Journal of Applied Fluid Mechanics, Vol. 14, No. 1, pp. 215-226, 2021.

Available online at www.jafmonline.net, ISSN 1735-3572, EISSN 1735-3645.

DOI: 10.47176/jafm.14.01.31472

Numerical Study of Single-Hole and Multi-Holes Orifice

Flow Parameters
M. Đurđević †, M. Bukurov, S. Tašin and S. Bikić
Department of Energy and Process Engineering, Faculty of Technical Sciences, University of Novi Sad, Trg
Dositeja Obradovića 6, 21000 Novi Sad, Serbia

†Corresponding Author Email: [email protected]

(Received February 12, 2020; accepted June 17, 2020)

ABSTRACT

Importance of accurate fluid flow measurement in industry is crucial especially today with rising energy
prices. There is no ideal measuring instrument due to numerous errors occurring during process of physical
quantities measurement but also due to specific requirements certain instruments have like fluid type,
installation requirements, measuring range etc. Each measuring instrument has its pros and cons represented
in accuracy, repeatability, resolution, etc. Conventional single-hole orifice (SHO) flow meter is a very
popular differential-pressure-based measuring instrument, but it has certain disadvantages that can be
overcame by multi-holes orifice (MHO) flow meter. Having this in mind, the aim of this paper is to help gain
more information about MHO flow meters. Both SHO and MHO gas (air) flow meters with same total orifice
area and the pipe area ratio β were numerically studied and compared using computational fluid dynamics
(CFD). Simulation results of 16 different orifices with four different β (0.5, 0.55, 0.6 and 0.7) were analysed
through pressure drop and singular pressure loss coefficient. Standard k-ε turbulence model was used as a
turbulence model. Beside singular pressure loss coefficient, pressure recovery as well as axial velocity for
both the SHO and MHO were reported. Results showed lower (better) singular pressure loss coefficient and
pressure drop as well as quicker pressure recovery in favour of the MHO flow meters. Also, centreline axial
velocity results were lower for MHO compared to corresponding SHO. CFD simulation results were verified
by experimental results where air was used as a working fluid. The influence of geometrical and flow
parameters on singular pressure loss coefficient was also reported and results showed that MHO hole
distribution did not have significant influence on singular pressure loss coefficient.

Key words: CFD; Orifice flow meter; Multi-hole orifice flow meter; Pressure loss coefficient.

NOMENCLATURE

A SHO opening area k turbulence kinetic energy

A1 MHO central opening area ṁ mass flow rate
A2 MHO peripheral opening area p absolute pressure
Ao total opening area (SHO and MHO) p1 gauge pressure upstream of the orifice
Ap pipe area Q volumetric flow rate
Cµ turbulent viscosity constant R universal gas constant
Cd discharge coefficient Re Reynolds number
D internal pipe diameter S strain rate parameter
d1 SHO and MHO central hole diameter T temperature
d2 MHO peripheral hole diameter Z axial distance
dc MHO homocentric circle diameter Z compressibility factor
e thickness of the orifice Zn dimensionless axial distance
E thickness of the plate ∆p pressure drop
Gk turbulent production µ dynamic viscosity
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

µt turbulent viscosity ε turbulence dissipation rate

pin upstream absolute pressure ξ singular pressure loss coefficient
Pn dimensionless pressure ρ density
Pz absolute pressure at axial distance σk, σε turbulent Schmidt number
α angle of bevel τ shear stress
β total orifice opening area and the pipe area υ average velocity through the orifice
ratio opening area
ε expansibility (expansion) factor υa centreline axial velocity

