friend's design is deliberately avoided, and is referred to as 'line losses'. The pressure loss (or energy loss, if you will) is greater than zero and only amounts to a loss in energy that your engine can expel in the form of thrust. I see what your buddy is getting at, trying to "employ" this sort of line loss to work as an advantage, but the bottom line is that this would simply increase these line losses.

There's a reason that SpaceX's Raptor 2 was a search to simplify design and reduce complexity. At the end of the day, all we want to do with our combusted flow is push it out, in a single direction, as fast as humanly possible. I'd venture to say that the platform you've drawn, in theory, would move right due to increased line losses on the "curved design" engines.

I'll spare you the technical details on how a throat functions to contain a normal shock as well, but trust me, if the flow leaving that blue tank is indeed encountering a normal shock, then the curved design would be a nightmare to cool and would likely encounter a thermal choke, which would just melt your engine.

mfb, PeroK and berkeman
The CO2 should actually simulate it reasonably well, as long as you achieve a choked throat and as long as your expansion isn't large enough that you start to see dry ice forming due to low temperatures. You'll need to be really careful that you're actually matching the mass flow rate though - depending on details, you could end up making more thrust just because one configuration has a slightly larger diameter, and in that case, you haven't learned anything.

The figure of merit for rockets is really thrust to weight (which is obviously going to be worse since this design requires more piping), and thrust to mass flow ratio (which basically tells you the efficiency). You're primarily going to care about the second, so you need to ensure that both drain the CO2 tank at about the same rate, or alternatively if you have a way to datalog the thrust over time, you can look at the total impulse (integrated thrust over time) vs the total mass of propellant consumed, which will give you the same data as thrust divided by mass flow rate.

Your picture shows a bell expansion as a straight lined cone.
If fact though, the cone has more of a curved shape.

What this does is allow the gases to have an isentropic expansion, within the bell, from the choked flow ( maximum flow ) in the throat area. With this type of expansion you get the most work out of the gas to propel your rocket.

With a short bell ( imagine cutting the length of the bell in the above picture closer to the nozzle ) the gases have not fully expanded in the short length and some of the 'pressure' is wasted and the gas just flairs out to the side. So, longer the bell the better to get the most work output from the gas.

Unfortunately, a longer bell adds weight to the rocket ( and a bit of extra friction ). So a compromise is done to limit the bell length to a value where adding extra length and the diminishing return of work output that goes with it, balances the extra work needed to accelerate the extra weight.

I forget the type of curve for the bell, perhaps hyperbolic, or parabolic, but you will notice that as the bell becomes longer and longer, it looks as of the bell becomes more and more straight lined, approaching that of a cylinder ( a regular pipe shape). If there was not that elbow in the right hand side, it could be said that it could be some sort of approximation to a long bell.

BTW, in regards to the picture, the flair on the right hand side is superfluous, since the expansion of the gas has already occurred somewhere within the 'pipe' length, hopefully starting right after the throat by design.

You could test one engine at a time with the same throat area, by putting the rig ( on wheels for ex ) up against a spring and see how much thrust is produced per engine ( spring compression ). That way the exact same throat area will be used.
I would test 3 engines.
1. short bell length such as on the left hand side of the picture.
2. long bell length, with an added straight 'pipe' attached to the short bell section in test #1
3. elbowed bell length, same 'pipe' length as in #2, but now curved.

That way you can tell if more/ less thrust is developed with additional length, and/or if the curve makes a difference .

berkeman
Thank you all for the comments and suggestions!

First of all I was thinking that by testing this with CO2 with air around it wouldn't simulate a rocket properly because quite a bit of force would be generated by CO2 pushing against the air, forming a sort of pressure bubble, but considering the right hand side has reduced pressure, the left part would still be more efficient.

I don't know how much force is generated in a rocket from forces other than the fuel being ejected (for example the gasses hitting the flair) but I am assuming it's not very much ...

I've just done a quick simulation of mass bouncing at 90 degrees vs. following the curve path and of course the 90 degree path is more efficient, I realize this is a simplification.

I've just done a quick simulation of mass bouncing at 90 degrees vs. following the curve path and of course the 90 degree path is more efficient, I realize this is a simplification.
Can you post your code that generates this simulation?

Can you post your code that generates this simulation?

It's an old program I have called Interactive Physics, and I was using a mass of 19 kg for the U-shape, and for the "fuel" there's a mass of 0.5 kg for each of the balls, and they are traveling at 10 m/s. On the left side they are traveling downward, and the right hand side of the simulation they are traveling at 180 degree away from each other.

The software is so old it only generates .avi files which in turn I converted to .gif :)

I have also made a similar simulation in Algodoo.

Are you trying to simulate behavior of a gas by tracking individual particles?

Are you trying to simulate behavior of a gas by tracking individual particles?

I don't know of any programs that would allow such simulation, at least nothing open source or free.

I was just trying to approximate Newton's 3rd law by using mass/velocity of ejected material,
as I don't see how a curved designed would be in any way more efficient, as was already mentioned.

I was using a mass of 19 kg for the U-shape, and for the "fuel" there's a mass of 0.5 kg for each of the balls
I asked because of that statement you made. Each of the balls implies individual treatment. The mathematics of gasses are thermodynamics and Navier-Stokes, not Newton's Laws.

