Advances in Engineering Software 41 (2010) 489–496

Contents lists available at ScienceDirect

Advances in Engineering Software

journal homepage: www.elsevier.com/locate/advengsoft

A MathCAD procedure for commercial pipeline hydraulic design considering

local energy losses
Gürol Yıldırım a,*, Vijay P. Singh b
a
Aksaray University, Engineering Faculty, Civil Eng. Dept. Hydraulics Div. Aksaray, Turkey
b
Department of Biological and Agricultural Engineering, Texas A and M University, Scoates Hall, 2117 TAMU, College Station, TX 77843-2117, USA

a r t i c l e i n f o a b s t r a c t

Article history: Using a power type equation for friction factor, this paper presents a design procedure which provides
Received 19 June 2009 accurate solutions for three types of pipe design problems (Types A–C) taking into consideration the
Received in revised form 11 September effect of local losses. The parameters introduced in the power type equation are related to the type
2009
and size of commercial pipes. Thus, several dimensionless physical numbers, obtained by suitably com-
Accepted 9 October 2009
Available online 11 November 2009
bining the variables relevant for the solution of Type B and Type C problems, are also introduced. For
solution of the general case of a Type B problem (sloping pipe with pumping power), a user-friendly
MathCAD procedure, which produces a consistent framework for analyzing and solving common pip-
Keywords:
Water distribution networks
ing-system applications problem, is also developed. In order to evaluate the accuracy of the proposed
Pipeline design procedure, several design examples are analyzed for three types of commercial pipes and a wide range
Commercial pipes of uniform pipe slope, and the results are shown as design curves. These curves have practical impor-
Pipe hydraulics tance, because they permit to quickly determine the values of required variables for a given pipe slope.
Pressurized flows The results of the proposed method are compared with those obtained from the methods existing in the
Steady-state analysis professional practice.
Software Ó 2009 Elsevier Ltd. All rights reserved.
MathCAD

1. Introduction
Analysis and design of water distribution networks have been
and will continue to be of considerable interest in hydraulic engi-
neering. Three types of design problems of single-lined commercial
pipes arise with the existing solution techniques: optimization of
the pump power (Type A), determination of discharge (Type B),
and design of the smallest pipe size or pipe length (Type C)
[21,24]. Since the friction factor is a function of relative roughness
of the pipe and Reynolds number, the conventional design method
requires a lengthy iterative process even for a single-lined set of
pipes and even without considering local losses. However, with
progress in technology and solution techniques, automated meth-
ods, which provide accurate and convenient alternatives to manual
methods, have become available and are preferred for solving dif-
ferent pipe-flow problems [13,4].
The friction factor for turbulent fluid flow in rough pipes has
been commonly estimated using the Colebrook–White (C–W)
equation [2,1]. However this equation, which relates the friction
factor to pipe roughness and Reynolds number (R), is implicit in
friction factor (f). Recognizing the implicit nature of the C–W
equation, f versus R charts were constructed [14] that allowed for
a graphical estimation of f when R and pipe roughness were
known. Since then, several approximations of the C–W equation
that are explicit in the friction factor have been proposed
[15,3,10,19,32,6,12,17,18,16,26,25,27–29]. Neglecting local losses
[22–24] Yoo and Singh [30] recently suggested a power law friction
factor for practical use and easy manipulation for a simple design
of single-lined pipes. Yoo and Singh [31] also proposed two meth-
ods for estimating the friction factor of commercial pipes; one
combining zero velocity points and the other combining friction
factors of smooth and rough parts of the pipe.
Considering the effect of local energy losses, the objective of this
paper is to present a design procedure, using a power form friction
factor, for three design cases of commercial pipes. The proposed
procedure produces simple but exact solutions for routine engi-
neering applications and is suitable for analysis and design of sin-
gle-lined pipeline networks.
2. Analysis of pipe design problems

* Corresponding author. Present address: Texas A and M University (TAMU), Bio.

and Ag. Dept. (BAEN), College Station, 201 Scoates Hall, 77843-2117 TX, USA.
Design problems of commercial pipelines with uniform slope
E-mail addresses: yildirimg3@itu.edu.tr (G. Yıldırım), vsingh@tamu.edu (V.P. can be analyzed in three main categories: (1) Type A problems,
Singh). (2) Type B problems, and (3) Type C problems. In Type A problems,

0965-9978/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.advengsoft.2009.10.007
490 G. Yıldırım, V.P. Singh / Advances in Engineering Software 41 (2010) 489–496

