How To Design Gravity Flow Water Sys
How To Design Gravity Flow Water Sys
Santiago Arnalich
How to design a Gravity Flow Water System
ISBN: 978-84-614-3744-3
All rights reserved. This manual can be photocopied for personal use if your economic situation
does not allow you to buy a copy. Otherwise, please consider buying one to support this kind of
initiative.
If you want to reproduce part of the contents of this book, contact us at the following email:
[email protected].
Cover photos: Rope and bucket. Illustration page 5: Arantxa Osés Alvarez
DISCLAIMER: The information contained in this book has been obtained from credible and
internationally respected sources. However, neither Arnalich - Water and Habitat nor the
author can guarantee the precision of the information published here and are not responsible
for any errors, omissions, or damage caused by the use of this information. It is understood
that the information published herein is without a specific purpose and under no
circumstances intends to provide professional engineering services. If these services are
W. H. Auden
1. Introduction 1
2. Piping 5
APPENDICES 91
I have a certain number of people who need water: What do I need to build?
I have a design proposal for a system: Is it going to work?
It’s common in cooperation projects to underestimate gravity flow systems and think
that they’ll work just on their own. Saying that a gravity flow system is going to work
simply because the water is flowing downhill, is like saying that a plane that’s losing
altitude is just going to land on its own: you’d better not be onboard! In both cases,
you could be in for a crash landing.
1. 2 SOME LIMITATIONS
This book is not intended to be everything to everyone. It has some necessary
limitations:
• It doesn’t show you how to choose specific materials for a system, install
piping, or trace out an optimum route for a pipeline. For this, have a look at
“Gravity Flow Water Supply” (S. Arnalich 2008).
8 CHAPTER 1. Introduction
• It also doesn’t show you how to use modelling software, even though it’s a
pretty good introduction. If you want to learn about this, try “Epanet in Aid;
How to calculate water networks by computer” (S. Arnalich 2007) and
“Epanet in Aid; 44 progressive exercises to calculate water networks by
computer” (S. Arnalich 2008).
You can read these books for free online, or pay for downloads or
printed copies here: http://www.arnalich.com/en/libros.html.
• It doesn’t deal with aqueducts or canals, which also work with gravity.
• It focuses on pipe sizing. In gravity flow systems there are other
components that require calculations, such as break-pressure tanks,
reservoir and sedimentation tanks. These are dealt with in “Gravity flow
Water Supply” (S. Arnalich 2008).
If this is your first gravity flow water system and you want to begin designing, or check
if a design will work, keep on reading: this book’s for you!
If you’re disciplined when working out the units, then you’ll spot most of these
mistakes before you have a nervous breakdown. Have a look at this example, a
3
conversion of m /h into l/s:
3 3 3 3 3
14 m /h = 14 m /h * 1m /1,000l * 3,600s/1h = 14*3,600/1,000 m *m *s/h*l*h
6 2
= 50.4 l*m / h *s
6 2
l*m / h *s?! If you’re like me and don’t have a clue what this unit of flow is,
something, somewhere, went a little bit wrong…
3 3 3 3 3
14 m /h = 14 m /h * 1,000l/1m * 1h/3,600s = 14*1,000/3,600 m *l*h/h*m *s
= 3.88 l/s
NOTE: Multiplying by 1h/3600s is the same as multiplying 1/1, as 1 hour and 3600
seconds are the same thing. If it’s easier, think of this: “1 hour has 3600 seconds.”
The result is a simple change of units.
Do what I say and not what I do! To make reading the exercises easier, I
sometimes leave out the units.
1. When water isn’t flowing in a pipe, the pressure is simply the difference in
height between the tap at the bottom and the surface of the water where it
enters the pipe at the top. It makes no difference the route taken by the pipe
on the way down.
12 CHAPTER 2. Piping
2. The pressure can be measured in metres (of column of water). 10 metres are
2
equivalent to 1 bar or 1 kg/cm .
3. When water flows through a pipe, some energy is lost due to the friction of
the water on the walls of the pipe.
4. Loss of pressure due to this friction can be expressed in metres per kilometre
of pipeline (m/km).
5. The smaller the pipe diameter, the more pressure is lost due to frictional
losses, as if the pipe was being “strangled.” (Make sure you read this point
1
carefully and understand it well !)