1. INTRODUCTION and Andersson in their paper (Erdal and Andersson,

1997) presented the effects of the mesh, boundary
Many industries, especially energy industry is conditions, the discretization schemes and the
relaying on accurate flow measurements. turbulence models on gas flow simulation through
Differential-pressure-based instruments are widely the pipe restriction. Good agreement between
spread measuring devices with estimates that at simulation results and experimental measurements
least 40% of industrial flowmeters in use nowadays authors achieved using standard k-ε turbulence
are differential-pressure-based where orifice flow model. Despite the use of few different turbulence
meter is the most popular one (Reader-Harris, models it was not possible to predict the accurate
2015). Installing an orifice plate inside a pipe pressure drop at the pipe restriction. Gan and Riffat
carrying fluid, causes a pressure drop, making it (1997) studied pressure loss characteristics of
possible to indirectly determine flow rate by square edged orifice and perforated plates in square
measuring pressure difference upstream and ducts. They used CFD simulations to predict
downstream of the orifice. This procedure is pressure distribution and (pressure) loss coefficient
described in standard EN ISO 5167-2:2012 where and to study effects of plate thickness. Ellman and
the flow rate is given by: Piche (1999) proposed a two-regime flow formula,
which has a smooth transition between laminar and
Cd ε d 2 π transition regimes for numerical simulations of fluid
ṁ = √2 ρ ∆p power circuits. Eiamsa-Ard et al. (2008)
4 (1)
√1 - β 4 investigated the effects of β (for 0.5, 0.6 and 0.8)
on the flow field of single-hole orifice using both
The parameter β is represented as ratio of square the standard k-ε turbulence model and Reynolds
root of total orifice opening area and the pipe area. stress model (RSM). Simulation results of both the
standard k-ε turbulence model and RMS are in good
Ao agreement with experimental measurements
.β = √ (2)
Ap (Eiamsa-Ard et al., 2008). Drainy et al. (2009) in
their numerical study of Zanker plate compared
Orifice flow meters were topics of numerous results of different turbulence models (standard k-ε,
experimental researches with different fluids and realizable k-ε, Reynolds Stress Model - RSM, large
flow regimes. Teyssandier and Husain (1987) Eddy Simulation - LES, and Detached Eddy
experimentally studied wall and orifice plate Simulation - DES) with pressure drop correlations
pressure gradients in air flows over a wide range of of BS EN ISO 5167-2:2003 while also studying the
Reynolds numbers from 21000 to 160000. Pressure effects of plate thickness and Reynolds number on
loss coefficients for square-edged orifices with β flow characteristics of Zanker plate. Numerical
ratios of 0.2, 0.3, 0.57 and 0.7 for Newtonian and methodology for predicting the calibration
non-Newtonian fluids in both laminar and turbulent coefficients of the orifice meter using water as a
flow regimes was experimentally determined by fluid was presented by Oliveira et al. (2010). Arun
Ntamba Ntamba and Fester (2012). Over the years et al. (2010) carried out CFD simulations to study
with growing application of computers in effects of different β (0.5, 0.6 and 0.8) and pipe
engineering, orifice flow meters became topic of diameter for non-Newtonian fluid. Hollingshead et
numerous numerical studies. In last two decades al. (2011) numerically studied discharge
CFD proved as a versatile tool for flow meters flow coefficients of different pressure differential devices
prediction (Singh et al. 2010). As a starting point (Venturi, standard orifice plate, V-cone and wedge
for the simulation of fluid flow through the orifice flow meters) at low Reynolds numbers. Shah et al.
plate, problems of flow simulation through the (2012) presented a comprehensive numerical study
restricted pipe can be observed. In one of such of flow through a single-hole orifice meter. Authors
papers Durst and Wang ( 1989) found a good match validated CFD simulation results with published
between the simulation results and the experimental experimental data of Nail (1991) and Morrison et
measurements. In numerical simulations, the k-ε al. (1993) as well as with pressure drop and energy
turbulence model was used, while the velocity was balance of their experimental data. Reader-Harris et
measured using the laser Doppler anemometer al. (2012) simulated contamination of the orifice
(LDA). The pressure drop caused by the fluid flow plate both by experiment and CFD simulations and
through the pipe restriction was not presented. Erdal good agreement between the two was obtained.

