(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

AOO-36600

AIAA 2000-3376

HIGH TEMPERATURE PERFORMANCE EVALUATION OFA COMPLIANT FOIL SEAL

Mohsen Salehi, Hooshang Heshmat

Mohawk Innovative Technology, Inc.®Albany, New YorkTel (518)862-4290 Fax. (518) 862-4293 e-mail: miti@albany.net

36th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit17-19 July, 2000

Huntsville, Alabama

For permission to copy or to republisn, contact the American Institute of Aeronautics and Astronautics,1801 Alexander Bell Drive, Suite 500, Reston, VA, 20191-4344.

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

2000-3376

HIGH TEMPERATURE PERFORMANCE EVALUATION OFA COMPLIANT FOIL SEAL

Mohsen Salehi, Ph.D.Member AIAA, STLE, ASME Associate Member

Hooshang Heshmat, Ph.D.Fellow ASME, STLE

Mohawk Innovative Technology, Inc.Albany, New York

Tel. (518)862-4290 Fax. (518) 862-4293e-mail: miti@albany.net

ABSTRACTThis paper presents performance of a novel non-contacting compliant foil seal (CFS) in a smallgas turbine simulator. The preliminarynumerical study on a CFS for the hot section ofa gas turbine engine was discussed in theprevious work by the authors. A hightemperature hybrid dynamic simulatorrepresentative of a small gas turbine enginespool was developed. The developed air turbinedriven hybrid simulator includes a compliant foilbearing (CFB) and a CFS in the simulated hotturbine section and an oil mist lubricated ballbearing in the cold compressor section of theengine. The hybrid simulator is successfullytested at speeds up to 56,000 rpm andtemperatures up to 600 °C . A solid lubricantfilm, PS304 developed by NASA, was used forthe CFB and CFS journals. The coating isemployed to provide a smooth contact surfacebetween the journal and CFS and to preventwear at startup/shutdown where thehydrodynamic lift off effects are minimal. TheCFS features the following three distinguishingcharacteristics: lower leakage flow ratecompared to Labyrinth and brush seal; capability

of handling larger shaft excursions; and theability to support dynamic loads. The CFSperformance at different operating speeds andtemperatures , and differential pressures wasinvestigated. In a similar condition, a leakageflow comparison was made among a Labyrinthseal, a brush seal and a CFS. The experimentalresults indicate superior performance of the CFSover the other two types of seals. Unlike brushseal, CFS showed no evidence of rub or inducedwear on the journal or seal surface.

INTRODUCTIONCompliant surface gas foil bearings (CFBs) asprecedent of the newly introduced compliant foilseal(CFS) have shown a great potential forreplacement of rolling elements bearings in highspeed rotating machineries such as air-cyclemachines (ACMs) employed for cabin coolingturbines/pressurization and auxiliary powerunits, Walton and Heshmat [1,2]. CFBs used inthese application demonstrate the advantages oflong service life with minimum maintenance andthe avoidance of contamination by elimination ofthe conventional oil lubrication system.Increasing demand for high efficient, more

Copyright © Mohawk Innovative Technology Inc. ® 1American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

durable, faster, longer lasting, and less requiredmaintenance machines has put a great emphasison development of more efficient seals.Currently brush seal and labyrinth seals arewidely used in gas turbine engines. The mainproblem associated with those bearings is thatthey don't benefit the structural compliance andtherefore in the case of shaft excursion severwear is on the surface of the shaft. Theresulted wear reduces remarkably theperformance of the seal.The basic elements of a CFS is illustrated in Fig.1. Spring bump configuration is shown in Fig.2. The performance of these CFBs and CFSsare based on the high pressure gas in a very thinlayer between the journal and the bearing or sealtop surface (shaft and top foil) governed by theReynolds equation for compressible gas. Thehigh pressure gas film between the shaft and thesurface of the seal (top foil) is due to the shaftrotation and hydrodynamic effect. Generation of

in these seals/bearings especially at highjournal rotating speed and high loads or at highambient temperature directly affect theirperformance. Two mechanisms are involved in

stiffness damping of the seal. Onemechanism is via the geometry top foil orspring bumps material which indicates the springsupport. The other mechanism is hydrodynamicfilm which results from the fluid film between theshaft (runner) and top foil [3-5], The highpressure between the shaft and the top foil

sufficient load capacity to separate theseal from the journal,

a non-contact continuous operation.Due to the compliance, the thickness

pressure are inter-related therefore aniterative Reynolds equation solution should beused to account for the localized seal surfacedeflection caused by the hydrodynamic pressureand/or axial pressure gradient. For performance

Shoulder

Top Foil

Spring Bump

Fig.l View of a compliant foil seal

Fig.l Configuration of a spring bump

enhancement of the CFS, the designer can tailorthe structural compliance and initial sealparameters.A detailed analysis of gas-lubricated foil seal ispresented by Salehi et al. [6]. In the previouswork [6], the governing equations and numericalanalysis of a CFS were discussed. Sampleresults of the analysis comparison with

were shown there. Instudy, performance of a CFS in a gasturbine engine is shown. Aperformance comparison among a CFS, aand a Labby seal is made. The effects of seals'surface on the shaft journals are also addressed.