Nomenclature

D pipe diameter (L) RN R for the design case of B2–3 (Eq. (13))
D0 base diameter (=1 m) (L) s uniform pipe slope
0
f Darcy–Weisbach friction factor s the friction slope or the slope of the energy-grade line
g acceleration due to gravity (L T2) SD dimensionless physical number (Eq. (15))
hT total head losses (L) S0D dimensionless physical number (Eq. (15))
hf head loss due to friction (L) SL dimensionless physical number (Eq. (15))
hl head loss due to localized disruption of the flow (L) T dimensionless physical number (Eq. (15))
HP total pumping head (L) V mean velocity in the pipe cross-section (L T1)
J dimensionless physical numbers (Eq. (11)) a empirical constant of power law friction factor (Eq. (4))
JL dimensionless physical number (Eq. (18)) b empirical constant (Eq. (4))
K dimensional unit conversion factor (Eq. (7)) c specific weight of fluid (K L3)
(T2+b L(3b+5)) x empirical constant (Eq. (4))
kf pipe friction loss coefficient f empirical constant (Eq. (4))
kl total local loss coefficient kJ empirical constant (Eq. (12))
L pipe length (L) kN empirical constant (Eq. (13))
M dimensionless physical numbers (Eq. (11)) kT empirical constant (Eq. (16))
N dimensionless physical numbers (Eq. (11)) dJ empirical constant (Eq. (12))
PW pump power (KL T1) dN empirical constant (Eq. (13))
p slope indicator which is taken 0, 1 and 1; for zero, dT empirical constant (Eq. (16))
downhill, and uphill slope cases r dimensionless constant (Eq. (9))
Q discharge (L3 T1) m kinematic viscosity of water (L2 T1)
R Reynolds number e absolute roughness of the pipe (L)
RJ R for the design case of B1–3 (Eq. (12)) e/D relative roughness

the required parameters are pump power (PW) and/or friction or Table 1
total energy loss (hf, hT); initially known parameters are pipe slope Three types of pipe design problems with subclasses (with existent/no existent design
cases).
(s), pipe diameter (D), pipe length (L), discharge (Q), and material
type of the commercial pipe (galvanized-iron, wrought-iron, tar- Problem Known quantities Required parameters Related
coated cast-iron, PVC, concrete, etc.). In Type B problems, the val- type Eqs.
ues required are Reynolds number (R), velocity (V) or Q and/or hf, A Discharge (Q), Pumping power (PW) and/or (1)–(9)
hT; known parameters are PW, and pipeline properties (pipe type, pipeline properties friction/total energy loss (hf,
s, D, and L); and for Type C problems, the required parameters hT)

are D or L, and/or hf, hT; known parameters are pipe type, s, Q, B PW, pipeline Reynolds number (R), (10)–(13)
and PW. properties velocity
(V) or discharge (Q) and/or
Each of the three types of design problems and their subclasses hf , h T
can be analyzed, depending on the design cases: neglecting or con-
C Pipe slope (s), pipe Smallest pipe size (D) or (14)–(18)
sidering local losses, hl, horizontal (s = 0) or sloping pipe, and with type, pipe
or without pumping power, PW. Table 1 synthesizes the three types Q, PW length (L) and/or hf, hT
of design problems (A–C) and their subclasses (A1, A2, A3, A1–3,
A2–3; B1, B2, B3, B1–3, B2–3; C1, C2, C3, C1–3, C2–3) with corresponding Subclasses Existent design cases No existent design cases
equations. Note the design parameters in this table; hl, s, and PW A, B, C s (N), hl (M), PW (J) –
are represented by their corresponding dimensionless physical A1, B1, C1 hl (M), PW (J) s (N)
A2, B2, C2 p = 1; s (N), hl (M) PW (J)
numbers (M, N and J), respectively, which will be defined in the
A3, B3, C3 s (N), PW (J) hl (M)
next section. A1–3, B1–3, C1–3 PW (J) s (N), hl (M)
Problems in categories A and B are analysis problems. Analysis A2–3, B2–3, C2–3 p = 1; s (N) hl (M), PW (J)
of Type A problems is direct, without iteration, even if any type of Note: Pipeline properties include the pipe diameter or length (D or L), uniform pipe
friction factor equation is used. However, iteration may be required slope (s), and pipe type; p: slope indicator (p = 0, 1, and 1; for zero, downhill and
for Type B problems. Type C is a sizing problem, which normally uphill slope, respectively); hl: local loss; M, N and J are dimensionless physical
requires more assumptions and more iterative computations to numbers related to the parameters hl, s and, PW, respectively, the subscripts of
subclasses illustrate existent design cases; (1): horizontal pipe case, (2): attractive
solve.
pipe flow case (with no pumping power), and (3): design case with neglecting local
In recent years, preprogrammed equation solvers have become losses, respectively.
available on certain more advanced programmable scientific cal-
culators and in some spreadsheet and mathematics software
3. Governing equations
packages, such as MathCAD, Matlab or Excel. Effective instructions
in hydraulic design and piping system analysis have constituted a
The governing equations for design of a single-lined pressurized
subject of several recent publications [9,8,11,5,7,20,21]. Note that
pipe with uniform slope can be expressed in general form as:
the December 2001 issue of the Journal of Hydraulic Engineering
was devoted to the topic of teaching hydraulic design. Conse- PW ¼ cQHP ð1Þ
quently, to avoid tedious trial-and-error procedures the auto- HP ¼ hT þ psL ð2Þ
mated solution techniques can be easily applied to obtain the
solution of either type of pipe design problems (Types B and C) V2 L
hT ¼ hf þ hl ¼ ðkf þ kl Þ ; kf ¼ f ð3Þ
[13,4]. 2g D
 G. Yıldırım, V.P. Singh / Advances in Engineering Software 41 (2010) 489–496 491

