12-1- 2005

Introduction

.

What follows are a compilation’s of posts from Paul Lamar’s Rotary Engine NewsLetter, his ACRE website as well as other sites, data from books, scientific reports, opinions gleamed from Mazda and other engineers and information that R.E.N. newsletter members have been kind enough to supply. The quest for making a reliable turbocompounding unit is ongoing. In the true nature of research many paths have been explored, as well as quite a few blind alleys.

After almost a year and a half of searching, theorizing, and eliminating a lot of unpractical and complex ideas, the convincing data was found in our hands all along when we had printed it out. I highly recommend you print out the papers I have noted so you can review them many times. I have taken the liberty to underline some information that was especially valuable to me, I apologize if this is annoying to some of you. This report will lead you to all of the proof you need that turbocompounding can work in aircraft and some vehicles. Actually since the early 1940s until present the concept is still in limited use. Our job is to find and test the parts that make it work in our rotary powered aircraft.

Turbo compounding is simply defined as the mechanism for allowing excess exhaust energy to be captured and applied back into the power shaft of an internal combustion engine. In order to be able to scavenge the excess power that a ‘charged’ or ‘boosted’ engine normally wastes, we first must understand the possible methods of first generating this ‘boost’. For a basic explanation of

turbochargers, and superchargers I recommend this website.

One of the first concepts we explored was electric turbocompounding. I have to admit I got suckered into the "if it is new it must be better" mindset. Upon further study it became apparent that to capture over 20Hp of energy the generator, controller, energy buffer, and motor would be too heavy and inefficient for an aircraft. Briefly, we attach a high speed permanent magnet (PM) generator between an exhaust turbine and the compressor. Controlling the high voltage energy output requires high capacity electronics and some sort of energy buffering device like a battery or ultra capacitors. The power is then delivered it to a PM pancake motor where the flywheel resides to save some weight. Sounds great. Ignoring the high prices of these components lets just add up inefficiency losses. In the order mentioned that would be 95% X 90% X 70% = about 60%. That means out of our target 20 Hp we are only transferring 12hp. Adding up the weights 15lb for the generator, 20+ pounds for the controller, about 35lbs+ for flywheel motor and the need for a some heavier. 60+ lbs. high voltage energy storage system. This is not including heavy wiring runs and the challenge of shielding the high frequency electromagnetic noise. Lovely idea but… not a practical system for a high speed aircraft. The turboelectric system does have merit in land based vehicles with the primarily need being to spin the turbine up rapidly to eliminate turbolag , black smoke etc., and the secondary need to minimize the power load on the primary alternator as indicated in this pdf file. This could save 3 to 5 hps under certain conditions.

Please print this out taking special note of the diagram on page 2.

on the left by the line that represents maximum turbine inlet temperature, and on the right by the line that represents maximum turbo speed. The various fueling ratio curves represent different engine power levels. A mechanical turbocompound with fixed geometry is limited to a single operating line. The addition of variable geometry to the turbocharger will increase the flexibility of the mechanical system."

Go to this link to see how a mechanical turbo compound system works on a diesel engine.

Here a second downstream power turbine captures remaining energy to return it to the crankshaft through gears and a fluid coupling.

Ah!! but will it all of this stuff work in an aircraft? Yes, much of the initial work was accomplished in the early 1940s. Go here and browse down the list to the find NACA report #786

Please print it out. If you are like me you might have to read it several times to catch the abundance of information.

A brief summary of #786 follows.

Performance of Blowdown Turbine Driven by Exhaust Gas of a Nine Cylinder Radial Engine.

By L. Richard Turner and Leland G. Desmon

Summary:

The engine power was decreased a maximum of 1% by the presence of the turbine when at the lowest turbine outlet pressure as compared with engine power delivered with a conventional collector ring discharging to an equal exhaust pressure. No engine power loss was imposed by the turbine with outlet pressures greater than 20inches of mercury absolute. The engine airflow was not affected by the presence of the turbine. At an engine speed of 2000 rpm and inlet-manifold pressure of 33.5 inches of mercury absolute, the turbine power varied from 9 % of the engine power with a turbine outlet pressure of 28 inches of mercury absolute to 21 % of engine power with turbine outlet pressure of 7.5 inches of mercury absolute."

