|
Post by Roger on Feb 26, 2019 19:59:13 GMT
I've decided to make this a separate thread since it's likely to get dipped into and updated occasionally and i'm not ready to make these yet. I will show the detail manufacture on my main thread though when the time comes. The purpose of this thread is to explore how to design injectors from information extracted from two sources as follows.. "All you need to know about miniature injectors and ejectors" - Bob Bramson "Miniature injectors inside and out" - D.A.G Brown Bob Bramson's book presents a theoretical approach, but I find the diagrams and explanations confusing. D.A.G Brown's book shows designs that you can copy, but with insufficient explanation as to how the sizes have been arrived at. Pretty much all the information you need is there between the two books if you can tease it out, but it's a struggle. I can't claim any experience or knowledge of the subject, all I'm trying to do is to pull together the fundamental principles of designing injectors so that anyone can design their own with a reasonable chance of success. Please point out any errors or inaccuracies so I can correct them. I'm trying to work all this out as I go along, so mistakes and misunderstandings are pretty likely! I'm sure many will think, what's the point? After all, there are published designs that work just fine or you can go and buy one. The trouble with this is that there are no scale injectors to copy, and besides, it's fun to figure these things out rather than attribute them to 'magic' and always be a victim of our own ignorance. I should say from the outset that there is contradictory information given in the two books, and any attempts to verify the dimensions given in the second book is fraught with difficulty. In the end, there is some leeway (but not much) in the geometry and sizes and clearly the designs shown work. It's probably fair to say that these have been empirically arrived at and fall close enough to the 'ideal' geometry for them to work. I'll try to work through each cone and how its dimensions can be determined. If errors or new information comes to light subsequently, I'll go back an edit each part as seems appropriate so that the information isn't fragmented. So to kick things off, here is the nomenclature I'll be using. 16 fl oz injector nomenclature by Anne Froud, on Flickr This is the 3D model I've created which is provides a powerful tool for experimenting with the geometry... 16 fl oz injector by Anne Froud, on Flickr ... it's created by drawing a sketch of the cross section of all of the cones and then creating a solid from them by rotating that about the centre line. The key thing to take home from this messy sketch is that certain geometries and proportions control the cones and gaps. For example, the Mixing Cone bore will grow if the gap to the Condensing cone is reduced. 16 fl oz injector cross section definition by Anne Froud, on Flickr I propose to look at this from left to right so let's start with the Steam Cone...You'll notice that there's no lead in on the LH end of the steam cone, it starts with a parallel throat diameter. Bob Bramson says this is all you need and I believe him. It's easier to make it this way, so that's what I'm modelling. D.A.G Brown shows a convergent/divergent nozzle here, complete with a bell mouth. In fairness to him, he does state that what he's using works and he sees no reason to change it. However, it adds more complication, so if it isn't necessary then it makes sense to leave it off. Another thing to appreciate is that the Steam Velocity increases from left to right in the Steam Cone. This is a Convergent/Divergent de Laval nozzle like you see on Rocket Exhaust nozzles, and it's because there will be Supersonic flow at the throat and from there on it's compressible flow. Although it's counter intuitive, the area increases but so does the velocity of the steam. An angle of 9 degrees seems to be used universally. The aim is to create a parallel jet of steam. (NB:- you can't apply bernoulli's equations to this, it's compressible flow. Reading D.A.G Brown's book on page 14, it doesn't appear that he appreciates this.) It does seem that the length of the Steam cone needs to be between 4 and 5 times the throat diameter, although at least one design uses less than this. Presumably the cone needs to be long enough for the velocity to increase, but not so long as to become unstable as it approaches the point where the flow eventually expands so much that the flow begins to turn subsonic again. The size of the throat in the Steam Cone drives the rest of the design because it controls the amount of energy available for any given pressure. (Bob Bramson shows a graph called Fig. 20 on Page 19 of his book which shows the relative sizes of the critical throat diameters for the Steam Cone, the Mixing Cone and the Delivery Cone for any size of injector.) I'm following Bob Bramson's recommendation to use End Regulation rather than the usual Annular Regulation. Annular Regulation results in a wafer thin end to the Steam Cone, and this is easily damaged and erodes away. It's also difficult to measure and control the annular gap. End regulation solves all these problems, all you have to do is know what the gap needs to be. The nose of the Steam Cone is blunt so it could potentially be made from something quite soft. (More on this later) That's enough for one session, I'll look at what I've discovered about the 'Combining Cone(s)' next.
|
|
|
Post by chris vine on Feb 26, 2019 23:25:28 GMT
Hi Roger,
That is a great project to get your teeth into. I am looking forward to seeing what you work out.
