enginehistory.org

[gryan]

The Ultimate Engine to fly would be ...........................(insert details) and would feature ...........................(insert favourite attributes and reasons).

Let's build a paper engine. I guess the best way to do this would be to start with an open to all contenders invitation. Given a fresh sheet of paper and the challenge to design the ultimate engine, what would evolve in 2005?

What better venue to ask this question at than the AEHS? Given all the knowledge that has been assembled in the last 50+ years and with a healthy dose of hindsight we should be able to think up some really impressive big piston engines. The design weaknesses and problems of past efforts can be avoided (as far as possible) and we can use to advantage of all that hard won experience from previous engines. Cooling troubles, combustion chamber issues, rod/stroke/bore ratios, RPM, balance and so on; all these matters can be addressed and the issues properly solved. Or can they?

There are likely to be some debates as preferences for particular design details come to the fore. Which way is better? Of course all the usual design decisions are there to be made and all the options (including the way-out ones) are available as well. Should the engine be an in-line or radial, air or water cooled, two-stroke or four-stroke, compound or not compounded? It could be opposed piston. Perhaps it could be spray cooled or use constant loss evaporative cooling, if either of those provide enough of an advantage. Should it have two, three, four, six or even seven valves per cylinder? What about sleeve valves; surely they are the best solution? There is nothing to prohibit oval pistons (as used by Triumph or more recently Honda). Rotary valves such as NSU's discs, Cross's, Aspin's, Negre's or Coates rotary valves are allowable (but how to make them work?). We could use Wankel's rotary engine as a valve (he did and he even employed it as a supercharger). The poppet still has a lot going for it (desmo?) and is a well known quantity which flows gas well. Then there's the issue of carburetion versus fuel injection versus direct fuel injection. Should the engine be integrated into the airframe (what about having it as part of a ducted fan system with waste heat recovery?) or should it be a power egg? Counter-rotating props; is that part of the deal? Reduction or direct drive? Plenty from which to choose. Perhaps we should stick with what is known and perfect it or is that too limiting?

Here's a basic requirement (we can tighten it up as we go if necessary).

We are looking for the ultimate piston engine design. The fuel is free although MOGAS might be preferred. Water injection is fine and power enhancers such as N2O and methanol are OK, but as they must be carried in the aircraft for the duration and as this one is not a Reno racer there may be a volume & mass issue right there. Let minimum cruise power be 1000 bhp continuous (more would be better, but this is here to avoid the use of automotive engines). TBO shall be minimum 1000 hours (although routine maintenance like plug checks etc. are OK, we just shouldn't be pulling down whole sections of the engine to deal with issues prior to 1000 hours- I guess the turbines have spoiled us!). SFC should be as parsimonious as possible in cruise and not too shocking at max power. Let's expect an endurance of six hours per flight and state the aircraft is a single-engine two-seat (tandem) monoplane requiring outstanding climb and acceleration, not forgetting good turn of speed (400+ mph). OK so it's a fighter (or a hot plane). An alternative use for this engine may be to have four of them in a larger plane. Whatever. It's a paper engine at this stage, a paper airplane is for another day.

Gentlemen, your ideas and knowledge. What would do the job in 2005?

Regards

Gerald

[jrussell]

Well, this ought to bring out a lot of arrows;). As for as very basic configuration, I propose that power to weight ratio is extremely important parameter to maximize. I am also assuming that we are talking about a PRACTICAL engine, not one which requires a elite team of Phd's to maintain. One which could be operated in a normal aircraft evironment ( NOT a SR-71 like device).Which leads to - #1 - Aircooled. Wartime, and more importantly, post war comercial experience, shows that liquid is a heavier, less reliable alternative. Way more maintenence hours. #2 - Turbochargers. Allows power at altitude. With modern turbo's and controls, this would allow the most flexiblity in other design choices. Which leads to #3 - Compound engine. The late 3350's had a bsfc of .36( from memory), which is most impressive. With current technology, like digital fuel and ignition controls, we should certainly be able to improve on this. #4 - Diesel. With a presumably low rpm engine, the combustion characteristics of diesel is an advantage. As well as the reality that 130/145 is not likely to reappear on the market. This also leads to a compression ignition engine, with it's favorable increase in thermal effiency. (Does anyone get the idea I am enamored with the RR Crecy?)
_________________
It runs best just before it blows!

[jjuutinen]

Good start except for air cooling. You can never achieve as uniform cooling with AC as with LC. What is more, IMHO the new engine should meet the strictest emission regs for ground vehicles and I doubt you can meet these with AC. And, given that how bloody rare are total cooling system failures in the millions and millions of cars trucks around today, the reliability factor does not really work in favour of AC engines. And you get far greater freedom for the engine installation with LC.

