Fluid Flow and

Rotating Equipment
S. JAIN, R. PATIL and A. GUPTA,
Reliance Industries Ltd., Jamnagar, Gujarat, India

Phenomenon of flow distribution in manifolds

Flow distribution is widely encountered in refinery opera-
tions, such as air distribution to burners, fluid distribution to
parallel heat exchangers and the distribution of reactants in re-
actors. Uniform distribution is of paramount importance in all
operations to achieve higher equipment efficiency and reliabili-
ty. Although design considerations for fluid distribution devices
are available, these may be overlooked by designers and lead to
flow maldistribution. Many engineers have difficulty in concep-
tualizing the phenomenon of distribution through a manifold,
depicted in FIG. 1, which is a header with a closed end and flow
outlet points along its length.
With uniform opening of all outlets, it can be believed that
the highest flow will be through the first outlet. However, this is
not the case and maximum flow through the outlet occurs near
the closed end of the header pipe. The reason for this flow behav-
ior can be explained by applying the standard Bernoulli equation
(Eq. 1) between the inlet and the closed end of the distributor.
1 1 2
Pin + ρVin2 + gZin = Pend + ρVend + gZend + h f
As the distributor is assumed to be horizontal, the effect of
gravity can be negated. Eq. 1 can be written in the form of pres-
sure at the closed end of the distributor, as in Eq. 3. For turbulent
flows, with dominance of inertial component, it can be easily
shown that Pend > Pin . Pressure increases gradually from the inlet,
and a sudden increase in the pressure occurs near the closed end,
as can be observed from the higher slope in FIG. 3 toward the end.
1
Pend = Pin + ρVin2 − h f (3)
2
From the discussion here, with the uniform-size openings on
the pipe of the constant cross-section, non-uniform distribution
of flow exists with outlet openings near the inlet having lower-
than-average flow and those toward the closed end having above-
1.0
0.8
(1)
Normalized velocity

2 2 0.6

Where P, f, V and hf denote pressure, flowrate, velocity and fric- 0.4

tional losses, respectively. Velocity will be close to zero near the
closed end of distributor; therefore, all kinetic energy will be 0.2
converted to pressure energy. With maximum pressure at the
end of the distributor and the same outlet pressure, openings at 0.0
the end will receive a higher flowrate. Flow through each open- 0.0 0.2 0.4 0.6 0.8 1.0
Normalized length
ing is proportional to driving force ∆P and, therefore, can be
depicted in the form of Eq. 2. FIG. 2. Velocity variation along the length of a manifold distributor.1
Fi ∝ Pi – Po (2)
1.0
Velocity is at its maximum at the entrance and decreases
toward the closed end of the distributor. Typical velocity pro- 0.8
file along the length of the manifold, evaluated in literature,1 is
Normalizer pressure

shown in FIG. 2. Change in velocity results in a variation of pres- 0.6

sure within the pipe and non-uniform flow distribution.
0.4
F1 , Po F2 , Po Fend , Po
0.2

Inlet flow 0.0

Vend , Pend 0.0 0.2 0.4 0.6 0.8 1.0
Fin , Pin Normalized axial length

FIG. 1. Schematic of a manifold pipe distributor with closed end. FIG. 3. Pressure drop along the length of a manifold distributor.1

Hydrocarbon Processing | NOVEMBER 2018 75

 Fluid Flow and Rotating Equipment
average flows. The cause is rising pressure through the ports
along the length for constant outlet pressure. An important de-
sign consideration can be deduced from Eq. 2: the key to achiev-
ing uniform flow is to attain the same pressure difference or driv-
ing force for flow through all outlet openings of the distributor.
Nonuniformity of distribution is caused by the effect of zero
velocity or pressure rise at the closed end. Therefore, any ap-
proach to enhance velocity or reduction in pressure along the
length of the pipe should help achieve uniform pressure drop.
Velocity can be increased by gradually lowering the cross-sec-
tional area of the main header. Higher velocities also translate
into a higher pressure drop due to friction. Alternatively, the size
of the outlet opening can be reduced along the length of the dis-
tributor to obtain equal outlet flows.
FIG. 4. Air distribution assembly for the combustion chamber in an SAR.
FIG. 5. Velocity profiles in an air distributor of a coker furnace.
FIG. 6. Geometry of an air header in a coker furnace.
76 NOVEMBER 2018 | HydrocarbonProcessing.com
Various geometric parameters in the design of manifolds can
be manipulated to achieve the goal of constant driving force,
such as a cross-sectional area of holes and main pipe, the number
and spacing of holes, the shape of the outlet manifold, etc. The
following measures are commonly employed when designing
manifold distributor pipes:
• Tapering the distributor pipe—With a decrease in the
cross-sectional area, velocity increases and a reduction
in pressure variation occurs along the length of pipe
from the Bernoulli equation.
• Nonuniform hole sizes/spacing—Variable hole sizing
compensates for pressure variation along the pipe.
Fabrication of such a pipe with variable hole sizes
might be difficult.
These methods create uniform distribution at the design stage.
However, it might be costly and time consuming to replace the
entire existing design. Some stop-gap solutions can be employed
to manipulate pressure drop to improve flow distribution.2 The
following case studies are practical examples of flow distribution-
related studies performed for refining equipment. While the first
case study emphasizes the impact of maldistribution, the second
illustrates that uniform flow can only be achieved when recom-
mended design practices are followed for manifold design.
Case study 1: Impact of air distribution on the combustor
in a sulfuric acid (H2SO4 ) regeneration (SAR) unit. This case
study involves the analysis of air distribution to burners in a com-
mercial SAR furnace. Air is distributed to three burners through
a manifold pipe that has three outlet ducts. An illustration of the
burner assembly is shown in FIG. 4. If the distribution was uni-
form, each of the burner arms would receive 33.33% of the total
air flow. Computational fluid dynamics (CFD) simulations were
performed for this unit, and it was determined that burners 2 and
3 receive more than average air flow, while burner 1 is starved
of air (TABLE 1). This finding confirms the previously discussed
phenomenon about flow bias toward the closed end of the mani-
fold. Flow maldistribution was detected in the existing design of
the distributor and was rectified with the help of design modifi-
cations. It was proposed to install restricting orifices in burners
2 and 3 to reduce the driving force for flow and thereby achieve
flow equalization. It can be noted that with the implementation
of modifications in design, equal flow was achieved in all burners.
TABLE 1. Flow distribution in an existing vs. modified design
Percentage of total flow of air, %
Burner number Existing design After design modifications
1 27.6 33.2
2 36.2 33.4
3 36.2 33.4
TABLE 2. Flow distribution to four passes in coker distributor
Pass number Percentage of total flow of air, %
1 24.4
2 24.8
3 25.3
4 25.5
 Fluid Flow and Rotating Equipment

