AIAA-02-3792

EVALUATION OF LARGE COMPLIANT FOIL SEALS UNDER ENGINE SIMULATED CONDITIONS

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

Hooshang Heshmat, Ph.D. Fellow ASME, STLE

Mohawk Innovative Technology, Inc., Albany, New York

Tel. (5 18) 862-4290 e-mail: trisalchi @"!albany.nct

ABSTRACT The results of recent design and development efforts on a large compliant air lubricated foil seal (CFS) to meet the NASA test engine simulator is presented. The tested seal is 152 mm in diameter with an effective length of 15.24 mm. This work is built upon the successful operation of a smaller scale CFS (72.1 mm diameter) that was reported in previous works. In order to increase the capability of the seal to stand the higher differential pressure. the seal was also modified. The modified seal showed better leakage performance and include a structure which is more robust. During the normal operation, the surfaces of the rotor and seal are separated via a thin high-pressure air film. A CFS features metallic bump foils that provide structural compliance. This feature allows for maintaining the non-contact operation in presence of thermal, centrifugal growth and excursion of the rotor. A dynamic seal test rig, representing a gas turbine engine simulator, was designed, built and tested for performance evaluation of a 6-inch diameter CFS. The test engine was supported by a grease-packed rolling element bearing and a

magnetic bearing. The magnetic bearing, in addition to serving as a support bearing, was employed to control the position of the rotor and to provide rotor operation under a desired eccentricity for CFS. The dynamic test engine is capable of operating at speed up to 20,000 rpm and temperature up to 1200 OF. The CFS performance at various operating speeds and differential pressures was investigated.

INTRODUCTION A compliant foil seal was introduced recently as a self-acting hydrodynamic mechanical component that performs the sealing action while it maintains no contact with the rotating component. The current compliant foil seal technology is developed based on the advanced technology of a compliant gas lubricated foil bearing [ 1-41. The recent studies [5-71 showed remarkable performance of a compliant foil seal in a high temperature section of a small gas turbine engine simulator. These CFS's have a potential long service life capability along with reliable performance and efficiency while requiring minimum maintenance.

1

American Institute of Aeronautics and Astronautics Copyright OMohawk Innovative Technology @Inc.

38th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit7-10 July 2002, Indianapolis, Indiana

AIAA 2002-3792

Copyright © 2002 by the author(s). Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

111 C;M 01. .I iiialtiiiictioii hI1i ;dI c o i i ~ i ~ ~ t i i i .i

fail, &iL> LO iii\tahilit;, Iriil?icclcI I)) ~ 1 1 ~ . ~ \ ~ ~ c ~ ~ . . I i I

con\entioiial seals, icicti Idbyrinth se;lls. the) lac h structural coiiipl ianw i ind there fore. 111 the case o f shaft excursion. sc~e t - weal. is jcetl or1 the surface of the r o t o r . The resultins ~ t : a r also reduces the perforniance of thc. seal. Ir1 contrast, the studies [3 ,5 -7] have shown that due to the mechaniwi o f hump foil,. the\e elements art: capable of proi iding additiorial dumping for the system and accommodate for the rotor excursion. The basic configiration of a ('FS is shown in f - ' i ~ t i r t . I . I'hc t h i r i high pi-e\mre gas f i lm ib

journal) and the top foil. With the proper choice o f structural compliance, the seal surface deforni'ition can hc. tailored to deforni in a manner that enhances the seal performance. For example, the compliant seal design can be tailored to form a variable axial converging wedge shape for improved sealing performance. In the previous work [7], computational analysis of a CFS was addressed. The combined hydrodynamic and structural governing equations were solved simultaneously using a finite difference method combined with a relaxation method. The analysis was primarily based on a laminar flow condition, in both axial and circumferential directions. The preliminary studies on the effect of turbulence were later addressed. In the later study, the governing equation of fluid flow and structural compliance was modified via turbulence functions which contained the Reynolds numbers of the flow in circumferential and axial direction. The effect of pressure on the viscosity was also considered; however, the thermal effects were not considered.

