http://faculty.washington.

edu/markbenj/CEE345/

CEE 345, Spring 2012

Part 2: Hydraulics and Open Channel Flow
Instructor:MarkBenjamin,335MoreHall;5437645,markbenj@uw.edu
Text:Munsonetal..,FundamentalsofFluidMechanics,6thed.

Pipe networks, pump selection, hydraulic transients

Flow dependence on depth & slope in rivers, streams,
culverts, storm and sanitary sewers

Weekly HW, due beginning of class; no late homework

One lab in HHL; no write-up, but data used in HW
Four Thursday sessions computer lab, HW, practice problems
Exam on hydraulics on Friday, May 18; open channel questions on final exam
 Our Plan Weeks 6 and 7
Review energy relationships in single pipes
Extend analysis to progressively more complex systems
Pipes in parallel or series
Pipe networks with known flow direction in each pipe
Interconnected pipe loops and reservoirs where flow
direction is not obvious
Consider key factors in selection of pumps to add
energy to fluid in a system
Consider some special cases of transients in pipe
systems cavitation and water hammer
 Overview of Pipe Networks
Pipe flow generally refers to fluid in pipes and
appurtenances flowing full and under pressure
Examples: Water distribution in homes, industry,
cities; irrigation
System components
Pipes
Valves
Bends
Pumps and turbines
Storage (often unpressurized, in reservoirs, tanks, etc.)
Energy Relationships in Pipe Systems
Energy equation between any two points:
E2 = E1 + hpump - hturb - hL , f

p2 V22 p1 V12
z2 + + = z1 + + + hpump - hturb - hL , f
g 2g g 2g
Analysis involves writing expressions for hL in each
pipe and for each link between pipes (valves,
expansions, contractions), relating velocities based
on continuity equation, and solving subject to
system constraints (Q, p, or V at specific points).
 Energy Losses in Piping Systems

Darcy-Weisbach equation for headlosses in

pipes (major headlosses):

l V2
hL = f
D 2g
Estimating f Graphically
 Trends in f
f declines with
increasing Re, e.g.,
increasing V at
fixed D.
In laminar region,
f = 64/Re

In turbulent region, for given e/D, f declines

more slowly than in laminar region;
eventually, the decline stops altogether.
 Mathematical Expressions for f
Colebrook and Haaland eqns yield good
estimates of f in turbulent flow
1 e D 2.71
= -2 log +
f 3.7 Re f

1
e D 6.9
1.11

= -1.8log
+
f 3.7 Re

Useful for calculations in spreadsheets or

special software for pipe flow analysis
 Understanding Headloss in Pipes
l V2
Darcy-Weisbach equation: hL = f
D 2g

For travel distance of one pipe diameterl (= D ):

Energy lost due to friction when
hL fluid travels a distance l equal to D
f = 2 =
V 2g KE of the fluid

f is the ratio of energy lost via friction (i.e., shear)

to the kinetic energy of the water when the
water travels a distance of one pipe diameter
 Example
Compare the velocity and pressure heads for typical
conditions in a street main:
V = 1.5 m/s; D = 0.5 m; p = 500 kPa
( 1.5 m/s )
2 2
V
= = 0.115 m
2 g 2 ( 9.8 m/s )2

p (1000 N/m2 )
( 500 kPa ) kPa
= 51.0 m
=
g 9800 N/m3

If f = 0.02, hL for each 0.5 m of pipe is 2% of the

velocity head, or 0.0023 m, corresponding to
0.0045% of the pressure head.
 Typical Pipe Flow Problems
Type I: Pipe properties (e, D, l) and V known, find
h L.
Determine f from Moody diagram or an
equivalent equation, and hL from the DW eqn

l V2
hL = f
D 2g
 Example
A 20-in-diameter galvanized pipe (e = 0.0005 ft) 2 miles long carries
4 cfs at 60oF. Find hL using (a) the Moody diagram and (b) the
Colebrook eqn.
Q 4 ft 3 /s
a) V= = = 1.83 ft/s
A p ( 1.67 ft ) 42

DV ( 1.67 ft ) ( 1.83 ft/s )

Re = = = 2.51x10 5

n 1.22x10 -5 ft 2 /s

e 0.0005 ft
= = 0.00030
D 1.67 ft

f = 0.017
 2 ( 5280 ft ) ( 1.83 ft/s )
2 2
l V
hL = f = ( 0.017 ) = 5.59 ft
D 2g 1.67 ft 2 ( 32.2 ft/s ) 2

