mems rotary eng.imece01

8/8/2019 Mems Rotary Eng.imece01

1/7

DESIGN AND EXPERIMENTAL RESULTS OF SMALL-S CALE RO TARY ENGINES

Kelvin Fu, Aaron J. Knobloch, Fabian C. Martinez,David C. Walther, Carlos Fernandez-Pello*,AI P. Pisano, Dorian LiepmannUniversity of California at BerkeleyBerkeley, CA 94720Phone: 510-642-6554, Fax: 510-642-6163Email:[email protected]

Kenji MiyaskaDepartment of TechnologyFukui UniversityFukui, JapanKaoru MarutaDepartment of Machine Intelligence and SystemsEngineeringFaculty of Systems Science and TechnologyAkita Prefectural University, Japan

ABSTRACTA research project is currently underway to develop small-scaleinternal combustion engines fueled by liquid hydrocarb ons. Theultimate goal of the MEM S Ro tary Internal Com bustion EngineProject is to develop a liquid hydrocarbon fueled MEMS-size rotaryinternal combu stion micro-engine capable o f delivering power on theorder of milli-watts. This research is part of a larger effort to developa portable, autonom ous power generation system with an order ofmagnitude improve ment in energ y density over alkaline or lithium-ionbatteries. The rotary (Wanke l-type) engine is well suited for thefabrication techniques developed in the integrated chip (IC)community and refined by the MicroElectroMechanical Systems(MEMS) field. Features o f the rotary engine that lend itself to MEM Sfabrication are its planar construction, high sp ecific power, andself-valving operation.The project aims at developing a "micro-rotary" engine with an3epitrochoidal-shaped housing under 1 mm in size and with a rotorswept volume of 0.08 m m3. To investigate engine behavior and designissues, larger-scale "mini-rotary" engines have been fabricated fromsteel. Mini-rotary engine chamb ers are approxim ately 1000 mm 3 to1700 mm3 in size and their displacements ran ge from 78 mm 3 to348 mm3.

A test bench for the mini-rotary engine has been developed andexperiments have been co nducted with gaseous-fueled mini-rotaryengin es to examine the effects of sealing, ignition, design, an d thermalmanagement on efficiency. Preliminary testing has shown net poweroutput of up to 2.7 W at 9300 RPM. Testing has been performedusing hydrogen-air mixtures and a range of spark and glow plugdesigns as the ignition source. Iterative de sign and testing o f the mini-engine has lead to improved sealing designs. These particular designsare such that they can be incorporated into the fabrication of themicro-engine.Design and fabrication of a first generation m eso-scale rotary

engine has been completed using a SiC m olding process developedCase Western Reserve University. The fabrication of the micro-rotengine is being conducted in U.C. Berkeley's MicrofabricatLaboratory.Testing of the mini-engine has lead to the conclusion that thare no fundamental phen ome na that would prevent the operation of micro-engine. Howev er, heat loss and sealing issues are key efficient operation of the micro-engine, and th ey must be taken iaccou nt in the design and fabrication of the micro-rota ry engine. Tmin i-rotary engin e design , testing, results and applications will discussed in this paper.NOMENCLATURE

Biehkksolid1nErvrcRtrestehemTmUVswethrhh

Biot NumberEngine EccentricityConvection CoefficientRatio of Specific HeatsConduction CoefficientLengthOutput Shaft SpeedCompression RatioCutoffRat ioRotor Generating RadiusResidence TimeChemical TimeTemperature (Melting)VelocitySwept Volum e (Displacement)Engine WidthTheoretical Compression RatioEquivalence RatioThermal Efficiency

Copyright 2001 by ASME

Proceedings 2001 ASME International Mechanical Engineering Congress and Expositio

November 11-16, 2001, New York, N

IMECE2001/MEMS-2392

EMS TOC


2/7

INTRODUCTIONMicroElectroMechanical Systems (MEMS)devices are mechanical elements constructed usingthe manufacturing techniques and m aterials used inthe integrated circuit / microchip industry [1,2].Typical ly, MEMS fabricat ion uses l i thography tomask the si l icon substrate, fol lowed by etching anddepos ition to create the high asp ect ratio features inthe substrates. The combinat ion of small scale andpotential ly inexpensive mass production is the Intakeprimary at t raction of the technology. Earlyproducts in the MEMS field (accelerometers andgyroscopes) have given way to other , more complex, engineeringsystems (head-mounted displays, opt ical communicat ion systems, andmicro-fluidic systems). As new manufacturing techniques andmaterials have been developed, M EMS-scale thermal devices are nowpractical [3]. These microscale engines have numerous applications,because the power output can be electr ical or mechanical ( torque).Applicat ions range from mobile electr ical power supplies (potent ialbat tery replacements) to mechanical power.

