ACTIVE MAGNETIC BEARINGS FOR OPTIMUM …
Transcript of ACTIVE MAGNETIC BEARINGS FOR OPTIMUM …
ACTIVE MAGNETIC BEARINGS FOR OPTIMUM TURBOMACHINERY DESIGN
Jerry Hustak, R. Gordon Kirk, and Kenneth A. Schoeneck Ingersoll-Rand Company
Phillipsburg, New Jersey 08865
The design and shop test r e s u l t s are given f o r a high speed e ight s t age centri- fuga l compressor supported by active magnetic bearings. r o t o r dynamics ana lys i s is presented with spec i f i c a t t e n t i o n given t o design consid- e ra t ions f o r optimum r o t o r s t a b i l i t y . ings i n ex is t ing machinery are discussed with supporting ana lys i s of a four s t a g e cent r i fuga l compressor. ments f o r successful machinery operat ion of e i t h e r r e t r o f i t o r new design turbo- machinery.
A br ie f summary of t h e
The concerns f o r r e t r o f i t of magnetic bear-
Recommendations are given on design and ana lys i s require-
INTRODUCTION
A decade ago new technology i n t h e form of dry gas seals w a s introduced i n t o t h e f i e l d of i n d u s t r i a l turbomachinery design t o minimize t h e c a p i t a l , operat ing, and maintenance cos t s associated with seal o i l systems (1). seal ing system i s gaining widespread recogni t ion because it has and is continuing t o demonstrate its superior mechanical, performance, and economic f ea tu res i n c e r t a i n appl icat ions. Now once again new technology i n t h e form of active magnetic bearings (AMB) is being introduced i n t o the marketplace f o r u se on individual turbomachinery. The f ea tu res of t h i s technology when applied t o turbocompressor design r e s u l t i n several economic, performance, and v e r s a t i l i t y improvements unavai lable t o t h e in- dus t ry at t h e p r e s e n t time. Active magnetic bearings used i n conjunction with dry gas seals and dry couplings now enable both t h e manufacturer and use r t o th ink i n terms of o i l - f r ee cen t r i fuga l compressors, c e r t a i n l y a dramatic change from only t e n years ago (2) .
Today t h i s type of
Patent a c t i v i t y on passive, active, and combination magnetic bearing systems spans 150 years. n e t i c systems because they were easy t o fabr ica te . It was later shown, however, that a passive magnetic suspension'for three axes of displacement i s unstable; a theory t h a t is sti l l v a l i d today. a French research firm, began inves t iga t ing t h e characteristics of both passive and active magnetic suspension systems f o r a satell i te flywheel appl icat ion. they developed a t o t a l l y active magnetic suspension system f o r t h e COMSAT comunica- t i o n s satellite. (S2M) t o f u r t h e r develop and commercially market active magnetic bearing (AMB) systems in t e rna t iona l ly ( 3 , 4 ) .
The bulk of t h e i n i t i a l inves t iga t ions centered on permanent mag-
I n 1969 Societe EuropeaMe d e Propulsion (SEP),
In 1970
I n 1976 SEP formed a new company Socie te d e Mecanique Magnetique
DEVELOPMENT CENTRIFUGAL COMPRESSOR WITH AMB
Figure 1 is a view of an e ight s tage , hor izonta l ly-sp l i t , back-to-back cent r i - fugal compressor equipped with magnetic r a d i a l and t h r u s t bearings and gas seals on test at t h e authors ' company (1980). pressor, o r ig ina l ly designed t o run a t 10,000 RPM (167 Hz) on hydrodynamic bearings,
The eight s t age r o t o r housed in s ide t h e com-
327
has s ince operated successful ly a t speeds up t o 13,000 RPM (217 Hz) on magnetic bearings. purpose of operating t h e r o t o r i n a pressurized environment over a wide range of pressures and flows from choke t o surge.
The compressor is shown at tached t o two closed loops constructed f o r t he
One unique f e a t u r e of t h i s compressor w a s t h e i n s t a l l a t i o n of t h e t h r u s t and journa l bearings located on t h e f r e e end of t h e ro to r d i r e c t l y i n t o t h e gas (n i t ro - gen) pressurized environment thereby eliminating t h e need f o r one s h a f t seal. To i l l u s t r a t e t h e concept of a nonlubricated cen t r i fuga l compressor a gas seal w a s chosen as t h e main s h a f t seal on t h e coupling end of t h e ro to r . some of t h e important design f ea tu res of t h e e ight s t a g e back-to-back ro to r while Figure 2 i l l u s t r a t e s the appearance of t h e f u l l y assembled test ro to r .