216
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

Elsaey et al. (2014) numerically studied turbulent while singular pressure loss coefficient  and
flow through fractal-shaped orifice with air as a pressure recovery were analysed. The parameter β
working fluid. Shaaban (2014) studied orifice and homocentric circle diameter dc were parameters
meter’s energy consumption optimization using that were varied by changing orifice geometry,
numerical simulations. Dong et al. (2018) studied while Reynolds numbers were varied by changing
measuring accuracy of the traditional stainless-steel air velocity. To analyse pressure recovery two
orifice flowmeter and improved carbide orifice nondimensional values were introduced, Pn and Zn
flowmeter by changing the entrance sharpness of that are defined as:
the two orifices using CFD. Morrison et al. (1994)
presented a slotted orifice flow meter as a Pz
replacement for conventional single hole orifice. Pn = (6)
Pin
Comparing experimental results of conventional
single-hole orifice and slotted orifice flow meter Z
Zn = (7)
proved that slotted orifice flow meter has much D
greater discharge coefficient and that its
substantially less sensitive to upstream flow
conditions. Barki and Ganesha (2014) compared 2. ORIFICE AND MULTI-HOLE
single-hole orifice and multi-hole orifice using ORIFICE DESIGN
water as a fluid in their CFD simulations. Besides
comparison of SHO and MHO, authors studied the Geometry of simulated MHO is similar to MHO
effects of multi-hole orifice plate hole distribution designed by Singh and Tharakan (2015), who
on flow rate, pressure drop, velocity and turbulence numerically studied MHO with same β ratio, while
intensity. Numerical analysis showed better flow demineralized water was used as a working fluid.
characteristics of MHO comparing to SHO. Singh Authors motivation for the current study was to see
and Tharakan (2015) numerically studied SHO and how these orifices will react in gaseous fluids. Out
MHO flow meters over a wide range of Reynolds of 16 orifice plates used in this study, 4 were single-
numbers using demineralized water as fluid. hole orifices (SHO) and 12 multi-hole orifices
Comparing SHO and MHO flow meters results (MHO), with 4 different β ratios (0.5, 0.55, 0.6 and
proved that MHO has better pressure recovery, 0.7) where for each β one SHO and three MHO
greater discharge coefficient and lower downstream were designed. MHO were designed in the way
velocity. having equal open area for flow as SHO with
In this paper authors present CFD simulation results corresponding β. The MHO plates were designed
of SHO and MHO flow meters singular pressure with eight smaller circular openings evenly
loss coefficient, flow patterns and pressure distributed on a homocentric circle in addition to
recovery. CFD simulation results were compared the central opening. The details of the different
with experimental results and good agreement was multi-hole and single hole orifice plates used for the
obtained. All analyses in CFD simulations and computations in this study are given in Table 1 and
experiments considered fluid (air) compressibility, shown in Fig. 1 a).
where fluid density was defined as:
In CFD model certain geometry approximations
p were made. Figure 1 shows approximated CFD
ρ= (3) model geometry without bevel angle α=30 ° (a) and
ZRT
geometry of experimentally tested orifices with
To calculate discharge coefficient Cd from Eq. (1) bevel angle α=30 ° (b).
expansion factor ε needs to be determined.
Expansion factor ε is determined experimentally, Table 1 SHO and MHO orifice dimensions
but for the MHO it wasn’t determined, so singular
D E d1 d2 dc
pressure loss coefficient  was introduced as a β [-] Orifice
[mm] [mm] [mm] [mm] [mm]
parameter for the analysis and comparison of SHO
and MHO. Expressions for singular pressure loss SHO 3.5 35.2 - -
coefficient are: MHO1 3.5 18.0 10.7 40.0
0.5 70.3
MHO2 3.5 18.0 10.7 45.0
∆p MHO3 3.5 18.0 10.7 52.0
ξ= (4)
0.5ρυ2 SHO 3.5 38.7 - -
and MHO1 3.5 21.0 11.5 43.0
0.55 70.3
MHO2 3.5 21.0 11.5 47.0
1-β4 MHO3 3.5 21.0 11.5 52.0
ξ= (5)
(Cd ε)2 SHO 3.5 42.2 - -
MHO1 3.5 23.0 12.5 42.0
The open literature contains no information on 0.6 70.3
MHO2 3.5 23.0 12.5 49.0
simulated MHO geometry for gaseous fluids. In MHO3 3.5 23.0 12.5 54.0
total 16 orifices with four different β (0.5, 0.55, 0.6 SHO 3.5 49.2 - -
and 0.7) were studied using commercial CFD code MHO1 3.5 28.0 14.3 48.5
Simcenter Star CCM+, Siemens. Input parameters β 0.7 70.3
MHO2 3.5 28.0 14.3 50.0
ratio, homocentric circle diameter dc (hole MHO3 3.5 28.0 14.3 53.0
distribution) and Reynolds number were varied,

217
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

a) CFD orifice model geometry b) Tested orifice model geometry

Fig. 1. CFD and tested orifice geometry.

MHO plate works on the same principal as SHO downstream of the orifice region to help simulation
causing pressure drop which is indirectly used to converge. There were three different regions (Fig.
determine flow rate. Although orifice flow meters 2) with Region 1 around orifice having finer mesh
are widely spread measuring instruments, there is of around 130000 cells and upstream and
no available standard covering MHO, but Eq. (1) downstream regions (Region 2 and Region 3)
can be used to determine the flow rate. having coarser mesh with around 10000 and 120000
cells respectively.
3. CFD MODEL AND NUMERICAL 3.2 Governing Equations
PROCEDURE
The governing equations for flow through SHO and
CFD simulations for total of 16 orifice pates were MHO are given by
carried out under fully developed laminar and Continuity equation:
turbulent conditions in a 70.3 mm internal diameter
and 3940.3 mm long pipe. Upstream section of 12 ∂ρ
+∇∙(ρv𝑖 )=0 (8)
D and downstream section of 44 D with orifice ∂t
plate thickness of 3.5 mm were modelled (Fig. 2).
Momentum equation:
To obtain the Reynolds numbers variation from 500
to 600000, air mass flow rate was varied from ∂
0.00051 kg/s to 0.61415 kg/s. (ρv𝑖 )+∇∙(ρv𝑖 v𝑖 )=-∇p+∇∙τ (9)
∂t
3.1 Mesh design Standard k-ε turbulence model was used in
numerous orifice CFD studies (Erdal and
Generating a high-quality grid is one of key issues Andersson, 1997; Shah et al., 2012; Singh and John
in CFD simulations as it governs the stability and Tharakan, 2015), hence authors chose this
accuracy of the flow predictions. For the present turbulence model for computations. Standard k-ε
study of SHO and MHO unstructured polyhedral turbulence model is given by:
grid was used with extruded regions upstream and

