ELSEVIER

Pressure Loss Characteristics of Orifice and

Perforated Plates
Guohui Gan • A study was conducted on the pressure loss characteristics of square-
Saffa B. Riffat edged orifice and perforated plates. Tests were carried out to determine
Institute of Building Technology, the pressure loss coefficient for thin plates in a square duct for a range of
University of Nottingham, Reynolds numbers. Computational fluid dynamics (CFD) was used to
Nottingham, United Kingdom NG7 2RD predict the loss coefficient, and the result was compared with experimental
measurement. The effect of plate thickness on the loss coefficient for the
orifice plate was studied using CFD. © Elsevier Science Inc., 1997

Keywords: computational fluid dynamics, orifice plate, perforated plate,

pressure loss coefficient, Reynolds number, tracer gas technique

INTRODUCTION flow emerging from a wide-angle diffuser and found that

the most uniform velocity distributions at the diffuser exit
Orifice plates are mainly used as a device of flow measure- were achieved using two perforated plates of 50% free
ment for fluid delivery systems. A literature review by area. According to Idelchik et al. [6], for the same free-area
Ward-Smith [1] shows that a great deal of work has been ratio the resistance created by a perforated plate placed in
done on the pressure drop characteristics of orifice plates a straight duct is the same as that of flow through an
in ducts. Cylindrical tube orifices have also been tested orifice plate. When interference exists between system
with the aim of achieving better measurement perfor-
components, however, the system resistance due to orifice
mance than that of orifice plates. Jorissen and Newton [2]
plates may differ from that due to perforated plates.
studied five cylindrical nozzles of various dimensions in a
Salvarli and Ward-Smith [7] showed that the performance
4-in. pipeline and found that the nozzles maintained a
of multihole orifice plates was, in general, less sensitive
constant coefficient of discharge at Reynolds numbers
than that of single-hole orifice plates to interaction ef-
much lower than standard pressure difference devices.
fects.
Use of an orifice plate for measurement of flow rate
The geometries and flow characteristics of orifice plates,
requires a long straight upstream duct because the up-
stream velocity profile affects the discharge coefficient of perforated plates, and cylindrical tube orifices are similar.
the plate [3] and also induces a large pressure loss through These constructions consist of a sudden sharp contraction
the device. In heating, ventilation, and air-conditioning of the flow area followed by a sudden enlargement. For a
(HVAC) systems, duct cross sections are generally large, thin orifice plate this occurs as single-jet flow, whereas for
so installation of such a plate results in considerable extra a thin perforated plate the fluid contracts during its pas-
fan power being required. Besides, straight sections of sage through the plate orifices and leaves the plate as
H V A C duct are generally not long enough to ensure the separate jets. The pressure losses for the thin plates are
flow conditions required for accurate air flow measure- thus associated with the entry into orifices and with the
ment using a pressure difference device. sudden expansion at their exit. For a thick orifice plate
On the other hand, perforated plates and screens are (i.e., cylindrical tube orifice), the total losses include not
often used to reduce flow nonuniformities in H V A C ducts only the sudden contraction and sudden expansion losses
or to create uniform flow diffusion from supply air outlets. but also friction losses in the orifice.
In chemical engineering, perforated plates are used, for Although a considerable amount of experimental infor-
example, as distillation trays to increase the heat and mass mation on the characteristics of orifice plates in pipe flow
transfer between fluids. In the design of perforated trays, is available in the literature, data for plates in noncircular
the resistance of the tray to fluid flow is of great impor- ducts are limited. The present investigation includes tests
tance. Many studies have been made to determine the to determine the pressure loss coefficient for an orifice
dry-tray pressure drop for distillation sieve plates, and as a plate in a square duct for a range of Reynolds numbers.
result a number of correlations are available in the litera- The pressure loss coefficient is compared for an orifice
ture for the pressure loss coefficient [4]. Sahin and Ward- plate and perforated plate under the same duct flow
Smith [5] investigated perforated plates for controlling the conditions. Computational fluid dynamics (CFD) is used

Address correspondence to Dr. G. Gan, Department of Architecture and Building Technology, Institute of Building Technology, University of
Nottingham, Nottingham, United Kingdom NG7 2RD.