" Preliminary calculations indicate that the greatest net power output can be obtained in a composite engine using a blowdown turbine when the blowdown turbine is used as a first stage of expansion and is followed by a steady–flow device in which further expansion is obtained."

Here is what the turbine looked like.

Three of these blowdown turbines were use in the long distance record breaking fuel efficient transoceanic "Connies" airplanes using the 18 cylinder turbocompounded Model R3350 TC engines.

Report # 822 can be found here. Please print it out.

Brief summary follows.

CALCULATIONS OF ECONOMY OF 18-CYLINDER RADIAL AIRCRAFT ENGINE

WITH EXHAUST-GAS TURBINE GEARED TO THE CRANKSHAFT

By RICHARD W. HANNUM and RICARD H. ZIMMERMAN

SUMMARY

Calculations based on dynamometer test-stand data obtained on an 18-cylinder radial engine were made to determine the improvement in fuel consumption that can be obtained at various altitudes by gearing an exhaust gas turbine to the engine crankshaft in order to increase the engine-shaft work. The calculations indicted that, for turbine and auxiliary supercharger efficiencies of 85 percent minimum net brake specific fuel consumption improvements of 0.357 pound per brake horse hour at an altitude of 10,000 feet and of 0.323 pound per brake horsepower hour

at 30,000 feet can be obtained by gearing the exhaust-gas turbine to the engine crankshaft and operating that engine at a speed of 2,000 rpm, an inlet-manifold pressure of 40 inches [5 psi boost] of mercury absolute, and a fuel-air ratio

of 0.063 [12:1].

The reduction in net brake specific fuel consumption that can be obtained if the exhaust-gas turbine supplies all the auxiliary supercharger power and if its residual power is transmitted through gears to the engine crankshaft as

compared with auxiliary turbo supercharging, is approximately 14 percent at an altitude of 10,000 feet and 21 percent at 30,000 feet. The net brake specific fuel consumption is a minimum is for engine exhaust gas pressure approximately 25 percent above inlet-manifold pressure and varies only slightly from that minimum for a range of exhaust gas pressure from 6 to 45 percent above inlet-manifold pressure.

INTRODUCTION

The use of an exhaust-gas turbine to drive a supercharger at high altitudes is an effective method of maintaining sea level engine power at altitude. Analysis has shown, however, that the waste energy of exhaust gases is recovered more effectively by maintaining an engine exhaust pressure higher than the minimum required for turbo supercharging and thus increasing the work output of the exhaust-gas turbine. The extra turbine power beyond that

required for supercharging Can be supplied to the engine crankshaft through suitable gearing (compound operation).

The purpose of the analysis reported is to determine the improvement in net brake specific fuel consumption that can be obtained if an engine is equipped with a geared turbine and supercharger as compared with the engine using a

standard turbo supercharger. The calculated values specific fuel consumption presented for an engine-turbine combination were based on NACA test data obtained on an 18-cylinder radial engine. Operating conditions for which the brake specific fuel consumption of the combination is a minimum are given. The required turbine-nozzle area is also calculated to indicate the size of turbine suitable for geared operation.

Because the engine, the turbine, and the supercharger have different characteristics, elements designed to give maximum efficiency at some operating conditions are incorrectly matched at other conditions. Provision must therefore be made to obtain satisfactory performance over the entire operating range. The problem of obtaining a wide operating range is briefly discussed. The investigation reported was conducted at the NACA Cleveland Laboratory in the fall of 1944.