One plea: If possible, can you please use SI units for everything? If I remember correctly, at least one of the books uses very old fashioned units which is a shame and makes calculations very difficult.
Yours Chris. PS, if anyone is addicted to old units, you have to remember that one BTU is the energy required to raise one pound of water through 1 degree Fahrenheit. And that 1 degree F, is 1/68 of the difference between freezing water and the temperature of Dr Fahrenheit's dog, (so I was told by my physics master), you can see that it isn't a perfect system...
|
|
|
Post by runner42 on Feb 27, 2019 5:53:40 GMT
Hi Roger,
an interesting topic, one that I have had little success with, both manufacture and purchase. Isn't a pity that simulation software doesn't exist, similar to valve simulation software, where one could vary aspects of the design to see what effect changes make to the overall performance. Or is there certain dimensions or characteristics which are fixed and immutable. I had a recollection that Julian said that the output cone angle must be 9 degrees, apologies in advance Julian if I have misquoted you.
How would you rate the performance of an injector, is it its ability to operate over a range of steam pressures or are there other characteristics like the ratio of the amount of water injected over the amount of steam required? In operation does all the steam after doing its work condense and contribute to the water being outputted from the injector, because at pickup the output from the overflow stops or nearly so, so it must be going somewhere?
Another point that has interested me is for a given boiler pressure which would be the pressure entering the injector say X psi, will it always produce a water pressure output greater than X psi to overcome the back pressure at the boiler clack and put water into the boiler?
Brian
|
|
|
Post by Roger on Feb 27, 2019 8:01:04 GMT
Hi Roger, That is a great project to get your teeth into. I am looking forward to seeing what you work out. One plea: If possible, can you please use SI units for everything? If I remember correctly, at least one of the books uses very old fashioned units which is a shame and makes calculations very difficult. Yours Chris. PS, if anyone is addicted to old units, you have to remember that one BTU is the energy required to raise one pound of water through 1 degree Fahrenheit. And that 1 degree F, is 1/68 of the difference between freezing water and the temperature of Dr Fahrenheit's dog, (so I was told by my physics master), you can see that it isn't a perfect system... Hi Chris, I'll mention the metric volumes and cross reference the Imperial capacities in the descriptions. Unfortunately, all of the design data only exists in imperial units, but that can be converted for our purposes. That's what i've done for the 3D model. All temperatures will be in Celcius. In reality, you only need to look up the throat diameters from that one graph for the delivery rate you require, and the rest is an exercise in geometry and drawing.
|
|
|
Post by Roger on Feb 27, 2019 8:05:41 GMT
Hi Roger, an interesting topic, one that I have had little success with, both manufacture and purchase. Isn't a pity that simulation software doesn't exist, similar to valve simulation software, where one could vary aspects of the design to see what effect changes make to the overall performance. Or is there certain dimensions or characteristics which are fixed and immutable. I had a recollection that Julian said that the output cone angle must be 9 degrees, apologies in advance Julian if I have misquoted you. How would you rate the performance of an injector, is it its ability to operate over a range of steam pressures or are there other characteristics like the ratio of the amount of water injected over the amount of steam required? In operation does all the steam after doing its work condense and contribute to the water being outputted from the injector, because at pickup the output from the overflow stops or nearly so, so it must be going somewhere? Another point that has interested me is for a given boiler pressure which would be the pressure entering the injector say X psi, will it always produce a water pressure output greater than X psi to overcome the back pressure at the boiler clack and put water into the boiler? Brian Hi Brian, I'll hopefully cover all of your points during the explanation of what I think these books say. There is leeway on angles, lengths and gaps, but both books seem to agree on a sensible sweet spot in the middle that works well, so that's what I'll use. It's a really interesting subject that has way too much mystery about it in my opinion. The idea is a simple one, but making it work well needs a bit of care and forethought as to how it works and where the problem areas are.
|
|
|
Post by 92220 on Feb 27, 2019 8:41:59 GMT
Hi Roger.