I asked because of that statement you made. Each of the balls implies individual treatment. The mathematics of gasses are thermodynamics and Navier-Stokes, not Newton's Laws.

Yes I realize that of course, but since a rocket can be very roughly simulated by throwing payload off that has a certain mass and velocity to create thrust, I figured we could now see what happens when the payload is thrown at 90 degree on a curved track, versus in the opposite direction of the object being propelled.

There's a simulation that someone did I believe in Algodoo that I have found on youtube, that simulates a rocket that's ejecting metal rods, one at a time, while the rocket is being propelled ... It's not meant to simulate thermodynamics of course but I believe when comparing the two different design layouts it should illustrate how each of them would behave.

I figured we could now see what happens when the payload is thrown at 90 degree on a curved track, versus in the opposite direction of the object being propelled.
How does that differ from a fluid in a pipe going around a 90 degree elbow?

How does that differ from a fluid in a pipe going around a 90 degree elbow?

I believe it would be no different, I think even a simulation using fluids would show the results.

Is there a good fluid simulating software available that could be used ...

Thank you all for the comments and suggestions!

First of all I was thinking that by testing this with CO2 with air around it wouldn't simulate a rocket properly because quite a bit of force would be generated by CO2 pushing against the air, forming a sort of pressure bubble, but considering the right hand side has reduced pressure, the left part would still be more efficient.

I don't know how much force is generated in a rocket from forces other than the fuel being ejected (for example the gasses hitting the flair) but I am assuming it's not very much ...

I've just done a quick simulation of mass bouncing at 90 degrees vs. following the curve path and of course the 90 degree path is more efficient, I realize this is a simplification.

View attachment 297471

Actually, here, you've inadvertently displayed the function of a nozzle. Your diagram on the left displays gas particles that are optimally expanded, so they flow directly normal to the bell inlet. Your diagram on the right is akin to the combustion chamber, or extremely under-expanded flow, as the gas particles attempt to expand outward.

But anywho, this model is incorrect. The thrust equation shown below displays that the force exerted by the rocket is equal to the mass flow of the propellant, multiplied by the exit velocity of the propellant. For your model above, both the exit velocity and mass flow are equal, thus the object should be moving at equal velocity in both models as the object incurs equal force in both models.

(Yes, I am aware the pressure correction portion of this equation has been omitted)

The equation above could also be restated as Thrust = the change in momentum of the gases leaving the engine. Since the mass of the flow isn't different or changing between models A and B, nor is the exit velocity different between models A and B, the two systems are generating equal thrust.

So AGAIN, to the original question: To maximize the force generated by an engine, you need to maximize the exit velocity! Therefore, since the thought-experiment design has losses in flow energy due to the line losses I had previously mentioned, it will generate less force.

escape75
To maximize the force generated by an engine, you need to maximize the exit velocity!
This is admittedly getting a bit more advanced than the original question, but at least in atmosphere, this is not correct. To maximize the force generated by an engine, you want the pressure at the exit plane of the nozzle to match the surrounding ambient pressure. You could achieve a higher exit velocity if you expanded the nozzle further, but this will cause the exit pressure to fall below ambient which causes the pressure thrust term to become negative, and the negative pressure thrust grows faster than the increase in momentum thrust from the slightly higher exhaust velocity. This condition would be known as an optimally expanded nozzle.

In space, of course, this just means to expand it as much as possible (until the extra mass of the longer nozzle outweighs the performance gains from the extra thrust), since the pressure at the exit plane can never fall below ambient.

escape75 and berkeman
This is admittedly getting a bit more advanced than the original question, but at least in atmosphere, this is not correct. To maximize the force generated by an engine, you want the pressure at the exit plane of the nozzle to match the surrounding ambient pressure. You could achieve a higher exit velocity if you expanded the nozzle further, but this will cause the exit pressure to fall below ambient which causes the pressure thrust term to become negative, and the negative pressure thrust grows faster than the increase in momentum thrust from the slightly higher exhaust velocity. This condition would be known as an optimally expanded nozzle.

In space, of course, this just means to expand it as much as possible (until the extra mass of the longer nozzle outweighs the performance gains from the extra thrust), since the pressure at the exit plane can never fall below ambient.
Yep, don't forget about expansion. But when we're dealing with a design like this... I'm just going to omit over-expansion from the equation altogether.

Actually, literally. Look at the thrust equation I shared earlier LOL

Yeah, as I said, this definitely is a bit out of scope relative to the original design lol. I just wanted to add the bit of extra nuance for any interested readers.

I've just done a quick simulation of mass bouncing at 90 degrees vs. following the curve path and of course the 90 degree path is more efficient, I realize this is a simplification.
This simulation only shows how a nozzle works, not the curved design you posted in the first post.
Perhaps set up a simulation where a particle is ejected directly into a straight 'pipe' and then out into space compared to one that is ejected into a 'pipe' that bends 90 degrees before opening into space?

This simulation only shows how a nozzle works, not the curved design you posted in the first post.
Perhaps set up a simulation where a particle is ejected directly into a straight 'pipe' and then out into space compared to one that is ejected into a 'pipe' that bends 90 degrees before opening into space?

I'm not sure the software can simulate this properly, after all the only interaction this simulation shows
is how the 2 masses interact with each other when they're either sent directly out (by 1st making contact
with each other - left side) or out on a curved path (right side).