[jrussell]

The reason I went with air cooling was as much to avoid the ducting losses and aerodynamic and structural issues regarding radiator placement, as it was a reliability issue. I wasn't much worried about a total system failure, as I was small nigling maintenence problems. I get to walk up and poke around a P-38 every weekend, and as much as I love the Allison engines, I wonder how much time was spent taking care of all those cooling system line joints. I have a friend who flew both P-38's and P-51's, I will ask how much of a problem they really were, in a combat environment. I think Porsche over the years proved that an air cooled engine can compete very well on a specific power basis, although on a state of the art engine like a formula 1 engine, liquid cooling is a necessity. Porsche only went to liquid cooled heads ( not barrels) on the 956, when they where producing 240 hp/liter. But are we going for this kind of extreme specific output, or one a "couple of notches" down the ladder? If we are going to through emmissions into the mix, I agree totally, that liquid is a must.
_________________
It runs best just before it blows!

[jjuutinen]

Liquid cooling is a must! But first we should decide whether we have SI or CI? I´d go for CI and add one requirement: multi fuel capability by running on either jet fuel, diesel oil or vegetable based diesel oils. I´d also have electronically controlled injection nozzles using common rail system. It would have a water cooled oil cooler (oil/water heat exchanger) like the Jumo 213 or the P-51H. I´d first try sleeve valve 4-stroke design. It would be a compoung engine.

[jrussell]

Why not a two stroke? It makes a lot of sense on a compound engine. It also helps on the power to weight issue. I definitly agree on the sleeve valve, but was holding back to see if anyone else was going to propose it.
_________________
It runs best just before it blows!

[jjuutinen]

Maybe a two stroke, if necessary!

[gryan]

There are two approaches I thought about.

For the first one I was thinking of a low pressure direct injection engine with hot bulb ignition. The hot builb approach allows the compression ratio of the engine to remain reasonable (low). Hence weight is not as high as required with a CI engine. The high temperature of the bulb ensures the charge ignites (you are not relying on the heat of compression to accomplish the trick). This also achieves multi-fuel capability.

Due to the moderate compression ratio there remains plenty of energy in the exhaust gas. This needs to be directed to undertake useful work. A method to do this is to employ a blow-down turbine. For best efficiency energy released to the engine coolant should be exploited as well. I was thinking of a scheme invented and patented by Alvin Lowi that uses the engine coolant in a type of Rankine cycle. Aside from capturing energy that would otherwise be wasted, this system could also take care of aircraft accessory loads and some of the the supercharger load at cruise.

The engine would be a two stroke and have sleeve valves exactly like the RR Crecy. It would be compounded so that it could operate in two modes. In cruise mode most of the work would be done by the piston engine with a blow-down turbine recovering lost energy. In "sprint" or "max climb" extra fuel would be injected downstream of the engine, immediately ahead of the turbine, and burned to extract more power. Vastly more power still is available if compressed air is by-passed around the engine and burned immediately prior to the turbine. Efficiency would be lowered dramatically. In effect this is a reheat scheme with much of the air undergoing a work cycle with a lower pressure ratio.

A better scheme would be to have the engine drive a multi-speed compressor or even have multiple compressors (engine air and by-pass air delivered at low pressure and high pressure respectively). It could be arranged that most of the compressor work come from the engine. Most of the air would by-pass the engine and be burned downstram. The pressure ratio would remain high. Better efficiency and lots of power for this scheme.

There would have to be counter-rotating props. They make the aircraft behaviour so much better!

As for layout. Easiest is a 90 degree V-12 for perfect balance. An alternative would be H-24 like a Sabre. I've been thinking about how this scheme could work for a radial but have yet to figure out tidy pipework routing for the exhaust (especially for a two row arrangement). It is possible the radial (liquid cooled oddly enough) would be the best package overall and the lightest. Not certain yet.

That's the first engine. Call this one "enhanced conventional." More later.

[kmccutcheon]

Thank you for starting this thread. I think it should provide some interesting insights. I know it has caused me to do some head scratching….

Approach
My approach to this problem is a pragmatic one. I would “like” to have an engine the size of a toaster that produces 10,000 horsepower at a specific fuel consumption of 0.1, lasts for 1,000,000 hours, requires no maintenance, no cooling, no lubrication and costs $10.00. Since I don’t live in a universe where the laws of physics allow such things, I have to settle with doing what I can rather than dreaming of what I can’t.

