Aspirator nozzle

Discussions about rockets, construction materials, adhesives, nozzles, nosecones and fin design.
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Aspirator nozzle

Post by cprobertson1 »

I came across a peculiar phenomenon (actually, it's only peculiar if you didn't expect it to be there!) while testing a 3D printed nozzle at the weekend.

The nozzle had a crack in it - and while it was discharging, instead of water leaking out of the crack, it instead pulled air in (via the venturi effect) - resulting in minor foaming of the jet.

I reckon if we 3D printed a nozzle to deliberately make use of this effect, we could draw in quite a lot of air (obviously at the expense of specific impulse).

Now, I've heard of adding soaps and foamers to rockets before in order to increase the "burn" time: but I'm rather curious about this method of using ambient air to induce the foaming (via the venturi effect) - effectively creating an aspirator nozzle.

My instinct is that by aspirating the mixture, we are retaining the jet in the nozzle rather than allowing it to escape - BUT - this will only increase burn time if the flow rate into the nozzle is lowered; in other words, I suspect we may simply be lowering the efficiency of the engine (using reaction mass at the same rate: but producing less thrust).

If you look at the attached drawing, it gives an idea of what I'm talking about: the jet of water from the rocket passes through a cone-shaped nozzle which draws air into the exhaust stream via the radial orifices: what I fear is that the majority of the thrust has already been created by the time it leaves the inner nozzle: meaning that the aeration of that jet only serves to create drag!

Anybody care to opine?

Ps - I can always build one and test it of course! But I haven't yet build my thrust-measurement jig, so it might be a while yet ;)
Attachments
venturi aspirator.png
venturi aspirator.png (392.28 KiB) Viewed 71 times
User avatar
anachronist
WRA2 Member
WRA2 Member
Posts: 244
Joined: Sun May 07, 2017 2:18 pm

Re: Aspirator nozzle

Post by anachronist »

Seems to me you'd be expending some of the energy of the mass flow to draw in air and mix the air into the stream.

I built a physics-based simulator to do things that other simulators wouldn't. One of the things I tried was to see the effects of nozzle diameter on altitude. For each diameter, the simulator would calculate the optimal fill volume for maximum altitude at a given pressure. Predictably, when the nozzle diameter got too small, the rocket wouldn't even fly (imagine the extreme case with a pin-hole for a nozzle).

Unexpectedly, it seemed that the best nozzle diameter was the the diameter of the rocket body. I recall the difference from a soda-bottle nozzle was small but not negligible. In this extreme case, the water simply acts like a piston in a cylinder, and the cylinder flies off the piston. This approximates a mortar being fired. After puzzling over this, it began to make sense that the most efficient nozzle diameter is when reaction mass isn't being wasted to lift itself, but instead is used 100% to lift only the rocket. In reality, such a rocket probably wouldn't fly as high due to the increased drag at the tail, but I didn't account for that. Anyway, at that point I resolved that I wouldn't spend my time on gardena nozzle designs that restrict the flow rate.

For a PVC rocket, you might consider a launcher with, say, a 1-inch PVC launch tube rather than the usual 1/2" launch tubes used in soda bottle designs. Your nozzle would be bigger then. Assuming you could seal the launcher against the nozzle, of course.
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Re: Aspirator nozzle

Post by cprobertson1 »

I think you've just discovered the spigot mortar! :D

I discovered something similar with a few simulators that threw a division-by-zero error when the nozzle was equal or close to the body diameter. Apparently my rocket will reach escape velocity a few microseconds after launch, and experience several tons of drag. If you plug it into a FEM simulation, my rocket will turn into a pancake, then a toroid, then an expanding disk of plastic fragments under these conditions!

Methinks that particular rocket simulator/model was missing something!

Right - so let's look at this logically - though I fear I may dunning-kruger myself in the foot here! The spigot-mortar design relies on a large impulse right at the start. Ignoring air density (we aren't going thousands of feet in the air most of the time!) - that leaves us with a light projectile that has accelerated a lot - after which it follows a typical ballistic trajectory.

The alternative is to have a lower impulse (a more traditional style rocket) which continues to thrust during the flight - resulting in a powered flight trajectory transitioning to a ballistic trajectory after burnout. Obviously for multi-stage designs, at the very least, the second stage must use this (as there isn't enough mass in the first stage to push against - if you spigot-mortar-in-flight you'll just throw the first stage retrograde rather than propelling the second stage prograde :P)

BUT - for a single stage rocket... hmm... I reckon if you plot the efficiencies for various designs (optimising for altitude) for both the spigot mortar and the conventional rocket design, you'll reach a point where the optimisations overlap.