-------------------------------------------------------

A key point to note in these reports is that at higher flight levels the turbo compound system is more efficient. There is a greater delta P between the exhaust pressure on the intake of the turbine compared to the ambient (outside) pressure. Obtaining a 14% decrease in BSFC in power at flight levels of 10,000 ft and high cruise power settings and boost levels slightly higher than sea level is nothing short of fantastic. For instance on a 300 hp engine plane cruising on a 75% power setting, this translates to imply roughly 225 hp is in use. If the BSFC is 0.48 then this translates to mean 108 lbs./hour or 18 gals/hr of gasoline are burned. If the BSFC were lowered 14% then the fuel burn would decrease to 0.41 or 15.4 gal/min. This is about the same or better than a TIO-540 engine at those power settings. At higher altitudes and similar power settings the benefits go up dramatically. Some of these benefits however are related to reduction of drag and increase in speed at higher altitudes. Obviously sleek planes that are designed to go high can benefit most. All things being equal the weight of the turbo compounding equipment won’t vary

that much for a 200 hp engine compared to a 300hp engine.

Ok fine you say, the concept has worked in different configurations in diesel powered land based equipment, old radials and inline piston high flying ancient aircraft engines. But … will it work in a modern day rotary. Sure. Why not? For this I’ll quote Paul Lamar

The rotary also has a very robust exhaust. Below is a chart showing the distribution of the waste energy of the rotary engine. We are interested in the enthalpy or kinetic energy available in the exhaust gas which makes up about 40% of the heat output of the rotary engine.

At this point we realized we had several challenges.

1. We needed to have a turbine or turbines capable of supercharging the engine as well as producing power from the exhaust.
2. We needed a light weight speed reduction device that could lower the high turbine rpms to near e-shaft speeds.
3. We needed a light weight simple constant variable transmission CVT to broaden the operational band.

For the first part of turbine design philosophy lets hear from Monty Roberts who’s job happens to involve turbine design.

"Gentlemen,

Some things to consider:

1) Radial inflow turbines as used for turbochargers do not have good characteristics for use as a power turbine. This has to do with the stalled torque values.

2.) Axial flow turbines are good for power turbines because they have a high stalled torque and operate like a torque converter.

3.) Turbines (both types) are typically designed to operate with choked flow through the nozzle.

4.) Turbines (both types) are designed to operate at a given wheel speed relative to gas flow. The solidity, inlet and exit angles, and airfoil type are all based on wheel speed and throughput at a given temperature ratio.

5.) All things equal higher wheel speed results in increased performance and decreased size.

_________________________________________________________________________

Monty again wrote:

"I would love to direct you to a source of turbines. Unfortunately you are going to have to roll your own. If you find a source I would love to know about it, as I need one myself. The problem (one of many) with turbomachinery is that the components must be matched to the application and to each other. It is very hard (impossible) to design a set of stock components. Radial inflow turbines seem to tolerate operation off design point without loosing too much efficiency. They can also be "trimmed" to give different characteristics using the same casting. You change the inducer and exducer dia. and the area ratio of the scroll to get different behavior. Radial turbines are also more tolerant of sloppy tip clearances. They just have lousy stalled torque characteristics, and the efficiency will suffer vs. a radial or axial turbine designed specifically for the application. Because of this "tune-ability", ability to use a series of stock castings, and less sensitivity to manufacturing tolerances, most of the small dia. high RPM turbocharger turbines are radial inflow.

With axial turbines there are also structural problems at high rpm and the flow rates needed for most turbocharger applications. Axial turbines are also more of a point design and there is no easy way to "trim" them for different applications. Axial and radial turbines, if designed specifically for an application are virtually identical regarding efficiency. They work best a different wheel speeds relative to flow and are not interchangeable in application. The other option is the "mixed" flow turbine which is a hybrid of the two designs. Having designed/developed several mixed flow devices, I cannot recommend this, unless you have CFD1, a lot of time, and a lot of money. Or really, really, good luck. I'll just say this: throw away the text books and the equations, because they are worthless. Add to this mix different percentages of reaction vs. impulse and you have a really exciting time ahead, ummm,,,,, I mean, rewarding design experience.