where PW (kg m s1) is the pump power in which a unit conversion For the solution of a Type B problem, the general governing equa-
factor can be applied to this result to express the power, horse- tion, Eq. (1), is restated in terms of the Reynolds number; substitu-
power (HP) or kilowatts (kW); c the specific weight of fluid tion of the power law friction factor given by Eq. (2) into Eq. (1)
(kg m3); Q the discharge (m3 s1); HP the total pumping head yields:
(m); hT, hf and hl the total, friction and local energy losses, respec-
8
tively (m), kf and kl the pipe friction loss and total local loss coeffi- MR3 þ aR3þb þ 2pN2 R  J3 ¼ 0 ð10Þ
cients, respectively, V the mean velocity in the pipe cross-section
p
!1=3
(m s1); D and L the pipe diameter and length, respectively (m), D ðgsÞ1=2 D3=2 1 PW gD2
M ¼ kl ; N¼ ; J¼ ð11Þ
V2/2g the kinetic energy head (m); s the pipe slope assuming to L m m cL
be uniform along the pipeline; p the slope indicator which is taken
0, 1 and 1 for zero, downhill, and uphill slope cases, respectively, where M, N and J is the dimensionless physical numbers [30]. Not-
and f is the Darcy–Weisbach friction factor. ing that in Eq. (4), one unknown is R when the discharge or velocity
is the required one parameter; then, the other physical numbers can
3.1. Power form friction factor (f) be determined depending on the initially known quantities.
For a simple design case of a pumping power on a horizontal
Yoo and Singh [30] have developed sets of pipe friction factor pipeline and neglecting local losses (Type B1–3: s = 0, N = 0; hl = 0,
equations neglecting local losses in the form of a power law that M = 0), the Reynolds number, RJ is directly computed from [30]:
allows explicit solutions for the aforementioned three types of pipe  1=ð3þbÞ
8 3
design problems. Based on comparison of the results of the power RJ ¼ kJ J dJ ; kJ ¼ ; dJ ¼ ð12Þ
pa 3þb
law with those of the Colebrook–White (C–W) and Hazen–
Williams (H–W) equations, they have reported that the overall For the case of a downhill sloping pipe with no pumping power and
error of the power law is 5.89%, while that of the C–W equation neglecting local losses (Type B2–3: p = 1, PW = 0, J = 0; and hl = 0,
is 11.24%, and that of the H–W is 15.25%. M = 0), the Reynolds number, RN, is given by
The power law friction factor for commercial pipes is given by  1=ð2þbÞ
[30]: 2p 2
RN ¼ kN NdN ; kN ¼ ; dN ¼ ð13Þ
 f a 2þb
D
f ¼ aRb ; a ¼ x ð4Þ
D0
VD 4Q 4Q 4.3. Sizing (pipe diameter or length) problem (Type C)
R¼ ¼ ; V¼ ð5Þ
m pDm pD 2 In order to design a pipe of diameter, D, a new dimensionless
where a, x, b and f is the empirical constants of power law friction
physical number, SD, in relation to discharge is also introduced.
factor as a function of pipe sizes (small: D < 25 mm; medium:
Rearranging Eq. (1) and Eq. (2), in terms of SD, one gets (with
25 mm < D < 1 m; and large: D > 1 m) and pipe type (galvanized-
D0 = 1 m):
iron, wrought-iron, tar-coated cast-iron and PVC); R the Reynolds
number; D0 the base diameter (=1 m); and m is the kinematic viscos-    b
ity of water (ffi106 m2 s1).
p2 4
T 5 S4D  x
fð1þbÞ f
SD S0D  kl S1
L ¼ 0 ð14Þ
8 p
!1=5
4. Methodology g 1=5 P W Q 2 mD 4
T¼  psQ 3 ; SD ¼ ¼ ;
m cL Q pR
4.1. Head loss and/or pump power problem (Type A)
mD0 mL 4L
S0D ¼ ; SL ¼ ¼ ð15Þ
Combining Eqs. (4) and (5) with Eq. (3) for the pipe friction loss, Q Q pDR
hf can be deduced in the power form as:
where T, SD, S0D and SL is the dimensionless physical numbers [30];
hf ¼ KQ 2þb ð6Þ and g is the acceleration due to gravity (m s2).
xð4=pÞ2þb L For Type C problems, all physical numbers except for SD (for this
K¼ ð7Þ case, SL is a known parameter) are determined using the initially
2g mb Df0 D5þb1
known values; and SD can be easily determined by using one of
where K is the dimensional unit conversion factor which can be ap-
the user-friendly software packages mentioned above. As a special
plied to express hf in meter (s2+b/m3b+5).
case, when local losses are neglected (Type C3: hl = 0, M = 0), one
The general form of a Type A problem (considering local losses)
obtains:
can be directly solved by Eq. (1) in conjunction with Eqs. (2) and
"  2þb #1=ð5þbfÞ
(3). For a simple design case of neglecting local losses (Type A3), x 4
PW can be computed as:
dT
SD ¼ kT T ðhl ¼ 0Þ; kT ¼ Sf
0D ;
2 p
PW ¼ rcQLðhl ¼ 0Þ ð8Þ
5
hf KQ 2þb dT ¼ ð16Þ
0
r ¼ s þ ps; s0 ¼ ¼ ð9Þ ½f  ð5 þ bÞ
L L
A derivation procedure similar to that presented above can be fol-
where r is the dimensionless constant which is the summation of
0 0 lowed to determine the pipe length, L (D is a known parameter,
the friction slope, s , and uniform pipe slope, s; s the friction slope
for this case) by introducing another physical number, SL:
or the slope of the energy-grade line (m/m); p and s is the slope
!
indicator and uniform pipe slope, respectively, defined earlier. 4 2J 3L  kl R2
SL ¼ ð17Þ
p aR3þb þ 2pN2 R
4.2. Discharge or velocity problem (Type B) !1=3
 1=3
1 4gP W D gPW D2
In order to determine the discharge (or velocity) through the JL ¼ ; or J L ¼ ð18Þ
m cpR cQ m2
pipeline, new dimensionless physical numbers are introduced.
492 G. Yıldırım, V.P. Singh / Advances in Engineering Software 41 (2010) 489–496

