See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/242138088

The History of the Darcy-Weisbach Equation for Pipe Flow Resistance

Article · October 2002

DOI: 10.1061/40650(2003)4

CITATIONS READS
154 21,033

1 author:

Glenn O. Brown
Oklahoma State University - Stillwater
87 PUBLICATIONS   1,204 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Use of Fly Ash for Phosphorus Trapping in Bioretention Cells View project

Beyond Pretty Pictures: Development of methods to quantify computerized tomography images of porous media View project

All content following this page was uploaded by Glenn O. Brown on 28 January 2017.

The user has requested enhancement of the downloaded file.

 ENVIRONMENTAL AND WATER RESOURCES HISTORY 34

The History of the Darcy-Weisbach Equation for Pipe Flow Resistance

Glenn O. Brown1

Abstract

The historical development of the Darcy-Weisbach equation for pipe flow

resistance is examined. A concise examination of the evolution of the equation itself
and the Darcy friction factor is presented from their inception to the present day. The
contributions of Chézy, Weisbach, Darcy, Poiseuille, Hagen, Prandtl, Blasius, von
Kármán, Nikuradse, Colebrook, White, Rouse and Moody are described.

Introduction

What we now call the Darcy-Weisbach equation combined with the

supplementary Moody Diagram (Figure 1) is the accepted method to calculate energy
losses resulting from fluid motion in pipes and other closed conduits. When used
together with the continuity, energy and minor loss equations, piping systems may be
analyzed and designed for any fluid under most conditions of engineering interest.
Put into more common terms, the Darcy-Weisbach equation will tell us the capacity
of an oil pipeline, what diameter water ma in to install, or the pressure drop that
occurs in an air duct. In a word, it is an indispensable formula if we wish to engineer
systems that move liquids or gasses from one point to another.
The Darcy-Weisbach equation has a long history of development, which
started in the 18th century and continues to this day. While it is named after two great
engineers of the 19th century, many others have also aided in the effort. This paper
will attempt the somewhat thorny task of reviewing the development of the equation
and recognizing the engineers and scientists who have contributed the most to the
perfection of the relationship. Many of the names and dates are well known, but
some have slipped from common recognition. As in any historical work, others may
well find this survey lacking in completeness. However, space limitations prevent an
exhaustive commentary, and the author begs tolerance for any omissions. As a final
note, to minimize confusion, standardized equation forms and variable symbols are
used instead of each researcher's specific nomenclature. Likewise, simple
replacements, such as diameter for radius, are made without note.

1
Professor, Biosystems and Agricultural Engineering, Oklahoma State University,
Stillwater, OK 74078; phone 405-744-8425; gbrown@okstate.edu
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 35

Figure 1. Moody diagram. (Moody, 1944; reproduced by permission of ASME.)

The Equation

The fluid friction between two points in a straight pipe or duct may be
quantified by the empirical extension of the Bernoulli principle, properly called the
energy equation,
V 2 p  V 2 p   p  p 
hl =  1 + 1 + z1  −  2 + 2 + z 2  ≈  1 + z1  −  2 + z 2  (1)
 2 g ρg   2 g ρg   ρg   ρg 
where hl is the fluid friction or head loss between positions subscripted 1 and 2, V is
the average velocity, g is the acceleration of gravity, p is the fluid pressure, ρ is the
fluid density and z is the elevation of the pipe. When analysis is limited to uniform
(constant area) flow, the velocity terms cancel, and the RHS is used. Note that Eq. 1
is not predictive unless all variables on the RHS are known. We must measure
pressures in a given pipe system at a specific flow rate to compute the losses. That is,
we have to build the system to determine how it will work.
Engineering design requires a relationship that predicts hl as a function of the
fluid, the velocity, the pipe diameter and the type of pipe material. Julius Weisbach
(1806-1871) a native of Saxony, proposed in 1845 the equation we now use,
fL V 2
hl = (2)
D 2g
where L is the pipe length, D is the pipe diameter, and f is a friction factor (Weisbach,
1845). Note that Eq. 2 only predicts the losses due to fluid friction on the pipe wall
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 36