Experimental Thermal and Fluid Science 1997; 14:160 165

© Elsevier Science Inc., 1997 0894- 1777/ 97/ $17.00
655 Avenue of the Americas, New York, NY 10010 PII S0894-1777(96)00041-6
 Pressure Loss of Orifice Plates 161

to study the influence of plate thickness on the loss pressure measurements and was calibrated against a pre-
coefficient of the orifice plate. cision nozzle in a wind tunnel. The measured static pres-
sure loss across the plate ranged from 50 to 250 Pa. The
THEORY estimated uncertainty in the measured pressure loss is
within _ 2%.
The pressure loss coefficient of an orifice plate or perfo- The mean velocity of air flowing through the duct was
rated plate, k, is defined as measured using the constant-injection tracer gas tech-
Aes nique. This method involves injecting a tracer gas (SF6) at
k = (1) a constant rate q (ma/s) at the entrance of the duct and
Pv' monitoring the concentration of tracer gas, C (ppm), near
where A P s is the static pressure loss across the plate (Pa), the end of the duct [8]. The mean air velocity, V (m/s), in
the duct with cross-sectional area A (m 2) is given by
Pv [ = ( 1 / 2 ) P V2] is the velocity pressure (Pa), p is the
fluid density (kg/m3); and V is the duct mean velocity q
(m/s). V = ~ - ~ X 10 6. (2)
For an orifice plate whose axis is coincident with that of
a duct or for a perforated plate with regularly spaced In order to maintain a constant and stable injection of
holes, the pressure loss coefficient is influenced by the tracer gas, a mass flow controller was used together with a
free-area ratio and the thickness/diameter ratio as well as storage tank as a reservoir. The concentration of tracer
Reynolds numbers. The free-area ratio is defined as the gas at the downstream end was sampled with the help of a
ratio of the total cross-sectional area within the orifices to pump, and the filtered gas was analyzed using an infrared
the cross-sectional area of the duct and is a measure of gas analyzer. The mass flow controller is accurate to _+ 1%
the extent to which the constriction obstructs the flow. according to the manufacturer's calibration. The gas ana-
The thickness/diameter ratio is the ratio of plate thick- lyzer was calibrated against a standard gas (SF6 of 200
ness to the diameter of orifice. ppm) with an accuracy of + 1 ppm. The overall uncer-
tainty in the measurement is estimated within + 5 % .
EXPERIMENTS To investigate the effect of flow rate on the pressure
loss coefficient, measurements were made for a range of
Experiments were conducted for a square-edged orifice duct mean velocities.
plate and a multiholed perforated plate of 0.002 m thick-
ness (Fig. 1). The diameter of the orifice plate is 0.239 m. Results a n d Discussion
The perforated plate has 145 uniformly spaced holes of
0.02-m diameter. When fitted in a square duct of 0.3 x 0.3 The measured mean velocity in the duct was between 5.1
m, both plates have a free-area ratio of 0.5. Figure 2 and 11.5 m / s , and the mean orifice velocity between 10.2
shows the schematic diagram of the test rig for the duct and 23.0 m / s . The corresponding Reynolds number for
with orifice plate. The axis of the orifice plate is coincident the orifice plate was 1.6 x 10 5-3.7 x 10 5. The Reynolds
with that of the duct. Tests were performed using air as number is based on the orifice diameter and the mean
the fluid. orifice velocity.
The basic parameters measured for the determination Figure 3 shows a typical static pressure distribution
of pressure loss coefficient are the pressure loss through along the duct fitted with the orifice plate. Pressure has
the plate and the duct mean velocity. The static pressure been nondimensionalized by the velocity pressure (Pv).
was measured using wall pressure tappings along the duct The pressure loss through the orifice plate was obtained
wall upstream and downstream of the plate together with by linear extrapolation of the pressure distribution from
a micromanometer. The precision of the manometer is 1 upstream and downstream to the constriction (solid
Pa. The manometer can be used for both velocity and straight lines in the figure).

I _

OOODO-D OOD
O 0 0 D O 0 0 0
000000000
O000000Q
QOO0Q0 Q
0 0O0-O
OOQQO
D~Q-O~OZO-OOQO
oooooOQQ
T OO OOO-O
OOQOOOOQ
O0000000O
OOO d
OOOOOOOO
000000000
QOOOOOOQ
OOQQO-OOQO
S°S°S°S°Soso;o;o;
0.3 r n
0.002 ra 0.3 m

Orifice plate Perforated plate

Figure 1. Dimensions of the orifice plate and perforated plate.