\ ! \ I ~ ' I ~ > L\ I r l l .I risicl t> l )c ' z ~ ' , l l . 1111: 1 4 ) t a ' i ' >\. \(<'1\i

foritled helween the surface or' ;I i'()toI ( 0 1 ;I

Fig. 1 Schematic uf compliant foil

The following sect ions describe t hc ex perimen tal ivork. 'The first sect ion describes the modification to the small seal ( 2 . 8 3 inch diarneter) and second section presents the design and fabrication of the 6 inch seal tzb[

rig.

EXl'LKl,\II~. \ 'I ' l l . \I 01th 72.14 mm (2.84 inch) Seal For testing of the small seal (72.14 mm), a dynamic simulator was built. A dynamic sirnulator was designed which featured a compliant foil bearing (CFB) and a CFS in one end (hot section) and a rolling element bearing in the other end (cold section). The ball bearing at the cold section was lubricated by an oil mist system. The schematic of the simulator with high temperature accessories and seal hardware is shown in Fig. 2. The CFB had a diameter of 50.8 mm ( 2 in) and a length of 38. I mm ( I .S in); however, the CFS had a diameter and effectibe length of 73,. I4 mm and 15.24 mm, respectively. The rotor weighed about 6.8 kg and was driven by the compressed air provided to the drive turbine. The rotor speed of one revolutionlminute was measured by a fiber optic sensor. Radial and horizontal motions of the rotor were measured via displacement probes. For tests at high temperature, heating elements were mounted inside the housing of CFS at equal

7 - American Institute of Aeronautics and Astronautics

-4 e Foil

1 I

-02-3792

L

a

Fig. 2 Schematic of the simulator for 2.84 inch seal

circumferential spaco. The inlet air was also heated using tube heaters. Thermocouples were installed to measure the temperature of CFBKFS, rolling element housing, inlet air and outlet air for the CFS housing, and oil used for lubrication of the ball bearing. The absolute pressure of inlet/outlet air to the CFS housing as well as the pressure drop across the CFS was measured for the seal leakage performance. For the high temperature tests, an extension tube was used to reduce the temperature of the air sensed by the pressure gauge sensors. More detail about this simulator can be found in [6]. Figure 3 shows a sample plot of the performance of the seal at low and high temperature. The nondimensional flow is defined as

T

T V

T T

a 0

e

AP (kPa)

Fig. 3 Nondimensional flow vs. Ap

where m is the mass flow rate, p is the density, D is the shaft diameter and AP is the differential pressure across the seal. The larger flow in higher temperature can be attributed to relative thermal expansion of the

3 American Institute of Aeronautics and Astronautics

A IAA-02-3792

I1

i,

RoLor +ul (rpm)

Fig. 4 Variation of non-dimensional mass with rotational speed

components and reduction in elastic modulus of the compliant elements and spring bumps. The seal showed similar performance in various rotor rotational speeds. The non- dimensional tlow variation with rotor speed is shown in Fig. 4. The tlow inside the seal is under two influences. The axial flow is called Poiseuille flow and the circumferential flow is called Couette flow. The combined effect of these two types of flows, which is a nonlinear combination, determines the pressure and film

OId seal housing

L- w 1

Fig. 6 Photo of 73 mm compliant foil seal hardware

thickness field inside the seal. A detailed discussion can be found in [6-81. With the preliminary small seal, maximum differential pressure of 40 psi was obtained. In order to conduct a seal test at high pressure, the seal housing section of the simulator and the seal hardware were modified. The old and modifled seal housings are shown in Fig. 5. There were two main modifications to the old housing; first, a flat end cap was mounted at the inlet section and second, an end plug was installed at the downstream of the flow.

Modified seal housing

Fig. 5 The original (old) seal housing and the modified housing

4 American Institute of Aeronautics and Astronautics

AI AA-02-3792

r. /,

2

IO - - -

i

' 1 I

i l i

, .w

f

Fig. 7 Enhanccmenf of conipliant foil seal perfonnancc for higher prcssure application