To address Paul's comments on radial inflow turbines. They will work over a very narrow operating range. Like a diesel. The Solar [turbine] unit also uses a radial inflow wheel. The PTO is directly off of the single shaft engine. It is confined to a very narrow operating range. This is acceptable for an APU. For your application I think you would be better off simplifying the PTO drive equipment and using an axial flow turbine. The more favorable off-design torque characteristics should make this possible. If you want to go to higher shaft speeds, you will probably have to use a radial inflow turbine. If you do you will either have to design it for cruise only and just accept that it is inefficient off design or go with a [CVT] to broaden the range. "

To change the operating point of the turbine, you change the inducer and exducer dia. and play with the gas inlet angle.

Mark Wrathall wrote

"I would suggest you start from scratch on the turbine. As Monty has pointed out, this will be far more likely to bear fruit. ……" "I suspect that automotive turbo's survive the high EGT's because they are only exposed to a relatively small mass flow and therefore low heat flow. They are able to reject this heat flow fast enough. As you up the duty cycle, although the EGT is not increasing, you are flowing more heat than the turbine can reject. Thus the need for Inconel or superalloy (MAR-M,RENE... etc). The combination of Wankel [high] EGT's, a high duty cycle, and manned aircraft expected TBO and reliability may well exceed the material capabilities of the super alloys and drive you to a cooled turbine (thus ending the project). Gas turbine engines only combust a fraction of the cycle air, and cool the combustion products significantly before they are passed across the first stage nozzle. We are not afforded that luxury.

For your design point I would see if Tracy could provide the following: At sea level, 65% RPM, the OAT, density, fuel mass flow, and air flow ratio. With this you can work out your mass airflow, and thus the HP the turbine delivers. You can then optimise the trade off between increasing turbine diameter simplifying it's design (and increasing it's cost) or higher speed gearbox (with higher cost).

Why not separate the tasks and use an aircraft turbo, and an axial flow free power turbine?

---------------------------------------------------------------

For calculating the horsepower output of a given turbine .Paul Lamar wrote:

"I had a mental block against this simple equation for quite some time. When I saw it in NACA Report 786 it hit me like a ton of bricks. Potential kinetic energy in the exhaust is M x V squared. Of course!. P sub t is the turbine net HP out.

I now have all the arithmetic tools I need to to size the turbines. Taking the turbine data from the Garrett new catalog and using the efficiency number (efficiency becomes a known in the equation) they provide will give us the HP from those turbines by working backwards. One can also size an axial flow turbine using the same tools."

Hisanori, an ex Mazda rotary engine fuel efficiency expert friend had this say about turbocompounding the rotary.

------------------------------------------------------------------------

Vance an ex rocket engine engineer had this to say about turbocompounding:

"Exhaust turbines are not free power. In order for the turbine to be useful you need a high "spouting" velocity. The 1700 fps sounds relatively reasonable since this would be well below sonic for this temp and molecular weight. Gas turbine initial nozzles are undoubtedly supersonic which requires a pressure ratio above critical (about 2.5) At sea level this would be a backpressure about 20psig (not good for the engine). As you go up in altitude this back pressure for sonic flow becomes more rational, and turbine power is greater, and turbo compounding begins to become worth the complexity and weight. This is just one more reason why turbo compounding has no place in automobile application. Exhaust volume, as you seemed to indicate, is a classic PV/WRT case using total temps (K 0r Rankine) and absolute pressures. The volumetric flows you are using have to reflect power levels and not just swept volume."

-----------------------------------------------------------------------

Paul Lamar has this to say

of fuel with hopefully better technology. If [recently] Caterpillar and Scania can get 10% more HP out of a diesel engine with a much cooler exhaust surely we can get 20% more HP out of a hot blooded rotary engine.

In answer to Paul Lamars’ question about the suitability of using "a 12 inch diameter blow down direct drive turbine running at 6000 to 7500 RPM or the use of a geared down turbocharger type turbine?