The ul t imate goal of the MEMS Rotary Internal CombustionEngine Project is to fabricate a micro-rotary 'Wan kel ' engine, whichproduces -10-10 0 mW o f mechanical power. The engine chamberwill fit in a cub ic millimeter, with a swep t volume o f 0.08 mm 3. Apower generat ion system using the micro-rotary engine would have5 t imes to 14 t imes the energy densi ty of a p rimary bat tery [4] (seeFig. 1). The MEM S scale "micro-rota ry" engine will be constructedprimari ly of Si l icon (Si) , Si l icon Carbide (SIC), and Si l icon Dioxide(SiO2). All parts subjected to high temperatures and stresses will bebuilt either using molded SiC [5] or a Si substrate with a thin SiCcoating [6].The research program is planned in progressive steps leading tothe fabrication of a MEM S-scale engine. A series of larger scale"mini-rotary" engines, w ith swept volumes o f 78 mm 3 to 348 mm 3,have been constructed to investigate design and combustion issues asthe engine scale is reduced. In addition, fabrication of the microrotaryengine has begun. This paper describes the chal lenges of bui lding asmall-scale rotary engine from test ing, design, and MEMS fabricat ionviewpoints and suggests solutions to these challenges.

t l t l t ~ t l

1 2 0 0 0

10000i=."~ 600O

4000

2000 -

230048O ~Q9

L i q u i d 2 0 % I f f fl c l e r ll Z ln C. dM I" B a t t e ry L l t l d u m I I ~ I d ka l l n eHyc~'ocarbonMtcro-Englne Baaery(octane)Figure 1. Comparison of energy densities betweeninternal combustion engines and primarybatteries.

ROTARY ENGINE LAYOUT AND DESIGNA rotary engine was selected for development as the basis of a

??

iJ)ark.,ire

Tla

Compression Ignition / Expan sion ExhaustFigure 2. Rotary Engine Operation

MEMS -scale power generat ion system due to several factors:- the planar design of the rotary engine lends i tself to MEM

fabrication- the rotor controls the t iming of the intake and exhaueliminat ing the need for the complex valve actuat ing and valt iming systems found in 4-stroke reciprocat ing engines- the power output is in the form of rotary motion of the sha

which is necessary for powering ei ther a micro-vehicle or electric generatorThe rotary engine consists of a t r iangular rotor rotat ing within epitrocho idal-shape d housing (see Fig. 2). The rotor centerlineoffset from the housing centerline by an eccentric, e. As the rorotates about the center of the epi trochoid, al l three apexes of the ro

remain in contact with the walls, forming three sealed chambers. large-scale systems, an elaborate seal ing system is used to preveleakage: 1) over the face of the rotor (face seals) and 2) around trotor ends (apex seals). The seals incorporate d into the present wowil l be addressed later in this paper.The rotary engine operates on a 4-stroke cycle, as illustratedFigure 2. As the rotor apex passes over the intake port, the increasvolum e in the chamber draw s in a fresh fuel / air charge (intake). Trotor continues to rotate, and the next apex closes the intake port acompresse s the fuel / air mixture (comp ression). The fuel / air mixtuis igni ted using a spark plug, glow plug, or compression igni tion. Tresulting pressure rise acts off the rotor axis due to the eccentricity,leading to a torque from which the shaft work is extracted (poweAfter the power stroke, the rotor apex uncovers the exhaust port. the rotor continues to rotate, the exhaust gases are ejected from tengine by the decreasing volume (exhaust). The rotor apex thuncovers the intake port again, taking in a fresh charge. Since thare three chambers formed by the rotor and epi trochoid, this cyoccurs three t imes for every revolut ion of the rotor . The rotor spins1/3 the speed of the output shaft, resulting in one power pulse every revolut ion of the shaft. References to engine speed in this pa

Table 1. Mini-Engine CharacteristicsMNI25 MN50 MN30

Estimated Pow er (W) 125 50 30Operating Spee d (RP M) 30,000 30,000 30,000Displacement (mm3) 348 138 78