162 Guohui Gan and Saffa B. Riffat

~ Mlcromanometer
Straightener/''-"-- k Ori~ee/perforated plate

I lqlanifold [ Flow direction

Static pressure tapping
f ulator
| Gas flow readout Flow ( ~ ~ Filter Flow meter

ank SF6Analyser I k.~ IlJl

Figure 2. Schematic diagram of the test

rig for the measurement of pressure loss
coefficient. T R A C E R - G A S I N J E C T I O N UNIT T R A C E R - G A S SAMPLINQ ~YSTEM

The measured pressure loss coefficient for the orifice free-area ratio. This contrasts with the finding by Idelchik
plate is 3.15 _+ 0.14. The measured value is lower than the et al. [6], that the resistance created by a perforated plate
data of Miller [9] for flow in pipes. This might be at- is the same as that created by an orifice plate. The higher
tributed to the lower blockage near the centerlines of the pressure loss coefficient for the perforated plate may be
duct where air velocity is higher than that near duct attributed to the poor downstream velocity distribution
corners. The minimum distance between the edges of the when air flows out of the plate to coalesce into a number
orifice and duct is about 0.1 duct hydraulic diameter. For of discrete jets [9]. This is because coalescence affects the
a circular duct with the same cross-sectional area (equiv- geometry of the flow pattern, in particular the coefficient
alent duct diameter of 0.338 m), the distance would be of contraction of the jets [10].
doubled and hence there would be higher blockage to flow
at high velocity. The explanation also implies that the loss CFD PREDICTION
coefficient depends on the velocity distribution in ducts
and consequently Reynolds number. For a uniform veloc- Computational fluid dynamics was used to predict the
ity distribution the coefficient in a circular duct could be pressure distribution and pressure loss coefficient for the
the same as that in a noncircular duct, given the same orifice plate for varying plate thicknesses. The predictions
free-area ratio. However, for the turbulent flow region were carried out using the C F D package F L U E N T [11]. In
tested, the effect of Reynolds number on the pressure loss F L U E N T , equations for the conservation of mass, mo-
coefficient was found to be insignificant (Fig. 4). mentum, and other transport equations such as the turbu-
The measured pressure loss coefficient for the perfo- lence model equations are solved using the control vol-
rated plate is 3.5. This is higher than that for the orifice ume method in a three-dimensional body-fitted coordinate
plate. It follows that a multiholed plate has higher resis- system. In the previous investigation [8] it was found that
tance to flow than a single-holed plate with the same for incompressible flows the most reasonable prediction of
pressure distribution in ducts is achieved using the stan-

4.0

¢.
c;
0

-1
L _ _ m~

.e,
r 3.5 J
It
3.0
-2
~Z
/~, 2.5
-3' A
~ w
&

-4' 2.0
-4 -2 0 2 4 6 8 150000 200000 250000 300000 350000 400000

Distance from the plate (m) Reynolds number

Figure 3. Measured static pressure distribution along the Figure 4. Effect of Reynolds number on the measured pres-
duct with the orifice plate (Re = 3.2 × 105). sure loss coefficient for the orifice plate.
 Pressure Loss of Orifice Plates 163

2. I -----r-
i
[

J ! i ! '
, i
I
d i
O _.. I[ !i
O.
0 Mean o f cross section [
Mean o f wall
¢0
0.
-2- i
I
"1" Centre line I
d , i

#
o4" ' L :

-6''

-8-
-3
i
-2
t
-1
l,
0 1 2
I
3 4 Figure 5. Predicted static pressure
distribution along the duct with the
Distance from upstream edge of orifice plate (m) orifice plate (Re = 3.2 x 105).

dard two-equation turbulence model with the quadratic the weighted average of the static pressures on the inner
upwind difference scheme. circumference of the wall for the cross section normal to
Because of the enormous grid cells required to repre- the centerline of the duct. The pressure for the cross
sent the geometry of the perforated plate with circular section (mean of cross section) is the weighted average
holes alone, it was not possible to model the whole duct of the static pressures for the whole cross section normal
system fitted with the perforated plate using the available to the centerline of the duct. It can be seen that upstream
computer (Silicon Graphics, Indy LC, 32 MB memory). of the plate there is a gradual decrease in pressure due to
Therefore, a C F D prediction was made for a perforated the friction at the duct wall. In the immediate neighbor-
plate simplified with square holes of the same free-area hood of the plate there are rapid variations of pressure.
ratio as the testing plate. This simplification is considered The static pressure away from the wall and orifice surfaces
reasonable because the measured loss coefficient for a declines suddenly in the orifice due to contraction of the
sharp-edged plate with circular orifices can be applied to air passage. However, there is a rise in the static pressure
the plate with square holes provided that the free-area on the wall surface just before the plate. Part of the
ratio is the same [9]. pressure recovers when air flows out of the plate as a
result of conversion of velocity pressure to static pressure.
RESULTS AND DISCUSSION The maximum recovery occurs at about two orifice diame-
ters downstream of the plate. Further downstream the
Figure 5 shows the predicted static pressure distribution pressure distribution again settles down to that associated
along the duct with the orifice plate used in the tests. The with duct wall friction. It can also be seen that at any cross
pressure for the wall at a cross section (mean of wall) is section in the inlet or outlet duct the static pressure is