That makes me wonder if we can't make a conventional rocket, and propel it out of a tube using compressed air - so that you get the mass of the rocket moving to begin with before even starting the traditional burn... in effect, shooting the rocket out of a gun before launch. In fact - I think a lot of land-and-sea-based missile launchers use a propellant charge and launch tube to similar effect xD

Bonus points if you turn the system into a discarding sabot launcher so you don't have to worry about retractable fins.

My current launch system (still in-development) uses a conical launch plate with an orifice in the centre - the rocket is held down on this, with the cone going up the throat of the nozzle - and an o-ring around the outside of the nozzle sealing it down. Seems to work pretty well so far - it meant the same launcher will be usable for up to 40mm diameter nozzles (allowing some clearance around the outside to avoid the water jet blowing the o-ring out during launch - at the expense of weight).

Bonus points for the hold-down clamps making the launcher look really cool!

----[br]

ANYWAY!

Aspirator nozzles :P

I think you may be right! Once I get my force-measurement apparatus set up (I'm thinking a see-saw type jig that the rocket is clamped to - as it goes up, it pushes on the force meter on the opposite side.) I'll confirm just how inefficient it actually is.

My launcher has an abort-valve which currently vents the rocket contents through an energy dissipator (actually, it's a 3D printed howell-bunger valve :P) to avoid eroding my garden during testing - plus, it looks cool. Adding an aspirator mechanism to that might increase the efficiency of the dissipator (that is, to say, it makes it more inefficient at forming a jet of water! Efficiency is inefficiency! I love it!)

I'll create a separate thread for the launcher at some point once I've got it prototyped a bit though - as it is definitely overengineered - just need to check the O-rings will hold for larger diameter nozzles (and make sure the launch clamps are strong enough to hold it down for a high-pressure rocket with a large nozzle). In fact, I'll make that thread just now - I might not have pics of it all working together yet, but it will certainly give you an idea of what I mean :P
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Re: Aspirator nozzle

Post by cprobertson1 »

It really is a pity one can't edit previous posts! (Is there an edit bussing that I've just missed?)

Anyway - found an MIT paper on water rocket calculations - totally going to get on this!
User avatar
anachronist
WRA2 Member
WRA2 Member
Posts: 244
Joined: Sun May 07, 2017 2:18 pm

Re: Aspirator nozzle

Post by anachronist »

That MIT paper is missing a lot of things, although their launch tube calculation is pretty sophisticated. The paper doesn't mention the adjustments needed to the Bernoulli equation, how to account for the effects of water vapor on temperature in an adiabatic expansion, calculating thrust from residual air pressure after the water is exhausted, air resistance, and so on. If you want something that takes into account all variables that are reasonably possible, you might want to look at mine.
I reckon if you plot the efficiencies for various designs (optimising for altitude) for both the spigot mortar and the conventional rocket design, you'll reach a point where the optimisations overlap.
I did that. Basically what you're referring to as a spigot mortar is the same as a rocket with no water, but a launch tube the diameter of the rocket body.
That makes me wonder if we can't make a conventional rocket, and propel it out of a tube using compressed air
That's basically what the launch tube does. It's a tube of compressed air inside the rocket. It gives the rocket a significant initial velocity boost with negligible leakage of water or air. Once the rocket clears the tube, the water thrust phase begins with the rocket already moving at a good clip.
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Re: Aspirator nozzle

Post by cprobertson1 »

That MIT paper is missing a lot of things, although their launch tube calculation is pretty sophisticated. ...... you might want to look at mine.
Aye, the MIT paper was a worksheet for a practical lab I believe :D Might you mind if I borrowed your workings/simulator? Is there a thread on here for it already that I've missed?
That's basically what the launch tube does. It's a tube of compressed air inside the rocket. It gives the rocket a significant initial velocity boost with negligible leakage of water or air. Once the rocket clears the tube, the water thrust phase begins with the rocket already moving at a good clip.
Ah, allow me to explain my thinking :D

A launch tube can only provide a working surface equal, at most, so the area of the nozzle: if the launch tube, however, covered the entire rocket such that the entire lower surface of the rocket (this assumes the nozzle is smaller than the body of the rocket) can be used to provide work.

You would, however, have an engineering problem in making sure the rocket burn starts just as the rocket leaves the tube - if it starts thrusting inside the tube the extra pressure isn't going to do much. You might be able to accomplish this by using a small diaphragm, and ensuring the pressure inside the muzzle is always higher than the pressure of gas inside the rocket (thus the diaphragm will rupture when the rocket leaves the tube and the pressure is lost.

