Вы находитесь на странице: 1из 8

lnt I Mech Sc~ Vol 20.

pp 729-736 0020-740317811001-0729150200/0
© Pergamon Press Ltd, 1978 Printed in Great Bntmn

THE PERFORMANCE PREDICTION OF FLUID


COUPLING WITH THE INTRODUCTION
OF A BAFFLE PLATE

A. WHITFIELD, R. SIVALINGAMand F. J. WALLACE


School of Engineering, Umversity of Bath, Bath, England

(Received 21 January 1978; in revised form 12 May 1978)


S.~mmary--The development of two theoretical models, the constant velocity model 2 and the
linear velocity distribution mode, 3 to accommodate the effect of the introduction of a baffle
plate on the torque characteristics of a fully filled fluid coupling is described. The predicted
torque characteristics, using both models, are compared with the experimental results of two
couplings with baffle plates

NOTATION
A flow area
C meridional velocity of the flmd
Co tangential velocity of the fluid
K empirical loss coefficient
K! empirical loss coefficient due to friction
KB empirical loss coefficient due to secondary flows
Kl empirical multiplier for circulation loss
KI empirical multiplier for incidence loss
th mass flow rate of the fluid
PL total power dissipation
Q volume flow rate of the fluid
RI inner radius of the coupling
R0 outer radius of the coupling
R• radial position of the centre of rotation of the c~rcnlating fluid
R radius of mean flow path
S slip (oJp - ¢Or)/~p
U linear velocity of rotating elements
¢ torque
ca rotational velocity of the circulating fluid
~op rotational velocity of the pump element
~ar rotational velocity of the turbine element
p density of circulating fluid

Su~ces
1 pump inlet station
2 pump exit station
P pump element
T turbine element

INTRODUCTION
A common technique used by fluid coupling manufacturers in order to reduce the
torque absorption capacity of their couplings is to introduce a baffle plate into the
coupling. This has the effect of blocking the flow path at the inner radius of the
coupling (see Fig. 1), thereby reducing the effective size. By this means the torque
characteristics of the coupling can be matched to the specific engine requirements. 1
Experimental results available from the manufacturers, show that the introduction
of a baffle plate leads to a larger reduction in stall torque than in torques at other
speed ratios. A typical experimental result, plotted on a non-dimensional basis, is
shown in Fig. 2. The effect of introducing a baffle plate is to reduce the stall torque
and to reduce the rate of change of torque with speed ratio at the stall point. It has
further been observed, experimentally, that if the baffle plate is made sufficiently
large, the maximum torque does not occur at the stall point. A typical experimental
result is illustrated in Fig. 3, here maximum torque transmission occurs at a speed
ratio of approx 0.3.
Recent attempts to develop theoretical techniques to predict the performance of
729
730 A. WHITFIELD, R. SIVALINGAMand F. J. WALLACE

T
R2 /(
Turl~ne
T
Ro

II h_ I
F]o. 1. Fluid coupling with baffle plate.

No Baffle Plate
...... Small B a f f l e Plate
0 002

o
O' 001

?
Z

0 02 04 06 OB 1-0
Speed Ratio
FIG. 2. I)iagramatic plot of non-dtmenslonal torque vs speed ratio.

0 002 i I

No Baffle Prate
.... Large Baffle Plate
E
0
I-
I
e"
0 0 001
C
G)
° o
E
?
c-
O
Z

0 04 06 0.8 10
Speed
FIG. 3. Diagramatic plot of non-dtmensional torque vs speed ratio.
The performance prediction of fluid couplings 731

fluid c o u p l i n g s 2"3 s h o u l d a l s o h a v e t h e a b i l i t y to p r e d i c t t h e e f f e c t o f i n t r o d u c i n g a
baffle p l a t e i n t o t h e fluid c o u p l i n g . C l e a r l y t h e baffle p l a t e h a s t h e e f f e c t o f a s p o i l e r
and increases the total circulation losses and reduces the effective area available for
flow circulation. The prediction procedures are essentially one dimensional and the
p o w e r d i s s i p a t i o n w i t h i n t h e c o u p l i n g is r e p r e s e n t e d t h r o u g h t h e u s e o f e m p i r i c a l l o s s
c o e f f i c i e n t s . T h e e f f e c t o f i n t r o d u c i n g t h e baffle p l a t e c a n , t h e r e f o r e , b e m o d e l l e d
theoretically by increasing these empirical loss coefficients. The mean streamline
t h e o r y o f R e f . (2) a n d t h e l i n e a r v e l o c i t y d i s t r i b u t i o n t h e o r y o f R e f . (3) a r e b o t h
e x t e n d e d t o a c c o m m o d a t e t h e baffle p l a t e a n d t h e p r e d i c t i o n s c o m p a r e d w i t h e x p e r i -
mental results.

