britishrocket's bipropellant GOX/ethanol thread

[Richard OConnell]

Good to see you back. I'm not going to lie, I was getting worried. When someone is building a rocket and disappears for a couple months, the imagination tends to wander... Glad you are safe

[Johansson]

Great to hear from you again Carl!

As the outside temps keep dropping and the Speed Weekend getting closer and closer I feel like it is time for me to blow some dust from my rocket engine build as well so any progress report from you will be great inspiration!

[britishrocket]

Thank you for that Richard.

Safety is my watchword. My time has been claimed by all the seemingly 5 minute jobs that need doing after buying a new house. The good news is that I now have a much better workshop.

I am hoping to get some work done on the engine in the next few weeks. I will be posting on my blog but I will let you know here too.

Best Wishes,

Carl.

[britishrocket]

Hello Anders,

Great to hear from you. I feel the onset of winter too. My new workshop will be much better in that regard. It wasn't much fun in my old one when it was -12 degrees celsius outside!

Best Wishes,

Carl.

[mator]

Summer holidays have finished))) I am glad to see all of you here. And I hope to see the progress of work and some interesting photos/videos)))
Carl, you live in the north of Britain?

[britishrocket]

Hello Mator my Russian Friend,

It is good to see you too. Yes I live in the North of Britain, in the Scottish Highlands.

I will be posting some more photos of the injector experiments and some video too, so look out for that.

The design for the configurable injector is complete. It allows to easily change the unit to try out different geometries, from gas centred shear, liquid centred shear, gas centred swirl or any other combination thereof.

I feel that the final design will end up being a liquid swirl on the annulus with a straight gas entry through the post. This so called "gas centred swirl" approach seems to be the current thinking in staged cycle combustion engines, where gasified LOX is the oxidiser.

[britishrocket]

Hello All,

I have just published another post on my blog concerning my Shear Coaxial Injector. This post deals with the physical principles and mechanisms that govern coaxial atomisation. The next post will show the actual experiments done. The analysis will use the principles presented in the present one.

You can see it at www.britishreactionresearch.blogspot.com

Thanks,

Carl.

[stevep]

Hey Carl,

Thanks for the update--always love reading about your research, please keep including it (it's a lot of work writing it all up, I know). I look forward to the cold flow results!

I've been rounding up some of the papers you reference and will be wading through them in the days ahead. I've done a massive amount of reading on impingement injectors but very little on coaxial ones, so the coax stuff is very interesting to me.

The other day I whipped up a small coax injector for my 50lbf thrust motor--it's somewhat smaller than yours with an annulus "thickness" of about .005" (0.13mm) and a center-post hole of about .07" (1.8mm). Some time ago I ran across a nitrous injector where the nitrous was run through the center hole and the fuel through the annulus, the reverse of what is commonly done--it makes for a "thin" annulus, but it's still manageable, so that's what I'm doing. I therefore noted with some interest that you had changed from your original design and were now, judging from your comment in post #42 above, running your gox through the center post. I'm sure there's an interesting story in there somewhere :-)

Two things that I've been wondering about and would like to hear your thoughts on are the following:

1. Everyone seems to use water for cold flow tests, but the surface tension of water is 3x that of most fuels (alcohols, kerosene, diesel, etc.) and I'm wondering if you've run across any studies that compare water breakup with the breakup of fluids with less surface tension--I'd think the results would be significantly different, but none of the formulas for drop size seem to take this into account.

2. Drop size and mixing measurements. I have been trying to come up with some low tech ways of measuring these. Mixing of two liquids is fairly easily measured but with a gas and a liquid it is much, much harder. Drop size, I keep thinking, should be measurable by simply measuring how far the drops go in still air--use the initial velocity and aerodynamic drag to compute how far drops of a given size should fly. I have also considered photographic means, but haven't done much with it so far.

Finally, a minor nit: in your blog you state "The Reynolds number of a fluid gives a measure of the ratio between aerodynamic and viscous forces. The Weber number relates aerodynamic forces to surface tension." I think the first "aerodynamic" got in there by accident and what you probably meant to say is that the Reynolds number is the ratio between *inertial* and viscous forces. At least that's the common characterization of it.

--Steve

[stevep]

Never mind about the surface tension--I see it is covered by the Weber number. I don't think the Weber number arises in the context of impinging flows....