2-

>
~- 0-
12.
0 Mean o f cross section
t5
°m
@ Mean o f wall
Jr Centre line
-2-

t~

-4-
Q.

-6- i Figure 6. Predicted static pressure distri-

-2 -1 0 1 2
bution along the duct with the perforated
D i s t a n c e f r o m u p s t r e a m edge of plate (m) plate (Re = 3.2 X 105).
164 Guohui Gan and Saffa B. Riffat

essentially uniform. The variations in static pressure over 3.6

the cross section occur only at about one orifice diameter
upstream and downstream of the plate. 3.4
Figure 6 shows the predicted static pressure distribution
along the duct with the perforated plate. The variation of It
static pressure over a cross section of the duct near the 3.2
perforated plate is negligible and is much smaller than
that near the orifice plate. By comparing Fig. 6 with Fig. 5, _o
3.0
it is seen that the pressure recovery takes place when air
flows through the perforated plate sooner than when it X

flows through the orifice plate. This demonstrates that 2.8

X
perforated plates can reduce flow and pressure nonuni-
X
formities in ducts or provide uniform flow patterns from 2.6
diffusers. 0 1 2 3 4 5
The predicted loss coefficient for the thin orifice plate is Plate thickness, t/d
3.4. This is about 8% higher than the measured value for
the orifice plate (k = 3.15) and is 3% lower than the Figure 7. Effect of plate thickness on the predicted pressure
measured value for the perforated plate (k = 3.5). The loss coefficient for the orifice plate.
predicted loss coefficient for the perforated plate is 3.56
and agrees very well with the measurement. CFD can The implication of the predictions is that cylindrical
therefore be used to predict the pressure loss coefficient nozzles are a better device for flow measurement than
for orifice plates with reasonable accuracy and more sig- thin orifice plates due to smaller pressure losses through
nificantly for perforated plates by modeling as plates with the device and also smaller Reynolds number effects [2].
simplified orifices such as square holes or approximately On the other hand, comparison of Fig. 8 with Fig. 5 shows
as simple orifice plates. that the maximum wall pressure drop across the plate
Having validated the CFD model, further predictions where pressure tappings are positioned decreases with the
were performed to study the effect of plate thickness on plate thickness. The degree of measurement precision is
the loss coefficient. thus smaller using a cylindrical nozzle than when using an
The variation of loss coefficient with the ratio of plate
thickness to orifice diameter, t/d, is shown in Fig. 7 and
Table 1. Effect of Plate Thickness on the Predicted Pressure
also in Table 1. It is seen that the loss coefficient de-
Loss Coefficient for a Square-Edged Orifice Plate with a
creases considerably when the thickness is increased to Free-Area Ratio of 0.5
half the orifice diameter. The minimum loss coefficient is
attained with a t/d ratio of about 1.5. When the thickness t/d k
increases further, the variation in loss coefficient is small.
0.008 3.4O
Stichlmair and Mersmann [12] also show that the mini-
0.125 3.25
mum loss coefficient is achieved at a t / d ratio between 1
and 2 for Reynolds numbers from 400 to 10 6. The pre- 0.25 2.86
dicted trend of the variation in loss coefficient with plate 0.5 2.73
thickness is similar to that given by Miller [9] for pipe 1.0 2.71
flows, but the predicted value for flow in the square duct, 1.5 2.63
even if increased by 8% for the underprediction, is again 2.0 2.66
lower than that for the pipe flow for the same plate 3.0 2.64
thickness and free-area ratio. 4.0 2.66