This means that my hypothetical engine will be based on what is practical. It must neither combine too much new technology nor push existing technology too far. It must use existing fuels that are at the airport today and will be for the foreseeable future. It must be something that is realistically achievable within reasonable budget and time constraints and something that people will want to install in existing airframes as well as new designs.

Given these constraints, let’s see what we have in our technological tool bag that we didn’t have 50 years ago: This includes better lubricants; better materials; digital modeling and simulation; digital data acquisition and control. There are also things we had 50 years ago that we don’t have now: Teams of experienced designers, testers, machinists and mechanics used to productively working with one another in piston engine development; managers willing to take risks; corporate executives willing to invest company resources in long-term visionary enterprises instead of pandering to the immediate gratification of shareholders. We can exploit the former and work around the latter.

Evaluation Criteria: In order to see how well we are doing with our endeavor, we need reasonable criteria for evaluation: I think there is no better single metric than mission specific weight (MSW), i.e. the combined weight of engine and fuel required for a period of cruising, divided by the takeoff power. I would suggest that our paper engine should have a MSW at least as good as the engines from 50 years ago.


Historical Perspective
It is instructive to see what worked 50 years ago when the best and brightest minds last pondered this problem. I have gathered what I believe to be a representative sample of engines from the pentacle of intermittent combustion aircraft engine development, and put them into a spreadsheet (please see http://www.enginehistory.org/Misc/NewEng_1.xls). The list is comprised largely transport engines which had to earn their keep in a dependable, affordable, reliable and maintainable fashion. I would like to have included others, such as the Rolls-Royce Crecy, the BMW 801, DB 601 and Jumo 211, but I could not rapidly locate enough verifiable data for a valid comparison. I do not think additional engines would change the picture anyway.

The weights used in the comparisons started as published dry weights. In order to fairly compare air cooled engines to liquid cooled engines, I added an allowance of 400 lbs to the weight of the liquid cooled ones to cover the additional weight of radiator, coolant and plumbing. I believe that, if anything, this figure is too conservative, but it is my best estimate from the data I have on cooling system components.

Results from the spreadsheet are also available for viewing as three Tables at http://www.enginehistory.org/Misc/NewEng_1.pdf

Sorting on the Specific Weight column, we get the following ordering (listed in increasing pounds per horsepower): 955C9HE1 (an uncompounded R-1820), Sabre VII, R-2800 CA17, R-3350-30WA (a compounded R-3350), 975C18CB1 (an uncompounded R-3350), Hercules 763, R-4360 CB2, Centaurus 568, V-1710-131(G3R), Merlin 724, Griffon 57, and Nomad NNm.7. (Table 1)

This ordering is not surprising. It clearly shows just how bad liquid cooled engines (with the exception of the Sabre VII) are in terms of specific weight. If they are to be competitive, they must either have very high specific power, or make up for the high installed weight with vastly superior fuel consumption.

Sorting of the 6-hr MSW column, we get the following ordering (listing in increasing pounds per horsepower): R-3350-30WA, Nomad NNm.7, R-2800 CA17, 955C9HE1, Sabre VII, Hercules 763, 975C18CB1, Centaurus 568, R-4360 CB2, Griffon 57, V-1710-131(G3R), Merlin 724. (Table 2)

Now that fuel consumption is an issue, the liquid cooled Nomad goes from dead last to near the top of the list. Also surprising is the appearance of old reliable uncompounded R-2800 and R-1820 near the top of the list. Near the bottom of the list is the Griffon, making one wonder how much money the RAF would have saved if Shakeltons had been Nomad powered…

Sorting of the 12-hr MSW column, we get the following ordering (listing in increasing pounds per horsepower): Nomad NNm.7, R-3350-30WA, R-2800 CA17, Hercules 763, 955C9HE1, Sabre VII, Centaurus 568, 975C18CB1, R-4360 CB2, Griffon 57, V-1710-131(G3R), Merlin 724. (Table 3)

This view of the data brings the Nomad to the head of the list and a sleeve valve engine into the top four. Interestingly, the R-2800 has not moved from the number three spot in any of these comparisons.


Toward the Ultimate Positive Displacement Aero-Engine Design

Fuels
It is noteworthy that ALL of these engines, except one, require fuel grade 100/130 or higher, with some requiring 115/145. I believe it a certainty that leaded fuel is going to disappear. When this happens, we will be stuck with motor fuel that has a maximum octane (not performance number) of 96. Referring to the chart (please see http://www.enginehistory.org/Misc/PNvsOct_2.gif), we see that using 96 octane fuel results in a knock-limited power decrease of about 15% for engines that require grade 100/130 gasoline, and a 30% power decrease for engines that require 115/145! Since we cannot expect to make up these kinds of power differences through any cleverness that I am aware of, and since more displacement will not improve on the MSW numbers already quoted, I believe that fuel concerns dictate a compression-ignition engine that burns Jet A.