I shall definitely follow this tread with interest! Injector design is a mystery to most of us. I have to make at least one of my injectors as nobody makes a working Type K outline injector.
Hi Brian.
Actually there is software available to help with designing injectors. I don't know anything about it but if you google software to design steam injectors you get a whole lot come up. Mind you, I would guess most of the software is commercial so would be expensive.
Bob.
|
|
|
Post by 92220 on Feb 27, 2019 8:50:11 GMT
Hi Roger.
I am a complete dunce as far as injectors go. Does condensation of the steam flow figure in any way in the operation of injectors? The reason I ask is: could an airline be used to test an injector? That may sound a silly question to most, but if air could be used for testing an injector, wouldn't it be so much easier than setting up (and possibly having to make) a boiler.
Bob.
|
|
|
Post by Roger on Feb 27, 2019 9:01:00 GMT
Hi Roger. I am a complete dunce as far as injectors go. Does condensation of the steam flow figure in any way in the operation of injectors? The reason I ask is: could an airline be used to test an injector? That may sound a silly question to most, but if air could be used for testing an injector, wouldn't it be so much easier than setting up (and possibly having to make) a boiler. Bob. Hi Bob, The answer is absolutely yes, in fact ALL of the steam needs to be condensed else the resultant column of water isn't solid enough to impart the necessary force. Air would probably make the ejector part work and pick up water, but it has nowhere to go after that. The steam effectively disappears, air can't. I'm afraid you'll need a test boiler, I will probably make one based on a kettle element, so watch this space.
|
|
JonL
Elder Statesman
WWSME (Wiltshire)
Posts: 2,906
|
Post by JonL on Feb 27, 2019 9:16:54 GMT
Really interesting to see someone experimenting rather than relying on old data solely. I look forward to reading this.
As a secondary point (and sorry if this deviates from the core of the subject a little) how do you machine the internal cones? Make a cutter of the correct angle?
|
|
|
Post by Roger on Feb 27, 2019 9:49:05 GMT
Ejector design...The Bramson book explains in detail about ejectors, but in a nutshell you need the two cones on the left to have throat diameters A and B in the ratio of between 1.38 and 1.45 (ie B/A=1.42 say) He states that anything less than 1.35 will not raise a vacuum at all. The angle of both cones is 9 degrees. Ejector diagram by Anne Froud, on Flickr Given that we have chosen the diameter of the Steam cone throat from the mentioned table, you can now draw this albeit not knowing how long the cones are or the gap between them. This is going to be a theme that repeats itself and why drawing this in CAD is a huge help. On my model, I name the Steam Throat Diameter and give it a value. The Condensing cone throat diameter is set as 1.42x that value. The cone angles are drawn and set to 9 degrees. I drew the Condensing cone at some arbitrary length which we'll come back to later. That can't be decided yet. You can see straight away something obvious though. Clearly the longer the Condensing cone, the larger the diameter on the left will be. The same goes for the steam cone, the right hand diameter will get bigger, the longer the cone. These cones will need to be long enough to allow the flows to develop, and we'll see more of that later. The main thing to note is that the outlet (RHS) of the steam cone needs to be smaller than the inlet (LHS) of the Condensing cone. You can picture extending the expanding cone of steam as it exits the Steam cone, and that must land inside the Condensing cone. We'll set that gap later. So to re-iterate what's going on here. The steam expands and accelerates all of the way through the Steam cone and then exits through the Condensing cone. Along the way it drags some of the air in the gap along with it, causing a vacuum. Now, if you imagine where this steam and air is going next, you can see why there's a gap between the Condensing cone and the Mixing cone. There's a lot of steam and air, a volume that's not going to be able to pass through the throat of the Mixing cone. Some will pass right the way through, but most will escape in the gap. So on this diagram, the green arrows show the air being drawn in and the blue arrows show where the Steam + Air is ejected. Without the gap after the Condensing cone, there would be back pressure and the steam would also exit in the reverse direction where the green arrows are. NB:-At this point there's very little condensing going on, this is wet steam and air going everywhere BEFORE the injector starts Ejector operation by Anne Froud, on Flickr
|
|
|
Post by Roger on Feb 27, 2019 9:54:03 GMT
Really interesting to see someone experimenting rather than relying on old data solely. I look forward to reading this. As a secondary point (and sorry if this deviates from the core of the subject a little) how do you machine the internal cones? Make a cutter of the correct angle? Just bear in mind that I'm reading books and interpreting at the moment. Soon enough I'll make a gaff that someone who actually knows what they're talking about will spot! The D.A.G Brown book shows in detail how these are made, but yes, a tapered reamer of the correct angle is usually made. I presume they also make a tool to create the bell mouth. Some sort of form tool made on the CNC would be my choice for making the bell mouth tool.