6. The material of which a pipe is made affects the amount of frictional loss.
Each kind of material has varying roughness (as well as a varying internal
diameter.)
7. A system will work if the pressure is between 10 and 30 metres at the taps,
and as long as the pressure at any point along the way is no less than 10
metres.
Designing a system is all about choosing pipes of the right diameter to make sure you
meet that requirement. By selecting different pipe diameters you can control the
amount of pressure in the system. That way, users aren’t exasperated by a miserable
flow coming out of a tap (due to insufficient pressure), nor are they showered by a
high-power water jet (due to excessive pressure).
1. Find the table for the given pipe material, nominal pressure and diameter.
For example, PVC 90mm at 10 bar (PN10).
1
Don’t get confused by the fact that if you squeeze a hose the water comes out faster. If
someone puts their hands around your neck and squeezes hard you get less blood to the head
and not more! This is one of the most frequent misunderstandings in cooperation projects.
2. Look for the required flow rate and read off the head loss. For example, a
flow rate of 1.25 l/s produces a 1 m/km head loss.
• Pipe is specified commercially with the internal diameter for metal pipe, and
the external diameter for plastic pipe (PVC or HDPE). A 25mm plastic pipe
and a 25mm metal pipe have different internal diameters.
Abbreviations:
E, Energy, or head. This is the energy in the form of pressure contained in a system
where there are no frictional losses.
J, Head loss factor. The energy lost as water travels through the pipe, per kilometre.
H, Head loss. The pressure lost as water travels through the pipe.
P, Pressure. The remaining pressure after frictional losses.
H = 15 m/km * 5 km = 75m.
What is the head loss in a PVC pipe of 10 bar, where the first 500m
are 110mm, and the final 300m are 90mm, at a flow of 3.6 l/s?
1. Looking at the PVC tables for 10 bar, we get 2.25 m/km for 110mm pipe: J110
= 2.25.
2. For 3.594 l/s, J90= 6.5 m/km (the head loss for 90mm pipe is 6.5 m/km).
3. The total head loss in a pipeline is the sum of the head loss in each
respective reach:
1. The static pressure with the tap shut is the difference in elevation:
P = 63m - 41m = 22m.
2. For 26.99 l/s, J200 = 4.75 m/km. The head loss is:
3. The pressure with the water flowing is the static pressure less the head loss:
Pay close attention to this exercise! It´s the first one in which you choose a pipe
diameter. Once you’ve gotten hold of it you´re well on your way.
2. The pressure that needs to be burnt off for the water to come out with 13m of
pressure is:
P = 17m – 13m = 4m
3. Those 4m are going to be burnt off over a distance of 4km, which means the
head loss is:
J = 4m / 4km = 1 m/km.
4. Look at the tables and find which diameter of PVC pipe has a head loss of
1m/km at a flow of 1.25 l/s. You´ll see that it´s 90mm.
2. 3 INTERPOLATING VALUES
Up to now you´ve been working with ready-made values, but it´s more likely that when
you look at the tables, the figure you need is somewhere in between the ones given.
To find the answer, you need to do what´s called a lineal interpolation. The generic
equation looks a bit muddled, but it´s fairly simple:
J x − J inf Q − Qinf
= x
J sup − J inf Qsup − Qinf
As you are going to be interpolating for even the simplest systems, draw up a
spreadsheet with Excel or use this one:
www.arnalich.com/dwnl/interpolator.xls
1. Look at the PVC tables. There´s no value that´s sufficiently close: it´s either
5.723 l/s or 6.71 l/s.
2. Using the formula you can work out the head loss:
J x − J inf Q − Qinf
= x
J sup − J inf Qsup − Qinf
J x − 15 6 − 5.723
=
20 − 15 6.71− 5.723
Jx = 15 + (20-15) (6 – 5.723)/(6.71-5.723) = 16.4 m/km
Using the spreadsheet calculator provided, you should get the same result:
The solution is to combine pipe sizes until you get the desired head loss. To work out
the required pipe lengths use the following reasoning:
The head loss for x meters in pipe A together with the head loss for x meters in pipe
B is the total required head loss.
If x is the length in meters of pipe A that needs to be installed, the length of pipe B is
d-x.
It´s very important to remember that the larger pipe diameter is almost always
installed first, such that the water travels through progressively smaller pipe sizes.