and the effect of fluid viscosity and does not include minor losses at inlets, elbows
and other fittings. While Weisbach presented Eq. 2 in the form we use today, his
relationship for the friction factor was expressed as,
β
f =α + (3)
V
where α and β are friction coefficients that could vary by pipe diameter and wall
material. Equation 3 was based on a relatively small data set. Weisbach reported 11
of his own experiments, while 51 measurements were taken from the reports of
Claude Couplet (1642-1722), Charles Bossut (1730-1799), Pierre Du Buat (1734-
1809), Gaspard Riche de Prony (1755-1839) and Johann Eytelwein (1764-1848).
Weisbach's publication covered most of engineering mechanics and arguably
set the standard for all later engineering textbooks. By 1848 it was translated and
published in America; a rather remarkable pace for the time. However, his text had
no apparent impact in France, the contemporary center for hydraulic research. This is
a curious situation since it is believed that Weisbach's interest in hydraulics
developed after visiting the Paris Industrial Exposition in 1839 and that he also
attended the 1855 Paris World Exposition. Perhaps since Weisbach's equation was
based mostly on their data, the French may have believed it provided no improvement
over the Prony equation in wide use at the time,
hl = (aV + bV 2 )
L
(4)
D
where a and b are empirical coefficients. While the exact values of the Prony
coefficients were debated, it was believed that they were not a function of the pipe
roughness.
A noteworthy difference between Eqs. 2 and 4 is that Weisbach developed a
dimensionally homogenous equation. Consequently, f is a non-dimensional number
and any consistent unit system, such as SI or English Engineering may be used. That
is not the case with Prony's. The roughness coefficients, a and b take on different
values depending on the unit system. To the modern eye, Prony's dimensionally
inhomogeneous equation is unsightly, but in 1840 there were no electronic
calculators, and in fact the modern slide rule was yet to be developed. Thus, Prony's
relationship that requires six math operations had a practical advantage over
Weisbach's that required eight. Additionally, it was standard practice for the French
to drop the first order velocity term, (aV) of Prony's equation to yield a roughly
equivalent relationship to Weisbach's that required only four math operations
(D'Aubuisson, 1834).
While Weisbach was ahead of most other engineers, his equation was not
without precedent. About 1770, Antoine Chézy (1718-1798) published a
proportionally based on fundamental concepts for uniform flow in open channels,
V 2P ∝ A S (5)
where P is the wetted perimeter, S is the channel slope, and A is the area of flow.
Chézy allowed that the proportionality might vary between streams. It is a simple
matter to insert a proportionality coefficient, C to yield,
V = C RS (6)
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 37

where R is the hydraulic radius given by, R = A/P. By introducing the geometry of a
circular pipe and noting that under uniform flow conditions S = hl/L, Eq. 6 is
transformed to,
4 L
hl = 2 V 2 (7)
C D
Equation 7 may be considered a dimensionally inhomogeneous form of Eq. 2.
1 C
Equating one to the other shows that = . Unfortunately, Chézy's work was
f 8g
lost until 1800 when his former student, Prony published an account describing it.
Surprisingly, the French did not continue its development, but it is believed that
Weisbach was aware of Chézy's work from Prony's publication (Rouse and Ince,
1957).
The Darcy-Weisbach equation (Eq. 2) is considered a rational formula since
basic force balance and dimensional analysis dictate that hl ∝ L D –1 V 2 g -1 .
However, the friction factor, f is a complex function of the pipe roughness, pipe
diameter, fluid kinematic viscosity, and velocity of flow. That complexity in f, which
results from boundary layer mechanics, obscures the valid relationship and led to the
development of several irrational, dimensionally inhomogeneous, empirical formulas.
Weisbach deduced the influence of roughness, diameter and velocity on f, but the
professional community apparently ignored his conclusions. In addition, the effect of
fluid properties was habitually neglected since water at normal temperatures was the
only major concern. It would take almost a hundred years and the input of several
others for f to be defined completely.