Table 1 summarizes
Before power is appl ied t o t h e bear ings t h e r o t o r is supported on two aux i l i a ry cage, dry lubr ica ted b a l l bearings located i n c lose proximity t o t h e AMB. ance between t h e ro to r and t h e inner race of t h e b a l l bearing is se lec ted t o prevent r o t o r contact with t h e AMB pole pieces o r t h e i n t e r n a l seals of t h e compressor while t h e ro to r is a t rest o r during an emergency shutdown. the AMB, i n t e r n a l seals, and aux i l i a ry bearing area are 0.012 in . ( . 3 mm), 0.010 in . (.254 mm) , and 0.006 in . (.15 mm) , respect ively. When power is applied t o t h e elec- t r o n i c cont ro ls t h e electromagnets levitate t h e ro to r i n t h e magnetic f i e l d and ro- t a t i o n of the dr iving source such as a motor or tu rb ine can be s t a r t ed . The sensors and cont ro l system regu la t e the s t r eng th and d i r e c t i o n of t h e magnetic f i e l d s t o maintain exact r o t o r pos i t i on by cont inua l ly ad jus t ing t o t h e changing forces on t h e ro to r . t h e aux i l i a ry bearings and r o t o r system are designed t o permit s a f e decelerat ion.
The clear-
Typical r a d i a l c learances i n
Should both t h e main and redundant f ea tu res of t h e AMI3 f a i l simultaneously,
The undamped critical speed map shown i n Figure 3 compares t h e standard f l u i d f i lm bear ing/o i l seal design t o t h e magnetic bearing/gas seal design f o r t h e 8 s t a g e development compressor. The magnetic bearing design increased t h e f i r s t r i g i d bear- ing mode by reducing t h e bearing span but decreased t h e second, t h i r d , and fou r th modes due t o t h e add i t iona l weight of t h e ferromagnetic j ou rna l sleeves and t h e l a rge r diameter t h r u s t co l l a r . ness and damping p rope r t i e s of t h e electromagnetic bear ings as a func t ion of r o t o r speed. have t o p a s s through t h r e e critical speeds and opera te approximately 20% above t h e t h i r d critical and 40% below t h e f o u r t h critical. and second criticals are r i g i d body modes only a s i g n i f i c a n t response at approximate- l y 8000 RPM (133 Hz) would be expected as t h e r o t o r passed through its t h i r d (free- free mode) critical. Subsequent unbalance f orced response ca l cu la t ions v e r i f i e d these expectations with acceptable operat ion of t h e compressor t o 14,000 R P M (233 Hz) (see Figures 4 and 5).
Superimposed on t h i s map are t h e as measured s t i f f -
For a design speed of 10,000 RPM (167 Hz), Figure 3 ind ica t e s t h e r o t o r would
Furthermore, since both t h e f i r s t
The damping required f o r optimum s t a b i l i t y may be a r r ived at by construct ion of
(22.8 N/pm), Figure 6 shows t h e movement of the Increased damping levels cause t h e f i r s t
The t h i r d mode increases i n s t a b i l i t y (87.7 N-sec/pm) but then decreases as damping i s in-
a Lund s t a b i l i t y map (5) from t h e ca lcu la ted damped c r i t i ca l speeds (5,6,7). 1st mode s t i f f n e s s of 130,000 l b / i n eigenvalues as t h e bearing damping varies. and second modes t o become c r i t i c a l l y damped. up t o a point of 501 lb-sec/in creased fu r the r . study w a s undertaken t o determine i f t h e t h i r d mode would go uns tab le a t i ts corre- sponding s t i f f n e s s of 229,000 l b / i n (40.1 N/llm). The r e s u l t s of that ana lys i s are presented i n Figure 7 which ind ica t e t h e t h i r d mode becomes c r i t i c a l l y damped as t h e damping is increased while t h e 1st mode damping would be a t an optimum f o r 645
For t h e
Since t h e f i r s t and second modes become c r i t i c a l l y damped, a second
328
lb-sec/in (112.9 N-sec/mm). t he increase i n bearing s t i f f n e s s , is similar t o r e s u l t s presented by Lund (5). These r e s u l t s i nd ica t e t h a t a l e v e l of 400-500 lb-sec/in be idea l f o r a l l modes up t o and including t h e fourth.
