Está en la página 1de 49

5/19/2008

1
Pipe Flow
Pipe Systems and Design PipeSystemsandDesign
PipeNetworkAnalysis
5/19/2008 1
Pipe Systems and Design
Twomajorconcerns:
Si th i ( f h t & t bl ) Sizethepipe(e.g.fromcharts&tables)
Determinetheflowpressurerelationship
Toanalysethesystem,e.g.tofindoutpumppressure
Byusingmanualorcomputerbasedmethods
Calculationsforpipelinesorpipenetworks
Can be er complicated for branches & loops Canbeverycomplicatedforbranches&loops
Basicparameters:pipediameter,length,friction
factor,roughness,velocity,pressuredrop
5/19/2008 2
5/19/2008
2
Pipe Systems and Design
Pipenetworkanalysis
Physicalfeaturesareknown
Solutionprocesstrytodetermineflow&pressure
ateverynode
Pipenetworkdesign
Variables are unknown Variablesareunknown
Trytosolve&selectpipediameters,pumps,
valves,etc.
5/19/2008 3
Pipe Systems and Design
Basic equations Basicequations
DarcyWeisbachEquation (forfullydeveloped
flowsofallNewtonianfluids)
C l b k Whit E ti (f t iti i )

=
g
V
D
L
f h
g
V
D
L
f p
2
or
2
2 2

ColebrookWhiteEquation (fortransitionregion):
*Theequationisimplicitinf (appearsonbothsides),so
iterationsarerequiredtosolveforf.

+ + =
f D
D
f ) / Re(
3 . 9
1 log 2 ) / log( 2 14 . 1
1

5/19/2008 4
5/19/2008
3
Pipe Systems and Design
Basicequations(contd)
HazenWilliamsEquation (alternativetoDarcy
Weisbachformula;empirical)
g) (
1
819 . 6
1.167 852 . 1

=
D C
V
L p
C =roughnessfactor(typically,C =150forplasticor
copperpipe,C =140fornewsteelpipe,C <100for
badlycorrodedorveryroughpipe)
5/19/2008 5
PipeFlow
5/19/2008 6
5/19/2008
4
5/19/2008 7
5/19/2008 8
5/19/2008
5
5/19/2008 9
5/19/2008 10
5/19/2008
6
5/19/2008 11
5/19/2008 12
5/19/2008
7
5/19/2008 13
5/19/2008 14
5/19/2008
8
5/19/2008 15
5/19/2008 16
5/19/2008
9
5/19/2008 17
5/19/2008 18
5/19/2008
10
5/19/2008 19
5/19/2008 20
5/19/2008
11
5/19/2008 21
5/19/2008 22
5/19/2008
12
5/19/2008 23
5/19/2008 24
5/19/2008
13
5/19/2008 25
5/19/2008 26
5/19/2008
14
5/19/2008 27
5/19/2008 28
5/19/2008
15
5/19/2008 29
5/19/2008 30
5/19/2008
16
5/19/2008 31
5/19/2008 32
5/19/2008
17
Table 6.1: Typical pipe roughness
(Reference: White, 1999)
Material Roughness, k (mm)
Glass smooth
Brass new 0 002 Brass, new 0.002
Concrete
Smoothed
Rough
0.04
2.0
Iron
Cast, new
0.26
0.15
Galvanised, new
Wrought, new
0.046
Steel
Commercial, new
Riveted
0.046
3
5/19/2008 33
5/19/2008 34
5/19/2008
18
5/19/2008 35
5/19/2008 36
5/19/2008
19
5/19/2008 37
5/19/2008 38
5/19/2008
20
5/19/2008 39
5/19/2008 40
5/19/2008
21
5/19/2008 41
5/19/2008 42
5/19/2008
22
5/19/2008 43
5/19/2008 44
5/19/2008
23
5/19/2008 45
5/19/2008 46
5/19/2008
24
5/19/2008 47
5/19/2008 48
5/19/2008
25
5/19/2008 49
MinorLosses
Inadditiontoheadlossduetofriction,there
are always other head losses due to pipe arealwaysotherheadlossesduetopipe
expansionsandcontractions,bends,valves,
andotherpipefittings.Theselossesare
usuallyknownasminorlosses(h
Lm
).
In case of a long pipeline the minor losses Incaseofalongpipeline,theminorlosses
maybenegligiblecomparedtothefriction
losses,however,inthecaseofshortpipelines,
theircontributionmaybesignificant.
5/19/2008 50
5/19/2008
26
Losses due to pipe fittings
2
V
K h
2
Lm
=
Type K
where h
Lm
= minor loss
K = minor loss coefficient
V = mean flow velocity
g 2
Lm
Exit (pipe to tank) 1.0
Entrance (tank to
pipe)
0.5
90 elbow 0.9
45 elbow 0.4
Table 6.2: Typical K values
T-junction 1.8
Gate valve 0.25 - 25
5/19/2008 51
Loss Coefficients
Use this table to find loss coefficients:
5/19/2008 52
5/19/2008
27
5/19/2008 53
5/19/2008 54
5/19/2008
28
5/19/2008 55
5/19/2008 56
5/19/2008
29
5/19/2008 57
5/19/2008 58
5/19/2008
30
5/19/2008 59
5/19/2008 60
5/19/2008
31
Example
Calculate the head added by the pump when the water
system shown below carries a discharge of 0.27 m
3
/s. If
the efficiency of the pump is 80% calculate the power the efficiency of the pump is 80%, calculate the power
input required by the pump to maintain the flow.
61 5/19/2008
Solution:
Applying Bernoulli equation between section 1 and 2