Liq. Fuel Con sum ption (ml/sec) 0.017 0.0068 0.0039CO 2 Output (mL/sec) 35 14 8

Heat Output (W) 486 193 109Rotor Diamete r (mm) 12.6 12.3 9.4

Width (mm) 9.0 3.5 3.6



3/7

refer to the output shaft rotational speed.The three primary engine parameters are: the rotor generat ingradius (R), the rotor width (w), and the ecce ntricity (e). Once thes e areestablished, all other engine parameters (displacement, theoreticalcompression rat io, maximum t ip veloci ty, and t ip angle) are alsoestablishe d [7,8]. Table 1 gives basic engine data for severalmini-engines developed for different potent ial applicat ions.

Table 2. Basic Rotary Engine Param eter CalculationsParameter Equat ion

Displacement [7] V = 3,J'3Rwe9,f3eR

TheoreticalCompressionRatio [8]

eth =1 +e2tr +eR(8cosO- 3x~)+ o'2R2 +12e2"3

J

3ewhere sin ~ = Rimum i.Ve loc ity [8] Umax - 1000

where ne = output shaft speed (rev/s)Table 2 lists the basic equations [7,8] used to calculate rotaryengine parameters. Some notes concern ing the compressio n ratioshould be made. High compression rat io in rotary engines requires alow R/e ratio, which means a small rotor radius and high eccentricity.For instance, an R/e ratio of 11.5 gives a theoretical compression ratio

of 30:1. Howe ver, the higher R/e ratio also results in a highermaximum tip angle (the angle at which the rotor apex is in contactwith the epi trochoid chamber) and higher surface to volume rat io atmaximum compression. The higher t ip angle is detrimental to seal ingeffectiveness, as well as seal durab ility [8]. The higher surface tovolume ratio in the combustion chamber will accelerate heat loss.Therefore, every rotary engine design is a compromise betweensealing effectiveness, surface to volume ratio, and compression ratio.The m ini engines have a R/e rat io of approximately 7.

The com pression ratio in a w orking rotary engine is also never ashigh as the theoretical comp ression ratio. All rotary engines have arotor cutout to al low the com bustion react ion to propagate at the onsetof the power stroke. The theoret ical compression rat io does not takethe rotor cutout into account, because the cutouts are different for eachengine.

A series of m ini-rotary engines o f intermediate size (78 mm 3 to348 mm 3) have be en bu ilt and used to stu dy combu stion, fluid, anddesign issues as the engine is reduced in size. The design of the mini-rotary mimics the proposed design of the micro-rotary. Each enginehas different power outputs and builds upon the lessons learned fromthe previous engine.COMBUSTION, THERMODYNAMICS, AND SCALING

The rotary engine cycle can be described by ei ther the Otto orDiesel cycle, depending on the igni t ion source (spark or compressionignition, respect ively) . The eff iciency of both the O tto cycle engineand Diesel cycle engine is dependent on the compression ratio. As can

be seen in T able 2, the compression rat io is dependent only uponrotor generating radius (R) and the eccentricity (e) rather than tlength scale of the engine. Therefore, decreasing the size of an engidoes not affect the theoret ical compression rat io or the theoret ithermal efficiency [9,10]1Otto Cycle: l]th,Otto = 1 k-1 (1rv

Diesel Cycle: l']th,Diesel - (2

For the theoretical compression ratio of 19, the theoretithermal eff iciency of a spark igni ted rotary engine (Otto Cycle)69%. For the same compression rat io, and assuming a cutoff rat io4:1 (typica l for a diesel engine), the theo retical thermal efficiency ocompression igni ted rotary engine (Diesel Cycle) is 56%. Howevthere are other factors that affect the design and efficiency of tengine due to the change from the macroscale to microscale. Issuinclude residence time vs. reaction time, increased heat loss, aquenching. Table 3 provides a qual i tat ive review of the effchanging the leng th scale has on these issues.