Case study 2: Air distribution to coker heater burners.
This study was performed for a commercial-scale furnace in a
coker unit. A CFD study was performed for the air distribution
configuration and determined uniform flow to all burners. This
study illustrates the role of air distribution geometry on the flow
distribution. Pipes in all passes are tapered to provide gradual
change in the area that, in turn, helps overcome the problem of
higher flowrates toward the closed end of the distributor.
In this configuration, air is distributed from the main header
to four passes, and then distributed to individual burners. Ve-
locity profiles obtained from CFD simulations are shown in
FIG. 5. Note that the velocity is approximately equal in all passes.
Flow through each pass is calculated from velocity profiles, and
an almost equal flow of air can be observed from TABLE 2.
Coker heater air header is tapered along the length, and pipe
in all the passes are tapered to achieve uniform distribution, as
shown in FIG. 6. Due to the gradual variation in area, equal resis-
tance can be provided for flow to different arms of the distribu-
tor; therefore, nonuniformity of flow distribution is minimized.
Findings from this case study again re-emphasize the impor-
tance of application of design recommendations.
Takeaway. The phenomenon of fluid distribution through the
manifold has been explained using basic chemical engineering
concepts. Various design aspects of manifolds must be consid-
ered to achieve uniform distribution. It is well known and ob-
served practically that flow maldistribution can reduce system
performance and efficiency, and has the potential to cause me-
chanical damage. These relevant case studies illustrate the flow
distribution problems encountered in refineries.
LITERATURE CITED
Wang, J., “Theory of flow distribution in manifolds,” Chemical Engineering Journal, 2011
Perry, R. H., Green, D. W. and J. O. Maloney, “Perry’s Chemical Engineers’ Handbook,”
7th Ed., 1997.
SHANU JAIN is a Research Scientist in the refining R&D group for Reliance
Industries Ltd. She previously worked at the Halliburton technology centre
as technical professional in cementing R&D. Her experience consists of multi-
objective optimization, computer-aided molecular design of ionic liquids,
computational fluid dynamics, non-linear regression, engineering software
development and polymer reaction engineering. She completed her BTech
in chemical engineering from the National Institute of Technology (NIT), Jaipur
and an MS degree from the Indian Institute of Technology, Bombay.
RAHUL C. PATIL is Lead Research Scientist in the refining R&D division of Reliance
Industries Ltd., Navi Mumbai, Maharashtra, India. He has 10 yr of experience
in debottlenecking, the application of computational fluid dynamics, process
modeling and the optimization of various petroleum refinery and petrochemical
units. He holds an MTech degree in chemical engineering from IIT Kanpur and
a BS degree in chemical engineering from ICT Mumbai.
AJAY GUPTA is Assistant Vice President in refining R&D at Reliance Industries Ltd.
in Jamnagar, Gujarat, India. He earned BS and MS degrees, as well as a PhD,
in chemical engineering from the Indian Institute of Technology, Delhi, India.
His experience includes the implementation of advanced process control
strategies in fluid catalytic cracking (FCC) and crude distillation units in several
refineries; the development of reactor models for various chemical processes,
including the effect of scale; and the application of computational fluid dynamics
for solving problems in petroleum refining and petrochemical industries.

US Gas Processing

ACTIONABLE INSIGHTS
Into US Natural Gas Facilities
The web-based GIS application includes detailed data on more than
500 US gas processing plants, 400 natural gas underground storage
locations, and 58 natural gas market hubs.

In-Depth Data
Such as key contact details, gas plant project/facility data
(including plant type/flow, BTU content, infrastructure,
storage capacity, plant status and ownership) for the
following layers:

• US gas processing plants • Natural gas underground

storage
• Natural gas pipelines
• Natural gas market hub
• Gas fields
• US compressor stations

Discover more at EnergyWebAtlas.com Hydrocarbon Processing | NOVEMBER 2018 77