218
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

Region 2 Region 1 Region 3

1213DD 44
44 DD

Fig. 2. Regions and Mesh.

t=300.18 K) were used to simulate flow through the

orifice for all meshes.
∂ ∂ ∂ μ ∂k
(ρk)+ (ρkvi )= [(μ+ t ) ] +Gk -ρε (10)
∂t ∂xi ∂xj σk ∂xj
∂ ∂ Table 2 Mesh sizes for grid analysis
(ρε)+ (ρεvi )=
∂t ∂xi Mesh Region 1 Region 2 Region 3
(11)
∂ μ ∂ε ε ε2
= [(μ+ t ) ] +C1ε Gk -C2ε ρ Mesh 1 17478 3330 37800
∂xj σε ∂xj k k Mesh 2 29375 3240 40800
Where C1ε, C2ε, σk and σε are standard k-ε turbulence Mesh 3 129389 9680 118680
model constants while Gk is turbulent production Mesh 4 298993 15340 182040
and it is modelled as:
2 2
Gk = μt S2 - ρk∇∙v𝑖 - μt (∇∙v𝑖 )2 (12) Figure 3 shows pressure recovery variation for four
3 3 types of chosen meshes. Two coarser meshes (Mesh
Turbulent viscosity is computed as: 1 and Mesh 2) have significant change in pressure
recovery pattern while finer meshes (Mesh 3 and
k2 Mesh 4) have no significant change in pressure
μt = ρCμ (13) recovery pattern.
ε
1.01
All governing equations were solved by segregated
solver while all solutions were considered to be 1
fully converged when each of the residuals was less
0.99
than 10-6.
Pn [-]

0.98
3.3 Boundaries
0.97
Air at 25° C was used as the working fluid. CFD
simulations were carried out over a wide range of 0.96
Reynolds numbers from 500 to 600000 while mass
flow rate boundary condition was specified at the 0.95
-20 0 20 40
pipe inlet. Flow split outlet boundary represents the Zn [-]
outlet of a duct and it was used to model flow exit
Mesh 1 Mesh 2
with conservation of mass. No slip boundary
Mesh 3 Mesh 4
condition was used for wall surfaces and high y+
wall treatment for near-wall modelling. This Fig. 3. Pressure profiles for different mesh size.
treatment assumes that the near-wall cell lies within
the logarithmic region of the boundary layer Convergence was assessed by monitoring pressure
(Siemens, 2019). drop results and comparing them with experimental
3.4 Grid Sensitivity Analysis pressure drop result (Fig. 4). The difference
between Mesh 3 and Mesh 4 is 1.01% for the
In order to find optimal mesh size for this study, pressure drop while the difference between Mesh 3
grid sensitivity analysis was performed. Four and experimental results is 0.88% compared to
different size meshes (Table 2) were studied where Mesh 4 where difference between simulation and
each had more cells than the previous. Same experimental results is 0.12%. Hence, Mesh 3 was
conditions (Q=166.84 m3/h, p=399 kPa and chosen for further simulations.

219
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

29 b) are showing cross-sectional velocities

Experiment downstream of the orifice, where velocities close to
27
Simulation the pipe wall for both SHO and MHO2 can be seen.
25 From these figures, it can be seen that MHO2 has
Δp [mbar]