Table 3. Effect of Decreasing Length ScalePr o p e r t y

Residence TimeChemical Rxn TimeQuenchingHeat Loss RateThermal Stresses

EffectNo changeNo changeIncreasesIncreasesDecreases

CommentsReside nce time ,,: 1/RPMAt same TWall effectLarger surface to volumeratio & wall effectDue to uniform T

Residence t ime (swept t ime of rotor through combustichamber) in a rotary engine is independent of size. Rather, residence t ime is related to the operat ing speed of the engine. Tpower stroke o f the rotary engine occurs over 90 of the rotor turn. 40,000 RPM, the fuel / air mixture has 375 ktsec to react before texhaust port is uncovered. At normal combustion temperaturhydroc arbons have a ch aracteristic reaction time o f 10 lasec [1Assuming tccs/tchem ---10 for complete react ion, the maximum reacttime limited engine speed for a rotary engine is 150,000 RPM. Sinthe micro-rotary engine is designed to operate at an engine speed40,000 R PM, the rea ction time is not a factor, if the temperaturekept elevated.As the length scale of the engine decreases, the surface areavolum e ratio increases as/-1. A Biot numbe r calculation [12]

hiBi = - - (3k solidBi = 3(104), assuming free convection and a conservative lengscale of 2.5 mm for the engine block. This result suggests that engine can be assumed to be operat ing with a uniform housitemperature distr ibution. The reduced temperature gradient betwethe hot (combustion / exhaust) and cold (intake) sides of the engwil l reduce the effect of thermal expansion of the housing.An isothermal housing implies that measures must be takencontrol the amount of heat added to the incoming fuel / air mixtu



4/7

While pre-heating the intake fuel / air mixture will aid in obtainingcombustion in the microscale, care must be taken to avoid autoignitionin the inlet port. A further concern is that the heat adde d to the intakemixture will reduce the thermal effic iency due to the reduced TH/Trat io of the cycle.Flame quenching occurs due to cooling of the react ion, as well asradical destruction at the walls [13]. The minim um distance in themicro-rotary engine occurs at the point of maximum compression,when the fuel /air mixture is compressed between the rotor cutout andthe side of the epi trochoid. The smallest l inear distance of themicro-rotary engine is well below the quenching distance for any fuel .It has been determ ined that a flame can be stabilize d in tubes wellbelow the quenching diameter with an external heat supply to heat thetube wall [14]. In a supporting experime nt, this strategy has beensuccessful ly employed using only the exhaust gas to h eat the tubewall [15]. Base d on these findings, the obvious solution to thequenching problem is to increase the temperature of the engine wall .The use of a SiC housing helps with this solut ion since there are nomaterial constraints (SiC Tm = 31000(2). In fact, quenching can betotal ly eradicated if the wall temperature can be maintained at theadiabatic temperature o f the fuel.If thermal m anagement is not suff icient to support gas phasecombustion, the combustion react ions can be enhanced throughcatalytic surface reactions. It should be noted that for catalyticreactions, the increasing surface-area-to -volum e ratio leads to overallhigher conversion eff iciency. For this purpose, the housing and rotorwalls can be c oated with a catalyst (e.g., Platinum).

From the point of view of combustion eff iciency, the cri t icalfactor in MEM S combustion systems is the increased heat loss at smalllength scales. Heat loss decreases combustion temperatures, whichleads to reduced reaction rate and quenching. Several strategies arebeing reviewed to reduce heat loss:

- Exhaust Gas Reeirculation (EGR ): A strategy that recirculatesthe combustion gases around the outside of the combustionchamber to decrease heat loss from the combustion chamber andpre-heat the incoming fuel-air mixture. The effect iveness of thisstrategy has been d emons trated [ 16]

- Stacking the engines to reduce the heat losses from the centerengine(s). The center engine(s) will operate at nearly adiabaticconditions due to insulation and the higher melting temperaturesof SiC

- An insulating channel filled with low conductio n material, suchas SIO2, surrounding the engineA signif icant problem facing large-scale rotary engines isleakage, either past the rotor tips or over the rotor face. Leaka gereduces the engine eff iciency (by reducing the compression rat io) andincreases the effect of incomplete combustion. Seal ing mechanismshave been created to reduce this problem in large-scale engines. Theseseal ing systems consist of spring-loaded tabs at the rotor apexes [7]and across the rotor face. The tabs maintain contact between the rotor

and epi trochoid walls and seal ing for the chambers. A pex seals havebeen fabricated in the mini-engine, shown to be effective, anddesigned for the micro-rotary engine. A face seal system has not yetbeen developed that is compatible with MEM S fabricat ion. However,i t should be noted that the smallest commercial ly avai lable rotaryengine ( the 5000 mm 30 . S. Graupner) does not use a face seal .MINI-ROTARY ENGINE TEST BENCHA t the scale of the m ini-rotary engine, there are no comm ercial lyavailable diagnostic engine test stands. Therefore, a test bench (Fig. 3)

was designed and fabricated to test the mini-engine operat ion. Ttest bench consists of an electr ic motor / dynamometer , opt ictachometer , igni t ion system (for spark plug use) , and f lywheel. magnet on the f lywheel is used to t r igger the spark. Igni t ion and spat iming is achieved with a Hall Effect sensor mounted on a rotary dand a spark igni t ion system manufactured by CH Electronics, InEngine speed is measured using a Monarch Instruments ACTtachometer with a ROS -5W remote optical sensor. The mini-enginerigidly coupled to the dyn amometer via a steel shaft.To test combustion in the mini-rotary engine, a pre-mix