The behavior of t h e 1st and 3rd modes, r e su l t i ng from
(70-87.5 N-sec/mm) would
The a c t u a l measured s t i f f n e s s and damping characteristics f o r t h e magnetic bearing are given i n Figure 8. ence when compared t o conventional f l u i d f i lm bearings. tilt pad bearings have c h a r a c t e r i s t i c s generated predominantly by operating speed, with l i t t l e inf luence from non-synchronous exc i t a t ions (8). The active magnetic bearing characteristics, shown i n Figure 8, t a t i o n regardless of operating speed. mize unbalance forced response, t h e damping characteristics a t subsynchronous exci- t a t i o n frequencies can be spec i f ied t o assure optimum s t a b i l i t y (3, 4).
Active magnetic bearings have an important d i f f e r - Typical preloaded f i v e shoe
are dependent on t h e frequency of exci- For a given s t i f f n e s s va lue se lec ted t o mini-
The test program out l ined f o r t h e compressor w a s d i rec ted toward confirming t h e ana ly t i ca l predict ions f o r t h e dynamic behavior of t h e r o t o r and experimentally dem- ons t ra t ing t h e r e l i a b i l i t y of t h e complete system under t y p i c a l operating conditions. The f u l l y assembled compressor w a s i n s t a l l e d on t he test stand and operated a t a max- imum discharge pressure of 600 psig (4.1 MPa) with speeds up t o 13,000 R P M (217 Hz). The r e s u l t s f o r a dece l as recorded a t t h e bearing probe loca t ions are given i n Figures 9 and 10. The r e s u l t s are i n general agreement with t h e predicted unbalance forced response r e s u l t s . Since t h e amount and loca t ion of t h e a c t u a l unbalance d i s - t r i bu t ions are never known, t h e amplitude of each damped response is d i f f i c u l t t o predict . with t h e test r e s u l t s .
The predicted peak response speeds are considered t o be i n good agreement
DESIGN EVALUATION OF A FIELD RETROFIT
The economic advantages of gas seals and/or magnetic bearings has prompted in- terest i n r e t r o f i t of ex i s t ing un i t s . must be given t o placement of cri t ical speeds f o r both main and back-up bearings, response s e n s i t i v i t y , and ove ra l l s t a b i l i t y considerations. The preliminary design study f o r a 4 s t age high speed cen t r i fuga l compressor w i l l i l l u s t r a t e in more d e t a i l t h e parameters that must be considered f o r t o t a l system dynamic analysis . The bas ic design parameters f o r t h i s ro to r are indicated in t h e second column of Table 1.
For e i t h e r r e t r o f i t o r new machinery, a t t e n t i o n
. The undamped critical speed map f o r t h e four s t age compressor with dynamic s t i f f - ness values is shown i n Figure 11. The magnetic bearing s t i f f n e s s is posit ioned such that t h e compressor must pass through t h r e e cri t ical speeds before reaching a maximum continuous operating speed of 14,500 RPM (241.7 Hz). Due t o t h e r i g i d body na ture of t h e second and t h i r d modes, t h e a c t u a l damped critical speeds w i l l occur from t h e f i r s t and fou r th modes as shown i n Figures 13 and 14 a t approximately 4300 (71.7 Hz) and 18,300 RPM (305 Hz) respect ively. constant f i r s t mode s t i f f n e s s of 86,300 l b / i n values. modes become c r i t i c a l l y damped as t h e bearing damping is increased. so r design, t h e f i r s t mode increases i n s t a b i l i t y as t h e damping increases up t o 225 lb-sec/in The damping va lue i n i t i a l l y supplied, 140 lb-sec/in (24.5 N-sec/mm), should be in- creased by 61% based on t h e r e s u l t s of t h i s ana lys i s .
Figure 15 shows t h e Lund s t a b i l i t y map using a (15.1 N/W) with v a r i a b l e damping
The f i r s t forward mode typ ica l ly goes uns tab le while t h e second and t h i r d For t h i s compres-
(39.4 N-sec/mm) but then decreases as t h e damping is increased fu r the r .