23 lower velocity close to the pipe wall compared to

SHO. These trends are same for all simulated
21
orifices, so for the corresponding SHO, MHO will
19 have higher pressure downstream which is leading
to lower pressure drop for same flow rate in favour
17
0 100000 200000 300000 400000 500000 600000 of MHO.
Cells [-] Figure 9 b) shows cross-sectional pressure
Fig. 4. Pressure drop comparison for different distribution downstream of the orifice for MHO2
Meshes. where pressure is not evenly distributed especially
close to the pipe wall. This can prove to be
important when placing MHO in the pipe especially
3.5 CFD Model Validation if pressure drop is measured 1 D upstream and ½ D
downstream of the orifice. Orientation of MHO
CFD model was validated with experimental data of smaller openings and position of pressure taps
SHO and MHO flow meters (Đurđević et al., 2019). should be taken into account, as this can lead to
Details of test facility with straight sections of 12 D inadequate pressure drop measurements and
upstream and 4 D downstream of the orifice are indirectly incorrect flow rate. Although downstream
shown in Fig. 5. Elster quantometers type QAe250 pressure differences close to the pipe wall are not
and type QAe650 were used for air flowrate that great, they should be taken in consideration.
measurement with former having a measuring range
(20-400) m3/h and latter (50-1000) m3/h, with an 4.1 Singular Pressure Loss Coefficient
accuracy of +/–1.5% over a scale of 20% to 100%
of Qmax while accuracy of +/–3% was over a scale Variations in singular pressure loss coefficient of
of 10% to 20% of Qmax. Upstream pressure was SHO and MHO plates are presented in Fig. 10.
measured with Yokogawa’s absolute pressure Four SHO (for β: 0.5, 0.55, 0.6 and 0.7) have
transmitter model EJA510A with measuring range pressure loss coefficient values of 2.304, 2.182,
(0-10) MPa and accuracy of +/–0.2%. Pressure drop 2.077 and 1.732, respectively. Comparing
over orifice plates was measured with Yokogawa’s simulation results it is evident that MHO have
differential pressure transmitter model EJA110A lower singular pressure loss coefficient than SHO.
with measuring range (0-50) kPa and accuracy of Greatest improvement in singular pressure loss
+/–0.065%. coefficient can be seen for orifice plate MHO3
with β ratios 0.5 and 0.55 and orifice plate MHO2
Figure 6 shows comparison of simulation and with β ratios 0.6 and 0.7. Comparing to SHO of
experimental results for singular pressure loss corresponding β ratios, these MHO have a decrease
coefficient of SHO and MHO1 for β=0.7. Good of 30.81%, 31.46%, 32.57% and 32.39% in
agreement between the two has been achieved with singular pressure loss coefficient for β ratios of 0.5,
maximum deviation being within 14% for SHO and 0.55, 0.6 and 0.7, respectively. Hole distribution is
7% for MHO1. These results are validating adopted not as significantly as β ratio affecting singular
CFD model. Therefore, this model can be used for pressure loss coefficient where decreases of the
further computations of SHO and MHO with coefficient influenced by hole distribution is by
different β ratios. 10.00%, 10.68%, 14.45% and 4.86% for MHO3
with β ratios 0.5 and 0.55 and MHO2 with β ratios
4. RESULTS AND DISCUSSIONS 0.6 and 0.7 respectively. This leads to a conclusion
that orifice plates MHO3 with β ratios 0.5 and 0.55
and orifice plates MHO2 with β ratios 0.6 and 0.7
Numerical procedure described in section 3 was
have optimal homocentric circle diameter in terms
used to simulate air flow through SHO and MHO.
of singular pressure loss coefficient.
Figure 7 shows predicted velocity scalar fields (Re
70000) for SHO and MHO2 for β=0.7. Both SHO In Fig. 11 singular pressure loss coefficents are
and MHO have equal open area, but MHO have plotted for all simulated orifice plates at different
more evenly spread area comparing to SHO where Reynolds numbers making it possible to estimate
there is only one central opening (Fig. 1 and Fig. 7). which orifices have equivalent singular pressure loss
With fluid passing through the SHO, large sized coefficients and indirectly discharge coefficients. It
eddies are generated downstream of the plate can be seen that with β ratio increase, orifices
comparing to MHO where smaller sized eddies are sensitivity to Reynolds numbers is decreasing while
generated downstream. MHO are less sensitive to Reynolds number change
than SHO. MHO singular pressure loss coefficient is
Both SHO and MHO2 have equal upstream
not as influenced by Reynolds number change like
velocities close to the pipe wall, as it is shown in
SHO, making in this way MHO more applicable over
cross-sectional velocity distributions upstream of
wide range of Reynolds munbers.
the orifice in Fig. 8 a) and 9 a). Figures 8 b) and 9

220
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

Fig 5. Test facility.

2.5

1.5
ξ [-]

0.5

0
0 100000 200000 300000 400000 500000 600000 700000
Re [-]

SHO β=0.7 exp MHO1 β=0.7 exp

SHO β=0.7 CFD MHO1 β=0.7 CFD

Fig. 6. Comparison of experimental and simulated results for SHO and MHO1 β=0.7.

Fig. 7. Downstream velocity scalar fields for SHO and MHO2 β=0.7 at Re 70000.

221
 M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

a) 1 D upstream b) ½ D downstream

Fig. 8. Cross-sectional velocity and pressure distribution for SHO β=0.7 at Re 70000.

a) 1 D upstream b) ½ D downstream
Fig. 9. Cross-sectional velocity and pressure distribution for MHO2 β=0.7 at Re 70000.