1 e D 2.71
b) Colebrook eqn: = -2 log +
f 3.7 Re f

F G H
9 e/D 0.0003 0.0003
10 Re 251000 2.51E+05
11 f 0.03 0.03
12 LHS =1/SQRT(G11) 5.774
13 RHS =-2*LOG(G9/3.7 + 2.71/G10*G12) 7.687
14 LHS - RHS =G12-G13 -1.913
 e/D 0.0003
Re 2.51E+05
f 0.017422
LHS 7.576
RHS 7.576
LHS - RHS 2.55E-07

2 ( 5280 ft ) ( 1.83 ft/s )

2 2
l V
hL = f = ( 0.0174 ) = 5.72 ft
D 2g 1.67 ft 2 ( 32.2 ft/s ) 2
 Typical Pipe Flow Problems

Type II: Pipe properties (e, D, l) and hL known,

find V.
Guess V, determine f and hL as in Type I, iterate
until hL equals known value, or

Solve Colebrook and DW eqns simultaneously to

eliminate V, yielding:
Solving Type II Pipe Problems:
Iterative Approach

l V2
hL = f
D 2g
 0.25
f = 2


gD 3 -1 2
e/ D

log
0.317 2 hL +


n l 3.7

1/2
Rearranged D-W eqn: V = 2hL gD

fl

2 gDhL e D 2.51n l
V = -2 log
+

l 3.7 D 2 gDhL
 Example
For the pipe analyzed in the preceding example, what is the largest
flow rate allowable if the total frictional headloss must remain <8 ft?
 Example
For the pipe analyzed in the preceding example, what is the largest
flow rate allowable if the total frictional headloss must remain <8 ft?

2 gDhL e D 2.51n l
V = -2 log
+

l 3.7 D 2 gDhL

Substituting known values, V = 2.19 ft s

p ( 1.67 ft )
2
ft 3
Q = VA = ( 2.19 ft s ) = 4.80
4 s
 Typical Pipe Flow Problems
Type III: e, l, V, and hL known, find D.
Several approaches, all iterative; e.g., Guess D,
determine V as in Type II, iterate until V equals
known value

Example
What diameter galvanized pipe would be required in the preceding
examples if a flow rate of 10 cfs was needed, while keeping the total
frictional headloss at <8 ft?
Solving Type III Pipe Problems:
Iterative Graphical Approach

l V2
hL = f
D 2g
 Solving Type III Pipe Problems:
Iterative Analytical Approach
What diameter galvanized pipe would be required in the preceding
examples if a flow rate of 10 cfs was needed, while keeping the total
frictional headloss at <8 ft?

2 gDhL e D 2.51n l
V = -2 log
+

l 3.7 D 2 gDhL

2 gDhL
e D 2.51n l

p D
2
Q = -2 3.7 + D
log


l 2 gDhL 4

 2 gDhL
e D 2.51n l

p D 2
Q = -2 3.7 + D
log


l 2 gDhL
4

g 32.2 g 32.2
hL 8 hL 8
l 10560 l 10560
eps 0.0005 eps 0.0005
nu 1.22E-05 nu 1.22E-05
D_guess 2 D_guess 2.206594
LHS = Q 10 LHS = Q 10
RHS 7.72E+00 RHS 1.00E+01
LHS - RHS 2.28E+00 LHS - RHS -8.22E-07
Dependence of hL on D and V

l V2
hL = f
D 2g
 Dependence of hL on D and V
In laminar region:
For a given pipe
64 l V 2 32ln '
hL = = V = klamQ
DV n
2
D 2 g gD
In turbulent region, when f becomes constant:

For a given pipe

2
l V
hL = f full = k full Q 2
turb D 2g turb

Under typical water distribution conditions, hL in a

given pipe can be expressed as kQn with n slightly <2.
 Example
For the systems analyzed in the first two examples, what value of n
causes the data to fit the equation hL = kQn?

n
hL ,2 kQ Q2
n
= = 2
n
hL ,1 kQ Q1
1

hL ,2 Q2
log = n log
hL ,1 Q1

log ( hL ,2 hL,1 ) log ( 5.72 ft / 8 ft )

n= = = 1.84
log ( Q2 Q1 ) log ( 4 cfs 4.8 cfs )
 Alternative Equations for Flow - Headloss
Relationships in Turbulent Pipe Flow
Hazen-Williams equation widely used for hL
as function of flow parameters for turbulent
flow at typical velocities in water pipes:
0.54
hL Aflow p D 2 4 D R
V = 0.849C HW R 0.63
Rh =
= =
h
l Pwetted p D 4 2

