Towards a two-stroke turbodiesel aero engine

[malc9141]

Hi:

Under "How do you control an ECU?", I briefly explained what I'm aiming at.
It's all very difficult because I find I'm working alone. Therefore, if anyone wants to help in any way, please get intouch. I'm near Sheffield but distance isn't an insuperable problem.

Many years ago, I was one of those amateurs who built racing cars, and looking back, what we did, and even what we failed to do, gives hope for mankind. Then, everything was more flexible. I could drive in the early hours and have a CVC shaft made up by breakfast time at Hardy-Spicer in Brum, and have it fitted by the evening, with my day's work fitted in between. Bit harder now.

In the back of my mind, I wondered why the two-stroke was so neglected.
I now know why (control of combustion needed the advent of sophisticated high pressure diesel systems) but many of these problems/objections can now be overcome, so it should have a place. But, because of its smoke and particles, it is outlawed for land and sea use.

The other problem is, any poppet valve is used, the camshaft rate is twice that on a four stroke, meaning, say, running at 3000 rpm implies a pretty carefully developed valve train.

But its use for a propeller driven aero engine is ideal.
Do we need such engines? Answer, emphatically yes. Piston enginers are cheap compared to jets, and a propeller is reliable, with certain provisos.
Jet engines are incredibly costly, and simply hitting a flying bird may effectively write off the whole plane by cost.

Initial considerations

Propellers generate thrust according to diameter and revs. But if you increase either revs or diameter, you increase the tip speed of the blades. As they approach sonic levels, bad vibration occurs. So there is a fundamental limit. Whatever you do, you are limited.
That means restricting propeller revs, effectively.

You can get more absolute power by running the engine faster and using a reduction gearbox. Two probs: g/boxes are either notoriously unreliable, or heavy. Both bad. (I don't know why a simple epicyclic g/box wouldn't be OK. I think it would be). But the prop is like a huge flywheel and the g/box between it and the engine, and may suffer.

It would good to have low engine speeds with direct drive. But to get the required power, you need high piston pressure. The two stroke gives that each 360o rather than each 720o. Double the power.

What are we after?

Let's say we chose (about) 2500 rpm, let's say 2400 rpm for simplicity.
That's a nice prop speed to avoid sonic vibration, given typical prop sizes..
Now let's look at what flies at present.
The beautiful Diamond twin aircraft has two A-series Mercedes turbo-diesel four-stroke engines. They weigh about 140 kg and yield about 130 bhp. And they are far too underpowered. So let's say we aim for 240 bhp. Let's think ambitiously of a 3.5 L, 8 cylinder engine (there are reasons for this).

Each cylinder volume is ~450 cc, giving 30 bhp.

Mean Piston Pressure is Power/(Revs*Vol) because Vol is area (which dermines Force) * Stroke (which takes us to Work done).

So, in Watts, Revs per second, Pascals and Metres,
we get
Mean Pressure=1250000 Pa or 12.5 bar.

So that's what we are after with any model that is going to be scaleable to our purpose.

I need to build a single cylinder two- stroke turbo diesel that will provide 12.5 bar at 2500 rpm.

I'll post again in weeks to come, detailing progress.

If anyone wants to get involved in any way, I'll be glad of help, as long as it's not a time waste.

Malc9141

(malcolm351@btinternet.com)

[waynne]

Good luck with the project and thanks for the update.

[malc9141]

Chapter 2

I was lucky in making friends with a Senior Lecturer at Sheffield Hallam University. He had several good ideas. He suggested a Lister-Petter gutsy engine to work from, I had had some dealings with that firm and rang for advice. We learned we wanted an AC1 and an afternoon on eBay located one. This is a 6 hp diesel, floor mounted, to drive things like barges, or concrete mixers.
I proposed getting some 3rd year students to help with the project.
The engine cost about £100 but on dismanting, carried out by one of the students, it proved to have a cracked piston and scratched big-end bearings. Another £100 to replace those, but quickly done. The student then dropped out of the project, getting no support from the University Machine Workshops.
Meanwhile, I designed the layout of the new cylinder-head.
Another student modelled the gas-flow, using Computerised Fluid Dynamics (FLUENT program) and we learned of a problem, which was readily designed out.
This was a big leap forward and on discussing and searching, it seemed we might have something quite novel. Now, the chance of coming up with something new in Engine Design must be close to zero, so I phoned and chatted to a veteran of diesel engine know-how who worked in the Glasgow Lister-Petter agency where I got the spares from. He described something crudely on the same lines, but I couldn't even find reference to that on searching.

I'll talk about this concept in my next post. It's quite exciting and a patent search is in progress.

Meanwhile, there was a lot of serious time-wasting. The students went off to find employment, and our proposal to get a Masters project going fell apart. There were University officers who would talk, but when it came to doing anything, it was always manana (tomorrow). I realised I had been misled.

So the University and I parted company.

Then an ex-student came forward, signed on at Cranfield for a Masters Degree in Mech Eng, and who had his own workshop. He had plenty of ideas, but unbelievably, did nothing. Absolutely nothing. Surrounded by lathes, TIG welders, turret milling machines etc, he did nothing. Another 8 months went by, with promises being excused.