(1)
+ + +

= + + +

2 1 L
2
2
2
2
p
2
1
1
1
H
g 2
V
z
g
P
H
g 2
V
z
g
P
P
1
= P
2
= P
atm
= 0 (atm) and V
1
=V
2
0
Thus equation (1) reduces to:
(2)
H
L1-2
= h
f
+ h
entrance
+ h
bend
+ h
exit
+ =
2 1 L 1 2 p
H z z H
From (2):
g 2
V
4 . 39
1 4 . 0 5 . 0
4 . 0
1000
x 015 . 0
g 2
V
H
2
2
2 1 L
=

+ + + =

81 . 9 x 2
V
4 . 39 200 230 H
2
p
+ =
5/19/2008 62
5/19/2008
32
The velocity can be calculated using the
continuity equation:
Thus, the head added by the pump: H
p
= 39.3 m
( )
s / m 15 . 2
2 / 4 . 0
27 . 0
A
Q
V
2
=

= =
in
p
p
P
gQH
==
P
in
= 130 117 Watt 130 kW.
8 . 0
3 . 39 x 27 . 0 x 81 . 9 x 1000
gQH
P
p
p
in
=

=
5/19/2008 63
Pipes in Series
When two or more pipes
of different diameters or
roughness are connected
in such a way that the in such a way that the
fluid follows a single flow
path throughout the
system, the system
represents a series
pipeline.
In a series pipeline the In a series pipeline the
total energy loss is the
sum of the individual
minor losses and all pipe
friction losses.
Figure 6.11: Pipelines in series
5/19/2008 64
5/19/2008
33
Referring to Figure 6.11, the Bernoulli equation can be written
between points 1 and 2 as follows;
(6.18)
where P/g = pressure head
z = elevation head
2 1 L
2
2
2
2
2
1
1
1
H
g 2
V
z
g
P
g 2
V
z
g
P

+ + +

= + +

z = elevation head
V
2
/2g = velocity head
H
L1-2
= total energy lost between point 1 and 2
Realizing that P
1
=P
2
=P
atm
, and V
1
=V
2
, then equation (6.14) reduces to
z
1
-z
2
= H
L1-2
Or we can say that the different of reservoir water level is equivalent to the
total head losses in the system.
The total head losses are a combination of the all the friction losses and the
sum of the individual minor losses sum of the individual minor losses.
H
L1-2
= h
fa
+ h
fb
+ h
entrance
+ h
valve
+ h
expansion
+ h
exit
.
Since the same discharge passes through all the pipes, the continuity
equation can be written as;
Q
1
= Q
2
5/19/2008 65
Pipes in Parallel
A combination of
two or more pipes
connected between connected between
two points so that
the discharge
divides at the first
junction and rejoins
at the next is
known as pipes in known as pipes in
parallel. Here the
head loss between
the two junctions is
the same for all
pipes.
Figure 6.12: Pipelines in parallel
5/19/2008 66
5/19/2008
34
Applying the continuity equation to the system;
Q1 = Q + Q
b
= Q (6 19) Q1 = Q
a
+ Q
b
= Q
2
(6.19)
The energy equation between point 1 and 2 can be written
as;

The head losses throughout the system are given by;