The 2000 hp Nomad
I believe that the Nomad was well on the way to being a groundbreaking engine, but got out-politicked by Rolls-Royce. So it goes. A smaller version of the Nomad, one that only produced 2000 takeoff horsepower, might fill the bill nicely. With the gift of hindsight, there are probably several things we could do to improve the Nomad. I shall approach this by systematically addressing its weak points and suggesting alternatives (or not) for each.

Cooling: Liquid cooling makes the Nomad heavy. I have never been able to shake the notion that liquid cooling for aircraft is a bad idea. The US Navy and most airlines were adherents to this philosophy. Liquid cooling adds weight and complexity. The systems are a pain to maintain. The coolants are messy and hazardous. Liquid cooling decreases reliability because it unavoidably increases parts count and the number of system interfaces. While automotive technology has improved reliability aspects of liquid cooling, I would guess that about one-third of the cars I see dead on the side of the road are there as a result of cooling system breakdown. I have always viewed liquid cooling as a rather clever way for aircraft engine designers to throw some of their most vexing problems over the wall for the airframers to solve through inspired use of pipes, ducts, valves, shutters, radiators, header tanks, dams, reservoirs, aqueducts, Archimedes’ screws and whatever else they could think of. It is for precisely this reason that I believe liquid cooling is a MUST for our Napier wannabe, despite its weight penalty and other baggage. Development of satisfactory air cooled cylinders is an extremely difficult, time consuming and expensive endeavor. Even if one makes an air cooled engine that cools perfectly at sea level, it won’t necessarily cool well at 30,000 ft. The cost of such development must be amortized over a huge number of engines, far more than we or anyone else will ever build. So despite its problems, the part of me that is a pragmatist thinks that liquid cooling is a must, and that we will have to mitigate its weight penalty in other ways.

Complexity: One thing that makes the Nomad so heavy is that Napier outdid itself “designing the simplicity out of it”. The Nomad II was simpler than earlier versions, but I still see lots of steel, and steel is heavy. Part (all?) of the complexity has to do with the fact the Nomad is a compounded engine. I believe that a compounded engine is a must if we are to achieve favorable MSWs. But I also can see that the means of combining power from the exhaust of the Napier is much more complex than in the Curtiss-Wright. Pratt & Whitney used an entirely different approach – one with no mechanical connection between the turbine and engine power train. With the P&W Variable Discharge Turbine (VDT), exhaust energy was used to drive the superchargers and produce jet thrust. I don’t claim to know what the right answer is, but I do claim to know the right methodology to discover the right answer, and that is to model all three approaches and select the best performer. This approach is a luxury that was unavailable to C-W, Napier and P&W

There is also the matter of the compressor. The axial compressor on the Napier is efficient and achieves a respectable pressure ratio. But it is also heavy. Again, modeling can give insight into whether the complexity of the axial compressor is warranted. It may be that we can do the job with off-the-shelf turbochargers.

Cylinders: Nomad cylinders are loop scavenged. Field experience with large-bore loop scavenged engines shows that differential heating of the cylinder ports leads to thermal distortion of the cylinder and piston scuffing. Uniflow scavenging largely circumvents this difficulty by uniformly heating the exhaust ports at one end of the cylinder. There are two approaches to uniflow cylinders. One populates the cylinder head with a plethora of poppet valves. This was the technique used by Bill Brogdon in Continental’s recent General Aviation Propulsion Intermittent Combustion Engine (please see US Patents 6,032,637 and 6,073,595). Another approach is that of the Rolls-Royce Crecy where the exhaust exits over the top of the sleeve. A fairly recent application of this appears in Greg Stevenson’s US Patents 5,088,285 and 5,183,014. I believe that the sleeve valve offers a reduction in complexity and weight.

Configuration: The flat-12 Nomad configuration produces inertial forces that induce crankcase flexing. Since we need only two-thirds of the cylinders (eight), we have the option of either a 90° or 135° V-8, which should be stiffer, lighter and have shorter couples. Again, modeling can help determine the ideal configuration.

Control: It is now possible to buy fiber-optic pressure transducers that are inexpensive and robust enough so that one can attach one to each cylinder. By sampling the pressure inside each cylinder 50,000 times per second, we can get a very good idea of what is going on during every cycle. Further, by controlling the fuel injection event, we can control the combustion in each cylinder on practically a cycle-by-cycle basis. This allows a level of control over engine performance in all regimes of flight that has never before been possible in intermittent combustion aircraft engines.