But I had found a machine shop in Sheffield which was willing to build the cylinder head. In the end, however, an order from, say, BMW would always push my job to the back of the queue, and I was lucky to get a block of aluminium squared off with a a few holes drilled at the right place. Nothing happened unless I went and stood there. Clearly, this was not going to be a goer.

The other people I found in Sheffield was a super Cylinder Head Repair firm, over in Wincobank. The boss was very enthusiastic and I learned details of where to source valves, guides etc. He said he should fit the valve seats, guides etc when the time came, but then he got too enthusiastic and wanted me to redesign everything. Where he was wrong, was in not keeping everything in his mind at the same time, so he wanted to take an easier path in one sense, overlooking the fact that it would cancel some other essential. He just wanted a diesel two-stroke. I wanted a certain power output. Twelve point five bar piston pressure, or bust.

Strange happening: a firm of ageing residual Master Cutlers outside Sheffield, open to visitors, proved to have a superb under-used workshop. "No, I could not use it, but perhaps a firm across town could help." They'd done a superb job making part of a huge candelabra for the city cathedral.
And there, finally, I met someone, half my age, highly intelligent, totally at home with his machine tools, who could usefully advance my ideas.

Next post:
The cylinder head design.

Malc

[malc9141]

Chapter 3 - cylinder head design, general.

If anyone wants to throw mud at this, go ahead. If I'm an amateur, there's bound to be stuff I get wrong.

First, if we want lots of power, we will burn O2 and have heat to get rid of.

Back in the "old days", English formula 1 engineers built engines with "large pots" and fewer cylinders because it "gave a simple rugged design." The opposition built multi-cylinder engines. Guess which engines burned valves out.

If we think of an individual cylinder, its volume (roughly) predicts the amount of air in it to burn. So for given revs, volume is proportional to heat generated (power). If we halve the volume, we get half the heat (power). But the combined area of the cylinder walls, piston top, and head-face are not halved. Volume is L^3, area L^2. So the heat "seen" by the walls is less, per unit area. Now heat isn't exactly temperature but it's proprtional to it. So I argue that a smaller cylinder is safer than a larger one. But you have half the power, so you need twice as many cylinders.
Remember, for an aero engine, reliabilty is paramount and overheating or burning exhaust valve seats must be designed out at the beginning.
Common experience tells us what a reasonable cylinder size is, but aircraft engineers are historically accepting of big pots, and petrol engine failures are all too common. (It's interesting how such failures always seem to be explained away as "fuel contamination" or "poor maintenance").

So about 450 cc for a pot seems right. For reasons I'll discuss later, for a turbo engine, we want an under-square configuration. In any case, we can't go for high revs.

That means a bore of, let's say, 80mm.
The most robust design is the Heron head, but since a two-stroke has almost total overlap of valve opening, the incoming air goes straight out of the exhaust. Worse, the stream of air actually impedes the residual exhaust gas from exiting.
What we want is: the standard explosive exit of exhaust gas, with if possible, the well known temporary vacuum created behind its momentum. The inlet valve opens at this point (of course, openings are not instantaneous) and the turbo-pressured air comes in. I argue that we want as smooth an entry as possible with minimum turbulence. (Turbulence is for the birds, and carburettors). We want the air to be as slow and smooth as we can get it (given we have about 7 milliseconds to get it in!). We do not want jets of high speed air. Why? Jets cause a local pressure drop and absorb the surrounding gas (exhaust fumes). It looks good, seeing high speed streams of air but it ends as a high volume, low velocity mixture of diluted oxygen. We want the air to come in like water, and expel the residual burnt gas as if it's oil. And then the exhaust valve closes.
To do this, we have to design a particular shape of intake port. This we did, and built a scale model with a glass partition, which showed the flow patterns when we pumped through high velocity air. We used children's fireworks sparklers to show the flow, and b/w photography to confirm the validity of the design.
I can't show the exact layout because it's the subject of a patent application, but it is simple and robust.
The nice thing is, the water cooling can be concentrated on the exhaust ports, and air cooling we believe will take care of the combustion chamber region. But above all, there is a low total surface area, meaning that heat stress in the head is minimised. (This is not at odds with my remark in para 4. That's heat per unit area, this is total heat pick-up).)
So we have a head with a simple surface, the combustion chamber under the cool inlet valves, and an air-flow pattern that scavenges well (perfectly?). We have done away with the usual inlet ports in the cylinder liner (jets, and bad ones at that), and have a smaller piston (no lomger doubling as a sleeve valve), lowering the whole engine, making it stiffer and lighter.

Next: Building the prototype head.

If anyone wants to get involved with is project, let me know.

Malc9141

[malc9141]

Chapter 4, April 08

A lot of time taken up trying to get the hang of the ultra-high pressure injection control. Plodding on.