H
L1-2
=h
La
= h
Lb
(6.20)
L
2
2
2
2
2
1
1
1
H
g 2
V
z
g
P
g 2
V
z
g
P
+ + +

= + +

L1 2 La Lb
( )
Equations (6.19) and (6.20) are the governing relationships
for parallel pipe line systems. The system automatically
adjusts the flow in each branch until the total system flow
satisfies these equations.
5/19/2008 67
Pipe Network
A water distribution system consists of complex interconnected pipes,
service reservoirs and/or pumps, which deliver water from the treatment
plant to the consumer.
Water demand is highly variable whereas supply is normally constant Water demand is highly variable, whereas supply is normally constant.
Thus, the distribution system must include storage elements, and must be
capable of flexible operation.
Pipe network analysis involves the determination of the pipe flow rates
and pressure heads at the outflows points of the network. The flow rate
and pressure heads must satisfy the continuity and energy equations.
The earliest systematic method of network analysis (Hardy-Cross Method)
is known as the head balance or closed loop method. This method is
applicable to system in which pipes form closed loops. The outflows from applicable to system in which pipes form closed loops. The outflows from
the system are generally assumed to occur at the nodes junction.
For a given pipe system with known outflows, the Hardy-Cross method is
an iterative procedure based on initially iterated flows in the pipes. At
each junction these flows must satisfy the continuity criterion, i.e. the
algebraic sum of the flow rates in the pipe meeting at a junction,
together with any external flows is zero.
5/19/2008 68
5/19/2008
35
Hardy Cross Methods: Step 1
By careful By careful
inspection we may
assume the most
reasonable
distribution of flows
in the pipe network p p
and make the first
guess of the flow
pattern.
5/19/2008 69
Step 1: Sign Convention in Pipe Network
Enter flows at nodes as positive for inflows and
negative for outflows negative for outflows.
Inflows plus outflows must sum to 0.
Enter one pressure in the system and all other
pressures are computed.
You do not need to use all the pipes or nodes.
Enter a diameter of 0.0 if a pipe does not exist.
If d i d d ll id b If a node is surrounded on all sides by non-
existent pipes, the node's flow must be entered as
0.0.
5/19/2008 70
5/19/2008
36
Step 1: How to Handle Minor Losses
Minor losses such as pipe elbows bends and Minor losses such as pipe elbows, bends, and
valves may be included by using the equivalent
length of pipe method (Mays, 1999).
Equivalent length (L
eq
) may be computed using
the following calculator which uses the formula
L
eq
=KD/f.
5/19/2008 71
Step 1: Summary of Minor Losses Calculation
Enter node flows, elevations, pressure.
Select Darcy Weisbach (Moody diagram) or
Hazen Williams friction losses.
Include minor losses by equivalent length of pipe.
5/19/2008 72
5/19/2008
37
Minor Losses
If you go by a rigorous method, f is the Darcy-
Weisbach friction factor for the pipe containing the Weisbach friction factor for the pipe containing the
fitting, and cannot be known with certainty until after
the pipe network program is run.
However, since you need to know f ahead of time, a
reasonable value to use is f=0.02, which is the default
value value.
5/19/2008 73
Hardy Cross Methods: Step 2
Write head loss condition for each pipe in the form:
H ( h ) K Q
n
H
i
(or h
L
) = K Q
n
n=2.0 for Darcy Weisbach losses
n=1.85 for Hazen Williams losses.
5/19/2008 74
5/19/2008
38
Friction Losses, H
Hazen Williams equation
Darcy Weisbach equation
5/19/2008 75
Hardy Cross Methods: Step 3
Compute the algebraic sum of the head losses
around each elementary loop,
H
i
(or h
L,i
) = K
i
Q
i
n
Consider losses from clockwise flows as positive,
counterclockwise negative.
Be careful about the common pipe sections shared
by two adjacent loops. by two adjacent loops.
5/19/2008 76
5/19/2008
39
Hardy Cross Methods: Step 4
Adjust the flow in each loop by a correction Q or
to balance the head in that loop and give to balance the head in that loop and give
K Q
n
=0
The heart of this method lies in the following
determination of Q . For any pipe, we may write:
Q=Q
0
+ Q
Where Q
0
is the assumed discharge and Q is the
corrected discharge.
5/19/2008 77
Hardy Cross Methods: Step 4
Binomial series gives: Binomial series gives:
If Q is small compared with Q
0
, we may neglect the
terms of the binomial series after the second one:
.....) ( ) (
1
0 0 0
+ + = + = =