American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)1 Sponsoring Organization.

Oil LubricatedBall Beating Drive Turbine

Foil Bearing

Fig. 3 Schematic of the simulator

EXPERIMENTAL WORKThe current experimental work presents theperformance of a CFS at high temperature in ahybrid CFB/CFS and rolling element bearingsimulator for the first time. This hybrid systemrepresents a small gas turbine engine simulator inwhich a rolling element bearing is used in thecold section and in the hot section a compliantfoil bearing and a compliant foil seal is installed.The rolling element bearing was a split ballbearing which required a continuous thrust loadfor the stable operation. The thrust load wasprovided by introducing air to the other end ofthe rotor(foil seal/fol bearing housing- Fig.3).

Test Apparatus and DescriptionThe preliminary work on testing of a sample ofcompliant foil seal (CFS) which was fabricatedat MiTi was addressed in the previous analyticalwork by Salehi et al. [6]. A dynamic simulatorwas designed which featured a CFB and a CFSin one end (hot section) and a rolling elementbearing in the other end (cold section). The ballbearing at the cold section was lubricated by anoil mist system. The schematic of the simulatoris shown in Fig. 3. In this application, one of theobjectives was to operate the foil bearing

approximately at 3 x 106 DN (D is the diameter inmm and N is the speed in rpm) with dynamicloads up to 10 times of the steady state loads.The CFB had a diameter of 50.8 mm (2 in) anda length of 38.1 mm (1.5 in), however, the CFShad a diameter and effective length of 72.14 mmand 15.24 mm, respectively. The rotor weighedabout 6.8 kg and was driven by the compressedair provided to the drive turbine. The totallength of the rotor was 46.86 cm. The CFS and

2500

20000 40000 60000 80000 SOOOOO

Rotor Speed (rpm)

Fig. 4 Stiffness characteristics of CFB

American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

1000

0 20000 40000 60000 80000 100000

Rota- Speed (rpm)

Fig. 5 Stiffness characteristics of the CFS

Mode No.= 1, Criltai Speed =6162 ipm

Mode No .= 2, Critical Speed = 15695 jpm

Mode No.= 3, Critical Speed = 69982 zpra

Fig. 6 simulator mode shaped

2.00

150

E roof-

Q 0.50

0.00

-0.50l

Ne

• lj £• * o ,j fj

--,-,-fsv

.........

ptweLog.Decremenflndi

Mode!

0 C v <J 0 0 0 -i 0 0 0 0 1 0 ^

Mode!

. . . . . . . . .

cates Instab

o 'J

;;"'----

-.%-;;;-"

llity ;

——— -,S.°J

«>°' '-

' ° 0fj fj ',

\ .........'16000 32000 48000 64000 80000

Rotational Speed (rpm)

Fig. 7 The stability map

CFB housing was equipped with holes equallyspaced cireumferentially for the cartridge heatersto be placed in them for the high temperaturetests. Each cartridge heater was providingapproximately 200 W and the housing couldaccommodate 16 heaters.Figures 4 and 5 are representative of thepredicted CFB and CFS dynamic characteristicsfor the simulator. A computer code developedin house [4] was used for calculation of thepredicted values. Preliminary analysis showedthat for a stable system within the range ofoperating speed, the CFB and CFS should beable to provide about 1.42 and 0.353 kN.s/mdamping.The finite element rotor analysis of the simulatorwas performed using Dyrobes program. Thefinite element analysis predicted the simulatormode shapes which are shown in Fig. 6. Therolling element bearing, foil bearing and foilare positioned in stations 2, 4 and 5,respectively. As it can be seem from the modeshapes, for the 1st mode, damping is provided bythe foil bearing and foil seal section of the rotorwhile in the 2nd mode most of the requireddamping is provided by the foil seal and partially

American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

Fig. 8 The simulator hardware

Fig. 9 The instrumentations and dataacquisition

via the rolling element bearing in station 2. Ashear damper assembly was used for the rollingelement bearing. In design development of thesimulator, it is essential to understand thecapabilities of the system with regards to thespeed and rotor vibrational motion and position.The stability map is shown in Fig. 7.