Knowing that expansion factor  is very close to 1, discharge coefficient comparing to SHO.
discharge coefficient Cd can be determined using Eq.
(5). It can be seen that with decrease of singular 4.2 Pressure Recovery
pressure loss coefficeint , discharge coefficeint Cd For all simulated orifices over the whole range of
increases, therefore MHO have higher (better) Reynolds numbers, MHO had faster pressure

222
 M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

3 3

2.5 2.5

2 2

ξ [-]
ξ [-]

1.5 1.5

1 1

0.5 0.5

0
0
0 100000 200000 300000 400000 500000 600000 700000
0 100000 200000 300000 400000 500000 600000 700000
Re [-]
Re [-]

a) β=0.5 b) β=0.55
3 3.0

2.5 2.5

2 2.0
ξ [-]

ξ [-]
1.5 1.5

1 1.0

0.5 0.5

0 0.0
0 100000 200000 300000 400000 500000 600000 700000 0 100000 200000 300000 400000 500000 600000 700000
Re [-] Re [-]

c) β=0.6 d) β=0.7
0
SHO MHO1 MHO2 MHO3
0 200000 400000 600000 800000
Fig. 10. Singular pressure loss coefficient for SHO and MHO.

β 0.7 MHO3
β 0.7 MHO2
β 0.7 MHO1
β 0.7 SHO
β 0.6 MHO3
β 0.6 MHO2
Re 1000
β 0.6 MHO1 Re 2000
β 0.6 SHO
β 0.55… Re 10000
β 0.55… Re 70000
β 0.55…
β 0.55 SHO Re 300000
β 0.5 MHO3
Re 600000
β 0.5 MHO2
β 0.5 MHO1
β 0.5 SHO

0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3 2.5

ξ [-]
Fig. 11. Singular pressure loss coefficients at different Reynolds numbers.

recovery compared to SHO. Figure 12 shows pressure recovery compared to MHO. With an
pressure recovery pattern for SHO and MHO with increase of Reynolds number to 3000, MHO with β
optimal homocentric circle diameter for all β ratios ratios of 0.5, 0.55, 0.6 have slightly greater pressure
at different Reynolds numbers. At low Reynolds recovery then corresponding SHO, while MHO
numbers (Re 1000) SHO have slightly greater with β=0.7 still has lower pressure recovery

223
 M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

1.000005
1.0000005

1
1

0.9999995 0.999995

0.999999 0.99999

Pn [-]
Pn [-]

0.9999985 0.999985

0.999998 0.99998

0.9999975 0.999975

0.999997 0.99997

0.9999965 0.999965
-15 -5 5 15 25 35 45 -15 -5 5 15 25 35 45
Zn [-] Zn [-]

a) Re 1000 b) Re 3000

1.05
1
1

0.95 0.9

0.9
0.8
0.85 Pn [-]
Pn [-]

0.8 0.7

0.75
0.6
0.7
0.5
0.65

0.6 0.4
-15 -5 5 15 25 35 45 -15 -5 5 15 25 35 45
Zn [-] Zn [-]

c) Re 300000 d) Re 600000

Fig. 12. Pressure recovery for SHO and MHO with optimal homocentric circle diameter.

compared to corresponding SHO. At lower This point of the flow stream where the diameter of
Reynolds numbers (1000 and 3000) pressure the stream is the least and fluid velocity at its
recovery both for SHO and MHO is significant as maximum is called vena-contracta. Beyond this
almost all lost pressure is recovered, but pressure point velocity is just decreasing to the approximate
recovery rate is insignificant. Pressure recovery at value it had at pipe inlet section just before the
higher Reynolds numbers (300000 and 600000) is orifice, which was the case for all simulated
greater for MHO compared to SHO for all orifices.
simulated orifices.
Table 3 Pressure recovery for higher Reynolds
Table 3 shows pressure recovery rates for higher numbers
Reynolds numbers and their improvements. It can
Re
be seen that with an increase of Reynolds number β
pressure recovery is decreasing. Greatest Orifice
[-] 200000 300000 600000
improvement of 7.7% was found for MHO3 with
β=0.5 at Reynolds number of 600000 while the SHO 89.4% 76.1% 58.5%
least improvement of 0.2% for MHO2 with β ratio 0.5 MHO3 91.3% 80.4% 66.2%
0.7 at Reynolds number of 200000. Improvement 1.9% 4.3% 7.7%
4.3 Axial Velocity and Vena-Contracta SHO 93.6% 85.5% 75.1%
0.55 MHO3 94.6% 87.7% 78.8%
The centreline axial velocity profiles for SHO and Improvement 1.0% 2.2% 3.7%
MHO are given in Fig. 13. In inlet section of the SHO 95.9% 90.9% 84.3%
pipe, flow is uniform i.e. centreline axial velocity is 0.6 MHO2 96.5% 92.2% 86.6%
unchanged. As approaching the orifice opening Improvement 0.7% 1.3% 2.3%
centreline axial velocity is increasing both for SHO SHO 98.3% 96.4% 93.8%
and MHO flow meters. Maximum centreline axial
0.7 MHO2 98.5% 96.7% 94.3%
velocity can be observed downstream of the orifice
Improvement 0.2% 0.3% 0.5%
while the pressure has it lowest value at this point.