where JL is the dimensionless physical number. Note that Eq. (17)
along with Eq. (18) can be directly used without an iterative
procedure.
6. Design examples
In the following, two design examples for different series pipe-
flow problems (Types B3 and C3) with a general design case (slop-
ing pipe with pumping power) are presented, respectively.

5. Computational example 6.1. Example 1 (Type B3: uphill sloping pipe)

In order to present the practicability the proposed design proce- It is required to supply water from a reservoir at an elevation of
dure, an original example [13] for the Type B problem of a downhill 410.4 m to another reservoir at an elevation of 433.2 m with a
sloping pipe case (with no pumping power) is selected. A tar-coated pump of effective power 58.5 kW. The length of the line of a tar-
cast-iron pipe connects two reservoirs having a difference in water
surface level of 6.1 m. The length of the pipeline is 365 m and has a
diameter of 0.46 m (=18 in.). The problem is: (a) determine the dis- a Q(G) (dow n) Q(W) (dow n) Q(T-c) (dow n)
charge when only pipe friction is considered; (b) determine the Q(G) (up) Q(W) (up) Q(T-c) (up)
change in the computed discharge when local losses for a sharp-
1.5 0.4
edged entrance, a fully open gate valve near the pipe exit, and the
pipe exit itself are also considered; and (c) determine the new dis-
charge for the case if the gate valve in part (b) was only 1/4 open.
1.2 0.3

(m3s -1)

(m3s -1)
(a) This is a Type B2–3 problem (p = 1: hl = 0, M = 0 and PW = 0,
J = 0); so, Eq. (13) can be directly used as follows:
0.9 0.2
(down)

(up)
Empirical parameters (x, f, b) for tar-coated cast-iron pipe of

Q
medium size (25 mm < D = 0.46 m < 1 m): x = 0.071,
Q

f = 0.255, and b = 0.13 (Table 1 of [30].

0.6 0.1
From Eq. (4): a = 0.071  (0.46/1)0.255 = 0.0865.
6:1
Slope of the pipeline, s ¼ 365 ¼ 0:0167 (p = 1; downhill
slope).
1=2
0:463=2 0.3 0
From Eq. (11): N ¼ ð9:810:0167Þ ¼ 1; 26325:1.
 1061=ð20:13Þ 0 1 2 3 4 5 6 7 8 9 10
2ð1Þ 2
From Eq. (13): kN ¼ 0:0865 ¼ 5:3615; dN ¼ ð20:13Þ
¼ 1:0695.
b Hp(G) (down) Hp(W) (down) Hp(T-c) (down)
Then finally, RN ¼ 5:3615  126325:11:0695 ¼ 1; 532250 is Hp(G) (up) Hp(W) (up) Hp(T-c) (up)
obtained.
6 20 200
From Eq. (5): Q ¼ p4Dm R ¼ p0:4610
4
 1; 532250 ¼
3 1 1
0:554 m s and therefore, V = 3.33 m s . 16 160
From Eq. (4): f = 0.0865  1, 5322500.13 = 0.0136.
0:071ð4=pÞð20:13Þ
From Eq. (6) with Eq. (7): K ¼ 29:8110
(m)

(m)
6ð0:13Þ
1:00:255 12 120
365
ð50:13þ0:255Þ ¼ 18:425 then finally, h ¼ 18:425
(down)

(up)
0:46
0:55420:13 ffi 6:1 m ðs0 ¼ 0:0167Þ, is computed.