Optical SparkTachometer TimingDia l Ex h au s t . . * 4 "\ I Tc / A, ,MassFinwElectricMotor/Dyno ControlFuel

" ~ l l ' ~ ~ 4 Air" . . . . . , . , ~ I ~ ~"Mini-RolaryElectronics ~ Sl~k Plu g Engine

Figure 3. Mini-Rotary Engine Test Benchhydrogen-air mixture is used rather than a l iquid hydrocarbon, givhydrog en's ease of igni tion. A gaseous fuel mixture is advantageofor this study because i t does not require the com plexity of a fcarburetion system. The gaseous fuel and air are mixed upstreamthe engine in a T-junction. The fuel is metered using valves arotameters upstream o f the T-junct ion.To measure engine power, a dynamometer has been developusing a Maxon TM brushless electronical ly commutated mo tor andrect ifier circui t . Power g enerated by the m ini-rotary engine spins dynamometer , which acts as a generator and produces electr ipower. The rect if ier circui t converts the dynam ometer 's three-phoutput to a DC v oltage potential. Rheostats are used to apply a loadthe dynamometer . The rheostats can be adjusted to produce appropriate load, based on the mini-rotary engine being tested.To cal ibrate the voltage across the dynamometer load mechanical power output , the dynamometer is driven by an electmotor while connected to a torque arm and load cel l . Figure 4 shothe cal ibrat ion curve for the Maxon 50 W brushless DC mot

MktI-Rotmy Englfle Oynamometer Calibral~fl (SOW Maxofl Motor)16 - Dyn~trt0rneteroad

12- ~ ~.g 1 0 -

4- "2-000 2 i i 1'o 1~ GenwtotP ~r 0h0Figure 4. Dynamom eter calibration chart for 50 WMaxon brushless electric motor.

4 Copyright 2001 by ASME


5/7

Voltage drops across the rect i f ier circui t and across the engine coi lswere taken into account. During engin e operation, the output voltageof the motor is measured and related to the power ou tput through thecalibration curve.Temperature measurements were also made upstream of theengine and at the housing near the combustion chamber using

thermocouple or infrared imaging. Pressure measurements upstreamof the engine were also m ade to accurately measure fuel and air f lowrates.MN30 MINI-ROTARY ENGINE DESIGN AND TESTINGThe MN 30 generat ion mini-rotary engine is shown in Figure 5.The basic engine is simple in design, consisting of 7 parts: front plate,epitrochoid housing, back plate, rotor, internal gear, spur gear, andshaft . Two bearings, mounted in the front and back plate, posi t ion theshaft . For speed o f fabricat ion, the mini-rotary engine is made fromsteel. The mo st accurate means of manufacture at this scale is throughelectrodischarg e machin ing (EDM) [2]. Note that the initial design ofthe MN30 engine does not include apex or face seals .

Figure 5. MN30 Mini-Rotary EngineThe MN30 mini-rotaryengine has a displacement of78 mm 3. Prelimin ary sealingtests determined that theengine suffered from 20%

leakage during operat ion dueto poor tolerances duringmanufacturing [17]. Fromthe leakage tests, it wasapparent that an apex seal ingsystem was necessary. Inorder to improve seal ing,slots were cut in the rotor apexes and apex seals inserted. The springsconsist of brass tabs with leaf springs formed from spring steel (seeFig. 6). Even these simple designs significantly improv ed leakage.Combustion tests were performed on the MN30 mini-engine withapex seals. In these tests, an electric moto r was used to rotate the