InstrumentationThe test rig was instrumented with several typesof sensors and probes to measure the necessarytest parameters. The simulator is shown in Fig.8. The instrumentations and data acquisitionsystem is shown in Fig. 9.

Some of the measurement parameters areexplained as followings:

Rotor Speed: The rotor speed wasmeasured by a fiber optic probe. The proberesponds to the once-per-revolution when a darkband marked on the rotor passed the face of theprobe sensor. A digital frequency countermeasured the output signal from the probesensor and after the conditioning, the final speedwas shown on a digital display in terms of rpm.

Rotor and Bearings Positions: Radialmotions were measured via two types of probes.For tests at Sow temperature, all the probes wereeddy current type measuring the X-Ydisplacements. For the high temperature tests,the capacitance type probes were initially used.These probes were mounted in the foilbearing/foil seal housing. It was found later thatthese probes are not functioning properly whenthe shaft is not perfectly grounded. Therefore,two low temperature probes were mounted inthe outside of the CFS housing. These probeswere protected by heat shield and were cooledwith air flow and were calibrated with the probesadjacent to the foil seal. Once the calibrationcurves were found, the high temperature probeswere not used and only these probes were usedfor monitoring the rotor position at the CFSstation.

Heating Elements : For performanceevaluation of the CFS and CFB at hightemperature conditions, two units of heatingsystem were adopted. In the first unit, twoshell&tube type heaters were installed prior tothe test section. Each heater had a capacity of3000 w. These two heaters were employed forelevating the inlet air temperature introduced tothe CFS/CFB housing. The outlet air from theseheaters were then mixed in a small chamber inorder to provide a more uniform inlet airtemperature to the test rig. In the second unit,

American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

Fig. 10 Compliant foil seal/bearingjournals

as explained before, 16 cartridge heaters wereinstalled around the CFS/CFB housing. Eachheater provided about 200 w.

T e m p e r a t u r e / P r e s s u r e / F l o wMeasurements: Thermocouples were to measurethe temperature of CFS/CFB, rolling elementhousing, inlet air and outlet air for the CFShousing, and oil used for lubrication of ballbearing. The temperatures were recorded duringeach test via a temperature chart recorder. Sixthermocouples were installed on the CFS andCFB, circumferentiaUy at 0, 135 and 220 angleswhere the 0 refers to the top of the bearing.These locations were chosen based on thehydrodynamic analysis of the bearing and sealtaking into account the initiation ofhydrodynamic pressure and locations with mosttemperature gradients.The absolute pressure of inlet/outlet air to theCFS housing as well as the pressure drop acrossthe CFS were measured for the seal leakageperformance. For the high temperature tests, anextension tube was used to reduce thetemperature of the air sensed by the pressuregauge sensors.The air flow was measured via accuraterotameters in the range of 0-50 CFM.

CFS/CFB Journals: As explained before,the CFS/CFB are non-contact components at the

. 100E+l

MAGML

P»R SP g ban LIN23,9kCPM . 452E+QML

1kHz.32SE+0 ML

Fig. 11 Coast down at 50 krpm - T = 25 °C

running speed where the hydrodynamic pressureis developed strong enough to support the rotorload. However, friction and wear does occurduring the start up and shut down where thehydrodynamic effects are minimal. To increasewear life and reduce friction torque at the shaft-to-foil interface, the seal and bearing journalsurfaces are coated either with electrolysis nickelor a solid film BaF/CaF Eutectic depending onthe operating temperature. For low temperatureapplication the foil top surface is also coatedwith a low friction Teflon. The Journals for theCFS and CFB are shown in Fig. 10. Theperformance of the coating is has beenextensively studied by NASA.

TEST RESULTS AND DISCUSSIONSIn order to verify the performance of the hybridsimulator, the rotor speed was increased withcontrolling the air to the turbine nozzle box.One of the method used for evaluation is knownas the coast down procedure. During the coastdown procedure, the speed of the rotor isreduced from the highest speed to zero byreducing the air to the driven turbine whilerecording the maximum rotor displacement(orbit). Figure 11 shows a sample of the

