Injectors

hi jason,

it is all very simple really.

the speed of the steam issuing from the steam cone nozzle is used to impart pretty much the same speed to the water it hits in the annular gap between end of steam cone nozzle and entrance of the combining cone. the latent heat in the steam imparts extra energy to what becomes a sort of emulsified mix of steam and water and then a jet of water as the combining cone decreases in size. the condensation of the steam in the water takes place in the first half of the combining cone. the speed is converted to pressure in the diverging delivery cone. nothing should emit from the overflow once the injector picks up unless very low pressure (30-20psi) when regulation of the water is required.

yes, it is all very clever!

to make a successful injector in miniature you dont have to know 'why' but 'how', and appreciate the nuances and pitfalls of what makes these things work. ive spent 30 plus years experimenting with these gadgets. all my locos have only 2 injectors that are utterly 100% reliable, and no hand pump or axle pump.

the design parameters are so well known that anyone ought to be able to make their own except that certain tolerances are crucial plus machining proceedures catch out the unwary. making the tapered reamers is also a big draw back. i am often sent 'problem' injectors to sort out due to damage to the cones or bad commercial examples. i have rectified a dozen or so for friends so far this year and have 6 waiting to attend to - one of which is a 'new' commercial product supplied by a well known ME supplier that simply will not work due to the cones being the wrong size proportionately and very badly made.

i have noticed a sharp increase in queries since John Cashmore and Gordon Chiverton died. Len Steel has taken over Gordon's injectors and are available via Pavier Steam. usual disclaimer.

cheers,

julian

Julian'sexplanation is on the right track. The steam is expanded through the steam cone and converts its pressure into momentum. The water then mixes with the steam, and condenses in the combining cone. Mix of water and steam then has lower velocity than the steam jet, but higher density and so has the same momentum. The delivery cone then exchanges the momentum back into pressure to push the water back into the boiler. If the water is too hot it won't condense, so it won't work. The latent heat of the steam doesn't contribute, in fact if it were lower we should be able to use hotter water. This is simplistic as it doesn't allow for the inefficiencies of the nozzle.

Edited By duncan webster on 08/10/2015 01:13:41

Perhaps we could note that the injector is one of the earliest inventions to come out of the aerospace industry, having been invented by Henri Giffard as a lightweight means of feeding water into the boiler of his airship in 1852.

Bear in mind that although all the heat energy that was in the steam does end up back in the boiler (neglecting any dribbles from the overflow) it is at a much lower temperature, having done work. For those who have studied thermodynamics, we can say that the entropy has been increased. If we used the injector to pump water to a height instead on into the boiler, we would find that we have to burn fuel to keep doing it, so work is being done and fuel is being burned, and the efficiency will not be 100% or anything like it. But still a very useful device.

John

For anyone not scared of the maths (or willing like me to glaze over and skip to the end) here's some work that calculates efficiency of about 85%. (it is in English):

uni-obuda.hu/journal/Bukurov_Bikic_Prica_37.pdf

This is a very good thread:-

Quote from Julian

to make a successful injector in miniature you dont have to know 'why' but 'how', and appreciate the nuances and pitfalls of what makes these things work.

How very true I made a batch of five the first one worked a treat so I just nocked the remainder out whoops not a one worked:- traced the problem to a change of brass bar that was a little softer than that what I used for the first:- result was that the D bit belled out the mouth of the steam cone when I cut it, so I changed the machining sequence to machine the OD last so taking out the bell mouth.

just for interest her's a few pictures of the process.

Stew

PS got the pics in the wrong place OK now

Edited By Stewart Hart on 08/10/2015 09:54:15

Nothing like a suitable analogy, much more going on in the boiler, the loop/injector is not a short circuit, unlike the diode in your example. In the boiler you have two elements, water which is uncompresable and steam which is. It is the different properties of fluids and gasses that allows the injector to work.

Injectors are very interesting devices as are diodes, let's not confuse the two 

 

Howard

I think Jason's making a similar point to mine. You aren't going to find normal steam wanting to come rushing out of a nozzle without a decent pressure difference across it to start with. The pressure between the nozzles must be atmospheric (same as the water).