Igni t ion Combus t ion Exhaus t

Figure 6. Modified MN30 Rotorwith Apex Seals

i i,

a b

Figure 7. Optical Images of Ignition in MN30 mini-enginea. High Spee d Image of H2-OiI-Air Ignitio nb. Infrared Image s of H2-Air Ignitio n

engine, while a stoichiometric H2-air or H2-C3Hs-air mixture wasupplied. The igni t ion system used was a spark plug. Igni t ion t iminwas changed from 15 degrees BTDC to 15 degrees ATDC. Figureshows optical images of combustion obtained while operat ing thengine with a plexiglass front plate. While the engine generated no npower output , there was a 50% reduction in the pow er required by thelectr ic motor to turn the mini-engine with combustion. Exhautemperatures also changed with spark t iming, with peak exhautemperatures at 7.5 degrees BTDC.MN125 MINI-ROTARY ENGINE DESIGN AND TESTINGBased on MN3 0 mini-engine testi ng, the following desigfeatures were incorp orated into the MN 125 mini-rotary engine:- Integrated f lexure apex seal system to test the applicabil i ty of th

system in MEMS fabricat ion- Exhaust gas recirculat ion to pre-heat the incoming fuel-a

mixture and heat the combustion chamber w ith exhaust gases- Increased size for better manufac turing toleranceThe displacement of the engine was increased from 78 mm 3 348 mm 3 to redu ce the effects of ma nufacturing tolerance. Thepitrochoid major and m inor axes were increased by a factor of 1.3while the width of the mini-engine was increased from 3.6 mm t9 mm to reduce axial heat losses. The increase in width also allowfor the use of commercial off the shelf (COTS) small-scale glow anspark plugs. Manufacturing tolerance of the MN125 engine waimprov ed from :t-0.2 mm to i-0.02 ram.The M N125 engine generated posi t ive output power operat ing oH2-air and using two different ignition sources: 1) a COT S spark pluand 2) a COTS glow plug. The experimental resul ts of tests with thspark igni tion wil l be discussed f i rst . The test procedure typicalconsists of the following: the MN 125 is initially brought up to speethrough the use of pressurized air to overcome fr ict ion. The sparsystem is energized and H2 introduced into the flow until sustainecombustion is obtained. The power generated by combustion resulin an increase in engine speed and voltage across the dynamometload. When the spark is energized, the voltage across the dynamometload, engine speed, and housing temperature all increase.During the experiment shown in Figure 8, the spark is turned oat three separate times for 5-10 seconds to achieve stead

4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .~ P o ~ e r- - -Speea3. 5

3So~k On

2.5{ 7 "~ 21

0. 5

0

. SPark On $Dark On J,,' = Soark On

~'" w e /v fL~

I1 rl25 50 75

8

4

100 125 150 175 200 225"t ime (see)

Figure 8. Engine power output and engine speed oMN125 mini-rotary engine operating on H2using a spark ignition.



6/7

J

non-combustion condit ions. Without combustion, there is a decreasein engine speed, measured power, and a lC/sec decrease in housingtemperature due to heat loss to the engine stand and convection.This difference in power output im mediately before the spark isterminated and after the engine reaches a steady state withoutcombustion (spark off) is the net power output of the MN125mini-engine. Figure 8 shows a plot of engine power and engine speed.For this part icular experiment , the net power output at 25 and 75seconds is 1.5 W an d 1.2 W, resp ectively. The stoichio metry duringthis time period is (I)=0.42. At 100 seconds, air flow is increased tomeasure power output at higher engine speeds. This leads to anincrease in engine pow er output to 2.7 W with a stoichiometry of(I)=0.37 due to the increased air flow rate. Spark timing for all threepower measurements is 0 degrees BTDC.Equivalence rat ios closer to stoichiometric wil l give higherpower, bu t flashback (related to leakage) prevents the expe riment fromreaching higher equivalence rat ios. The experiment was terminatedbecause of lubricat ion breakdown due to elevated engine temperature.In a separate test , the breakdown temperature of the lubricant wasmeasured to be Tbreakdown : 275C.The net power ou tput as a funct ion of the peak engine speed for aMN125 m ini-engine is compiled in Figure 9. The plot indicates al inear r ise in power with engine speed consistent with a relat ivelyconstant torque output. Th is indicates the residence time is greater thanthe reaction time. The s catter in the data is due to sligh tly differentoperating conditions during each test. Stoichiom etries for these testsvary betw een q~= 0.4 ~= 0.04, w hile the spark timin g is m aintained at 0degrees BTDC. The maximum pow er obtained by the engine is 2.7 Wat 9300 RPM with rhh of less than 0.5%. It should be noted that this isthe resul t of prel iminary tests of the M N 125 engine. Further tests wil lincorporate thermal management, improved seal ing, and optimizedinlet / exhaust ports.Tests have begun using an O.S. Engines Type RE glow plug asthe ignition source. Initial tests with H2-air have also show n positivepower generat ion. The glow plug uses plat inum or a platinum-compo site for the heater coil. Since the platinum acts as a catalystduring engine operat ion, power is only needed to star t the engine.This result is critical for implem entation in the micro-ro tary engine, asa spark igni t ion system would require too much energy for a ME MSsystem.