Propeller: This is going to be a tough one. A prop that can absorb 2000 hp is going to be huge and will require some kind of reduction gearing because, even at the sedate maximum of 2000 rpm, we are still more than twice as fast as we need to be. Since our engine is intended to power a relatively small airplane, a contra-rotating prop would both reduce diameter and improve aircraft handling. But the contra-rotating reduction gearing is heavy and complex. Additionally, we in the US have never built a truly successful one. Another option is a small diameter ducted fan with no reduction, but this would require completely new aircraft designs. I reckon we might wind up with a contra prop driven through a Griffon-like gearbox.


Whew! That’s all for now.
_________________
Kimble D. McCutcheon

[jjuutinen]

Interesting points! I do have one question, however. Did you take into account the lesser octane/PN requirement of liquid cooled engines? And does that PN chart take into account what Heron wrote about "mild" engines? I.e. that in mild engines the Lean PN becomes the rich PN.

As for power/weight ratios, I calculated some figures of engine intallation weight/power ratios of wartime US fighters based on the data given in Francis Dean´s America´s Hundred Thousand. The results were pretty surprising as the installation weight/power ratios were very closely matched. e.g. using the conservative 67" boost for the Merlin, the Mustang´s installation weighed less than 9% more than that of the Hellcat. If I had used ratings that were approved for the Mustangs in 1944, the result would have in fact been in favour of the Mustang.

And as for coolants being messy and hazardous, is water really messy and hazardous? After all, e.g. in most of the southern states (AL, CO, AZ, TX etc) of the US you could perfectly well use pure water throughout the year.

As for reliability, I have never seen a car which has failed its driver due to cooling failure. On the other hand, I have seen plenty which have had an electric failure.

What is more, liquid cooling is THE choice for lorries, tractors and other similar heavy vehicles. In vehicles with which any unplanned stoppage causes loss of consinderable money.

[kmccutcheon]

[jjuutinen]

I have to check AHt first, but the installation weight I calculated includes cooling system, propeller, auxiliaries etc.

I don´t know what is the first sheet as I cannot open that file.

As for why air cooled engines dominated in the airlines, you well know why:
-most engines in airline service were of US origin->most US engines were air cooled ones->it doesn´t take much analysis to conclude why the AC engines dominated the scene. If we consider the 3 major combatants, UK, USA and Germany (in the field on engine design), it is pretty clear that Germany and the UK were definitely in the LC camp as all late war German project engines had liquid cooling and the most powerful British engines were also liquid cooled (Eagle, Sabre, Crecy). Even in the States the highest power engine, the XR-7755, had liquid cooling.

Please enlighten me: if the ambient temperature at a location is allawys above 0 deg C, how can water freeze? Once it has warmed up during ground running, there is plenty of energy to keep it from freezing even when flying at high altitude.

As for the PN, what is the lean/rich PN rating of standard unleaded Mogas?

[kmccutcheon]

[jjuutinen]

Well, that table seems to be a good general survey. But, e.g. the mission profile should be more specific as the time spent at high power is crucial due to the great difference in rich mixture SFC at high power settings. Also for fairness the power data should be with the same fuel for all engines.

I checked out the AHT and the cooling system weight was given as 663 lbs.

BTW, the Merlin data is quite conservative. The V-1650-9 was cleared for 2240 hp with ADI.

For the Pn conversion table to make any sense at all, it is absolutely essential to know the lean/rich PN of Mogas! E.g. if the the fuel is rated at 95/110, this means (based on Heron´s thesis) that an air cooled engine is at severe disadvantage when cruising at lean mixtures vs. the liquid cooled one.

Why should I shut down the engine in flight in a single engined aircraft? If such a need arose, I´d believe that freezing of the coolant would be the least of my concerns...

[jrussell]

With modern turbo's being capable of a pressure ratio of 5.0, this seems like an area that would give us a big advantage over Napier and the others. It also seems that PW's approach gives more flexibility and less development problems than a directly coupled design. The ironic point is, Napier is one of the market leaders in turbochargers ( can one of there new designs save one of their old ones?). I agree with Kim that reality forces the use of Jet A for this one. It also seems that a 90 degree V-8 is the best compromise as far as overall packaging if the approach is taken to just lop off 4 cylinders from a Nomad, although this leads us straight into the firing order dilemma - a "flat" crank vs a 90 degree one?
_________________
It runs best just before it blows!