Laminar Flow

By the 1830's the difference between low and high velocity flows was
becoming apparent. Independently and nearly simultaneously, Jean Poiseuille (1799-
1869) and Gotthilf Hagen (1797-1884) defined low velocity flow in small tubes
(Hagen, 1839; Poiseuille, 1841). In modern terms they found,
L V
hl = 64ν 2 , (8)
D 2g
where ν is the fluid kinematic viscosity. Note however that neither Poiseuille nor
Hagen used an explicit variable for the viscosity, but instead developed algebraic
functions with the first and second powers of temperature. The most important aspect
of Poiseuille's and Hagen's results was their accuracy. While the restriction to small
tubes and low velocity was realized, theirs were the first fluid- friction equations to
achieve modern precision. When compared to one another, Hagen's work was more
theoretically sophisticated, while Poiseuille had the more precise measurements and
looked at fluids other than water. An analytical derivation of laminar flow based on
Newton's viscosity law was not accomplished until 1860 [Rouse and Ince, 1957].
Darcy (1857) also noted the similarity of his low velocity pipe experiments
with Poiseuille's work. "Before seeking the law for pipes that relates the gradient to
the velocity, we will make an observation: it appears that at very- low velocity, in
pipes of small diameter that the velocity increases proportionally to the gradient."
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 38

Later he showed explicitly that his newly proposed pipe friction formula would
reduce to Poiseuille's at low flow and small diameters. He noted that this was a "…
rather remarkable result, since we arrived, Mr. Poiseuille and I, with this expression,
by means of experiments made under completely different circumstances."
Osborne Reynolds (1842-1912) described the transition from laminar to
turbulent flow and showed that it could be characterized by the parameter,
VD
Re = (9)
ν
where Re is now referred to as the Reynolds number (Reynolds, 1883). The most
widely accepted nominal range for laminar flow in pipes is Re < 2000, while
turbulent flow generally occurs for Re > 4000. An ill-defined, ill-behaved region
between those two limits is called the critical zone. Once the mechanics and range on
laminar flow was well established, it was a simple matter to equate Eqs. 4 and 9 to
provide an expression for the Darcy f in the laminar range,
64
f = (10)
Re
It is unknown who was the first person to explicitly state Eq. 10, but it appeared to be
commonly recognized by the early 1900's. Equation 10 is plotted on the left side of
Figure 1.

Turbulent Flow

In 1857 Henry Darcy (1803-1858) published a new form of the Prony

equation based on experiments with various types of pipes from 0.012 to 0.50 m
diameter over a large velocity range (Darcy, 1857). His equation for new pipes was,
L  β   β' 
hl =  α + 2  V + α '+  V 2  (11)
D  D   D 
where α, β, α ' and β' are friction coefficients. Darcy noted that the first term on the
RHS could be dropped for old rough pipes to yield,
L β " 2
hl =  α "+  V (12)
D D
where the coefficients α" and β" would have different values than for new pipes.
Contrary to existing theory, he showed conclusively that the pipe friction factor was a
function of both the pipe roughness and pipe diameter. Therefore, it is traditional to
call f, the "Darcy f factor", even though Darcy never proposed it in that form.
J. T. Fanning (1837-1911) was apparently the first to effectively combine
Weisbach's equation with Darcy's better estimates of the friction factor (Fanning,
1877). Instead of attempting a new algebraic expression for f, he simply published
tables of f values taken from French, American, English and German publications,
with Darcy being the largest source. A designer could then simply look up an f value
from the tables as a function of pipe material, diameter and velocity. However, it
should be noted that Fanning used the hydraulic radius, R instead of D in the friction
equation. Thus "Fanning f" values are only ¼ of "Darcy f" values. The Fanning form
of the equation remains in use in some fields, such as heat exchanger design, where
non-circular conduits are common.
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 39

