MATHEMATICAL MODELLING OF A publication of IIETA

ENGINEERING PROBLEMS
ISSN: 2369-0739 (Print), 2369-0747 (Online)
Vol. 4, No. 4, December 2017, pp. 162-167
DOI: 10.18280/mmep.040404
Licensed under CC BY-NC 4.0 http://www.iieta.org/Journals/MMEP

Design and modelation of piping systems by means of use friction factor in

the transition turbulent zone
Yanán C. Medina1*, Oscar M.C. Fonticiella1, Osvaldo F.G. Morales2

Center for Energy Studies and Environmental Technologies, Universidad Central de las Villas,
Santa Clara, Cuba
Technical Sciences Faculty, Universidad de Matanzas, Matanzas, Cuba

Email: [email protected]

ABSTRACT

In this paper a new model is presented for design and modelation of piping systems. This work
results from recent investigations on pipes friction factor. It provides an empirical solution for the
solution of the three basic problems found in the design and evaluation of pipe systems, which in
conventional cases require tedious iterative trial and error processes. The proposed solutions are
valid in the same interval as the traditional methods used, and in all cases the average error
computed never exceeds 2% with respect to traditional iterative methods. The research was done
with a regression analysis between kinematic viscosity, relative roughness, flow rate, friction
factor, and others factor, using experimental data reported by different authors, establishing
comparison with the Swamee-Jain solution for this problems types concluding that between new
model and the most universally used there are not signified differences without is lightly better.

Keywords: Explicit Equation, Darcy Friction Factor, Flow in Pipes, Pipe Diameter.

1. INTRODUCTION friction factor, depending on relative roughness of the

numerical value of Reynolds's dimensionless number, the one
Many of the current problems of engineering involve the flow that the friction factor begins to be persevering may be
of fluids in pipes, this brings associated that the determination determined by the following expression obtained by (Camaraza:
of the friction factor in these systems is of vital importance for 2011)
an adequate study of the process in question. Friction is an
element that occurs in any flow regime, either in single phase 542,38
Re CRIT 2  (2)
or in two phases. For the determination of Darcy's friction factor, e d 1,13
one of the most widespread works in the world is the well-
known Coolebrook-White Equation, which in a simplified form Table 1. Equivalent roughness values for new commercial
is described by the following expression: pipes

e  Material Roughness e (mm)

 2 Log  d  
1 2,51 (1)

Glass , plastic 0,0005 - 0,01
f  3.7 Re f
  Concrete 0,05 - 9
Wood stave 0,5
In Equation (1) e is the pipe wall equivalent sand grain Rubber smoothed 0,01 – 0,06
Commercial steel 0,04 – 0,08
roughness, d Equivalent inner tube diameter of pipe , f in the Cooper or brass tubing 0,01 – 0,05
Darcy-Weibach dimensionless friction factor; Re  Gd  is Cast iron 0,15 - 0,35
Galvanized iron 0,07 – 0,23
the Reynolds number, (with G being the mass flux and  the Wrought Iron 0,025 – 0,046
dynamic viscosity). Stainless steel 0,002 – 0,004
However it is known and confirmed at the same time by
experiments, than at the totally wrinkled zone the numerical The absolute roughness value e , included in the above
value of Reynolds's dimensionless number stops influencing the Equations, varies with the material of the conduit and with the