The optimum damping f o r s t a b i l i t y w a s a l s o ca lcu la ted by an approximate method
329
using the modal (see Tab le 1). high K r a t i o s ) :
m a s s , r i g i d bearing cr i t ical frequency, and ac tua l bearing s t i f f n e s s The equation from Reference (9) can be wr i t ten a s follows (val id f o r
70417.7 Kg 1.356 x lo-’ Ncr {G + N2cr
Example f o r 4-stage 1st mode:
Co = 1.356 x lo-‘ x (5423) x (146 + (70417.7) x (86300)/(542312)
= 259 lb-sec/in
This ca lcu la t ion gives an answer 15% higher than the ac tua l optimum damping f o r t h i s K r a t i o of 1.41.
Figure 16 shows a comparison of s t a b i l i t y versus aerodynamic exc i ta t ion between the conventional f l u i d f i lm design and t h e magnetic bearing r e t r o f i t design. crease i n s t a b i l i t y due t o t h e magnetic bearings moves t h e fog dec from near zero t o a value of 1.41.
The in-
CONCLUSIONS AND RECOMMENDATIONS
The capab i l i t y of an active magnetic bearing system t o support a f l e x i b l e turbo-
During t h e 350 hours of accumulated operating time f o r t h e development compressor ro tor and simultaneously inf luence its v ibra t ions has been successful ly demonstrated. compressor supported by magnetic bearings t h e following observations have been made:
1. The ro to r behaved in a s t a b l e manner a t a l l times when accelerating/deceler- a t ing through its f i r s t t h ree critical speeds.
2. The ro to r behaved i n a s t a b l e manner while undergoing surge cycles a t maxi- mum discharge pressure.
3. The ro to r w a s ab l e t o s a t i s f y commonly accepted v ibra t ion amplitude and cri t ical speed amplif icat ion criteria a t a l l operating speeds up t o 13,000 RPM (217 Hz). Speeds beyond t h i s point w e r e l imited by impeller stress considerations.
The ana lys i s of t he two compressors has c lea r ly shown t h e advantages of adjust- ab le bearing s t i f f n e s s and damping t o achieve minimum response s e n s i t i v i t y and opt i - mum s t a b i l i t y . zero t o 1.41 f o r t h e 4-stage compressor could not be accomplished by conventional f luid-f i l m bearings.
For example, t h e increase i n s t a b i l i t y from a log dec of approximately
The following recommendations can be made f o r t he design and ana lys i s of magnet- i c bearing suspension turbomachinery:
1. Bearing s t i f f n e s s should be selected by evaluation of shaft s t i f f n e s s r a t i o with typ ica l placement a t t h e beginning of t h e 3rd mode ramp on t h e undamped cr i t ical speed map.
2. Bearing damping should be specif ied t o give t h e optimum growth f ac to r s , with consideration given t o a l l modes below maximum operating speed.
330
3. Consideration must be given t o t h e next mode above operating speed ( typ i - c a l l y t h e 4 t h mode) t o assure adequate separat ion margins.
4. All clearances i n t h e bearings and seals should be se lec ted t o avoid ro to r / s t a t o r contact (i.e., rubs) f o r normal expected operating conditions.
5 . The machinery must be engineered t o g ive a minimum of 10% separat ion margin on any continuous operating speed f o r operat ion on t h e aux i l i a ry bearings.
REFERENCES
1. Schoeneck, K.A. and Hornschuch, H., "Design Concept of a High Speed-High Pressure Ratio Centrifugal Compressor," ASME Paper 75-PET-4, Presented at Petroleum Mech. Engrg. Conf., Tulsa, OK, Sept. 1975.
2. Schoeneck, K.A., "The Application of G a s Sea ls and Magnetic Bearings t o Centr i f - ugal Compressors," Pac i f i c Coast G a s Assoc. Transmission Conference, Spokane, WA, A p r i l 18-19, 1985.
3. Haberman, H., "The Active Magnetic Bearing Enables Optimum Damping of F lex ib le Rotors," ASME Paper 84-GT-117.
4. Haberman, H. and Brunet, M., "The Active Magnetic Bearing Enables Optimum Control of Machine Vibrations," ASME Paper 85-GT-22, Presented a t Gas Turbine Confer- ence, Houston, TX, March 18-21, 1985.