Meanwhile, here's some updating of the cylinder head. A diagram of the layout looks simple, and it is. I believe that's the beauty of it.

url='http://www.torquecars.com/forums/vbimghost.php?do=displayimg&imgid=53'][/url]

It's an oversimplified drawing. But remember in a two-stroke both inlet & exhaust overlap hugely. So clean air tends to go up the exhaust spout (note Z). Here, by creating a barrier to airflow towards the exhaust, spent gas is pushed out first. (Don't mention exhaust tuning - impractical in an aeroplane).

But (I hear you say) there's a barrier to a portion of the air trying to get in. True, there has to be (note Z), and does air flow mainly to the side opposite the exhaust, but also out at the side of the valves - and also, not shown - there's a clam shaped bleed (expanding air slightly) on the central side of the inlet valve, letting air go straight down.
So plenty air gets in but shouldn't mix with the exhaust. It should force the spent gas out.

To test this, I made a scale model in wood. It isn't as crude as it looks - the plasticine seals blunt the image. The flow is shown by sparks (which aren't quite like air, but make a good foto). Sparks bounce around. When I sift flour in, the flow is seen beautifully but it doesn't foto well.

You'll notice there's a stagnant zone under the inlet valve head. With modification with plasticine, I could show that seeds, stuck on pins with butter, were whisked off. So there is airflow under the inlet valve also. Other Modifications didn't help much.
Keep it simple.



In the picture, the shovel pushing sparks rightwards is the inlet valve and the tract on the left is the exhaust. The horizontal "barrier" is the ex valve head. All shapes were profiled, tho' the picture doesn't show this. The bottom is the piston top, and a small ramp on this was helpful.

As I get the hang of adding fotos, I'll give some pictures of the head being machined in stages.

Malc

[malc9141]

22 April 08

I'm spending a lot of time trying to get the hang of controlling the injection. I think it's going to be possible.

I show here the prototype cylinder head. It was machined from solid alum alloy. Four valves, with machined tracts. The cylinder size is 325cc and the inlet valves are 29mm od the exhaust 25mm.
The springs were from a Yamaha - and I'm just going to use the (smaller) inside one of the two springs per valve. (The Yammie motor revs to God-knows-what so the dual springs are very very stiff to stop bounce).

When the valve closes, if it slams against the seat, that's bad for the seat, so I think a minor preload is OK. Which is what I've got. The inlet valves are from a VW 20-valve engine, exhausts from a Fiat of some sort. It's a case of needing certain valve stem lengths to suit the angle of the tract + then spring length has to be fitted to suit.

Camshaft - I'll do that next time.

The inlet and exhaust tracts can be seen. The larger inlets (second foto) drop straight down from on top - the two holes on top are the where the inlet manifold will fit. The exhaust holes are at a 40 degree angle so are oval - their manifold would fit into these exit holes.

[HDi fun]

To MALC9141 - why are you trying re-invent the round wheel.

That round wheel of which the Greeks and Romans were eminently proud?

I, for one, am not willing to take a craft airbourne powered by a single cylinder engine.

Whether or not it's fuelled with diesel, petrol, corn oil or potato peelings. One cylinder is not enough.

I wish you luck with your ambitious project; I think you're on a hiding to nowhere; and I think that it won't work.

Having said that, I hope that it does come to fruition,

Kind regards,

Paul Anderson.

[malc9141]

One cylinder is not enough.

Ha Ha. I wouldn't fly in a plane with a 1950s petrol engine, either. Like your average Lycoming.

Actually, I've written earlier on these posts quite a lot about the concept. I personally couldn't build a aircraft engine. I doubt if anyone after, what? - 1910 - could.

An aircraft engine should be designed just for that purpose - not a modified car engine.
There is a three cylinder engine which is not bad but I would like to see a V6 or V8. About 3 L capacity. But to build anything like that is beyond anyone other than an established manufacturer.

No, I wanted to convince a particular manufacturer that a certain design would have advantages that seem to have been overlooked. Central to this "proof of concept" is the inlet valve arrangement (which in my foto does not show final machining but never mind - - later) and gas flow. But there are six other advantages that follow and no disadvantages.

I can test the effectiveness of the idea (prove or disprove - I can't predict if that will work) best with an uncomplicated single cylinder engine. If the idea is right, then I can argue that "this is the way forward." (I doubt if anyone will listen which is the way of the world). But there is a manufacturer - & I know those guys well enough that they just might be interested.

But do you know, it cost Diamond/Centurion nearly a million Euros just dealing with the EU Regulators - proving that their motor is ultra reliable. Not manufacturing - proving it was safe. If it breaks down, the AA is a long way off!

So, no, I'm not building an engine to fly a plane. I'm building an engine which - if it does what I hope - will allow an aero engine to be designed using this concept.

All ze best

Malc

[HDi fun]

Rather than a dual acting piston, how about have two pistons directly opposed in the same bore with a central injector. A bit like a single cylinder from one of the Deltic units of Napier fame. British Rail made good use of these in the Class 55's.

These had two eighteen cylinder (thirty-six) piston engines on a common crankshaft with the generator in the middle.

72 pistons - on a two stroke.

http://rowla.dyndns.org/justin/img/piston_deltic320.mpg

[waynne]

As an aside! Can you theoretically do a rotary diesel engine?