Q nQ Q K Q Q K KQ H
n n n n
1
0 0

+ =
n n
QKnQ KQ H
5/19/2008 78
5/19/2008
40
Hardy Cross Methods: Step 4
For a loop H
i
(or h
L i
) = K
i
Q
i
n
=0: For a loop, H
i
(or h
L, i
) K
i
Q
i
0:
We may solve this equation for Q :
0
1
0 0
= + = =

n n
L
QKnQ KQ h H

= = =

0
1
0
1
0 0
1
0
0
/ Q h n
h
KQ n
Q KQ
KnQ
KQ
Q
L
L
n
n
n
n
0 0 0
/ Q h n Q n Q
L
#Sum the numerator algebraically with due account of each sign
#Sum the denominator arithmetically
5/19/2008 79
Hardy Cross Methods: Step 4
It indicates that clockwise flows may be
id d d i l k i l d considered as producing clockwise losses, and
counterclockwise flows, counterclockwise
losses.
This means that the minus sign is assigned to
all counterclockwise conditions in a loop, all counterclockwise conditions in a loop,
namely flow Q and lost head h
L
.
5/19/2008 80
5/19/2008
41
Friction Losses, H
The calculation procedure gives you a choice of
computing friction losses H using the Darcy computing friction losses H using the Darcy-
Weisbach (DW) or the Hazen-Williams
(HW) method.
The DW method can be used for any liquid or gas
while the HW method can only be used for water at
temperatures typical of municipal water supply temperatures typical of municipal water supply
systems.
n=2.0 for Darcy Weisbach losses
n=1.85 for Hazen Williams losses.
5/19/2008 81
Iteration
After we have given each loop a first correction, te we ave g ve eac oop a st co ect o ,
the losses will still not balance, we need to repeat
the procedure, arriving at a second correction, and
so on, until the corrections become negligible.
5/19/2008 82
5/19/2008
42
Step 5: Pressure Computation
After computing flowrate Q in each pipe and loss After computing flowrate Q in each pipe and loss
H in each pipe and using the input node elevations
Z and known pressure at one node, pressure P at
each node is computed around the network:
P
j
= (Z
i
- Z
j
- H
pipe
) + P
i
node j is down-gradient from node i.
= fluid density [F/L
3
].
5/19/2008 83
Assigning clockwise flows and their associated head losses are
positive, the procedure is as follows:
Assume values of Q to satisfy Q = 0.
Calculate H
L
from Q using H
L
= K
1
Q
2
.
If H
L
= 0, then the solution is correct.
If H
L
0, then apply a correction factor, Q, to all Q and repeat from
step (2).
For practical purposes, the calculation is usually terminated when H
L
<
0.01 m or Q < 1 L/s.
A reasonably efficient value of Q for rapid convergence is given by;
(6.21)

=
Q
H
2
H
Q
L
L
5/19/2008 84
5/19/2008
43
Example
A pipe 6-cm in diameter, 1000m long and with =
0.018 is connected in parallel between two points M
and N with another pipe 8-cm in diameter, 800-m long
and having = 0.020. A total discharge of 20 L/s g g
enters the parallel pipe through division at A and
rejoins at B. Estimate the discharge in each of the
pipe.
5/19/2008 85
Solution:
Continuity: Q = Q
1
+ Q
2
(1)
2
2
2 2
1
2
V ) 08 . 0 (
4
V ) 06 . 0 (
4
02 . 0

+

=
Pipes in parallel: h
f1
= h
f2
S b tit t (2) i t (1)
4 4
074 . 7 V 778 . 1 V
2 1
= +
V
08 . 0
800 x 020 . 0
V
06 . 0
1000 x 018 . 0
gD 2
V L
gD 2
V L
2
2
2
1
2
2
2 2
2
1
2
1 1
1
=
=
Substitute (2) into (1)
0.8165V
2
+ 1.778 V
2
= 7.074
V
2
= 2.73 m/s
) 2 ( V 8165 . 0 V
2 1
=
5/19/2008 86
5/19/2008
44
73 . 2 x ) 08 . 0 (
4
V A Q
2
2 2 2