224
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

35 compared to SHO. This could lead to a conclusion

30 that for MHO flow meters moving downstream
25 pressure tap location closer to the orifice plate
(vena-contacta) greater pressure drop could be
υa [m/s]
20

15
recorded. In this way it would be possible to use
pressure measuring instruments of lower quality
10
than if downstream pressure tap location was on
5
D/2 distance from the orifice plate.
0
-1000 0 1000 2000 3000 4000

Position [mm] 5. CONCLUSION

a) β=0.5
In this study SHO and MHO with four different β
30 ratios (0.5, 0.55, 0.6 and 0.7) were numerically
25 studied over a wide range of Reynolds numbers
with air as a working fluid. Numerical model was
20
compared to experimental results and good
υa [m/s]

15 agreement was achieved.

10
SHO and MHO singular pressure loss coefficients
5 were studied by varying structural parameters like
hole size (β ratio) and hole distribution. Simulation
0
-1000 0 1000 2000 3000 4000 results showed a significant decrease of MHO
Position [mm] singular pressure loss coefficient compared to SHO
b) β=0.55 by 30.81%, 31.46%, 32.57% and 32.39% for β
ratios 0.5, 0.55, 0.6 and 0.7 respectively. Not so
25
significant influence was recorded on singular
pressure loss coefficient by MHO hole distribution.
20 Beside singular pressure loss coefficient, pressure
recovery for both the SHO and MHO was reported.
υa [m/s]

15
Comparing to SHO faster pressure recovery was
10 recorded with all simulated MHO. MHO had
5
greater pressure recovery compared to
corresponding SHO at Re>3000. This can be
0 attributed to MHO geometry, as they have more
-1000 0 1000 2000 3000 4000
evenly spread open area compared to SHO and in
Postion [mm]
that way causing less of an obstruction in fluid
c) β=0.6 flow. Greatest pressure recovery was recorded for
MHO with β=0.5 as well as pressure recovery
16
increase with β ratio increase. Lower velocities
14
downstream of the MHO are leading to pressure
12
drop decrease thus requiring better pressure
10
υa [m/s]

measuring devices compared to SHO for the same

8
flow rates. Centreline axial velocity analyses
6
showed that MHO flow meters have lower axial
4
2
velocity compared to SHO as well as different
0
vena-contracta position.
-1000 0 1000 2000 3000 4000
Position [mm] Orifice flow meters are still very popular measuring
d) β=0.7 flow instruments due to its affordability, ease of
use, reliability, simplicity, maintenance etc., but
they also have disadvantages like low pressure
recovery. Presented results showed advantages of
Fig. 13. Centreline axial velocity for SHO and MHO compared to SHO in singular pressure loss
MHO at Re 70000. coefficient and pressure recovery and in this way
put MHO one step closer of becoming a drop-in
replacement for the SHO, but further research is
All simulated MHO flow meters had lower needed.
centreline axial velocity compared to corresponding
SHO. Decrease in centreline axial velocity was
16.33%, 15.75 %, 14.38 % and 12.9 % for orifices
ACKNOWLEDGMENTS
with β ratios 0.5, 0.55, 0.6 and 0.7, respectively.
Hole distribution influence was insignificant to The authors gratefully acknowledge all support
maximum centreline velocity. from GasTeh d.o.o. Inđija that allowed use of their
test facility as well as manufacturing all necessary
From Fig. 13. it can be seen that, vena-contracta orifices and parts for the study. This paper is a
position is different for SHO and MHO flow result of research within the project TR31058, 2011
meters. For all simulated orifices, MHO vena- – 2019, financially supported by the Ministry of
contracta position was closer to orifice plate

225
M. Đurđević et al. / JAFM, Vol. 14, No. 1, pp. 215-226, 2021.

Education, Science and Technological 14(2), 160–165.

