Using Formula Le  0.

347 

1  0.216 Fc
L  d 
L : Line Length (m)
d : Nominal Pipe Diameter (in)

Very Complex Manifolds Fc = 4.0

Manifold-type piping Fc = 2.0
hemical Engineering
(March 16, 1987 Normal Fc = 1.0 UNIT

, Page 85~86) Long, Straight-run Fc = 0.5 INTERCONN

OSBL, Utility supply line Fc = 0.25

 €nGin«mnGf€nTUR€

HOW TO PREDICT
PRESSURE DRO
BEFORE DESIGNING THE PIPIN
Yessure-drop estimates Grant S. Brown, Rust International Corp.
ieveloped from correia- In the design of process and power plants, it often is necessary to estimate
ions of pipe fittings are the pressure drop of a piping system before all the required information is
jseful for preliminary available. A typical example is in the preliminary phase of plant design,
after the flowsheets and plot plans have been worked out but before the
izing and costing of piping has been designed in detail: The engineer is supposed to deliver a timely
uid-flow systems, and and reasonably accurate estimate of the pressure drops in the various process
iven for ordering pumps. and utility lines. But while the overall line lengths can be estimated from the
plot plans, the number of fittings is unknown, and those typically cause as
much pressure drop as the pipe itself.
Fortunately, there is a way around this vexing problem. 'A correlation
developed by W. B. Hooper [7] accurately predicts the number and type of
valves and fittings per unit length for piping of various sizes. These predic­
tions, combined with the known line lengths and published equivalent-length
factors, provide an accurate estimate of the true equivalent length and,
therefore, the pressure drop. The author has used this approach for four years,
and has found it extremely useful. A similar approach has been used for
estimating economic pipe diameters [2\
Predicting the number of valves and fittings
Hooper’s correlation predicts the number and type of valves and fittings in a
pipeline as a function of the pipe diameter and a complexity factor, Fc. The
correlations are based on data from 46 projects executed at Monsanto Co.;
these ranged from small to very large piping systems, and included carbon-,
galvanized- and stainless-steel piping. Typical values for Fc are: 4 for very
complex manifolds; 2 for manifold-type piping; 1 for normal piping; Vz for long,
reasonably straight piping runs; and 14 for utility supply lines, outside battery
limits. This article converts the data presented by Hooper into equations for
estimating pressure drop.
The accuracy of the estimates depends on the value chosen for Fc, which is a
matter of engineering judgment. This judgment grows rapidly with experience,
and can also be developed systematically by analyzing existing piping systems
or finished designs: Fc can be back-calculated from the known lengths and
pressure drops. Or, Fe can be estimated from the description above (or the
figures in [j]), and then used to check whether the design data seem
reasonable.
According to Hooper, the number of pipe fittings per foot of straight pipe is:
Nf — 0.35 FJD1/2, where D is the nominal pipe diameter in inches. Combining
Hooper’s breakdown of fittings by type with a typical equivalent length (.Le/D)
for each type of fitting [S\ yields a weighted-average equivalent length for all
the fittings:
Type of Fraction of Weighted
fitting total Le/D Le/D
Flanges 0.45 0 0
90-deg elbows 0.25 16 4.0
Tees 0.15 40 6.0
Reducers 0.10 15 1.5
45-deg elbows 0.025 16 0.4
Plugs and caps , 0.025 0 0
Weighted-average LJD 11.9