5. Lund, J.W., "S t ab i l i t y and Damped Critical Speeds of a Flex ib le Rotor i n Fluid- Film Bearings," J. of Eng. f o r Industry, Trans. ASME, Se r i e s B, 96, No. 2, May (1974).
6. Bansal, P. and Kirk, R.G., "S tab i l i t y and Damped Crit ical Speeds of Rotor-Bearing Systems," J. of Eng. f o r Industry, Trans. ASME, 98 Se r i e s B, No. 1, February (1976).
7. Kirk, R.G., "S tab i l i t y and Damped Critical Speeds - How t o Calculate and In t e r - p r e t t he Results," CAGI Technical Digest, Vol. 12, No. 2, 1980.
8. Wilson, B.W. and Earnet t , "The Effect of Eigenvalue-Dependent , T i l t Pad Bearing Charac te r i s t ics on the S t a b i l i t y of Rotor-Bearing Systems", University of Virginia, Report No. UVA643092/MAES5/321, January, 1985.
9. Barrett, L.E., Gunter, E.J., and Allaire, P.E., "Optimum Bearing and Support Damping f o r Unbalance Response and S t a b i l i t y of Rotating Machinery, Trans. ASME, J. of Engr. f o r Power, 1978.
331
al M
a U
m & 3
0
F
a, M
a
U
m U
d
w 5
n
v)
w
d
s H m v * m
U
d
0 3
v! 3 al &
1
U
a
rl 0
8 k4 6 . &
al U
al
d 2 PI
n
h
+
* N
W
0 0
VI 2
n
h
h
+
N
W
0 0 0
m
+
n
;
W
E L
0
a al
al a
m L
2 0" d
U
a &
PI
n
h
m
N
+
W
m
h
N
n
m
h
\o
m
W
h
N
m
n
z W
5 L
B L
U
d
%I
PI
(d
U
8
n
0
\o
m
m
W
m
m
U
m
n
m
\o
N d
W
h
m m U
W
1 E: d L
L
9 2 a
m
d
I.I a
al m
n
m
VI N
d
v
N
m
m U
n
N 0
m d
W
0
m U
h
1 W d 9 8 $
d L
L
M
4
U
tu a
n
m
u- 2
W
00 N
h
n
m
8 W
0 0
N
+
1 W d d L
L
M
c a al 3 0
a
W M
E
d
rl a
1
0
U
E a
n
* N
+
W
m W
N
N d
n
d
d
\o
W
m
w
m
m
9
n
0
m
N
W
m w
0
\o
d
n
o\
m
\o
W
VI w VI \o
m
n
\ 3 z E
d
W
n
+
9
4
W
+
9
d
n
U
h
0
W
* h 0 h
i! n
d .rf W
*
d
El
s 0
d .)
L
3 m
W
al 0
w
tu d
U
m
n
N
\o
h
W
0
0
m
n
m
N
Q: +
W
m
m
U
3 W G
d I.)
.. &
al U
al
a
d
e n
3 s a
m I c
n
0
PI m W
\o
\o
m
n
VI I-
m
+
W
m
m
h
1 W 0
d L
L
&
a)
U
al
d
rl
a c & 1
d n
n
m
m 0
m
W
m
N
U
VI
n
N
m
N
m
W
+
tr) m m
h
L4 W E .)
&
z" L
rl
a U 1-I U
d
&
U
2 d
& a al m 3 M
d
PG U
m
&
d
F
n
I-
\o
+
h
W
0
0
m
U
n
m m
N
d
W
0 0
h
h
n
L4 W
E L
+
m
PI
h
a a,
a) a
m
a 0 a
m L
2 d 24 a
al P
I U
m
&
d
k
4
n
0
VI 0
m
W
0 0
m 2
n
h
d
h
N
W
0 0
m
4
n
2 E W
p:
N
m P
I F
a a,
al a
r/J
al ).)
L
9 a m
2 $ 8 al
PI a 0
al m
n
m
+
d
W
U w m
\o
03
n
m
N
N
W
VI w 0
m
+
n
N I
w
* m
* W m
VI N
n
N
I w 2
a
W
U
N
m
n
e a 25
0
al m
W
332
EG 0
E-r 0
d W
2 $l E-r X
W H
W
ORIG
INAL PAGE IS
OF PO
OR Q
UA
LITY
D
w n
N
W
e4 B H
Er
3 3 0 H
F
m
w
333