Q1.85 1
hL = 10.7l 4.87 1.85 Coefficients shown are for SI units; for BG
units, replace 0.849 by 1.318 and 10.7 by 4.73.
D CHW
 Comparison of Equations for Transitional and
Turbulent Curves on the Moody Diagram
D-W H-W* Manning*

hf 1 0.849CHW Rh0.63 S 0.54 1 0.67 0.50

2 gD Rh S
V l f n

= 2g D 0.50
S 0.50
f -0.50 = 0.354 D 0.63 S 0.54CHW = 0.397D 0.67 S 0.50 1
n

Q p 2 g 2.50 0.50 -0.50 2.63 0.54 1

D S f 0.278 D S CHW 0.312D 2.67 S 0.50
4 n

hL (=S*l)
8 2 l l 1 l 1
Q 5 f 10.7Q1.85 10.3Q 2

p g
2
D D 4.87 1.85
CHW D 5.33 n2

*
CoefficientsshownareforSIunits(Vinm/s,andDandRhinm);forBGunits
(ft/sandft),replace0.849by1.318;0.354by0.550;0.278by0.432;10.7by
 Energy Losses in Bends, Valves, and
Other Transitions (Minor Losses)
Minor headlosses generally significant when pipe
sections are short (e.g., household, not pipeline)
Caused by turbulence associated with flow transition;
therefore, mitigated by modifications that smooth
flow patterns
Generally much greater for expansions than for
contractions
V2
Often expressed as multiple of velocity head: hL = K minor
2g
K is the ratio of energy lost via friction in the device of
interest to the kinetic energy of the water (upstream
or downstream, depending on geometric details)
Energy Losses in Contractions

V22
hc = kc
2g

All images from Finemore & Franzini (10e, 2002)

 Energy Losses in Expansions

( V - Vc )
2

hx =
2g

V 2 Vc2 V 2 Vc2
hx,discharge = - hx,discharge = -
2g 2g 2g 2g

All images from Finemore & Franzini (10e, 2002)

 Energy Losses in Expansions
Conical diffuser

( V1 - V2 )
2

hcone = kcone
2g

k, rough
k, smooth

All images from Finemore & Franzini (10e, 2002)

Energy Losses in Pipe Fittings and Bends

V2
hb = kb
2g

All images from Finemore & Franzini (10e, 2002)

 Example
A 5-in-diameter pipe with an estimated f of 0.033 is 110 feet long
and connects two reservoirs whose surface elevations differ by 12
feet. The pipe entrance is flushed, and the discharge is submerged.
(a) Compute the flow rate.
(b) How much would the flow rate change if the last 10 ft of the
pipe were replaced with a smooth conical diffuser with a cone
angle of 10o?

5" = 0.417 ft
 l V2 V2 l V2
( a) hL ,tot = hL , pipe + hL ,minor = f + ( 0.5 + 1) = f + 1.5
D 2g 2g D 2 g

2 ghL ,tot 2 ( 32.2 ft/s 2 ) ( 12 ft )

V= = = 8.70 ft/s
fl
+ 1.5 ( 0.033) ( 110 ft ) + 1.5
D 0.417 ft

p ( 0.417 ft )
2

Q = VA = ( 8.70 ft/s ) = 1.19 ft 3 /s

4
 ( b) hL ,tot = hL, pipe + hL ,entrance + hL ,cone + hL ,exit

( V1 - V2 )
2 2 2
l1 V V V22
= f + kentrance
1
+ kcone
1
+ kexit
D1 2 g 2g 2g 2g

D2 = D1 + 2 Lcone tan 5o = 0.417 ft + 2 ( 10 ft ) ( 0.0875 ) = 2.17 ft

2 2
V2 D1 0.417 ft
= = = 0.0370
V1 D2 2.17 ft

From graph, for a smooth, 10o cone, kcone = 0.175

 ( V1 - V2 )
2 2 2
l1 V V V22
hL ,tot = f 1
+ kentrance + kcone
1
+ kexit
D1 2 g 2g 2g 2g

100 ft V12 V12

= ( 0.033) + 0.5
0.417 ft 2 g 2g

( V1 - 0.037V1 ) ( 0.037V1 )
2 2

+ ( 0.175 ) + 1.0
2g 2g

V1 = 9.49 ft/s

p ( 0.417 ft )
2

Q = V1 A = ( 9.49 ft/s ) = 1.29 ft 3 /s

4