During the early 20th century, Ludwig Prandtl (1875-1953) and his students
Theodor von Kármán (1881-1963), Paul Richard Heinrich Blasius (1883-1970) and
Johann Nikuradse (1894-1979) attempted to provide an analytical prediction of the
friction factor using Prandtl's new boundary layer theory. Apparently, Blasius (1913)
was the first person to apply similarity theory to establish that f is a function of the
Reynolds number. From experimental data he found for smooth pipes,
0.3164
f = (13)
Re1/ 4
which is now referred to as the Blasius formula and is valid for the range 4000 < Re <
80,000. Using data from Nikuradse, the entire turbulent flow range is better fit by the
relationship,
1
f
( )
= 2 log Re f − 0.08 (14)

Equation 14 has been referred to both as von Kármán's (Rouse, 1943) and Prandtl's
(Schlichting, 1968). It is plotted on Figure 1 and labeled "Smooth Pipes".
Rough pipes offered additional challenges. At high Reynolds number in rough
pipes, f becomes a constant that is only a function of the relative roughness, ε/D,
where ε is the height of the interior pipe roughness. Similar to the smooth pipe
formula, von Kármán (1930) developed an equation confirmed by data collected by
Nikuradse (1933),
1 ε
= 1.14 − 2 log   (15)
f  D
The horizontal lines on the right of Figure 1 plot Eq. 15 for various ratios of ε/D.
The transition region between laminar and fully turbulent rough pipe flow was
defined empirically by detailed measurements carried out by Nikuradse (1933) on
pipes that had a uniform roughness created by a coating of uniform sand covering the
entire pipe interior. His data showed clear trends that could be explained by the
interaction of the pipe roughness with the fluid boundary layer. However,
measurements by Colebrook and White (1937) showed that pipes with non- uniform
roughness did not display the same transition curves. For commercial pipes White
(1939) showed the transition region could be described by,
1 ε 9.35 
= 1.14 − 2 log  + (16)
 D Re f 
f  
Equation 16 is plotted in Figure 1 for various ratios of ε/D in the region labeled
"Transition Zone".

Integration

It would wait for Hunter Rouse (1906-1996) in 1942 to integrate these various
formulas into a useful structure. He noted unambiguously, (Rouse, 1943) "These
equations are obviously too complex to be of practical use. On the other hand, if the
function which they embody is even approximately valid for commercial surfaces in
general, such extremely important information could be made readily available in
diagrams or tables." Using published data he showed Eq. 16 was a reasonable
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 40

function for commercial pipe. Rouse then developed a diagram (Figure 2) that
presented Eqs. 10, 14, 15, and 16 in a systematic and somewhat intricate fashion. The
primary vertical axis plotted 1 / f , the primary horizontal axis plotted Re f , and
secondary axes plotted Re and f. Equations 15 and 16 were plotted for various values
of relative roughness. In an open corner, he also provided a convenient list of pipe
roughness. Rouse's original contribution in addition to the overall synthesis was
defining the boundary between the transition and fully turbulent zones as,
1 ε Re
= (17)
f D 200
Equation 17 is plotted on both Figures 1 and 2.

Figure 2. Rouse diagram. (Rouse, 1943; reproduced by permission of IIHR.)

Lewis Moody (1880-1953) was in the audience when Rouse presented his
paper. Moody felt that Rouse's diagram was "inconvenient" and decided to redraw
Rouse's diagram "in the more conventional form used by Pigott, ..." (Moody, 1944).
Moody's paper was primarily an instructional lecture, and as he said, "The author
does not claim to offer anything particularly new or original, his aim merely being to
embody the now accepted conclusion in convenient form for engineering use."
Moody acknowledged previous researchers, and reproduced figures from both
Colebrook and Rouse.
It should be noted that Moody's diagram is more convenient to use when
finding hl with known Q and D. However, Rouse's diagram allows a direct, non-
iterative solution for Q with known hl and D. Thus, each has its advantages.
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 41