MN50 MINI-ROTARY ENGINE DESIGN AND TESTINGThe MN50 mini-rotaryengine has been designed andfabricated for operation withliquid fuel (Fig. 10). Thisengine has a displacement of138 mm 3. Design features o f theMN50 include a new rotorprofi le for improved seal ing andcompression and a l iquid fuelcarburetor.

Engine temperature affectslubricat ion and performance ofthe engine. As enginetemperature increases, enginelubricants break down, whichleads to increased leakage across the apex seals and reducecompre ssion ratio. The use of a liquid fuel with incorporated lubricais necessary.Initial testing of the MN 50 m ini-rotary engine using H2-air hcommenced. Posi t ive power output of the MN50 engine operating oH2-air has been observed. Wh en the spark is turned on and off, thengine speed responds accordingly. Test ing of the MN50 mini-engiwith l iquid hydrocarbon (Methanol-Nitromethane) fuels wil l fol low.

Figure t0. MN50 Mini-RotEngine with LiqFuel Carburetor

MEMS-ROTARY ENGINE FABRICATIONDesign and fabricat ion of the MEM S-scale rotary engine (micrrotary engine) has begun. A larger scale "meso-rotary" engine wfabricated at the Case W estern Reserve Universi ty (CWR U) MEMResearch Center using Si and SiC. The CWR U micro-engine hasdisplacement of 1.2 mm 3 and a rotor diameter of 3 mm, compared the ul t imate goal of 0.08 mm 3 displacement and 1 mm rotor diametThe larger engine was fabricated to test the l imits of the CW RU Sfabrication process.The rotor is made of molded SiC [18] from a three-wafer Si mo(see Fig. 11). The Si housing is fabricated from three separa500 mm wafers deep reactive ion etched (DRIE) to form the featurand then fusion bonde d together (see Fig. 12).

3

25

2g

1

05

020 ~

Figure 9.3O00

4

4000 ~ 60(30 7000 80 ~ 9000 10000RPM

Net power ou tput of Gen II mini-rotary engineoperating on H2-air and using spark ignitionIO = 0.4 + 0.04~

Figure 11. CWRU Micro-Rotary EngineLeft: Three.wafer Si moldRight: Molded SiC rotorThe CWRU micro-engine highlights some of the diff icul t ies fabricating an engine using MEMS techniques. Even though ttolerance obtained in MEM S is typical ly on the order of microns, tpercentage tolerance when compared to the length scale is high. Fohigh precision system such as the rotary engine, the tolerance requir

is on the order of 0.1% (


7/7

i

one in Figure 12 for the Si ]n l= Exhausthousing due to a 20 Ixmmisalignment. Wo rk on thenext generation CWRUmeso-engine has begun withthe aim of highermanufacturing tolerances.Fabrication of the 1 mmdiameter micro-rota ry enginehas commenced. Themicro-rotary engine will befabricated using bulk Sietching techniques at the U.C. Berkeley Microlab with a SiC thincoating for protection against high stresses, temperatures, andchemicals [6].

G e a rE pi t ro . . . . .