--Steve

[britishrocket]

Hello Steve,

Thank you for your kind comments regarding my blog. I am trying to get all of my research notes posted and effectively "catch up" to where the project currently stands in terms of the development of the injector. It does tend to take longer than I would like, and when I have free time I have to admit I'd rather be spending it in the workshop, doing the practical stuff!

Early on in the life of the project I looked at impinging jet injectors and decided against them. I wasn't keen on having to drill angled holes at small diameters with the attendant problems of stream mis-impingement and hydraulic flip. The coax injector seemed more amenable to fabrication, plus it fitted in with my use of GOX and the need for benign chamber wall conditions.

My use of the term "Aerodynamic" was not accidental. Reynolds number is the ratio of dynamic pressure to shearing stress. Aerodynamic force is directly proportional to dynamic pressure. Hence my use of the term.

Your idea for gauging the drop sizes based on how far they fly is interesting. The only problem with this in terms of the coax injector is that the droplets tend to coalesce as they collide and so very often the droplet size distribution shows larger drops further downstream of the injector face.

I must have in excess of 30 papers regarding coaxial injectors and air blast atomisation. The only one I can think of that takes differing surface tension into account is CR120936 by Falk and Burick. They show an expression, based on earlier work (Ingebo) for correcting drop sizes based on density, viscosity and surface tension.

I agonised over droplet sizes quite early on in my research. I soon realised that the coaxial shear injector is capable of producing extremely small drop sizes indeed. The final iteration of my shear coaxial design, shown in one of my earlier posts here, produced a spray so fine that it almost looked like a fog. Drop size decreases with decreasing surface tension and so I was not too worried about using water as a simulant.

I have looked into the various methods for measuring droplet size and mixing. As you say there is no simple means of measuring the latter. Various laser/optical techniques can be used to measure mixing (e.g. Salgues and Mouis et al). Regarding droplet size, I corresponded with Dr. Matt Stickland, of Strathclyde University in Glasgow. He confirmed that laser phase doppler anemometry is the current weapon of choice for particle size measurement. Needless to say, this equipment is very expensive, in excess of £100k.

Dr. Stickland also put me on to another method that on the face of it should be within reach of the amateur workshop:- Shadow sizing. The idea here is to illuminate the spray from behind with a flash gun and take a photograph of it with a short depth of focus. The resulting image will be a shadow of the spray frozen by the flash. Using software or manual means (and some perseverance) the droplet sizes can then be measured.

I haven't been too worried about the droplet sizes in my experiments thus far as they have quite patently been very small. I have gained the most information from carefully studying slow motion video of the cold flow experiments. That said, it would be extremely interesting to co-relate actual drop measurements with calculated predictions. So I have just about got everything together that I need to set up a rudimentary shadow sizing system and I do intend to try it. I also have a piece of software that I think can be pressed into service to perform the drop measurements.

As you mention, the next iteration of my injector design will be gas centred. I am also using swirl on the liquid portion, as well as a few other means to make for more efficient and earlier atomisation. The idea behind having the gas delivered through the centre is to keep the oxidiser away from the walls by totally surrounding it in the liquid fuel. In most liquid centred shear coax injectors the fuel is the gas (e.g. Methane in Falk and Burick).

I am extremely interested in your adoption of the coaxial injector. It is an excellent choice when one of the propellants is a gas or can be easily gasified. I hope to get my shear coaxial injector cold flow tests up on the blog next week. I also have to complete a repair on the power feed of my milling machine, but hopefully I'll fit it all in.

Thanks again for your interest in my blog and your kind words. Wishing you well with your own project,

Carl.

[britishrocket]

An Addendum,

Post 39 on this thread shows the image of the shear coax injector running.

Best wishes,

Carl.

[stevep]

Hey Carl,

I'd have replied sooner, but as you can see, I've been busy :-)

Aerodynamic: OK, I follow what you're saying.

Drop sizes: Yes, I studied your photo before I made my comments. Unfortunately it's hard to see just how fine the mist is from the photo. My concern with using water was that if the drop size was too large with water, what assurance was there that a substance with less surface tension would produce drops *small enough*--clearly the drops would be smaller, but I couldn't convince myself that they'd be small enough.

Question: did you see a large difference between using argon and shop air?

Having done some rudimentary, proof-of-concept photography of drops captured on a clear plastic, I'm very interested in seeing/hearing about your efforts in this area.