CHEMICAL ENGINEERING/MARCH 16,1987

 .€AGM€€ftH)G F€ATUR€

_ u-« (Le/D)/L = 12 ID + LWoFJDV* + 2.592 FJD (2

10
or expressed as:
0)
a / LJL - 1 + (0.347/3^ + 0.216)fc (:■
ag
.c
CD Complexity factor (Fc) /4 Eq. (3) is plotted in the figure, for various values of tl
'£ 8
/ complexity factor.
« t /
/ Applying the method
%7
ftc / Using Eq. (3), one can calculate the total equivalent length
® 6 r the straight pipe plus the valves and fittings, for a giv<
c
S 2
length of straight pipe. As an example: A new plant is t
25
Q)
“7*
s' have a 6-in.-dia. carbon-steel pipe traversing a distance of
a
£a>4 ft (L = 300) and carrying 500 gal/min of water (p = 62
c
tD 1 Ibm/ft3, y = 6.72 X 1(H Ibm/ft-s). The piping is expected
c 3 be of normal complexity (Fc = 1). What frictional pressu i
0>
m 1/2 drop is predicted?
> 2 The first step is to estimate the total equivalent lengt
3
CT
(D
1/4 From Eq. (3):
<d
o Le = 300 [1 + (0.347 (6)1/2 + o.216)(l)] = 620 ft
O The velocity of the water (v) is about 5.67 ft/s, and II
0.5 0.75 1 4 6 8 10 16 24 36
Nominal pipe diameter (D), in. Reynolds number (pvd/y, where d is diameter in feet)
about 263,000. Assuming an’absolute roughness of 0.0018
Figure — The equivalent-length adjustment factor the Darcy friction factor (/) for the pipe is about 0.01 i
(Le/L) depends on the diameter and complexity of the piping
Using the Darcy equation, the frictional pressure drop (A
psi) is therefore:
The above value for tees is based on half of the tees having
AP = (fLtJd)(p ri72 #C){1/144) = 4.6 psi
flow through the run, and half having flow through the
branch. where gc is the gravitational constant (32.2 ftdbm/lbf*
Likewise the number of valves per foot of straight pipe is: Again, nonfrictional pressure drops due to entrances, exi
Ar„ = 0.09 FciD. Combining Hooper’s breakdown of types of changes in elevation and throttling valves must be added
valves with an equivalent length for each type {3\ yields a this figure.
weighted average for all the valves: What if Fc were 2 instead of 1? According to the figure,
Type of Fraction of Weighted total equivalent length per unit length of straight pipe woi
valve total Le/D Le/D have been about 3.1 (instead of 2.00). Lt, would have been
Block 0.90 20 18 ft (instead of 620 ft), and AP would have been 6.9 psi (ins
Check 0.08 50 of 4.6 psi).
Globe 0.02 340 6.8 In conclusion, the author has found this method to be vc
useful for preliminary estimates of pipe size and pressi
Weighted-average L(JD drop, for checking final design calculations to see whell
It is assumed that two-thirds of the block valves are gate they are reasonable, and even for ordering pumps ;
valves, and one-third are butterfly valves. The check valves compressors in cases where the frictional pressure dro|
are assumed to be the swing type, and the globe valves are relatively small compared with the total pressure drop.
assumed to be fully open. Throttling valves are not account­ Mark Upowicz, Edi
ed for in this correlation. Instead, throttling pressure drops
References
are added to other head losses such as changes in elevation,
1. Hooper, W. R,, Predict fittings for piping systems, Chvm. Eng., Ma
and entrances or exits, and simply added on to the frictional 1!»2, pft. 127-12!).
pressure drop. 2. Culler, D. W., Pipe-sizing economics, Chvm. Eng., May 2X, 19S4, pp
110.
3. Westaway, 0. R., ami Jxwmis, A. W., efts., “Cameron Hydraulic D;
Predicting the pressure drop Ingersoll-Rand Corp., Woodcliif Lake, N.J., 1!)77.
This analysis provides estimates for the number of valves
and fittings, and the equivalent length per valve or fitting.
Combining all these relationships yields a single equation for The author
Cirunt S, Brown is a stair engineer in the Proc
the total equivalent feet of pipe per foot of straight pipe: Stall' Dept. of Rust International Corp., 441 Smith
St., Pittsburgh, PA 15222, where ho is involved ii
(L,JD)/L = 12iD + U.9(0.3f)Fc/Z)U2) + 28.8(0.09^0) (1) design of industrial power plants. He has had 17 y
nf experience, and has worked on such project
where L is the length of straight pipe (ft) and D is the
•*r-
L- 3 « ffuidized-hod boiler plants, and liquefied-gas term
and peakshaving plants. He has a R.S. degree
nominal pipe dia. (in.). Here, the first term accounts for the

A
Case Western Reserve University, and an M.S.
Massachusetts Institute of Technology, both in
straight pipe, the second for the fittings, and the third for the chanieal engineering, lie is registered in PennsyK
valves. Eq. (1) can be simplified to: \A and is a member of ASMK.

86 rilt'IMK'Al, KNOINKKlilN’d/MAkni 1(1, W?