2-

Of
> wJ
~L
D.
-2-'
d
P
-4,
u~
u~

D.
-6-,

t=4d

Figure 8. Predicted wall static pressure along -8 -j

the duct with orifice plate 4d thick (Re = 3.2 × -3 -2 -1 0 1 2 3 4
105).
Distance from upstream edge of orifice plate (m)
 Pressure Loss of Orifice Plates 165

orifice plate. If the variation also applies to a p e r f o r a t e d Pv velocity pressure, Pa

plate that requires low pressure losses, the plate should be q tracer gas injection rate, m 3 / s
thicker than the hole diameter, but then holes need to be Re Reynolds n u m b e r ( - V o D / v ) , dim.ensionless
drilled rather than punched. t thickness of orifice, m
V duct m e a n velocity, m / s
PRACTICAL SIGNIFICANCE
V0 mean orifice velocity, m / s
O n e of the objectives of this work was to establish a Greek Symbols
technique for predicting the pressure loss through orifice 1, kinematic viscosity, m Z / s
plates and p e r f o r a t e d plates and o t h e r duct fittings. This p density of air, k g / m 3
study has shown that C F D can be used to predict the
pressure loss coefficient for these types of plates and that
in making this prediction the g e o m e t r y of holes for a REFERENCES
p e r f o r a t e d plate can be simplified. The practical applica- 1. Ward-Smith, A. J., Pressure Losses in Ducted Flows, Butterworths,
tions of C F D include the effect of thickness, free-area London, 1971.
ratio, and cross-sectional a r e a of an orifice or p e r f o r a t e d 2. Jorissen, A. L., and Newton, H. T., Discharge Measurements by
plate on the pressure loss as well as flow and pressure Means of Cylindrical Nozzles, Trans. ASME, 74(7), 825-835,
interactions between a plate and c o m p o n e n t s upstream 1952.
and downstream of the plate. 3. Morrison, G. L., Hall, K. R., Macek, M. L., Ihfe, L. M., Deotte,
J. R. E., and Hauglie, J. E., Upstream Velocity Profile Effects on
Orifice Flowmeters, Flow Measure. Instrum., 5(2), 87-92, 1994.
CONCLUSIONS
4. Lockett, M. J., Distillation Tray Fundamentals, Cambridge Univ.
The following conclusions can be drawn from this study: Press, Cambridge, UK, 1986.
5. Sahin, B., and Ward-Smith, A. J., The Use of Perforated Plates
1. A t a free-area ratio of 0.5, the pressure loss coefficient to Control the Flow.Emerging from a Wide-Angle Diffuser, with
of a p e r f o r a t e d plate in a square duct is higher than Application to Electrostatic Precipitator Design, I, J. Heat Fluid
that of an orifice plate. Flow, 8(2), 124-131, 1987.
2. F o r turbulent flow with Reynolds numbers of 1.6 × 6. Idelchik, I. E., Malyavskaya, G. R., Martynenko, O. G., and
105-3.7 × 10 5, the influence of Reynolds n u m b e r on Fried, E., Handbook of Hydraulic Resistance, 2nd ed., Hemi-
the pressure loss coefficient is negligible. sphere, Washington, DC, 1986.
3. By increasing the plate thickness to 1.5 times the orifice 7. Salvarli, H., and Ward-Smith, A. J., Effects of Bend and Orifice
diameter, the pressure loss through an orifice plate can Plate Interactions in Internal Flow Systems, ASME-FED, 170,
be r e d u c e d substantially. 25-30, 1993.
8. Gan, G., and Riffat, S. B., k-Factors for HVAC Ducts--Numeri-
This work is funded by the U.K. Engineering and Physical Sciences cal and Experimental Determination, Building Services Engi-
Research Council. neering Research and Technology, 16(3), 133-139, 1995.
9. Miller, D. S., Internal Flow Systems, 2nd ed., British Hydrome-
chanics Res. Assoc., Cranfield, UK, 1990.
10. Ward-Smith, A. J., Internal Fluid Flow, Clarendon Press, Oxford,
NOMENCLATURE
UK, 1980.
A cross-sectional area of duct, m 2 11. FLUENT User's Guide, Fluent Inc., Lebanon, NH, 1995.
C concentration of tracer gas, p p m 12. Stichlmair, J., and Mersmann, A., Dimensioning Plate Columns
for Absorption and Rectification, Int. Chem. Eng., 18(2), 223-236,
d d i a m e t e r of orifice, m
1978.
k pressure loss coefficient, dimensionless
Ps static pressure, Pa
APs static pressure loss, Pa Received October 14, 1995; revised March 26, 1996