The seal hardware was also modified by incorporating an additional foil ring. The modified seal hardware is shown in Fig. 6. To verify the performance of the enhanced small seal, a static test was conducted in which the seal leakage flow was measured vs. differential pressure. The results for non- dimensional mass flow were compared with results corresponding to the old seal. The comparison is shown in Fig. 7. As the results show, with the enhanced seal a higher

differential pressure was obtained. The old seal at differential pressures higher than 30 psi showed a sudderi increase for the flow rate: however, the modified seal displayed semi linear behavior during the range of differential pressure tested. Dynamic Engine Simulator Rig for 150 mm Seal Test The dynamic test rig was employed to conduct tests which are close to conditions faced in an engine. The test rig consisted of a rotor, a rolling element bearing (REB) housing, an active magnetic bearing (AMB) and a seal housing. The rotor had a total length of 85 cm (33 .5 inch), weighed about 60 kg (132 Ibs) and was made of Inconel 718. The rotor was supported via the REB and AMB. The test rig is shown in Fig. 8. The REB employed was a deep grove ball bearing with 30 mm ID and 62 mm OD, 23.8 mm length, type G, grease- packed. The bearing is rated for 12,900 N (2,900 Ibf) static load and 18680 N (4,200 Ibf ) dynamic load. The ball bearing housing made of AIS1 1045 included a squirrel cage for holding the ball bearing and shear damper. The shear damper was employed to provide

Fig. 8 The dynamic test engine simulator for I50 mm seal 5

American Institute of Aeronautics and Astronautics

A I AA-02-3792

i' a *

* - I- A

Fig. 9 The dynamic simulator test rig

structural damping for the ball bearing during the test operation. For prospect higher speeds, the REB was accommodated with oil mist lube. Two thermocouples were installed on the back of the ball bearing for measuring its temperature. The system hardware is shown in Fig. 9. The magnetic bearing in this test rig was used t o provide the orbit control process while the rotor is running on the ball bearing. The control of orbit via AMB allows the rotor to run stably about an orbit with a predetermined eccentricity. This control is used to evaluate the performance of the seal when the rotor is operating off from center of mass. The magnetic bearing was capable of applying radial load. A Hetro-polar bearing with 8 poles was chosen to meet the load and space limitations. Finite Element Magnetic Analysis was used to determine the size of the poles and to assess the flux levels in the material to prevent magnetic saturation that would limit the load capability. In designing the AMB, the following parameters were used:

I

I 1

L

Fig. 10 Magnetic bearing inside the housing

1 . Rotor diameter: 150 mm (6 i n ) 2. Air gap: 0.508 mm (20 mil) 3. Pole face: 101 mm x 50.8 mm 4. Lamination material: Hyperco 5. Bias coil current: 10A 6. Control coil current: 10A

The magnetic bearing is shown in Fig. IO. The flux levels in the material averaged about 15 kilogauss, which is below the knee on the B-H

6 American Institute of Aeronautics and Astronautics

AI A A-02-3192

Drive motor Coupling

Fig. 11 Drive motor connection to the rotor \*ia mechanical coupling

curve for M 19, the material used for the bearing laminations. In the laboratory, the shaft was set up on load cells and a load test of the bearing was performed. The current in the control coils was varied from 0 to 7 amps and the load from the load cells was recorded. An equivalent load of up to 8000 N (1800 Ibf.) could be obtained. The rotor was driven via an electrical motor. The rotor drive motor was installed adjacent to the ball bearing housing, and axially to the rotor via a coupling. This coupling was axially soft and stiff along the radius. Figure I I shows the drive motor connection to the rotor via the mechanical coupling. The drive motor was a three-phase motor with about 60hp of power and a maximum rotational speed 30,000 rpm. The measurement and instrumentation included displacement probes, load cells, speed probe, thermocouples, flow meter, pressure transducer, mass flow meter and volume flow meter. For ambient temperature application, eddy current probes were used and for high temperature seal testing, special probes were used which could tolerate temperatures as

high as 425 "C (800 "F) for short time operat ion. The load cells were used to measure the load applied to the rotor via the magnetic bearing as well as the load which is required to move the rotor for a specific displacement. The load cells were Piezo type load washers. The rotor speed was measured by an optical type probe. The probe works based on generating a pulse for each complete revolution of the rotor. The probe is connected to a pulse counter and a conditioner. For each revolution of the shaft, a pulse is generated, and for speed calculation, these pulses are counted vs. time, which represents the speed of the rotor. Type K thermocouples were employed, which could be used for temperatures up to 593 "C ( 1 100 OF). Three types of sheaths were used with these thermocouples: a. Teflon; b. fiberglass and c. Inconel. The fiber glass and Inconel type are normally used for high temperature. Mass flow meter was employed to measure the seal flow. The mass flow meter was equipped with a Coriolis Elite type sensor. The nominal flow rate range is 0 to 83 kg/min (182 Ibdmin). The mass flow accuracy was 0.35% of the flow rate. The mass flow repeatability was 0.2% of flow rate. For this type of flow meter, there existed a minimum flow known as zero stability flow. For the accuracy range noted, the flow should be greater that this. The zero stability flow for this mass flow meter was 0.008 kg/min. The sensor for this type of mass flow meter can stand up to 100 bar (1460 psi). The volume flow was measured via rotameter. The flow range was 0-50 cfm. This flow meter was used for measuring the flow rate through the ball bearing for cooling purposes. The actual mass flow was corrected by measuring the pressure of the flow.