162
technology of its manufacture. In this paper, the values 2. MATERIALS AND METHODS
recommended in the Cuban’s NC-176-2002 are shown in the
table 1 2.1 Development of problems type I.
The tubes used in industrial facilities are different from those
used in the experiments in the sense that the roughness of the The problem type I is the simplest of all because the results
former is not uniform and it is difficult to give an accurate sought are obtained directly. The friction factor is determined
description of it. Table 1 gives values of the equivalent and then the Darcy Weibash relationship is implemented to
roughness for some commercial pipes, but it must be borne in obtain the pressure drop. The only problem that limits the
mind that these values are for new pipes and the relative precision of the final results is the determination of the friction
roughness of these can be increased with the use as a factor, since depending on the model used and the assumed
consequence of corrosion, the accumulation of scale. and the conditions for obtaining the friction factor, it is the degree of
precipitation. As a result, the friction factor can be increased by precision.
a factor of 5 to 10 Currently in the literature consulted and available are known
It is worth mentioning that the precision of the calculations more than a dozen expressions for determining the friction
on friction in the tubes, can vary with the somewhat factor. A large part of them allow obtaining the friction factor
unpredictable change in the roughness and the friction factor, as an explicit variable based on known variables, thus
due to the accumulation of sediment and corrosion on the walls facilitating the calculation and simplifying the analyzes,
of the tube with the time of use of the conduit . This however they have the disadvantage that their areas of
accumulation not only increases the surface roughness, but also applicability is lower with respect to Coolebrook-White’s
reduces the effective diameter of the tube and can lead to a Equations (1) and that the accuracy of the results is also
potentially large increase in the friction factor after the tube has sacrificed due to having a lower correlation index. Table 2
been in service for a long period, so in these cases the roughness shows a summary of the most widespread at present in the
values given in the table can contain errors that can compute up literature consulted and available.
to 70%.
In the literature consulted a remarkable group of explicit 2.2 Development of problems type II and III.
Equations for the calculation of the Darcy friction factor f in
smooth and rough pipes for turbulent regime. For the In the problems type II, diameter is given, but the flow rate is
determination of the pressure losses in pipes or systems of these, unknown. A good assumption for the friction factor in such case
it is required to estimate the friction factor. To this end, iterative is obtained from the fully turbulent flow region for the given
solutions such as Equation (1) can be used, however their use roughness. This is true for large Reynolds numbers, which is
requires an appreciable calculation time, especially when often the case in practice. After the flow rate is obtained, the
studying large water distribution networks. A quick solution friction factor can be corrected with the Moody diagram,
that currently enjoys great acceptance is the use of the Moody Equation (1) or similar from table 2, and the process is repeated
diagram, especially in previous decades, however its use has until the solution converges (usually, only a few iterations are
two important drawbacks: needed for convergence to three or four digits of precision).
1- The precision of the results is affected by reading errors in In problems type III , the diameter is not known and therefore
logarithmic scale the Reynolds number and the relative roughness cannot be
2- it is not possible to apply in computer-aided simulations calculated. Consequently, calculations are started with the
In the design and analysis of piping systems that involve the assumption of a pipe diameter. Then the pressure drop
use of the Moody chart (or the Colebrook Equation), is it usually calculated for the assumed diameter is compared to the
encounter three types of problems, in which the fluid and the specified pressure drop, and the calculations are repeated with
roughness of the pipe are assumed to be specified in all cases. I) another pipe diameter iteratively until convergence.
Determining the pressure drop (or head loss) when the pipe To avoid tedious iterations in load loss, flow rate and
length and diameter are given for a specified flow rate (or diameter calculations, Swamee and Jain proposed in 1976 the
velocity); II) Determining the flow rate when the pipe length following explicit relationships, whose use throws average error
and diameter are given for a specified pressure drop (or head with respect to the values obtained through the iterative process
loss) and III) determining the pipe diameter when the pipe described in the previous paragraphs.
length and flow rate are given for a specified pressure drop (or a) For the problems type 1 (Determining the pressure
head loss) drop)
In the engineering processes that require the analysis of piping
systems, there are generally three basic types of problems to be The Equation (12) is validate for
solved, in which it is necessary to assume the specific pipe 3  103  Re  3  108 and 0,01  e d  106
roughness. These three basic problems are reduced to:
I. Determine the pressure drop for the circulation of a required In Equation (12) w is the flow rate, g is the gravity
flow (or velocity), the diameter and length of the pipe being acceleration, L is the pipe length, ν is the kinematic viscosity.
known.
II. Determine the allowable flow rate for a preset pressure drop, b) For the problems type II (Determining the flow rate)
the diameter and length of the pipe being known.
III. Determine the diameter of the pipe required for a preset
pressure drop, the length of the pipe and the flow rate flowing
through it being known.