American Institute of Aeronautics and Astronautics

(02000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

MAGML

0 PtfR SP A hon UN IkHzX| 43. BKCPM is . 121E+1 ML

Fig. 12 Coast down at 45 krpm-T = 595 °C

16-

• T = 25°CT T = 595°C

so-

e6-

4 = 45,000 rpm

0 20 40 60 80 100 120

AP(kPa)Fig. 13 Non-dinmensional flow vs AP

simulator coast down at the maximum speed ofabout 50,000 rpm. This test was performed atthe room temperature (-25 °C). As shownthere, the simulator performs well and a smoothdamped operation can be seen.A seal with large initial clearance was designed

and fabricated for the high temperature tests.The coast down at 595 °C is shown in Fig. 12.This temperature was the average of the threethermocouples installed on the seal housingbeneath the spring bumps (Fig. 1). As it can beseen from the figures, the orbit size in the hightemperature test is much larger than that for thelow temperature test. This can be attributed tothe lower stifihess and damping provided mostlywith the CFB and CFS. As the temperatureincreases, the elastic modulus of the springbumps reduces and stiffness and dampingcharacteristics are function of temperature.Another effects is due to thermal growth andinterface thermal stresses which can result in anon-uniform growth of the rotor compared tothe CFS/CFB housing. The units for the Y axisin the two above figure is Mil (1 mil = 0.025mm). The leakage test for this seal at twotemperatures is shown in Fig. 13. The non-dimensional flow, *P is defined by the followingcorrelation:

mp°-5£>2AP°-5 (1)

Where m is the mass flow rate, p is the density,D is the shaft diameter and AP is the differentialpressure across the seal.In these two tests the pressure before the sealand after the seal were reading from the pressuresensors. The pressure a t the flow meter wasalso recorded in order to make the necessarycorrection for the density of the air as a functionof pressure. The calibration for the rotameterwas also considered based on the factoryrecommendation. As it mentioned before, thisseal was not optimized with regards to springbumps and compliancy.A test was conducted with employing a Brush

7American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

Fig, 14 Tested brush seal

I

Cross Brush Seai - H25, Interf. = 2 mB (r)CFS _..-

= 40,000ipm „- ' •"'= 20°C ,'*'

0 2 4

journal Diameter = 2.84 in AP(Psi)

Fig.15 CFS vs brash seal

seal. The Brush seal was made by Cross Co. inUK and it was designed for the current journalsize (72 mm). The bristles were made of Haynes25, contained 945 wires per circumferential cm.The bristles were positioned at a 40 ° angle. Thebrush seal is shown in Fig. 14. The results ofthis test was compared against the similar testwith the compliant foil seal and the result isshown in Fig. 15. Thenon-dimensionalizationofthe result was based on ratio of leakage flow ofthe CFS to the minimum leakage flow whenbrush seal was installed. The limited differentialin the brush seal was due to the large leakageflow and higher AP across the brush seal was notobtainable.The brush seal was also tested several times (thetotal operating time was about 4 hrs) up tospeed of about 41,000 rpm. The coast down onthe brush seal is shown in Fig. 16.The test rig was disassembled after the test andit was found that sever wear was formed on theseal journal. The wear tracks are shown in Fig.17. Unlike brush seal, CFS showed no evidenceof rub or induced wear on the journal or sealsurface.

BRUSH SEALIkHz Ai AC/0,3V ZV S. UK 832/16 DL'AL Ik

SP B hen UNtiStCPM Vj . I34EHML

M SP A hon UN '.^z% iZ.8kCPM Jj .U4E+i«L

Fig. 16 Coast down on the brush seal

A test was performed with a CFS to address theeffect of speed for a specific differential pressureacross the seal. For this test, a differential

8American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

Fig. 17 Brush seal wear track on the sealjournal

1 5 -

fe

1

ESoZ

= 315°C

0 10000 20000 30000 40000 50000

Rotor speed (rpm)Fig. 18 Flow variation with rotor speed

pressure of 85 kPa was maintained across theseal. After reaching the steady state condition ata minimum speed of 10,000 rpm, then the rotorspeed was increased to the next speed level. Ateach speed level, the flow was adjusted to keepthe differential pressure constant. The flowthrough the seal was then measured and the

0.14

0.12

0.10 i

0.08

0.06

0.04

D=36.5 mm TL = 15 mm / Labby Seal

Foil Seal

50 100 ISO 200 250 300 350 400 450

Differentia! Ptesuie(kPa)Fig. 19 Comparison of leakage flow in two seals

necessary correction was made due to pressureand temperature. The results are shown in Fig.18. The results are shown based on the non-dimensional flow parameter defined by Eq.l. Asit can be seen the flow across the sea! is notchanging with the speed of the rotor. Thisphenomena was explained bases on the analysisin [6]. It can be said that with increasing thespeed, the effect of hydrodynamic pressureincreases which the pressure

at the boundaries, however, thestronger hydrodynamic filed also reduces theeffect of cross flow which is directly related tothe seal leakage flow.A prototype compliant foil seal with 36.5 mmdiameter, 15 long 0.02 clearancewas compared to a comparable labyrinth sealwith clearance of 0.127 mm. For both seal oneside was pressurized and the other side was keptat atmospheric condition. A pressure (0-60 psi) an acrylic precision rotameter (0-40SCFM) were for the pressure and flowmeasurement, respectively. The results is shownin Fig. 19. It is seen that the foil seal hassignificantly lower leakage flow rate. More

American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization.

information on the performance of the labyrinthseal can be found in [7,8].