7 American Institute of Aeronautics and Astronautics

AI AA-02-3792

which are assembled together. These riietallic seals provide good gas sealing for pressures up to 6,800 kPa ( IO psi). These seals can operate at continuous temperature of 538 "C (IO00 "F) and maximum temperature of 638 "C. The seal housing is also equipped with two holes approximately 90" apart for instal I i ng the displacement probes. There are also two tap holes for mounting two pressure transducers and measuring the pressure before and after the seal. The foil seal housing included equally spaced circumferential holes for placing cartridge heaters in order to test the seal at a high temperature. The objective is to reach the temperature of 482 "C (900 O F )

for seal.

Fig. 12 The seal housino onen with front

The pressure transducers used were Piezo tjpe. The range ot pre~sure has from 0-200 psi. The pressurt. transducer was rated for operating temperature as high as 250 p4.F

The foil seal housing shown in Fig. 12, was made of PH- 17-3. The housing had two wings on both sides through which i t was mounted on a base via high strength screws. Each wing was approximately I 1.7 cm (4.61 in) long and 13.7 1 cm (5.394 i n ) wide and 4.44 cm (1.74 inc) thick. This base was then bolted and clamped to a heavy base table. This table was serving as reference surface on which magnetic bearing, rolling element bearing and seal housings were mounted. The seal housing had outer dimensions of 32.715 cm (12.8 in) as OD, 14.70 cm (5.787 in) length. The seal housing was equipped with eight equally spaced circumferential inlet air connections. The high pressure air from a compressor after passing through the inline mass flow meter and a flow control valve, was introduced to the housing via those air connections. Inside the seal housing, an flow distributor ring was fitted in order to direct the flow around the compliant foil seal. All the mating surfaces such as this ring, the foil seal assembly and the seal housing cap are sealed with metallic seal in order to eliminate any flow through the mating surfaces and parts

Foil seal housing

IZl,Sl!l,l'S and 1)151:1. SSI( )hS In this section, the results of final tests for 72 mm seal and preliminary results for 150 mm seal are presented. The performance enhancement of the modified seal for tolerating higher operating pressure and exhibiting better tlow was shown in Fig. 7. After modification of the seal, the test engine simulator was run to ensure the stable operation of the rotor for various operating speeds. The plot of rotor orbit size vs. rotational speed is shown in Fig. 13. The peak point in the graph indicates the approach to rotor critical speed. As seen, the rotor orbit is very small (0.001 3 inch). The Performance of the modified seal vs. various pressures for different speed is shown in Fig. 14. This test was conducted in the test engine simulator for a 72 mm seal. The semi linear behavior of flow with differential pressure across the seal can be seen. It is also seen that the rotation of the rotor has a minimal effect of the leakage flow. The rotor speed has influence on both the axial pressure

8 American Institute of Aeronautics and Astronautics

AI A A-02-3792

r

m Jr g=- P,,D 001'

- _ c i -

.