HP
8 80
HP

(b) This is a Type B2 problem (p = 1: hl – 0, M – 0 and PW = 0,

J = 0); local loss coefficients for a sharp-edged entrance and 4 40
pipe exit are: kent. = 0.5 and kex. = 1.0, and for fully open gate
valve, kvalve = 0.2 [13], Table 2.5). Therefore, the total local 0 0
loss coefficient kl = (0.5 + 1.0 + 0.2) = 1.7 is found. 0 1 2 3 4 5 6 7 8 9 10
From Eq. (11): M ¼ 1:7  0:46 365
¼ 2:142  103 .
From Eq. (10): (2.142  103)  R3 + 0.0865  c
2.87 10 hf(G) (down) hf(W) (down) hf(T-c) (down)
R  (3.189  10 )  R = 0.
hf(G) (up) hf(W) (up) hf(T-c) (up)
Using MathCAD Equation-Solver: R = 1.415  106 then,
Q = 0.511 m3 s1 and V = 3.1 m s1 are computed. This result 200
15
indicates that discharge has decreased by 0.043 m3 s1, a bit
under 8%, as a consequence of considering local losses. Using
150
Eqs. (3)–(7), one can obtain the following:
f = 0.0137, hf = 5.25 m and hl = 1.7  3.12/(2  9.81) = 0.82 m
(m)

10
(m)
(down)

(hT = 6.1 m). 100

(up)

(c) A similar procedure as presented in part (b) is repeated,

hf

hf

except that the gate valve is only 1/4 open. For this case, 5
50
the valve loss coefficient has increased from 0.2 to 17.0
[13], Table 2.5) since it will cause more head loss than the
pipe friction term. Replacing 0.2 in part (b) by 17.0 0 0
(kl = 18.5, M = 0.0233), computations were redone 0 1 2 3 4 5 6 7 8 9 10
using MathCAD Equation-Solver and the values found (Zero-Down-Up) Slope, s (%)
were: R = 9.15  105 then, Q = 0.33 m3 s1, and
Fig. 1. Results for the design variables [(a): discharge, Q, (b): total pumping head,
V = 1.98 m s1, f = 0.0145, hf = 2.32 m and hl = 3.69 m
HP, and (c): head loss due to friction, hf] for Different Pipe Slopes (s = 0%  s = ±10%),
(hT = 6.1 m). It is remarkable that there is a decrease of about and for Three types of commercial pipe [G: galvanized-iron, W: wrought-iron, and
one third from the discharge computed in part (b). T-c: tar-coated cast-iron], using data of Example 1 (s = 1.25%, uphill slope).
 G. Yıldırım, V.P. Singh / Advances in Engineering Software 41 (2010) 489–496 493

a f(G) (dow n) f(W) (dow n) f(T-c) (dow n)

f(G) (up) f(W) (up) f(T-c) (up)

0.015 0.02

0.019

0.014
0.018

0.017
(down)

(up)
0.013
0.016

f
f

0.015
0.012

0.014

0.011 0.013
0 1 2 3 4 5 6 7 8 9 10

b R(G) (down) R(W) (down) R(T-c) (down)

R(G) (up) R(W) (up) R(T-c) (up)

4 1.2

3 0.8
(10 )
6

(10 )
6
(down)

(up)
R
R

2 0.4

1 0
0 1 2 3 4 5 6 7 8 9 10
(Zero-Down-Up) Slope, s (%)

Fig. 2. Results for (a): friction factor, f, and (b): Reynolds number, R; for Different Pipe Slopes (s = 0%  s = ±10%), and for three types of commercial pipe [G: galvanized-iron,
W: wrought-iron, and T-c: tar-coated cast-iron], using data of Example 1 (s = 1.25%, uphill slope).