163
 Table 2. Somes empirical Equations for the determination of Darcy's friction factor f

Author Equation Validity range

2
  Re  4000  Re  108
Alshul(1962) f  1,8Log   (2)
  0,135 Re e d   6,5  0,05  e d  106
e 0, 9 
5000  Re  108
 6,81  
 2 Log  d  
1
Pavlov(1975)  3.7  Re   (3)
f
  0,01  e d  106
 e
5,74  5000  Re  108
 2 Log  d  0,9 
Swamee-Jain 1
 3.7 Re  (4)
(1976) f
  0,01  e d  106
 e  1,11

 6,9  5000  Re  108
 1,8Log   d 
1
 
Halaand (1982)   (5)
f   3.7 

Re 

0,01  e d  106

e e 
 13   5000  Re  108
Log  d 
1 5,02
Zigrang (1982)  2 Log  d  (6)
f  3.7 Re  3.7 Re   0,05  e d  107
  
1
1  8 12 1  12

 8   3 
f  Re   A  B  2 
16 4000  Re  108
Churchill (1973)    (7)
   16 0,05  e d  107
   37,53 
A  2,457 Ln
1
 ; B 
 0, 9
 Re 
   7   0,27 e 
   Re  d 

  Re  
e 4,518Log   
1
 2 Log  d   7   4000  Re  108

  0, 7  
Barr (1981) (8)
f 3,7  1 0,05  e d  107
 Re 1  Re 0,52 e 
  29 d 
0,08 e d   0,265e d   66,69 e d 
0 , 225 0, 4
f  5000  Re  108
Wood (1961) Re N (9)
0,05  e d  107
N  1,778e d 
0 ,134

 e d 
2
 95 96,82  5000  Re  108
f   2 Log    
Madanilli (1981)
  3,7 Re 0, 983
Re  (10)
0,05  e d  107
2
  e  
f   2  Log     0,568 
  d   
4000  Re  108
 47,6 3,256 e d  e d 
Camaraza et al. 1,15

2, 2
(11)
(2011)   8,49 Log  2    0,05  e d  107
 Re Re 16 ,598 

 
  Re Log Re 2,5 e d 1,1  0 , 01

 
2
 
w L   e d  d    gdp  e d 1,784 
2 0,9

p  1,07  Ln   4,62    (12) w  0,965 d 2 Ln  (13)

gd 5   3,7 w   L  3,7 gd 3 p 
 
 L 

164
 Equation (13) is validate for 4  10 3  Re  108 and 4,94 gd 5 p  e 43,12 2 L  
0,05  e d  107 . In Equation (13) ∆p is the head loss in the w Log     0,568  (16)
L   d gd 3 p  
duct.  

c) For the problems type III (Determining the pipe The Equation (16) is validate for 4  10 3  Re  108 and
diameter) 0,05  e d  107
d) k) For the problems type III (Determining the pipe
0, 04
 2 4, 75 5, 2
 diameter)
1, 25  Lw  9, 4  L 
d  0,66e    w    (14)
  gp   gp    ew 7 , 6  0,11L 
1, 9
 0, 4 w3,76  0,14L 
2 , 08 0 ,1

d        (17)
 p   p 
3, 8 2 , 08
 g g 
Equation (4) is validate for 3  103  Re  3  108 and
6
0,01  e d  10 The Equation (17) is validate for 3 103  Re  3,2 108 and
0,05  e d  10 6
Note that in the expression (12); (13) and (14) all quantities
are dimensional and the units simplify to the desired unit. Note that in the expression (15); (16) and (17) all quantities
are dimensional and the units simplify to the desired unit.
Figure 1 shows the correlation of 2284 experimental points
3. PROPOSED EXPLICIT EQUATIONS with the obtained Equation (17) to solve problems type III
(Determining the pipe diameter) in logarithmic coordinates,
A data set of 3418 exact values of d was generated by
solving numerically the Swamme-Jain Equation (14) for
3  103  Re  3,2  108 . For every value of Reynolds number,
e was change in the range 0,05  e d  107 . The previous
ranges of Re and e correspond to 3 10 3  Re  3 108 and
0,01  e d  10 6 . The coefficients for the two Equations
presented in this paper were development with the Least
Squares Method in tkSolver.
Absolute relative errors were estimated by:

d  d SJ
E
d SJ Figure 1. Correlation of 2284 experimental points with the
Equation (17)
where dSJ is the diameter value obtained by the Swamme-Jain
Equations (12), (13) y (14).
According the authors [19], which is reaffirmed in [20],
Equation (14) has an average margin of error of approximately
2%. Recently [Babajimopoulos and Terzidis, 2013] found that
the mean error is approximately 2.75%. The authors of this
article agree with this criterion when finding a mean error
value of Equation (14) equal to 2.8%
The authors got three expressions for the determination of
head losses , the flow rate and the diameter calculation , whose
use generate a of half error regarding values that were
obtained through the iterative process described in previous
paragraphs, in addition to possess a more ample specific field.
The so-called expressions are:

i) For the problems type 1 (Determining the pressure Figure 2. Correlation enter adimensional Reynold’s number
drop) and medium error obtained with the uses of Equation (16)
0,25 w 2 L (15)
p  2
In the figure 2 is shown the correlation between
    d   dimensionless Reynold’s number and the medium error
gd 5  Log e d   23 ,155    0,632
1,11
obtained when using Equation (16) was made a comment
   w   beforehand about the divergence among these models is little,
but the question arises as to which of these models best
The Equation (15) is validate for 3  103  Re  3  108 and represents the experimental data that originated or was used for
0,05  e d  10 6 its validation.
The calculation of the relative error when validating
j) For the problems type II (Determining the flow rate) Swamee -Jain's Equation (14) is shown in figure 3. The

165
analysis made to this Equation evidences a maximum error facilitates the engineering calculations, besides extending the
equal to 3,8 %, while experimental available data correlate domain of applicability of the same ones to possess a zone of
with a  2.8% in 82.4 % of the experimental points. validation more extensive, so that its work is recommended in
The calculation of the error relative when validating process of engineering calculations.
Equation (17) is shown in figure 4. The analysis made to this The mathematical performance of the explicit proposed
Equation evidences a maximum error equal to 3.92 % , while model is:
experimental available data correlate with a  2.5% in 88,1 %
2 , 08 0 ,1
of the experimental points. It can be verified that the Equation  ew 7,6  0,11L 
1, 9
 0, 4 w 3,76  0,13L  
(17) provides results that are closer to the available d      
 p   p 
3,8 2 , 08
experimental data, so it is considered to be more accurate,  g g 
besides having a more specific field.
Validate for 3 103  Re  3,2 108 and 0,05  e d  10 6
The analysis made to this Equation evidences a maximum
error equal to 3,92 % , while experimental available data
correlate with a  2,5% in 88,1 % of the experimental points.
It can be verified that the Equation (8) provides results that are
closer to the available experimental data, so it is considered to
be more accurate, besides having a more specific field.

5. ACKNOWLEDGMENT

This work was supported by Doctoral Research Program of

Universidad Central de las Villas, Cuba.