K I

t

D &=284m T -m-BO'F L-0.6lnsh

. .

o * *

? 9 2 1 0

Y a

7 I 5 __ __- - _ L - _ - ~ 0

I) 20 40 60 m loo l r )

Differen~al Ressure [Ri j

Fig. 14 Leakage flow vs differential pressure for static test and various 1uta SjWt'J\

Fig. 15 Flow factor for static and dynamic conditions

Where T is the operating temperature in Rankine, P, is the upstream pressure in psi and D is dic diutnc.ter in inch. The mass flow rate is in Ibdsec. A large seal was built similar to configuration of the modified small seal. The challenge in fabrication of the seal included proper formation of the top foil and additional foils which overall provided the compliant foil seal

9 American Institute of Aeronautics and Astronautics

AIAA-02-3792

I [) - . ~- --

0. I

0 0

6 mcn seat Room temp lesl Static

Fig. 16 Six inch seal static performance vs. differential pressure

integrity. The radius of curvature on the seal top foil should match with that of the seal housing. The result for the preliminary static test with a large seal is shown in Fig. 16. As the results show, two regions can be identified from the plot. In the first region, the seal flow increases linearly with the differential pressure at a higher rate compared to the second region where the seal flow has lower variation with the differential pressure. By increasing the pressure of upstream, the flow velocity at the downstream approaches sonic and therefore, this would reduce the rate of flow rate.

CONCLUSION The progress on the development of a compliant foil seal was presented in the form of testing, improvement of the existing seal hardware and design, and fabrication of a dynamic test rig for experiment with a larger scale. The existing seal hardware limiting pressure was increased from 50 Psia to 115 Psia. The seal leakage flow was greatly improved. The linear behavior between the differential pressure and seal leakage flow was

observed. The performance of the seal is a strong function of the differential pressure; however, the seal performance varies minimally with variation of the rotor speed.

ACKNOWLEDGMENT The authors would like to extend their great appreciation to NASA GRC for their support of this research. Special thanks are reserved for Margaret Proctor and Dr. Bruce Steinetz. The help of colleagues M. Tomaszewski and D. Johnson is appreciated.

CFl3 CFS D Pa Lbf

m

P 5

NOMENCLATURE

complaint foi 1 bearing compliant foil seal journal (rotor) diameter pascal force (in pounds)

mass flow rate flow factor mass density

10 American Institute of Aeronautics and Astronautics

AIAA-02-3792

AP d i ffe re n t i al p re ssiire

5 . REFERENCES

1 . Heshmat, H., Walowit, J., and Pinkus, 0. "Analysis of Gas-Lubricated Compliant Journal Bearings." J. Lubr. Tech. Trans. ASME 105, no. 4 (1983):

2. Heshmat, H. and Hermel, P. "Compliant Foi 1 Bearing Technology and Their Application to High Speed Turhoniachinery." 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. Ku, C.-P.R., and Heshmat, H. "Compliant Foil Bearing Structural Stiffness Analysis Part 11: Experimental Investigation." ASME paper No. 92-'Trib-6, to be presented at STLUASME Joint Trib. Conf., San Diego, CA, October 18-21, 1992, Journal of Tribology, Trans. ASME Vol. 1 15, no. 3 (1993): 364-369.

4. Heshrnat. H. "Advancements In The Performance of Aerodynamic Foil Journal Bearings: High Speed and Load Capability." ASME Paper No. 93-Trib-32, Presented at the STLE/ASME Tribology Conference, October 24-27, 1993 New Orlems, L,oiii4ana, ASklF TIMIS., J. of Trib.,

647-655. 6.

7 .

8.

Vol. 1 16, No. 2 April 1994, PP 287- 295. Salehi, M., Heshmat. H., 2001, "Performance of a Complaint Foil Seal in a Small Gas Turbine Engine Simulator Employing a Hybrid Foil/Ball Bearing Support System", STLE Transaction, Vol. 44, No.3., pp.

Salehi, M., Heshmat, H., 2000, "High Temperature Performance Evaluation of a Compliant Foil Seal", ALAA 2000-3376, 36th AIAA/ASME/SAE/ ASEE joint Propulsion Conference& Exhibit, July 17- 19, Huntsville, Alabama. Salehi, M., Heshmat, H., Walton, J., and Cruzen S., 1999, "The Application of Foil Seals to a Gas Turbine Engine", AIAA paper 99-2821, 35th AIANAS ME/S AE/ASEE joint Prop. Conference& Exhibit, June 20-24, Los Angeles, CA. Salehi, M., Heshmat, H., 2001, "Analysis of a Compliant Gas Foil Seal with Turbulence Effects", AIAA

ASEE joint Propulsion Conference& Exhibit, July 18-1 I, Salt Lake City, Utah.

458-464.

2001-3482, 37th AIAA/ASME/SAE/

1 1 American Institute of Aeronautics and Astronautics