Pw (Galvanized iron) This is a Type B3 problem, and computation is done as follows:

Pw (Wrought iron)
400 Pw (Tar-coated cast-iron)
s ¼ ð433:2  410:4Þ=1824 ¼ 0:0125 ðp ¼ 1:0; uphill slopeÞ

300 From Eq. (11): M = 0, N = 109,251.2 and J = 189,328.3 are computed

and substituted into Eq. (10), then finally we obtain: 0.0865R2.87 +
2.387  1010R1.728  1016 = 0. Using MathCAD Equation-Solver:
PW (kW)

200 R = 5.9  105 then, Q = 0.213 m3 s1 and V = 1.28 m s1, f = 0.0154
0
and hf = 5.1 m and therefore, s = 5.11/1824 = 2.8  103, are com-
puted. It should be noted that in order to check the accuracy of re-
100
sults obtained from computations, Eqs. (8) and (9) can be directly
0
used as follows: r = s + ps = 2.8  103 + 1.0  0.0125 = 0.0153;
0
then PW = 0.0153  1000  0.213  1824/102 = 58.5 kW (HP = 5.1 +
0 1 2 3 4 5 6 7 8 9 10 1.0  0.0125  1824 ffi 28.0 m).
(Zero-Up) Slope, s (%) On the basis of the algorithm discussed, computations of Exam-
ple 1 are also extended to encompass three types of commercial
Fig. 3. Results for pumping power (kW), PW, for different uphill slope ranges
(s = 0%  s = ±10%), and for three types of commercial pipe [G: galvanized-iron, W:
pipes (tar-coated cast-iron, galvanized-iron and wrought-iron) of
wrought-iron, and T-c: tar-coated cast-iron], using data of Example 1 (s = 1.25%, medium size (25 mm < D = 0.46 m < 1 m), and for various slope
uphill slope). combinations; the results are shown in Figs. 1–3. Empirical param-
eters of power form friction factor given by Eq. (4) for three pipe
types are [30]: tar-coated cast-iron: x = 0.071, f = 0.255, and
coated cast-iron pipe (x = 0.071, f = 0.255, and b = 0.13) is b = 0.13 (a = 0.0865); galvanized-iron: w = 0.034, f z = 0.282,
1824 m, and it has a diameter of 0.46 m (=18 in.). Neglecting local and b = 0.08 (a = 0.0423); and wrought-iron: w = 0.067,
losses, compute the discharge in the pipeline. f = 0.135, and b = 0.12 (a = 0.0744).
494 G. Yıldırım, V.P. Singh / Advances in Engineering Software 41 (2010) 489–496

Design curves, shown in Figs. 1–3, have practical importance, be- lope increases more deviation in the values of friction factor
cause they permit to quickly determine values of design parameters (Fig. 2a) is observed.
for a given pipe slope. It should be noted that for Figs. 1 and 2 the
slope ranges in percent are indicated on the x axis (starting from 6.2. Example 2 (Type C3: uphill sloping pipe)
0% to ±10%); and values of design parameters for different downhill
(on dotted line) and uphill (on straight line) slope ranges are individ- The data of an original design problem presented by Yoo and
ually represented on the first axis and second y axis, respectively. Singh [30] are selected here for the sake of comparison: the eleva-
Fig. 1a–c illustrates the variation in discharge, Q, total pumping tion difference between two reservoirs is 5 m, and the total pipe
head, HP, and head loss due to friction, hf, respectively, whereas length including the equivalent length for secondary losses is
Fig. 2a and b shows the values of friction factor, f and Reynolds 2 km. It is required to supply water from the reservoir located be-
number, R, versus different ranges of downhill and uphill slopes. low to the reservoir located above with a discharge of 0.2 m3 s1
On the other hand, values of pumping power, PW, for three pipe with a pump of effective power 10 kW. Compute the minimum
types and for different ranges of uphill slope are also plotted in diameter of tar-coated cast-iron pipe.
Fig. 3. As shown in Figs. 1 and 2, for the three pipe types, as the This is a Type C3 problem, and computational steps are pre-
down slope increases the discharge, the Reynolds number, and sented below:
the head loss due to friction increase (with decreasing f); whereas
the total pumping head decreases. In other words, the values of Q, 1. s = 5/2000 = 0.0025 (p = 1.0, uphill slope);
R, and hf decrease (with increasing f); whereas the values of HP in- 2. From Eq. (15):
crease with increasing upslope. !1=5
Examination of the results indicates that (Figs. 1 and 2) for the 9:811=5 10  102  0:22
T¼   1:0  0:0025  0:23 ¼ 82; 405:6;
three pipe types, variations in Q, R, hf, and HP coincide, except for a 106 1000  2000
little change, for a wide range of pipe slope; whereas as the ups- 106  1:0
S0D ¼ ¼ 5  106 ;
0:2
a Tar-coated cast iron
D (dow n) D (up)

0.5 0,99 1

0.9
0.458
0.45

0.429
0.8
(m)

(m)
0.409
(down)

0.4 0.401 0.7

(up)
0.393
D
D

0.381
0.618 0.37
0.362
0.6
0.354
0.35 0.347
0.341
0.505 0.335 0.5
0.458

0.3 0.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

b L (dow n) L (up)

5000 1000 1000

4656

836.3
4000 800

718.4
671.1
(m)

(m)

629.7
(down)

3000 600
(up)