THEORETICAL ANALYSIS
Both the mean streamline theory, 2 and the linear velocity distribution model) are extended and
compared. The technique used to accommodate the baffle plate is sunilar for both theoreucal models. The
blockage effect is included by assuming that the fluid circulates between the baffle radius, Rs, and the outer
radius of the coupling, R0 (see Fig. 1). The theoretical model then considers a fully filled coupling of these
reduced dimensions. The increase in energy losses due to the baffle plate are allowed for by multiplying the
,ncidence loss at pump inlet by a parameter KI and by mulUplying the circulation loss by the parameter K1-

M e a n streamline theory
The Euler equation for turbomachines when applied to the pump element of the fired coupling2'3 yields
the expressmn for torque as

r = rh(Up2R2- UrIRD (1)

which can be rearranged to give

~--- m (¢o~R22- o~p(l - S ) R I 2) (2)

The flow rate m of equation (2) is obtained by considering the energy dissipation. The total energy
dissipated within the coupling is given by

PL = T(~op- ~or). (3)

With equation (2) the specific power loss can be derived as

e L = mPL
= too2S(Rz 2 - (1 - S ) R I 2 ) . (4)

For any specified input speed too and slip S the specific power loss PL can be calculated from equation (4).
This power loss is equated to the internal losses considered. These are classified as an incidence loss and a
circulation loss Following Refs. (2) and (3), the incidence loss is given by"

1
PII = KI ~ [(top - t o r ) R d 2 (5)

at the pump inlet, and by

I 2
Pn = ~ [(~oo- toT)R2] (6)

at turbine inlet.
The multiplier /(i introduced into equation (5) ~s a simple empirical means of increasing the incidence
loss at pump inlet due to the introduction of the baffle plate.
The circulation loss is considered to be proportional to the square of the velocmty2

pf = K ! 1 K C 2 (7)

where K is an empirical parameter as given m Ref. (2) and K f is an additional multipher introduced to
accommodate the additional losses, if any, introduced by the baffle plate.
The total specific power loss can be equated to that given by equation (4) to yield

l (oh, - oar)2(KiR, 2 + R22) + K f 1 K C 2 = to02 S(R22 _ (1 - S)R, 2) (8)

which can be rearranged to give the circulation velocity as

C2 = ~ K K ! [R22 - ( 1 - S ) R t 2] - ~ [KtRi 2 + R22]. (9)


732 A. WmTFmL~, R. SIVALINGAMand F. J. WALLACE

The slip at which the circulation velocity and hence the flow rate is a maximum through each element of the
coupling can be obtained by differentiating equation (9) with respect to slip S to yield

dc t°ez [2Rz2 + 4 S R I 2 - 2Ri 2 - 2SKtRI' - 2SR22]. (IO)


2 C d-S = K K t

By equating d c l d S to zero equation (10) can be solved for slip to give

Rz2 - RI 2
S = R2Z _ Ri2(2 _ KD" (11)

For the standard coupling without a baffle plate it is usual to assume that K~ = 1 (see Ref. 2) and for this
case equation (11) predicts that the flow rate will be a maximum at S -- 1, i.e. at the stall condition. If the
incidence loss is increased by the introduction of a baffle plate, i.e. Kl is greater than 1, equation (11) shows
that the maximum flow rate will occur at slip rates less than one.
The experimental data, however, do not give information with respect to flow circulation rates but
indicate the effect of the baffle plate upon the torque-slip characteristic. Equation (2) gives the torque equation
as

¢ = p . Ac[~ooR22 - oJo(1 - S)R1 z] (12)


which can be differentiated with respect to slip S to yield

de - z dc 2
" ~ = p . A C . topRi + p A - ~ (topR2 - too(1 - S)RI2). (13)

Substituting for d c l d S from equation (10) and for C from equation (9), equating to zero yields a quadratic
equation which can be solved for S

S ' R I 2 ( 4 R I 2 - 2Rz2 - 2 K I R I 2) + S(Rz 2 - RI2)(SRI z - R22 - K I R I z) + (R22 - Ri2) 2 = 0. (14)

Figure 4 presents a plot of equation (14) for a 482 mm (19 in.) fluid coupling. This figure shows the slip at
which the torque will be a maximum for any assumed incidence loss parameter KI. When Kt ~< 1.32 the
maximum torque occurs at stall, i.e. at S = 1. If K~ is increased further the maximum torque occurs at
reduced slip magntudes. It should be noted that the circulation loss parameter K r does not influence the slip
rate at which the torque maximum occurs