A cork *might* work as well - again you'd need to make sure it doesn't release when the muzzle pressure isn't there though - otherwise you'll just have a regular launcher that happens to be in a tube :P
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Re: Aspirator nozzle

Post by cprobertson1 »

Dang! I done it again!

I meant to ask - do you reckon an anti-vortex plate right above the nozzle orifice (that is, on the inside of the tank) may improve efficiency?

When I was testing the aspirator nozzle yesterday (now used as a dissipator), I noticed that the water in the tank had a habit of forming a vortex - which allowed some of the working gas to pass through the nozzle.

Would it actually help, or does the acceleration of the rocket counteract the effect (it may only be a problem when the tank is stationary, after all - I can't see inside it while it's in flight!)
User avatar
anachronist
WRA2 Member
WRA2 Member
Posts: 244
Joined: Sun May 07, 2017 2:18 pm

Re: Aspirator nozzle

Post by anachronist »

cprobertson1 wrote: Wed Jul 25, 2018 9:42 am You would, however, have an engineering problem in making sure the rocket burn starts just as the rocket leaves the tube - if it starts thrusting inside the tube the extra pressure isn't going to do much.
This is not really a problem. There will always be slightly more pressure in the launcher than the rocket as it rises. The volume of my launcher apparatus is about the same as the volume of air in my bottle. The bottle rises so fast that water doesn't really get down inside the tube until the rocket has cleared it. Even so, I have a valve in the end of my tube (a glass marble in a space bored out wider to fit it, with a wooden stick penetrating two holes at the top of the tube to contain the ball). At my last launch the marble managed to break past its barrier and blow out of the tube.
I meant to ask - do you reckon an anti-vortex plate right above the nozzle orifice (that is, on the inside of the tank) may improve efficiency?
I think so. The rocket undergoes a huge acceleration -- I calculated that mine reaches nearly 200 G. In this paper, equation 20 shows that the height of the vortex funnel gets smaller as acceleration (g) goes higher... but that's for a spinning vortex being driven by a spinner, not by fluid flow through an orifice. this web page suggests that the same is true for a vortex formed by drainage; the size of the funnel is inverse the acceleration.

But is 100 gravities or so enough to ignore the formation of a vortex? I don't really know.

I've always wanted to film a super-slow-motion close-up of the first few meters of a launch so I can see what's actually happening in the bottle. My own simulator, as well as others, make an assumption that the water stays level during thrust, and I know that's probably not quite right, although may be a decent approximation.
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Re: Aspirator nozzle

Post by cprobertson1 »

Hmmm.... My DSLR can manage up to 30 seconds of 1000 FPS recording - BUT there's a caveat (several, actually).

First of all it only uses the middle of the sensor - and it records at - I think it's 400x300px - so if we turn it to portrait mode we can keep the rocket in shot for slightly longer.

Two ticks while I grab some numbers from a simulator - and... okay, so for a typical 2L bottle rocket, burnout occurs after 66ms...

So we'd capture a maximum of 66 frames, IF the rocket stays in frame. Burnout occurs at 1.7m in this example - so clearly it won't be in frame :P Let's assume you have the camera set so it captures the first meter of movement.

If we assume a constant acceleration equal to the launch acceleration given by the sim, then:
a= 632.3 m/s^2
t=??
s(displacement)=1m

As per s=1/2at^2, t=0.056...ish. So we should, in principle, capture as many as 56 frames, assuming the actual acceleration doesn't change too much in that timeframe. Played back at 25fps will give us a little over 2 frames of footage - covering approximately 5/6 of the burn period (minus the air blast). If you placed the camera a bit further back you could manage to capture it all - at the expense of probably not being able to see what is going on. You could also mount the camera a little higher up to capture different "windows" in the burn period.

Totally doing that once I get things built up and tested :D
User avatar
anachronist
WRA2 Member
WRA2 Member
Posts: 244
Joined: Sun May 07, 2017 2:18 pm

Re: Aspirator nozzle

Post by anachronist »

My own simulator shows (for a 200 g empty-weight 2L rocket filled to 0.6L at 100 psi on a 200mm launch tube) that the rocket clears the launch tube at T=36ms, burnout is at T=91ms (the difference being the total burn time of 55ms, close to what you got), at which time the rocket has reached 1.8m and is going 52m/s and experiencing maximum acceleration of 125 G at burnout. It should be simple enough to get a 1.5 meter tall region in the camera view.