560.4
L
L

504.9
2433
459.4
421.4
2000 2000 389.2 400
361.6
1647 337.7

1245
1000 200
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
(Zero-Down-Up) Slope, s (%)

Fig. 4. Results for (a): pipe diameter, D, and (b): pipe length, L; for Different Pipe Slopes (s = 0%  s = ±1%), and only for tar-coated cast-iron Pipe, using data of Example 2
(s = 0.25%, uphill slope).
 G. Yıldırım, V.P. Singh / Advances in Engineering Software 41 (2010) 489–496 495

3. From Eq. (16): with those obtained from the existing literature. It can be con-
cluded that the software based-design technique is simple, easy
"  20:13 #1=ð50:13þ0:255Þ to apply and sufficiently accurate in comparison with the tradi-
0:071 4
kT ¼   ð5  106 Þ0:255 ¼ 0:31; tional trial-and-error procedures for different types of design
2 p problems of commercial pipelines, in routine hydraulic engineer-
5 ing applications.
dT ¼ ¼ 0:976;
½0:255  ð5  0:13Þ
SD ¼ 0:31  82936:50:976 ¼ 4:94  106 ;
Acknowledgments

4. Then finally from Eq. (15):

The Scientific and Technological Research Council of Turkey
(TUBITAK) is acknowledged for supporting by the fellowships
SD Q 4:94  106  0:2 and grant program (2219) during the studies in the Texas A&M
D¼ ¼ ¼ 0:99 ffi 1:0 m; University/BAEN. The author also thanks to the Editor and the
m 106
anonymous reviewers for providing constructive reviews and crit-
icism on the content of the paper, in advance.
5. From Eq. (6) with Eq. (7): K = 1.887 therefore, hf = 1.887 
0
0.220.13 = 0.093 m (f = 0.0141); s = 0.093/2000 = 4.65  105;
6. To check the results from Eqs. (8) and (9):
References
5 3
r ¼ 4:65  10 þ 1:0  0:0025 ¼ 2:55  10 ; [1] Colebrook CF. Turbulent flow in pipes, with particular reference to the
3 transition region between the smooth and rough pipe laws. J Inst Civil Eng,
PW ¼ 2:55  10  1000  0:2  2000=102 ¼ 10 kW: London 1939;11:133–56.
[2] Colebrook CF, White CM. Experiments with fluid friction in roughened pipes.
Moreover, the present example can be restated assuming the Proc Roy Soc, Ser A Math Phys Sci 1937;161(904):367–81.
pipe length L is a required parameter, and the pipe diameter [3] Churchill SW. Friction factor equation spans all fluid-flow regimes. Chem Eng
1977;84(24):91–2.
(D = 1.0 m) with the remaining parameters (Q = 0.2 m3 s1,
[4] Finnemore EJ, Franzini JB. Fluid mechanics with engineering applications. 10th
s = 0.0025, PW = 10 kW) are initially known for this case. Computa- ed. Mc-Graw-Hill; 2002. p. 268–82.
tion steps by using Eqs. (17) and (18) are: [5] Gottfried BS. Spreadsheet tools for engineers: EXCEL 2000Version, New York:
McGraw-Hill; 2000.
 n  0:255 [6] Haaland SE. Simple and explicit formulas for the friction factor in turbulent
1. R ¼ p4Q
Dm
40:2
¼ p1:0106 ¼ 254; 648; a ¼ x DD0 ¼ 0:071  1:0
1:0 pipe flow. J Fluids Eng 1983;105:89–90.
1=2 3=2 1=2 3=2 [7] Hodge BK, Taylor RP. Piping-system solutions using Mathcad. Comput Appl Eng
¼ 0:071; N ¼ ðgsÞ m D ¼ ð9:810:0025Þ 106
1:0
¼ 156; 604:6; Educ 2002;10(2):59–78.
 1=3  1=3
[8] Huddleston DH. Spreadsheet tools utilized to introduce computational field
2. J L ¼ 1m 4gP WD
cpR ¼ 1
106
 49:81101021:0
1000p254;648
ffi 36; 848; SL ¼ p4 simulation concepts to undergraduate engineering students. Comput Educ J
 3
  3
 2002;12(1):6–11.
2J L
aR3þb þ2pN2 R
¼ p4  0:071254;64830:13236;848
þ21:0156;604:62 254;648
ffi 0:01 [9] Huddleston DH, Alarcon VJ, Chen W. Water distribution network analysis using
Excel. J Hydraul Eng, ASCE 2004;130(10):1033–5.
ðkl ¼ 0Þ; [10] Jain AK. Accurate explicit equations for friction factor. J Hydraul Div, ASCE
3. L ¼ SLmQ ¼ 0:010:2 ¼ 2000 m. 1976;102(5):674–7.
106 [11] Jewell TK. Teaching hydraulic design using equation solvers. J Hydraul Eng,
ASCE 2001;127(12):1013–21.
On the basis of the procedure discussed above, the data of [12] Keady G. Colebrook–White formula for pipe flows. J Hydraul Eng, ASCE
Example 2 (Type C) are also applied for a wide range of pipe slope 1998;124(1):96–7.
[13] Larock BE, Jeppson RW, Watters GZ. Hydraulics of pipeline
(starting from 0% to ±1%), only for a pipe type of tar-coated cast- systems. Florida: CRC Press LLC; 2000.