|
|
|
Post by chris vine on Feb 27, 2019 10:08:30 GMT
Hi Bob,
The steam is vital to how the injector works. It is the condensation of the steam into water which gives the feed water the punch to get into the boiler. The steam is travelling very fast when it hits the water. Then it condenses onto the cold water and, in being slowed from high velocity to almost stopped, gives its huge momentum to the feed water. It is the law of conservation of momentum.
Air won't condense and so there is no large change in velocity and so it won't impart a momentum change to the feed water.
Hope that helps?! Chris.
|
|
|
Post by Roger on Feb 27, 2019 11:30:46 GMT
Continuing with the design process, we looked up the Steam Cone throat diameter A, and that set the Condensing Cone throat diameter B via the ratio. The Mixing cone...The Mixing cone throat diameter (referred to in Bramson as Combining cone throat diameter since it's then end of the combined pair that make up the Combining cone) is also looked up on the graph. The angle of the Mixing Cone is unsurprisingly the same at the Condensing Cone ie 9 degrees. You can see that as we nail down these throat diameters and angles, the possible lengths of each cone start to become limited. Again, the power of laying this out with CAD can't be understated. You can draw one long cone at 9 degrees and the Condensing and Mixing cones must lie on that cone. In other words, the cones are effectively one cone with a gap between them. Again, the Mixing cone is drawn at some arbitrary length to start with. So, looking at what happens when things start to get going, we have the following in the diagram below. The Steam (Yellow arrow) draws water (Light Blue arrow) into the Condensing cone where it's completely condensed into water, giving up its Kinetic energy to the water. The steam is travelling at Supersonic speed and then within a short distance it's slowed down enormously. The water + condensate exits the Condensing cone and leaps the gap into the Mixing cone because it's travelling very quickly. The result of that is that it drags air from the gap with it which we don't want. Air mixed with the water will make it elastic and it won't provide enough force to enter the Delivery cone. The solution is to provide a non-return valve to stop the flow of air (Green arrows) from mixing with the water + condensate. As an aside, it's clear that the hotter the water, the less quickly it can condense the steam. If the water is very hot, it won't complete the process and there will be elastic bubbles of uncondensed steam right the way through to point C. That's not going to be able to enter the delivery cone, it MUST be a solid column of water, even if it's really hot. Injector starting by Anne Froud, on Flickr So where are we on the road to designing the injector? Three diameters A,B & C are now fixed as are the angles. There are already limits on the lengths of the various cones though. Clearly if there is to be a gap at B between the Condensing and Mixing cones, there's a limit to how long the Mixing cone can be ie how far you move left from throat diameter C before there isn't a gap at all. This seems a sensible place to decide on what that gap needs to be. It has to be wide enough to allow for all the Air and Steam to pass while it's starting, but after that, there doesn't need to be a gap at all. The Bramson book states that this is too small on many injectors and goes on to say what those should be for different sizes of injectors at the bottom of Page 22. Given that we accept the figures suggested in the book, that completely defines the Mixing cone dimensions. That's because we know the diameter of the Mixing cone throat (C) and also the diameter of the Condensing cone throat (B) as well as the angle and the gap. We still don't know the length of the Condensing cone though. Interestingly, he says that the jet of water contracts slightly as it leaves the Condensing cone so it ought to be able to enter the Mixing cone without touching the edge where the cone starts. I understand that breaking that edge (the Linden secret?) can really help make injectors work. This may not be necessary if the geometry of the cones including the gap is good enough.