The frame rate isn't the issue, it's the shutter speed. Even just need ONE frame that clearly shows the surface of the water in the bottle during flight would be useful.
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Re: Aspirator nozzle

Post by cprobertson1 »

Ah, I didn't use a launch tube when I plugged it into the sim :P

Getting a good frame of the water should, in principle be fairly easy to do (provided there isn't significant motion blur) - the problem is getting enough light on the subject, and getting the camera close enough to capture detail (if we are only using the centre of the sensor, the field of view decreases accordingly - to capture the entire 1.5 meters you'd need it back quite a bit - at which point the resolution is the problem.)

The shutter speed is just the inverse of the shutter rate (t=1/f : f = 1/t): you can't vary one without the other varying inversely - 1000fps = 1/1000 seconds per frame. Period <<get it! It's a frequency joke! No? Not funny? Well... fair enough xD

The time taken to capture a frame must be equal to 1/f or less, which poses a problem - as cameras don't read the entire sensor simultaneously.

For a 400x300 sensor size that's 120'000 pixels that need scanned before the next frame can be taken - depending on how fast the rocket moves the entire picture may be come skewed (per line) resulting in an effect similar to what you'd see if you pointed a camera at a CRT television.

Another way of looking at it is by asking how far the rocket can move in 1/1000 seconds: because I can't be bothered doing maths properly, I'm just going to divide the distance by the time taken, and we get around 2cm per frame: this is a problem, as it takes around 1/1000 (or less) for the camera to scan the sensor - by the time it reaches the end of the sensor the rocket will have moved up 2cm. You can counteract this somewhat by angling the camera downwards, but even then the problem is still there - just we've changed the direction of the distortion. It will, however, hopefully keep the surface of the working fluid in-shot longer than before!

That's a worst-case scenario though - I suspect it reads the sensor in less than 1/1000s, and then spends the rest of the time storing it in ram - after you've captured the footage the video is transferred from ram into your actual storage (an SD card in my case).

We shall see soon, hopefully!
User avatar
anachronist
WRA2 Member
WRA2 Member
Posts: 244
Joined: Sun May 07, 2017 2:18 pm

Re: Aspirator nozzle

Post by anachronist »

One way to minimize blurring is to film it so that the rocket is coming at the camera or going away from the camera. I found a couple of videos that do this and the level of the water is clearly seen.

This first video (you need to view it full screen) shows the water staying rather level during the entire time the rocket is in the frame, which covers most of the burn.


This second video, at the 20 second mark, appears to show a funnel forming in the water just before the rocket leaves the frame. Hard to tell for sure though. At the 55 second mark there's another launch that shows the water staying reasonably level.

The first video I'd consider definitive though, as it shows the rocket during the majority of its burn.

These launches don't use a launch tube. I expect the presence of a launch tube leaves a cylindrical column of air inside the water as the rocket leaves the tube, and that would probably mess things up a bit. Certainly a launch tube would contribute to turbulence as it exits the nozzle.
cprobertson1
Advanced Member
Advanced Member
Posts: 30
Joined: Wed Jul 11, 2018 5:48 am

Re: Aspirator nozzle

Post by cprobertson1 »

Excellent! I'll have a look at those tonight after I go swimming :D (they are blocked by the firewall just now)

I'd be very interested in comparing the dynamics of a launch-tube to a non-launch tube rocket under high speed footage actually - and the effect of adding baffles, anti-vortex plates, and other dampers and devices to the water column (obviously at the expense of weight). I suspect such modifications would improve nozzle efficiency - but the results may be insignificant compared to the weight added by the devices :P

My suspicion regarding the launch tube would be in line with yours - in that it will create turbulence throughout the water column. I suspect it's wake won't draw in too much air from the top though - rather I reckon it will serve to agitate the top of the water column, after which the compressed air will keep the water in contact with it - the rest of its journey will be underwater. I might be wrong though :P Will need to test!
User avatar
anachronist
WRA2 Member
WRA2 Member
Posts: 244
Joined: Sun May 07, 2017 2:18 pm

Re: Aspirator nozzle

Post by anachronist »

cprobertson1 wrote: Wed Jul 25, 2018 9:42 am Aye, the MIT paper was a worksheet for a practical lab I believe :D Might you mind if I borrowed your workings/simulator? Is there a thread on here for it already that I've missed?
I apologize for missing this question. I know this is an old thread. You can find my simulator at https://www.nablu.com/p/water-rocketry.html
-A