Figure 3. Correlation of experimental data with Swamee-Jain REFERENCES

Equation (14)
[1] Giustolisi O., Berardi L., Walski T.M. (2011). Some
explicit formulations of Colebrook–White friction
factor considering accuracy vs. computational speed,
Journal of Hydroinformatics, Vol. 13, No. 3, pp. 401-
418.
[2] Bombardelli F., Garcia M. (2003). Hydraulic design of
large-diameter pipes, J.Hydraul.Eng., Vol. 129, No. 11,
pp. 839-846.
[3] Brkić D. (2011a). New explicit correlations for
turbulent flow friction factor, Nucl.Eng.Des., Vol. 241,
No. 9, pp. 4055-4059.
[4] Brkić D. (2011b). Review of explicit approximations to
the Colebrook relation for flow friction, Journal of
Petroleum Science and Engineering, Vol. 77, No. 1, pp.
Figure 4. Correlation of experimental data with Equation 34–48.
(17) [5] Clamond D. (2009). Efficient resolution of the
Colebrook equation, Ind Eng Chem Res, Vol. 48, No. 7,
By the results obtained, the Equation (17) shows with an pp. 3665-3671.
average error value of the order of 2,5%, which is lower than [6] Camaraza Y. el at., (2010) , Ecuación explí cita para el
the 2,8% error obtained by using Equation (14). Another aspect cálculo de factores de fricción en la zona de transición
of interest is the fact that the experimental data used in the del régimen turbulento, Tecnologí a Química, Vol. 30,
validation of the models fit better to Equation (17) than to No. 1, pp. 76-83.
expression (14), these reasons are solid elements by which the [7] Colebrook C.F. (1939). Turbulent flow in Pipes, with
authors of this article recommend the expression (17) for the reference to the transition region between the smooth
evaluation of piping systems, considering it accurate and and rough pipe laws, J. Inst. Civil Eng, Vol. 11, No. 4,
reliable. pp. 133-156.
[8] Danish M., Kumar S., Kumar S. (2011). Approximate
explicit analytical expressions of friction factor for flow
4. CONCLUSIONS of Bingham fluids in smooth pipes using Adomian
decomposition method, Communications in Nonlinear
An explicit model was obtained to determine the diameter Science and Numerical Simulation, Vol. 16, No. 1, pp.
of a duct, using a regression, with the best possible adjustment 239-251.
to the experimental data that gave rise to it and better than the [9] Diniz V.E. M.G., Souza P.A. (2009). Four explicit
best recognized model, the Swamee - Jain Equation. The new formulae for friction factor calculation in pipe flow,
Equation obtained allows to reach a greater precision,

166
 Transactions on Ecology and the Environment, Vol. [15] Jain K. (1976). Accurate explicit equation for friction
125, pp. 369-380. factor, Journal of Hydraulics Division, ASCE, Vol. 102,
[10] Fang X., Xu Y., Zhou Z. (2011). New correlations of pp. 674-677.
single-phase friction factor for turbulent pipe flow and [16] Li P., Seem J.E., Li Y. (2011). A new explicit Equation
evaluation of existing single-phase friction factor for accurate friction factor calculation of smooth pipes,
correlations, Nucl.Eng.Des, Vol. 241, No. 3, pp. 897- Int.J.Refrig., Vol. 34, No. 6, pp. 1535-1541.
902. [17] Romeo E., Royo C., Monzón A. (2002). Improved
[11] Barr D.I.H. (1977). Discussion on accurate explicit explicit equations for estimation of the friction factor in
equations for friction factor, J. Hydraul. Div. Am. Soc. rough and smooth pipes., Chem.Eng.J., Vol. 86, No. 3,
Civ. Eng., Vol. 103, No. 3, pp. 334-337. pp. 369-374.
[12] Gulyani B.B. (2001). Approximating equations for pipe [18] Sonnad J., Goudar C. (2006). Turbulent flow friction
sizing, Chemical Engineering, Vol. 108, No. 2, pp. 105- factor calculation using a mathematically exact
108. alternative to the colebrook–White equation.,
[13] Imbrahim C. (2005). Simplified equations calculate J.Hydraul.Eng., Vol. 132, No. 8, pp. 863-867.
head losses in comercial pipes, The Journal of [19] Swamee P.K., Jain A.K. (1976). Explicit equations for
American Science, Vol. 1, No. 1, pp. 1-2. pipe flow problems, J.Hydraul.Eng. ASCE, Vol. 102,
[14] Haaland S.E. (1983). Simple and explicit formulas for No. 5, pp. 657-664.
the friction factor in turbulent pipe flow, J.Fluids Eng., [20] Swamee P.K., Rathie P.N. (2007). Exact equations for
Vol. 105, No. 1, pp. 89-90. pipe-flow problems, Journal of Hydraulic Research,
Vol. 45, No. 1, pp. 131-134.

167