But the point is that without the condensation of some of the steam releasing energy, you have a problem with the laws of thermo, mainly the second one, if you want to return the steam (and the additional water) to the boiler. That phase change has got to be one of the key differences between the injector and a diode!

The optimal conditions are so far from being intuitive that it's no surprise it's such a challenge to get scaled down versions to work correctly.

But we're not trying to put steam back into the boiler, we're putting hot water back in. The following simplified sum ignores nozzle inefficiencies:

If you expand steam from 80 psig through a properly designed nozzle to atmospheric it will have a speed of 1493 ft/sec.

If you then mix this steam with say 10 times it's own weight of water it will have a speed of 1493/(10+1)=136 ft/sec, but the flow is now water. Using Bernoulli equation if we then slow this flow down to rest it will have a pressure of density*speed^2/2*g) = (62.4*136^2/2*32.2) = 17922 lb/sq.ft = 124 psi, which leaves enough to cope with nozzle inefficiency etc. Nothing to do with latent heat, as Howard says it is to do with compressibility.

If you really want to make your brain hurt try an exhaust steam injector where steam at a few psi was used to feed water into the boiler against full pressure. I'm sure I'm not skilled enough to make a live steam injector, but can I challenge these demigods who can to make an exhaust injector?

Edited By duncan webster on 08/10/2015 14:47:21

Edited By duncan webster on 08/10/2015 14:48:29

It would be more like using a 'joule thief' circuit to charge its own battery, the electricity would go round and round, and the battery would gradually get flatter due to the inefficiencies in the system.

In fact its a good analogy - the injector reduces kinetic energy to create extra pressure to get water into the boiler, the joule thief reduces current and generates extra voltage.

Neil

hi duncan,

the principal difficulty with an exhaust injector for a miniature loco is in removing the oil so that it does not go into the boiler. LBSC claimed he had made one that worked, but for the above insurmountable problem.

an exhaust injector in miniature would achieve little or nothing so far as improved efficiency of a miniature loco is concerned due to frequent closing of the regulator. for exactly the same reason exhaust injectors were not fitted to shunting locos etc in fullsize.

cheers,

julian

I never understood why steam locos (almost) never used condensors when almost any efficient stationary engine would have one..

Neil

Got a bit careless with mu brackets in earlier post try

density*speed^2/(2*g) = 62.4*136^2/(2*32.2) = 17922 lb/sq.ft = 124 psi,

Jason,

you can't simply apply ideal gas laws to steam as it is a vapour, ie it can be liquified by increasing the pressure alone. However at pressures above the critical point (210 bar, 374C) the latent heat becomes zero, which has all sorts of good effects on cycle efficiency. Some power stations work up in this range, I wouldn't recommend trying it on your next model loco.

The bizarre effect is that liquid water cannot exist at above 374C, which sounds daft until you remember that liquid oxygen cannot exist above -119C. This implies that as long as the temperature is above 374C you can have steam at whatever pressure you want.

Only if they are English props, boyo

with miniature injectors we unfortunately have to deal with lots of bad practice and faulty accessories.

in the last 9 months ive made new steam cones for someone who used far too strong a solution of citric acid to clean his injectors which a club member of his told him needed cleaning after every run. substantial amounts of the ends of the steam cone nozzles had disappeared as a result!

one friend sent me 2 excellent Gordon Chiverton injectors saying they wouldnt work. i tested them and found nothing wrong with them. his description of attemping to operate and what happened at the overflow was better than most and i suggested his steam valves/passages through same were too small which proved to be the case.

another friend sent me 2 injectors to sort out. one of these injectors had been cleaned by poking a drill through the steam cone - unfortunately it had also gone through the combining cone! the other one had something poked up the overflow and damaged the entry to the delivery cone. replacing combining cones is never satisfactory in my experience so i sent him a new one, tested, and sorted out the other one with new steam cone and delivery cone, also tested. 'wont work' was the subsequent report'! turns out his loco cab obstructed the steam valve handles so that they couldnt be opened more than 200 degrees ie less than a whole turn!

i have frequently had people ask me 'why doesnt my injector work?' 'have you fitted a filter to the water supply side' i ask. 'No' is the reply!

i understand there is something in the current ME about injectors - i havent bought my copy yet - so will be intrigued to see what it contains!

cheers,

julian