Monty wrote:

"Short answer-I don't know. Long answer- it depends. You are probably going to have way too much in the way of tip/boundary layer loss with the large dia. turbine. The flow area will be very small relative to the leakage path. In addition the pressure ratio you can support with an axial design is less than the radial inflow design. The large dia. will be very heavy, but you are replacing a flywheel with it. The only way to know for sure is do a design study on both options and compare the numbers. If you do a 12 in design, to make the blading work, you will probably have to use a partial flow design where only a portion of the blades see exhaust gas, rather than the whole annulus. More leaks, more weight. This is done on some tip driven fan designs to make it work. It is not the best way to go and can only be justified where the inefficiencies and weight gain are overcome by getting rid of a gearbox, and other complexity. My guess is the large dia. is not going to work well, especially at part throttle cruise. I remain open minded, but not very optimistic. Turbines like to spin fast-just not too fast ;-)."

____________________________________________________________________________

"I think you could make one of the larger turbocharger turbines work. You will probably have to trim the wheel to match your situation, and try to get one with blading that is "close" to what you need. Then use a nozzle guide vane rather than a scroll to get the thing to work where you want it to. Lots of testing, testing, testing."

Note this design becomes more analogous to a Francis hydroelectric turbine which is also a hybrid and is closer to my field of experience.

http://en.wikipedia.org/wiki/Francis_turbine

In summary about the turbine; for initial testing it makes sense to modify what is available and those are reaction radial flow turbines like the turbines used on turbochargers. To optimize the design we should probably go with an impulse axial turbine or a hybrid of the two.

Here is an incredibly small, high performance (75 to 80%) efficient supercharger, "driven by a planetary friction drive using rollers and special oil that is said to achieve a transmission efficiency of over [96] percent." At first glance we knew the traction drive was what we were looking for.

It has a step down ratio of nearly 9.5 :1 and is capable of transferring 50hp + continuously and can have power put in or taken out of either end.

I contacted the company immediately, went to their office in Copenhagen and purchases one unit with the understanding that I would keep them informed on our project. The 12.5 pound unit is a superb piece of engineering with one slight problem. The input shaft turns the opposite direction of the output . This was done for a several reasons; one of them to lower production costs and the other to reduce centrifugal loads on the annulus ring. For ease of installing the unit near the front of the rotary engine it is best that the input shaft and output shaft both turn in the same direction. A minor setback but Paul is in the middle of designing a new unit now, and I am pondering how to use it as is. The input shaft is tiny, about 1/4" (7mm) in diameter. The output torque from the power turbine however is very low at high speeds. It has been roughly calculated to be 1.57 ft.lbs using an off the shelf unmodified radial turbine. The traction drive is rated to 120,000 rpm on the input side but we would most likely not use it over 100,000 rpm when coupled with a high speed turbine. Less if our turbine has a higher torque output. At 100,000 rpms with a torque of 1.57 the power potential is 29.89 Hp.

The transfer efficiency is 96%. So we are down to 28.7 Hp.

The CVT.

Perhaps not needed in actual flight application if the turbocompound system becomes a narrow band use design, however a C.V.T. is deemed necessary for R&D. Somehow we have to take the torque and power from the traction drive back to the driveshaft. A strong belt can transfer power at efficiency levels of 98%. Roughly we are returning 28.1 hp. to the e-shaft. This would be only 12.5% of 225 horsepower. A conservative estimate.

Having our traction drive redesigned will allow us step down ratios of about 10.5 :1 This means that our 100,000 input rpm will be reduced to a 9,524 rpm output.

To match it closer to the engine speed (perhaps 7,000 rpms) we will have to use different size pulleys and a ribbed belt or a C.V.T.

Again Gordon came through with some excellent diagrams and photographs of a belt driven C.V.T. commonly used on motor scooters and motor cycles. He also made an excellent suggestion to use a highly efficient Rotrex supercharger to boost the engine and use a dedicated power turbine to turbocompound the engine. While I understand it, this is not a ‘new’ idea at all, the high efficiencies of this supercharger has lead me to reconsider the benefits of using the turbocharging for boost. True the supercharger must take its power directly from the e-shaft, but its high efficiency (close to 80% with model # C30-94) at the air requirement ranges I needed for a P-ported 13B at altitude, combined with a much cooler air output tend to enhance the merits of using it. I must add this is pure speculation at this point in time and can only be confirmed by testing. Check out the "Rotrex technical datasheet C30" document on their website. Weight increase issues aside, this setup would allow a power turbine to be designed specifically for that function alone.