A Rose by Any Other Name

The naming convention of the Darcy-Weisbach's equation in different

countries and through time is somewhat curious and may be tracked in the
contemporaneous textbooks. Generally, French authors have identified any
relationship in the form of Eqs. 2 or 4 as "La formule de Darcy". The friction factor
may be listed as either f or as Darcy's Number, Da. An early English text, (Neville,
1853) identified Eq. 2 as the "Weisbach Equation", but later authors have generally
adopted the French terminology. Surprisingly up to the 1960's, German authors
either gave it a generic name like "Rohrreibungsformel" (pipe formula) or use the
French jargon. However, almost all German authors now use "Darcy-Weisbach".
The equation's designation has evolved the most in America. Early texts such
as Fanning (1877) generally do not name the equation explicitly; it is just presented.
In the period around 1900 many authors referred to Eq. 2 as Chézy's or a form of
Chézy's (Hughes, and Stafford, 1911). However, by the mid-century, most authors
had again returned to leaving the equation unspecified or gave it a generic name.
Rouse in 1942 appears to be the first to call it "Darcy-Weisbach" (Rouse, 1943). That
designation gained an official status in 1962 (ASCE, 1962), but did not become well
accepted by American authors until the late 1980's. A check of ten American fluid-
mechanics textbooks published within the last eight years showed tha t eight use the
Darcy-Weisbach naming convention, while two continue to leave the equation's name
unspecified. While variations across oceans and languages are to be expected, it is
disappointing that a single nomenclature for Eq. 2 has not been adopted after 157
years, at least in the United States.
Rather ironically and contrarily to the practice with the equation name, the f
versus Re diagram is universally credited to Moody, and the contributions of others
are seldom acknowledged. This was a sore point with Rouse (1976), and he wrote of
their 1942 meeting (in third person),

"After the Conference, Lewis Moody of Princeton suggested using the

latter variables (f and Re) as primary rather than supplementary, as in
the past, but Rouse resisted the temptation because he felt that to do
would be a step backward. So Moody himself published such a plot,
and it is known around the world as the Moody diagram!"

In his writing, Rouse used the exclamation point very sparingly, thus the intensity of
his opinion is apparent.

Closing Comments

With Moody's publication, practitioners accepted the Darcy-Weisbach

equation and it is dominant in most engineering fields. Its results are applied without
question, which may not be appropriate for all conditions. Rouse (1943) showed that
the Eq. 16 is only valid for pipes with interior roughness similar to iron. Spiral or
plate fabricated pipes had substantially different functions. Statements of true
accuracy are rare, but based on his personal judgment that pipe roughness is difficult
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 42

to define, White (1994) has stated the Moody chart is only accurate to +15%. Thus, it
is surprising that the diagram has not been modified or replaced over the last 58 years.
Efforts have been made to streamline the procedure and eliminate the manual
use of graphs. This difficultly is responsible for the continued use of less accurate
empirical formulas such as the Hazen-Williams equation. While they have a limited
Reynolds number range, those irrational formulas are adequate for some design
problems. Therefore, the most notable advance in the application of the Darcy-
Weisbach equation has been the publication by Swamee and Jain (1976) of explicit
equations for pipe diameter, head loss and the discharge through a pipe, based on the
Colebrook-White equation. Swanee and Jain's formulas eliminate the last advantages
of the empirical pipe flow equations. Thus, because of its general accuracy and
complete range of application, the Darcy-Weisbach equation should be considered the
standard and the others should be left for the historians. Liou (1998) presented an
interesting discussion on the topic.
By necessity this was a brief survey of the historical development that focused
solely on the Darcy-Weisbach equation and the Darcy friction factor, f. Additional
theoretical background on f is presented by Schlichting (1968), while an excellent
historical overview that includes other pipe friction formulas is provided by Hager
(1994).