--Steve

[britishrocket]

Hello All,

I recently published a new post on my blog showing some of the shear coaxial injector tests I carried out recently, and some photographs of the sprays. I'm using a very useful piece of software for measuring the droplet sizes in the pictures.

Most of the following description of what I've been up to is culled from an email I sent to Steve, so, apologies Steve for having to read my meandering jottings twice!

I had my air pressure up at 100psi so I had to accept a set choked mass flow for the air and use a needle valve to reduce the water flow to achieve some verisimilitude. This way I was sure of a decent velocity ratio. What I found was generally what the theory predicts, and what you have largely discovered yourself. Droplet size diminished with increasing velocity ratio and momentum flux ratio. I also found it to be a strong function of water orifice size. No real surprises there.

The sprays produced contained a wide variety of particle sizes, from fogs of miniscule particles that I could not resolve, right up to a few at 0.5mm. The vast majority of resolvable drops were in the 200 - 350 micron range. There was what appeared to be an intact or undisrupted length at the start of the spray; it was hard to tell if this was a highly turbulent flow of water or a very dense cluster of droplets. Some of the pictures favour the latter theory. Increasing velocity ratio or momentum flux ratio decreased this length. I also witnessed the spray pulsing, and this can be seen in the photographs. The cone angle increases and decreases with the pulsing. I am really interested in your finding that the pulsing caused liquid mass flow variation. I didn't have a gauge on my water supply so I couldn't tell if the pressure (ergo mass flow) was fluctuating with the pulsing. In the set up I trialled two different recess lengths and increasing the recess increased the pulsing. I believe there may be some sort of recirculation of the air flow happening which affects the water flow.

The really interesting things started happening when I introduced swirl. I swirled the gas through the centre and delivered the water straight through the annulus first. The swirled air ripped into the water stream and produced a compact spray but with coarse droplets. No pulsing. Next the water was swirled through the centre. I had to open up the needle valve quite considerably due to the low Cd of the swirler I was using. A deep recess (2Dl) meant that the thin film produced by the swirler was stabilised by the annulus wall before being hit by the gas. Ideally the swirled liquid should be delivered via the annulus, but this was the only way I could trial this configuration with my current test injector. The result was a marked improvement. The intact length was non existent, the drops were so small I could not resolve them (i.e. "fog" type spray) and there was no pulsing; the swirl seemed to stabilise everything.

This is what I have been trying to get to in the blog. The swirl is the key in this system. Using the pressure of the liquid and tangential channels, a thin film is produced that rapidly ligaments and forms droplets. These are then further atomised by the high velocity gas.

The next thing I plan to do is to explore the swirl concept further and develop better swirl inducers. I intend to do this by cutting some two start trapezoidal threads. I also want to look into the mass flow fluctuation you saw, though as I say the telltale pulsing was not apparent with the swirl. Perhaps the air was slowed when it hit the film of water... Anyway I have some fantastic little stainless steel turbine flowmeters I got from a manufacturer in the south of the UK. I'm going to try to set up a flow indication system based around one of these...and also include a gauge this time.

As I mentioned I have a large amount of data to get through and I am trying to establish how far my observations correspond to the theory. Preliminary results show general agreement. That said, I'm not sure that I want to spend a huge amount of time calculating droplet distributions, as I would far rather be getting on with machining some new parts for a swirl test set up. I now at least have a technique to gauge injector performance and I have confirmation of my earlier experimental results.

More to follow (eventually) best wishes all,

Carl.

[britishrocket]

Just a quick update for you all. I have gone a little further down the road towards a swirl prototype by manufacturing some swirl inducers.

This was done by generating a two start metric trapezoidal thread on a section of round bar. The metric trapezoidal thread form is similar to acme, though with a thread angle of 30 degrees as opposed to 29 degrees.

The photo shows a swirl inducer that has just been machined, along with the cutter. The thread is of 2mm pitch, meaning the lathe was set up to cut a 4mm pitch to produce the first start, and then the compound was set over another 2mm to put the second start in between the first.

What I want to do is to have the liquid swirled through the annulus, thereby creating a thin film which will then be atomised by the core fed oxygen stream. High momentum flux ratio and high swirl are the keys to this system.

[britishrocket]

For some reason I couldn't upload the photo, so I have put it on my blog...

www.britishreactionresearch.blogspot.com

Thanks.