Development, Republic of Serbia.
Hollingshead, C. L., M. C. Johnson, S. L. Barfuss
and R. E. Spall (2011). Discharge coefficient
REFERENCES performance of Venturi, standard concentric
orifice plate, V-cone and wedge flow meters at
Arun, N., S. Malavarayan and M. Kaushik (2010). low Reynolds numbers, Journal of Petroleum
CFD Analysis on Discharge Co- Efficient Science and Engineering 78(3–4), 559–566.
During Non- Newtonian Flows Through
Orifice Meter, International Journal of Morrison, G. L., R. E. Deotte Jr, G. H. Nail and D.
Engineering Science and Technology 2(7), L. Panak (1993). Mean velocity and turbulence
3151–3164. fields inside a β=0.50 orifice flowmeter,
AIChE Journal 39(5), 745–756.
Barki, M. and T. Ganesha (2014). CFD Analysis
and Comparison of Fluid Flow Through A Morrison, G. L., K. R. Hall, J. C. Holste, M. L.
Single Hole And Multi Hole Orifice Plate, Macek, L. M. Ihfe, R. E. Deotte Jr and D. P.
International Journal of Research in Advent Terracina (1994). Comparison of orifice and
Technology 2(8), 6–15. slotted plate flowmeters, Flow Measurement
and Instrumentation 5(2), 71–77.
Dong, J., C. Jing, Y. Peng, Y. Liu, H. Ren and X.
Liu (2018). Study on the measurement Nail, G. H. (1991) A study of 3-Dimensional flow
accuracy of an improved cemented carbide through orifice meters. Ph. D. thesis, Texas
orifice flowmeter in natural gas pipeline, Flow A&M University, USA.
Measurement and Instrumentation 59, 52–62. Ntamba Ntamba, B. and V. Fester (2012). Pressure
Drainy, Y., K. Saqr, H. Aly, M. Nazri and M. Jaafar Losses and Limiting Reynolds Numbers for
(2009). CFD analysis of incompressible Non-Newtonian Fluids in Short Square-Edged
turbulent swirling flow through zanker plate, Orifice Plates. Journal of Fluids Engineering
Engineering Applications of Computational 134(9).
Fluid Mechanics 3(4), 562–572. Oliveira, N. M. B., L. G. M. Vieira and J. J. R.
Đurđević, M., M. Bukurov, S. Tašin and S. Bikić Damasceno (2010). Numerical Methodology
(2019) Experimental research of single-hole for Orifice Meter Calibration, Materials
and multi-hole orifice gas flow meters, Flow Science Forum, 660–661(November 2015),
measurement and instumentation. 70, 101650. 531–536.

Durst, F. and A. B. Wang (1989). Experimental and Reader-Harris, M. (2015) Orifice Plates and
numerical investigations of the axisymmetric, Venturi Tubes. Springer International
turbulent pipe flow over a wall-mounted thin Publishing, Switzerland.
obstacle, in Symposium on Turbulent Shear Reader-Harris, M., N. Barton and D. Hodges
Flows, 7th, Stanford, CA, Aug. 21-23, 1989, (2012). The effect of contaminated orifice
Proceedings. Volume 1 (A90-35176 15-34). plates on the discharge coefficient, Flow
University Park, PA, Pennsylvania State Measurement and Instrumentation 25, 2–7.
University 1989, 10.4.1-10.4.6.
Siemens (2019).User Guide STAR CCM+.
Eiamsa-Ard, S., A. Ridluan, P. Somravysin and P.
Promvonge (2008). Numerical investigation of Shaaban, S. (2014). Optimization of orifice meter’s
turbulent flow through a circular orifice, energy consumption, Chemical Engineering
KMITL Science and Technology Journal 8(1), Research and Design 92(6), 1005–1015.
43–50.
Shah, M., J. Joshi, A. Kalsi, C. Prasad and D.
Ellman, A. and R. Piche´ (1999). A Two Regime Shukla (2012). Analysis of flow through an
Orifice Flow Formula for Numerical orifice meter: CFD simulation, Chemical
Simulation, Journal of Dynamic Systems, Engineering Science 71, 300–309.
Measurement 121(4), 721–724.
Singh, R. K., S. N. Singh and V. Seshadri (2010).
Elsaey, A., A. Abou El-Azm Aly and M. Fouad CFD prediction of the effects of the upstream
(2014). CFD simulation of fractal-shaped elbow fittings on the performance of cone
orifices for flow measurement improvement, flowmeters, Flow Measurement and
Flow Measurement and Instrumentation 36, Instrumentation 21(2), 88–97.
14–23.
Singh, V. K. and T. John Tharakan (2015)
Erdal, A. and H. Andersson (1997) ‘Numerical ‘Numerical simulations for multi-hole orifice
aspects of flow computation through orifices, flow meter, Flow Measurement and
Flow Measurement and Instrumentation 8(1), Instrumentation 45, 375–383.
27–37.
Teyssandier, R. G. and Z. D. Husain (1987).
Gan, G. and S. B. Riffat (1997). Pressure loss Experimental Investigation of an Orifice Meter
characteristics of orifice and perforated plates, Pressure Gradient. Journal of Fluids
Experimental Thermal and Fluid Science Engineering 109(2), 144–148.

226