L i n e a r velocity d i s t r i b u t i o n m o d e l
In this model the fluid is assumed to rotate about the mean radius R~ with a constant angular velocity oJ,
consequently the flow velocity increases linearly with distance from the mean radius. The detailed
equations developed for this model are fully described in Ref. (3).
The development here to accommodate the effects of a baffle plate is similar to that described above, Le. the
incidence and circulation losses are increased by the uses of the multipliers KI and K I.
From Ref. (3) the torque equation, for the particular case of a coupling with straight blades, can be
written as

¢ = ox.%(BI - (1 - S)B2) (15)

70 r I I I

484mrn dla Couphng


6 0i \ 254mrn dla Baffle Plate
\ Constant VelocityNodel
E 5C \\ i .... Linear Velocity Nodel
0

0 4C
\
\
\
\
\
c.)
c 3C
\ \
o

U
C
20

0.1 02 03 04 05 06 07 08 09 10
Slip
l~o. 4. Plot of incidence loss coefficient vs slip at which the maximum torque occurs.
The performance prediction of fluid couplings 733

where
B | = 2 c r p [ R 0 5 Ro4RM F R 5 ]
4 -~J

and
R/RM ~_R ~ ]
4 --~'1

and the total power dissipated is given by

PL = ~ p 2 S [ B i - (1 - S)B2]. (16)

This total power dissipation is equated to that bqven by the empirical loss models. From Ref. (3) the
incidence loss at pump inlet is given by

KI
P I | ----~ B2 ¢acap2S 2 (17)

where KI is now introduced to increase this loss due to the baffle plate.
At turbine inlet the incidence loss is given by

BI 2c,2
PI2 = " ~ COOp ,3 . (18)

The circulation loss was considered to be composed of two parts, one due to friction and the other due to
secondary flow losses. The total circulation loss being given by

Pc = co3(Kg+ KB) = Kr¢o 3. (19)

With the introduction of the baffle plate this loss was empirically increased to

Pc = KrKT¢°3. (20)

By summing equations (17), (18) and (20) and equating to equation (16), the circulation velocity ~o can be
derived from

[o:s,.,- s,,- F,B, +


To find the slip at which the torque is a maximum equation (15) is differentiated to yield

-- ~
dr wwpB2+ top(Bl - ( 1 - S)B2) -doJ
~ =O (22)
dS

~-~is obtained by differentiating equation (16)

zE 400

D
cr
0
~- 200 ~ ~

0
500 1000 1500 2000
Input Speed rev/rnln
F3G. 5. Torque absorption characteristics of 482 mm diameter coupling with 254 mm diameter
baffle plate.
734 A. WHITFIELD, R. S[VALINGAMand F. J. WALLACE

with respect to S and when substituted into equauon (18) y~elds

2 B 2 S 2 ( 2 B 2 - B ~ K I - B1) + S ( B I - B 2 ) ( S B 2 - B 2 K I - B t ) + ( B i - B 2 ) 2 = 0. (23)

This equation, which gives the slip at which the torque will be a maximum for any specified incidence loss
parameter KI, is plotted in Fig. 4. The resulting curve is similar in form to that obtained from equation (14)
of the previous section.

PRESENTATION AND DISCUSSION RESULTS


Experimental results were available from the manufactures for two fluid couplings with baffle plates,
these are shown in Figs. 5 and 6. Fig. 5 presents a coupling where the maximum torque is shown to occur at
the stall point, whereas with the coupling of Fig 6 the maximum torque is shown to occur at a speed ratio
of 0.5
In Figs. 7-9 the theoretical results are compared with the experimental torque characteristics. In these
figures the experimental curves of Figs. 5 and 6 have been transformed to a plot of non-dimensional torque
against speed ratio and the experimental results are presented as discrete points. Fig 7 presents the
theoretical results obtained from the constant velocity model With K~ and Kf both equal to one the
predicted stall torque is in excess of that achieved experimentally. The theoretical stall torque has been
made to agree with that obtained experimentally by increasing/(i or K s whilst maintaing the other equal to
one Of these results, shown on Fig. 7, the case with K / = 1.0 and Kz = 1.36 shows the best agreement with
the experimental results, although this cannot be considered to be good. Better agreement could be obtained
t~y further increasing KI and reducing K s, however ~t was not considered satisfactory to have Kf less than
one as this suggests a reduction m the circulation loss with the introduction of the baffle plate The
increased turbulence caused by the introduction of the baffle plate would be expected to increase the
circulation loss and lead to a value of Kr > 1. For the results shown in Fig. 7 a value of K i > 1 and
Kt < 1-36 could be apphed to give the correct stall torque with the predicted curve lying between the full
and chain dotted lines of Fig. 7. This, however, does not improve the correlation with the experimental
results. The multiplier K s is multiplying an already uncertain friction loss parameter K, equation (7), or Kr,