Figure 12. CW RU Micro-RotaryEngine Housing

CONCLUSIONPower MEMS is an emerging field that may ultimately produceportable power generation systems with an order of magnitude higherenergy density than com mon batteries. The combination of low unitcost through mass production and high energy density through l iquidhydro carbon fuels show s great promise.A research project has begun with the objective to design,fabricate, and obtain positive power output from a micro-rotaryinternal comb ustion engine. A series o f mini-rotary engines have beenmanufactured and tested to investigate the effect of scale reduction onsealing, fluid, and combu stion issues. The mini-rotary engines havebeen operated on gaseous H2-air using both spark and glow plugs asthe ignition system. Positive pow er output of up to 2.7 W at~9,300 RPM has been obtained from the MN125 mini-rotary engine.Further refinement of the mini-engine is necessary to increase engineefficiency through improve d sealing and thermal management. Engineoperation is affected by engine temperature and fuel-air stoichiometry.Richer mixtures are necessa ry for improved power output. Howev er,thermal management schemes are necessary to prevent pre-ignition inthe inlet port. The results obta ined to date with small-scale eng inesindicate that there are no fundamental phenomena that would preventthe deve lopment o f the micro-rotary engine.Fabrication of the micro-rotary engine has begun with a largerscale m eso-engine fabricated at Case W estern Reserve University.The challenges facing the fabrication of the micro-rotary engineinclude manufacturing tolerances, wafer alignment, assembly, an dmaterials. Fabrication of the micro-rotary engine has begun at U.C.Berkeley.ACKNOWLEDGEMENTSThis work is supported by DARPA through Grant NumberDABT63-98-1-0016. The authors would like to thank ProfessorMehran Mehregany and Dr. Chris Zorman o f the MEM S ResearchCenter at Case Western Reserve University and Ms. Miki Yamada ofthe University of California at Berkeley. The authors wo uld also liketo thank Marone y Comp any for their assistance in the fabrication ofthe mini-rotary engines.REFERENCES

1. Kovacs, G. T. A., 1998, Micromachined TransducersSourcebook, Mc-Graw-Hill , New York.2. Madou, M., 1997, Fundamentals ofMicrofabrication, CR CPress, B oca Raton.

3. Waitz, I.A., Gauba,, G., Tzeng, Y-S., 1996, "Com bustors Micro-Gas Turbine Engines", Proceedings o f the ASMAerospace Division.

4. Linden, D., 1995, Handbook o f Batteries 2 nd ed., McGraHill, New York.5. Mehrega ny, M., Zorma n, C. A., 1999, "SIC MEMOpportunities and Challenges fo r Applications in HarEnvironm ents", Thin So lid Films, 355-35 6, pp. 518-524.

6. Stoldt, C. R., Carrar o, C., Ashurst, W. R., Fritz, M. C., GaD., Maboudian, R., 2001, "Novel Low-Temperature CVProcess for Silicon Carbide MEMS", Proceedings, acceptto Tran sduc ers 2001 , International Solid-State Senso rs aActuators Conference.7. No rbye, J.P., 1971, The Wankel Engine: DesiDevelopment, Applications, Chilton Bo ok Co.8. Ansdale, R.F., Lockley , D.J., 1969, The Wankel RC EnginA.S. Barnes and Com pany, South Brunswick.9. Van Wylen, G. J., Sonntag, R. E., 1986, Fundamentals Classical Thermodynamics, 3rd ed., John Wiley and SonNew York.10. Cengel, Y. A., Boles, M. A., 1994, Thermodynamics, Engineering Approach, McGraw-Hill , New York.11. Borma n, G. L., Ragland, K. W ., 1998, CombustiEngineering, McGraw-Hill, Boston.12. Incropera, F., DeWitt, D., 1996, Fundamentals o f Heat aMass Transfer, John Wiley & Sons, New York.13. Glassman, I., 1996, Combustion, 3rd ed., Academic Pre

San Diego.14. Zamash chikov, V. V., 1997, "Experimental InvestigationGas Combustion Regimes in Narrow Tubes", Combustiand Flame, 108, pp. 357-359.15. Fu, K., Knobloch, A. J., Cooley, 13. A., Walther, D. CFernandez-Pello, C., Liepmann, D., Miyasaka, K., 200

"Microscale Combustion Research for Applications MEMS Rotary IC Engine", Proceedings, accepted to 20National Heat Transfer Conference.16. Cooley , B., 1999, ," Exploring the Limits of Microsc

Comb ustion" Ma ster's thesis, U niversity of CalifornBerkeley, CA.17. Fu, K., Walther, D. C., Fernandez-Pello, A. C., LiepmanD. and Miyasaka, K., 1999, "Preliminary Report on MEMRotary Internal Combustion Engine", Proceedings, WesteStates Section Fall Meeting.18. Rajan, N.; Me hregany, M.; Zorman, C. A.; Stefanescu, Kicher, T. P., 1999, "Fabrication and Testing Micromachined Silicon Carbide and Nickel Fuel Atomizfor Gas Turbine Engines", Journal

MicroElectroMe chanical Systems, 8, No. 3, pp. 251-257.


mems rotary eng.imece01

Documents