CONCLUSIONS

AP Differential pressureNon-Dimensional flow

This paper addresses the applicability of a non-contact compliant gas film riding foil seal in asmall gas turbine engine simulator. The paperalso addressed the performance of a hightemperature hybrid bearing/seal in the simulator.The compliant foil seal and compliant foilbearing were used in the hot section of thesimulator where a rolling element bearing wasused in the cold section of the engine simulator.The rotor-dynamic performance of the hybridsimulator was shown at different temperatures.The simulator was successfully operated up to56,000 rpm on the compliant foil bearing androlling element bearing. The performance of thecompliant foil seal was compared with a similarbrush seal installed on the simulator. The resultsshowed the superior performance of thecompliant foil seal over the brash seal.Additionally, sever wear track was observed onthe seal journal from the brush seal bristles,however, no evidence of rub or wear was foundon the journal or the surface of the compliant foilseal.

AcknowledgmentThis work was sponsored by NASA andMohawk Innovative Technology Inc. (MiTi®).The authors would like to express their thanks toNASA for their support.

NOMENCLATURECFB Compliant foil bearingCFS Compliant foil sealD Journal (rotor) diameterm Mass flow rateN Rotor speedp Density

REFERENCES1. Walton, J.F. and Heshmat, H.,1994,"Compliant Foil Bearings For Use In CryogenicTurbopumps." Proceedings of Advanced Earth-to-Orbit Propulsion Technology ConferenceHeld at NASA/MSFC May 17-19,1994, NASACP3282Vol.l, pp. 372-381.

2. Heshmat, H. and Hermel, P. "Compliant FoilBearing Technology and Their Application toHigh Speed Turbomachinery." The 19th Leeds-Lyon Symposium on Thin Film in Tribology -From Micro Meters to Nano Meters, Leeds,U.K., Sep. 1992, D. Dowson et al. (Editors),Elsevier Science Publishers B.V., (1993), pp:559-575.

3. Heshmat, H., 1991, "A Feasibility Study onthe Use of Foil Bearings in CryogenicT u r b o p u m p s . " p r e s e n t e d a tAIAA/SAE/ASME/ASEE 27th Joint PropulsionConference, June 24-26, Sacramento, CA,Paper No. AIAA-91-2103.

4. Ku, C.-P.R., and Heshmat, H.,1992,"Compliant Foil Bearing Structural StiffnessAnalysis Part I: Theoretical Model IncludingStrip and Variable Bump Foil Geometry,"Journal of Tribology, Tram. ASME Vol. 113,no. 2 ,pp. 364-369.

5. Walowit, J.D., and Anno, J.N., 1975, ModernDevelopment in Lubrication Mechanics, AppliedScience Publishers, Ltd. London.

6. Salehi, M., Heshmat, H., Walton, J.F. andCruzen, S., 1999, "The Application of Foil Sealsto the Gas Turbine Engine," presented at the 3 5th

10American Institute of Aeronautics and Astronautics

(c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)1 Sponsoring Organization.

AIAA/ASME/SAE/ASEE Joint PropulsionConference and Exhibition, June 20-24, LosAngeles, CA, paper No. AIAA 99-2821.

7. Hendricks, R.C., Griffin, T.A., One, T.R.,Csavina, K.R., Pancholi, A., and Sood, D.,1994a, "Relative Performance ComparisonBetween Baseline Labyrinth and Dual-BrushCompressor Discharge Seals in a T-700 EngineTest," 39th ASME International Gas Turbineand Aero-engine Conference, The Hague,Netherlands, June 13-16, 1994.Paper No. 94-GT-266

8. Rhode, D.L., Demko, J.A., Traenger, U.K..,Morrison, G.L., and Sobolik, S.R., "Predictionof Incompressible Flow in Labyrinth Seals",Journal of fluids engineering, March 1986, Vol.108, pp. 19-25.

11American Institute of Aeronautics and Astronautics