8oc 50 0 9o 95 I
•75 96 97
60£

zE 40C

E 200
#_
0
0 500 1000 1500 2000 2500
Input Speed rev/mln
FIG. 6 Torque characteristics of 450 mm fired coupling with 216 mm baffle plate

f
Constant Velocity Model~--- KI = 1'0 KF = 1 0
0 010 /--KI=136 KF=I 0
L--
K I = I 0 KF=I 84
0 00~ - - - - - - . _ °

-5o

*~ 0004

:~ 0 002

0 02 04 06 08 10
Speed Ratio ~ T / ~ p
FIG. 7 Comparison of theoretical and experimental results for a 482 mm diameter coupling
with a 254 diameter baffle plate
The performance prediction of fluid couplings 735

Lr~ear Velocity ModelIi--- KI = 114 K F = 1.00


0.010 KI=16 K F = 1.00
Q;
D KI = 1'14 K F = 1 70
0 008 O Experimento[
Results
Jtm
O00~ ---......

0004
"O

z
c
o
I
0002 -%
0 0.2 04 06 08 1'0
Speed Rabo co T/CO
P/

~6. 8. Comparison of theoretical and experimental results for a 482 mm diameter coupling
with a 254 mm diameter baffle plate.

0 006

l=r
i..
t~
o
13.
0°°°'00,
0 0 003 , o
r- 3
E
"I0
0-002 " " ~
!
¢-
o 0 001 -- Constont Velocity Mode[
Z
o Expenmentol Results
i I
0 0 2 0-4 0.6 0 8 10
Speed Rabo ~oT/cop
F3G. 9. Comparison of theoretical and experimental results for a 450 mm diameter coupling
with a 216 mm diameter baffle plate

equation (20). The friction loss parameter used was identical to that used with a 482 mm diameter coupling
without a baffle plate. 3 The introduction of a baffle plate effectively reduces the size of the coupling and if
this modifies the effective hydraulic diameter and flow path length a change in the magnitude of K and K r
can be anticipated. Current work is aimed at calculating the friction loss separately by calculating the
hydraulic diameter and flow path length and correlating it with Reynolds number. Until the friction loss
parameter can be fixed with confidence little can be gained by multiplying it by a factor Kf which differs
little from unity.
Fig. 8 shows the theoretical results obtained with the linear velocity distribution model in a similar
manner to that given in Fig. 7. By setting K~ equal to 1.6 and K / = 1-0 the predicted stall torque agrees with
that obtained experimentally and reasonable agreement is maintained at other speed ratios. The peak torque
is predicted to occur at a speed ratio of approx. 0.15 and not at the stall point as given experimentally.
However, the difference in torque magnitudes is small and it would have been difficult to observe this peak
experimentally.
Fig. 9 presents the theoretical and experimental results obtained with the 450 mm diameter coupling. For
both theoretical models K / w a s set equal to one and Ki set to make the predicted stall torque agree with
that obtained experimentally. It can be seen that the hnear velocity model shows good agreement with the
experimental results at all speed ratios.

CONCLUSIONS
By empirically increasing the incidence loss at pump inlet by a parameter K, the
effect of the introduction o f a baffle plate has been modelled. The results obtained
736 A. WHrrFIELD, R. SIVALINGAMand F J. WALLACE

with the linear velocity distribution model show better agreement with the experi-
mental results than those given by the constant velocity model.
Sufficient experimental results are not at present available, and it is not possible to
draw detailed conclusions and correlations between the empirical parameter Kt and
the effective baffle plate size. There is a need for a detailed systematic study of the
effect of baffle plate size on fluid coupling performance.
REFERENCES
1. F. J. WALLACE, M. S. GHADIRI-ZAREH and J. M. MYERS, I. Mech. Engrs Conf. on Land Transport
Engines~Economic vs Environment, Paper C22/77 (Jan. 1977).
2. J. W. QUALMANand E. L. EGBERT, S A E Trans. Workshop Meeting, 2nd F_An,pp. 183-197 (Adv. in Eng,
Vol. 5). SAE, New York (1973).
3. F. J. WALLACE, A. WHITFIELD and R SIVALINGAM,Int. J. of Mech. Sci. 20, 335-347 (1978).