= =
Q
2
= 0.0137 m3/s
From (2):
V
1
= 0.8165 V
2
= 0.8165x2.73 = 2.23 m/s
Q
1
= 0.0063 m3/s Q
1
Recheck the answer:
Q
1
+ Q
2
= Q
0.0063 + 0.0137 = 0.020
(same as given Q OK!)
5/19/2008 87
Example
For the square loop shown, find the discharge
in all the pipes. All pipes are 1 km long and
300 mm in diameter, with a friction factor of
0.0163. Assume that minor losses can be
neglected.
5/19/2008 88
5/19/2008
45
Foranalysis,basiccriteriashouldbe
followed:
Thetotalflowenteringeachpoint(node)must
equal to the total flow leaving that joint (node) equaltothetotalflowleavingthatjoint(node).
Q
in
=Q
out
Theflowinapipemustfollowthepipesfriction
lawforthepipe.
h
f
=kQ
n
niscoefficientdependonequationused
Thealgebrasumoftheheadlossinthepipe
networkmustbezero.
h
f
=0
5/19/2008 89
LoopMethod
Steps:
1. Assumeflowsforeachindividualpipeinthenetworkbut
Q=0 ateachnode.
2. CalculatestheheadlossthrougheachpipewithDarcyor
Hazen William HazenWilliam.
3. Namethepipenetworktoseveralloop.
4. Findthealgebrasumoftheheadlossesineachloopinthe
pipenetwork.
5. Calculate(nhf/Q) foreachloop(ignorepositiveor
negative).
6. GetQ valueforcorrection
Q= hf
(nhf/Q)
7. Calculatethenewflowrate
Qnew =Q+Q
8. RepeatthisstepuntilQ=0 .
5/19/2008 90
5/19/2008
46
NodeMethod
Thismethodusedwhenpressureheadatcertainnode/point
knownbutflowrateandpressureheadatseveral(different)
node/pointshouldbeobtained.Thistypeofproblemgenerally
requiresatrialanderrorsolution.
St Step
1. Assumeanelevationofthehydraulicgradeline(head)atnodeJ.
2. CalculatetheQvalueineachpipeusingHJ (assumebefore).Use
DarcyWeisbachorHazenWilliamformula.
3. CalculateQin =Qout fromnodeJ.
4. CalculateQ/nhfforeachpipethenfind(Q/nhf).
/ / 5. Calculatehf= (Q/(nhf/Q)) forcorrection.
(Darcy n=2 HazenWilliam n=1.85)
6. CalculateHJnew =HJ hf
7. Repeatstep2 5untilhf0
5/19/2008 91
Solution:
Assume values of Q to satisfy continuity equations all
at nodes.
Th h d l i l l t d i H K
1
Q
2
The head loss is calculated using; H
L
= K
1
Q
2
H
L
= h
f
+ h
Lm
But minor losses can be neglected: h
Lm
= 0
Thus H
L
= h
f
Head loss can be calculated using the Darcy-Weisbach
equation equation
g 2
V
D
L
h
2
f
=
5/19/2008 92
5/19/2008
47
Q
x 77 . 2
A
Q
77 . 2 H
81 . 9 x 2
V
x
3 . 0
1000
x 0163 . 0 H
g 2
V
D
L
h H
2
2
2
2
2
L
2
L
2
f L


= =
=
= =
First trial
554 ' K
Q ' K H
Q 554 H
3 . 0 x
4
A
2
L
2
L
2
=
=
=


Pipe Q (L/s) H
L
(m) H
L
/Q
Since H
L
> 0.01 m, then correction has to be applied.
AB 60 2.0 0.033
BC 40 0.886 0.0222
CD 0 0 0
AD -40 -0.886 0.0222
2.00 0.0774
5/19/2008 93
Second trial
s / L 92 . 12
0774 . 0 x 2
2
Q
H
2
H
Q
L
L
= =

=
Pipe Q (L/s) H (m) H /Q
Since H
L
0.01 m, then it is OK.
Thus, the discharge in each pipe is as follows (to the nearest integer).
Pipe Q (L/s) H
L
(m) H
L
/Q
AB 47.08 1.23 0.0261
BC 27.08 0.407 0.015
CD -12.92 -0.092 0.007
AD -52.92 -1.555 0.0294
-0.0107 0.07775
Pipe Discharge (L/s)
AB 47
BC 27
CD -13
AD -53
5/19/2008 94
5/19/2008
48
5/19/2008 95
5/19/2008 96
5/19/2008
49
Thank You
5/19/2008 97

También podría gustarte