References

ASCE (1962). Nomenclature for hydraulics. ASCE, New York, 501 pages.

d'Aubuisson, J. F. (1834). Traité de l'hydraulique à l'usage des ingénieurs, Pitois-

Levrault, Paris (in French).

Blasius, H. (1913). Das Ähnlichkeitsgesetz bei Reibungsvorgängen in Flüssigkeiten,

Forschungs-Arbeit des Ingenieur-Wesens 131. (in German).

Colebrook, C. F. (1939) "Turbulent flow in pipes with particular reference to the

transition region between the smooth and rough pipe laws." Proc. Institution Civil
Engrs., 12, 393-422.

Colebrook, C. F. and White, C. M. (1937). "Experiments with fluid- friction in

roughened pipes." Proc. Royal Soc. London, 161, 367-381.

Darcy, H. (1857). Recherches expérimentales relatives au mouvement de l'eau dans

les tuyaux, Mallet-Bachelier, Paris. 268 pages and atlas (in French).

Fanning, J. T. (1877). A practical treatise on water-supply engineering, Van

Nostrand, New York, 619 pages.

Hagen, G. (1839). "Über die Bewegung des Wassers in engen zylindrischen Röhren."
Pogg. Ann., 46, 423-442 (in German).

Hager, W. H. (1994). "Die historische Entwicklung der Fliessformel." Schweizer

Ingenieur und Architekt, 112(9), 123-133. (in German)
 ENVIRONMENTAL AND WATER RESOURCES HISTORY 43

Hughes, H. J. and Stafford, A. T. (1911). A treatise on hydraulics, Macmillan, New

York. 501 pg and 5 plates.

Kármán, Th. von (1930). "Mechanische Aehnlichkeit und Turbulenz." Proc. Third
International Congress for Applied Mechanics, C. W. Oseen and W. Weibull eds.,
Stockholm. Vol. 1, 79-93 (in German).

Liou, C. P. (1998). "Limitations and proper use of the Hazen-Williams equation." J.

Hydraulic Engrg., ASCE 124(9), 951-954.

Moody, L. F. (1944). "Friction factors for pipe flow." Trans. ASME, 66:671-678.

Neville, John, (1853). Hydraulic tables, coefficients and formulae. John Wale,
London, 224 pages.

Nikuradse, J. (1933). "Strömungsgesetze in rauhen Rohren." Forschungs-Arbeit des

Ingenieur-Wesens 361 (in German).

Poiseuille, J. L. (1841). "Recherches expérimentales sur le mouvement des liquides

dans les tubes de très-petits diamètres." Comptes Rendus, Académie des Sciences,
Paris 12, 112 (in French).

Reynolds, O. (1883). "An experimental investigation of the circumstances which

determine whether the motion of water shall be direct or sinuous and of the law of
resistance in parallel channel." Phil. Trans. of the Royal Soc., 174:935-982.

Rouse, H. (1943). "Evaluation of boundary roughness." Proceedings Second

Hydraulics Conference, Univ. of Iowa Studies in Engrg., Bulletin No. 27.

Rouse, H. (1976). Hydraulics in the United States, 1776-1976, Iowa Institute of

Hydraulic Research, Univ. of Iowa, Iowa City, 238 pg.

Rouse, H. and Ince, S. (1957). History of Hydraulics, Iowa Institute of Hydraulic

Research, Univ. of Iowa, Iowa City, 269 pg.

Schlichting, H. (1968). Boundary-layer theory, Translated by J. Kestin, 6th ed,

McGraw-Hill, New York, 748 pg.

Swamee, P. K., and Jain, A. K. (1976). "Explicit equations for pipe-flow problems."
J. Hydraulics Division, ASCE, 102(5), 657-664.

Weisbach, J. (1845). Lehrbuch der Ingenieur- und Maschinen-Mechanik, Vol. 1.

Theoretische Mechanik, Vieweg und Sohn, Braunschweig. 535 pages (in German).

White, F. M. (1994). Fluid mechanics, 3rd ed., Mc-Graw Hill, New York. 736 pg.

View publication stats