|
|
RGR 60130
Statesman
Posts: 747
Member is Online
|
Post by RGR 60130 on Feb 27, 2019 15:47:39 GMT
Interestingly, he says that the jet of water contracts slightly as it leaves the Condensing cone so it ought to be able to enter the Mixing cone without touching the edge where the cone starts. I understand that breaking that edge (the Linden secret?) can really help make injectors work. This may not be necessary if the geometry of the cones including the gap is good enough. I believe this phenomenon is the 'Vena Contracta' as seen in connection with orifice plates. I've a feeling that Bernoulli's mass flow equation etc. are applicable to injectors as well as orifice plates. There's plenty of light reading on the subject out there for the dedicated / insomniacs. Reg
|
|
Midland
Elder Statesman
Posts: 1,870
|
Post by Midland on Feb 27, 2019 16:33:05 GMT
Roger Brave man. If you can design an injector that works under a wide range of pressures, temps and dirt etc, you will be our hero!!! D
|
|
stevep
Elder Statesman
Posts: 1,070
|
Post by stevep on Feb 27, 2019 17:14:09 GMT
Roger,
I don't know whether you think that it should be included in your reference work, but Laurie Lawrence wrote a series of articles in Model Engineer about making injectors.
I made about a dozen, following his instructions, and they all worked (some better than others).
|
|
|
Post by 92220 on Feb 27, 2019 17:22:15 GMT
Hi Roger. I am a complete dunce as far as injectors go. Does condensation of the steam flow figure in any way in the operation of injectors? The reason I ask is: could an airline be used to test an injector? That may sound a silly question to most, but if air could be used for testing an injector, wouldn't it be so much easier than setting up (and possibly having to make) a boiler. Bob. Hi Bob, The answer is absolutely yes, in fact ALL of the steam needs to be condensed else the resultant column of water isn't solid enough to impart the necessary force. Air would probably make the ejector part work and pick up water, but it has nowhere to go after that. The steam effectively disappears, air can't. I'm afraid you'll need a test boiler, I will probably make one based on a kettle element, so watch this space. Hi Roger.
Thanks for explaining that. I'll have to make one then.
At least I will be able to make use of it afterwards, for making my cappuccinos in the workshop!!
Bob.
|
|
|
Post by 92220 on Feb 27, 2019 17:35:05 GMT
Hi Roger.
I forgot to add....I had my ceramic fibre delivered today. I did a quick test but have to do a serious test with a piece of red hot steel tomorrow. With 2 layers of 3mm ceramic fibre 'paper', I brought a steel pan of water to the boil, on the induction hob, and the ceramic fibre didn't reduce the power reading. I lifted the bowl another 1/4" and that was when the power was lost, so a decent insulating layer is possible and still have full magnetic induction power.
When the water was hot, I took the bowl and ceramic fibre off the hob and rather gingerly put my hand underneath. I finally rested the bowl on my hand and it only felt warm...definitely not hot, so the insulation is protecting the hob surface. Tomorrow I will try getting a piece of steel red hot on the hob and keep my fingers crossed that it doesn't pack up. I'll report back when I'm sure it all works as I am hoping it will. One thing I might have to do is get a large circular steel plate to cover the complete hob. Up to now I've used odd rectangular pieces of steel sheet and they have all buckled slightly, possibly because they are not circular to match the magnetic field of the hob.
Bob.
|
|
|
Post by Roger on Feb 27, 2019 18:08:05 GMT
Roger Brave man. If you can design an injector that works under a wide range of pressures, temps and dirt etc, you will be our hero!!! D Hi David, I think this has already been done by D.A.G Brown. If you make an injector accurately as per his instructions and dimensions you should get exactly that. The problem is that if you decided to do your own thing and make something that isn't precisely to a published drawing, eg a scale injector or one with End Regulation, that's not so easy. Although the information to do this is contained in Bob Bransons book, I don't think it's presented clearly enough to be readily accessible. Hopefully the thread will rectify that.
|
|
|
Post by simplyloco on Feb 27, 2019 18:51:31 GMT
I believe this phenomenon is the 'Vena Contracta' as seen in connection with orifice plates. I've a feeling that Bernoulli's mass flow equation etc. are applicable to injectors as well as orifice plates. There's plenty of light reading on the subject out there for the dedicated / insomniacs. Reg I did Bernoulli on my army Clerks of Works course when studying steam power generation. The injector maths were 'orrible! John
|
|