iron, and the complete results for the pipe diameter are shown in [14] Moody LF. Friction factors for pipe flow. Trans ASME 1944;66:671–84.
Fig. 4a. Furthermore, the values of the pipe length for different [15] Moody LF. An approximate formula for pipe friction factors. Trans ASME
1947;69:1005–6.
slope ranges, using the same data except for D = 0.4 m [16] Ozger M, Yıldırım G. Determining turbulent flow friction coefficient using
(Q = 0.2 m3 s1, s = 0.0025, PW = 10 kW), with Eqs. (17) and (18), adaptive neuro-fuzzy computing technique. Adv Eng Software 2009;40(4):
are also shown in Fig. 4b. Fig. 4a illustrates that as the downslope 281–7.
[17] Sonnad JR, Goudar CT. Constraints for using Lambert W function-based explicit
increases the pipe diameter decreases, whereas the pipe length in-
Colebrook–White equation. J Hydraul Eng, ASCE 2004;130(9):929–31.
creases (dotted lines). In other words, the pipe diameter increases [18] Sonnad JR, Goudar CT. Turbulent flow friction factor calculation using a
and the pipe length decreases (straight lines) with increasing ups- mathematically exact alternative to the Colebrook–White equation. J Hydraul
Eng, ASCE 2006;132(8):863–7.
lope (Fig. 4b).
[19] Swamee PK, Jain AK. Explicit equations for pipe-flow problems. J Hydraul Div,
ASCE 1976;102(5):657–64.
[20] Weiss PT, Gulliver JS. What do students need in hydraulic design projects. J
7. Summary and conclusions Hydraulic Eng, ASCE 2001;127(12):984–91.
[21] Yıldırım G. Hydraulic analysis and direct design of multiple outlets pipelines
laid on flat and sloping lands. J Irrig Drain Eng, ASCE 2006;132(6):
Using a power form friction factor, a design procedure is devel- 537–52.
oped for three design cases of single-lined commercial pipelines. [22] Yıldırım G. An assessment of hydraulic design of trickle laterals considering
effect of minor losses. Irrig Drain (ICID J) 2007;56(4):399–421.
General equations improved in this procedure provide accurate
[23] Yıldırım G. Analytical Relationships for designing multiple outlets pipelines. J
solutions for the general case of three types of design problems Irrig Drain Eng, ASCE 2007;133(2):140–54.
considering the effect of local losses. The user-friendly MathCAD [24] Yıldırım G. Determining inlet pressure head incorporating uniformity
parameters for multioutlet plastic pipelines. J Irrig Drain Eng, ASCE
procedure is also carried out for solving the general case of a Type
2008;134(3):341–8.
B problem. The proposed procedure is applied to several design [25] Yıldırım G. Discussion of ‘‘Turbulent flow friction factor calculation using a
examples for three types of commercial pipes and for a wide mathematically exact alternative to the Colebrook–White equation”. By
range of uniform pipe slope, and the results are shown as design Sonnad and Goudar. J Hydraul Eng, ASCE 2008;134(8):1185–6.
[26] Yıldırım G, Özger M. Neuro-fuzzy approach in estimating Hazen–Williams
curves for practical purposes. These curves provide an opportu- friction coefficient for small-diameter polyethylene pipes. Adv Eng Software
nity to compare the complete results of the proposed method 2009;40(8):593–9.
496 G. Yıldırım, V.P. Singh / Advances in Engineering Software 41 (2010) 489–496

[27] Yıldırım G. Computer-based analysis of explicit approximations to the implicit
Colebrook–White equation in turbulent flow friction factor calculation. Adv
Eng Software 2009;40:11:1183–90.
[28] Yıldırım G. Simplified procedure for hydraulic design of small-diameter plastic
pipes. Irrig Drain (ICID J) 2009;58(3):209–33.
[29] Yıldırım G. Discussion of ‘‘simplified method for sizing laterals with two or
more diameters” by Muluneh Yitayew. J Irrig Drain Eng, ASCE, in press.
[30] Yoo DH, Singh VP. Explicit design of commercial pipes with no secondary
losses. J Irrig Drain Eng, ASCE 2004;130(5):437–40.
[31] Yoo DH, Singh VP. Two methods for the computation of commercial pipe
friction factors. J Hydraul Eng, ASCE 2005;131(8):694–704.
[32] Zigrang DJ, Sylvester ND. Explicit approximations to the solution of
Colebrook’s friction factor equation. AICHE J 1982;28:514–5.