Gordon wrote:

"…..What would have made things even better (speaking of past turbocharging systems) is compressors that could be stepped up to a higher speed in order to achieve better adiabatic efficiency -- more boost with less heating of the charge. But the gearboxes weren't up to it -- and still aren't. That's the neat thing about the Rotrex supercharger. It gives up very little to a turbocharger in terms of efficiency because it is able to spin just as fast, thanks to its ingenious drive mechanism."

"Comparing a turbocharger to this very efficient supercharger we see that a turbocharger at low altitudes is not going to be at its best either -- the pressure difference between the exhaust and atmosphere is not going to be as advantageous as at altitude. This means more back pressure putting a slight drag on engine output. How much compared to the high-efficiency supercharger? That's the question but I wouldn't lose any sleep over it.

Definitely on the turbine wheel for testing but Rotrex has informed me that they are working on a variable speed supercharger that they will bring to market in the summer of 2006. This would allow us to raise the supercharger speed at higher altitudes producing more air without paying a horsepower loss down low. Another possible method is to size the compressor for design level high altitude and ‘unload’ the compressor at lower altitudes. This could be done via an internal air by-pass valve.

For changing the rpm outputs from the traction box to the e-shaft we may want to keep the input rpms. slightly over the engine rpms. so we only need a very small range of ratios from say 1:1.3. to 1: 1:15 We assume the engine will tolerate slight mismatched rpms. Why? This is actually what happens when you turn on your car air conditioner when you are driving up a hill. Or vary the load on the aircraft engine by changing the propeller pitch angle. The engine doesn’t stop it just slow downs and adjusts for the new load. The converse is also true. Turn your air conditioner off and the engine speeds up. Turbocompounding then is like turning off several air conditioners or more precisely stated adding the power these devices would normally consume.

Again from Gordon,

"Above is the front pulley on the scooter and it is mounted to the engine　shown in low ratio. On the exhaust turbine, this would be mounted to the traction drive box. Centrifugal force causes the roller weights to act on the angle plate, which causes the pulley halves [variator] to come together forcing the belt outward, resulting in a bigger effective pulley diameter as seen below:

"Below is the pulley at the other end. On the scooter, this is mounted on the back wheel and on the turbo compound system, this would be mounted at the e-shaft."

"You can see that this scooter pulley also has a clutch, but this is not needed, unless you wanted a way to disengage the power turbine. The two pulleys work together to keep belt length and tension equal as the ratios change. Below are shown both pulleys and the belt at the low speed setting:"

You can tailor the action of the CVT by changing the contra spring and roller weights. The idea is to dial in the action for both the turbine pulleys to match the operational envelope. Paul has suggested using a computer and electric servos to control the pulleys, and this will obviously provide the ultimate in programmability, but you could dial in the system quite nicely I would bet with just the mechanical means. The pulleys from some of the bigger scooters (350cc or bigger) should be able to handle the power we're talking about. Here's a picture showing the components of the rear pulley form a small Aprilia scooter:

In any case some type of rubber belt must deliver torque from the traction box to the e-shaft. With little penalty in weight that belt could easily be the simple scooter CVT system. An exhaust bypass valve may be lighter and safer then an unloaded turbine using a clutch.

Present project status:

The 13B test engine in Japan is torn down and undergoing rebuild with P ports, lightweight high compression rotors, two additional spark plug holes as well as water flow enhancements. A custom designed hybrid turbine is being studied with the purpose of being attached to a non modified ROTREX traction drive.

The quest for creating a reliable turbo compound system fitting the Mazda series of rotary engines is ongoing. I would like to thank Paul Lamar and all the people who have assisted this project with their advice, opinions, and good will.