In general it makes no sense to use smaller pipes followed by larger ones. If new
connections are made in the future or the demand increases, a small diameter pipe
placed too soon in the pipeline will strangle the rest of the system.
Which HDPE pipe would you install for 20 meters of head loss
over 2 kilometres at a flow of 2 l/s?
J63= 15 m/km. For 2 km you lose 15m/km * 2km = 30m. Too much!
J90= 2.5 m/km. For 2 km you lose 2.5m/km * 2km = 5m. Too little!
2. As you may have suspected, the solution lies in combining pipe sizes:
Ja*x + Jb(d-x) = D
J63*x + J90(d-x) = D
15x + 5 – 2.5x = 20
When you get more familiar with the process, you´ll end up using trial and
error with the formulas in the spreadsheet, rather than using these equations.
Either way, take it one step at a time.
Here, the important thing is the last part of the last principle we looked at. Take a
moment to look it over again and refresh your memory.
“[The pressure]… at no point along the pipeline should be less than 10 meters.”
For practical purposes, if you manage to get over the main obstacles by over 10
meters, you don´t have to worry about the rest of the pipe line:
In this profile, there are 2 critical points, the high points A and B at 400 and 800m.
We´re going to deal with them in the following exercise.
Which HDPE pipe would you use to get to 1,400 meters with 2 bar
of residual pressure at a flow of 4 l/s?
1. Look for the critical points. Note that B is not really that critical, since the
elevation 10 meters above is 30m, and the elevation at which we want to
deliver the water is 10m + 20m = 30m (remember, 2 bar is equivalent to
20m). The exit point and B are therefore really at the same height.
2. Let´s work out the first reach of pipe to get to A with at least 10m of pressure.
In other words, with 46m + 10m. The maximum head loss will be:
3. In the tables, J90= 9 m/km. The head loss in that reach is:
4. For the second reach, the start point is 60m – 3.6m = 56.4m of energy. Since
the end point is 30m, the required head loss is:
5. In the tables, J90= 9 m/km. To find the value of J63 we need to interpolate:
J63*x + J90(d-x) = D
So we´ll use 400m + 584m = 984m of 90mm HDPE, PN 10, followed by 416m of
63mm HDPE, PN 10 (in that order!)
If you are tempted to run off and design pipelines now, be patient and
finish this chapter. There are still some things that have deliberately not
been dealt with to keep the exercises simple, and which could scupper
your efforts so far.
You´ll soon discover the reason we´re using such unorthodox units: they let us
visualise everything at the same scale.
In this section, you´ll learn how to see your calculations. You´ll see the profile which
shows the energy of the water in the pipe, together with the topographic profile. The
vertical distance between these 2 profiles indicates the pressure.
The hydraulic grade line (HGL from now on) represents the energy the water has. It
started with 60 meters of energy, all of which are due to the elevation. The pressure at
any particular point is the difference between the hydraulic grade line and the
elevation of the terrain itself.
As energy is lost due to friction in the pipeline, the HGL begins to incline. This incline
is the J you have calculating so far. In the first reach of 90mm pipe, up until it reduces
at 984m, the frictional loss was 9 m/km. After that, it inclines further with the reduction
to 63mm, reflecting the head loss of 50.78 m/km.
Now is a good moment to read sections 1.4 and 1.5 of the theory book.
Section 1.5, The hang glider analogy, presents you with a more intuitive way
of understanding the calculations. Don´t skip this bit! You can carry on with
the Initiation flight in section 1.6, which is very similar to exercise 8.
Take note: it´s easy to make the mistake in thinking that the HGL is where the
pipe actually goes. Remember, the pipe follows the terrain profile!
The path taken by a pipe of a given length between one point and the next, does not
affect the pressure at either of those points. In other words, the outlet pressure of a
pipe is the same whether it´s rolled up, in a dead straight line, or laid out in whichever
way possible.
1. Define the outlet point and look for the critical areas. The outlet is the
elevation, 618m, together with the required residual pressure, 1 bar: 618m +
10m = 628m.
Pay close attention to the scales of the topographic survey. At first glance,
the profile can give you a false idea that there´s a big drop and that there
won´t be any critical points. If you adjust the vertical scale, you can see
the situation more clearly:
You can understand the need for precise calculations once you see the profile at the
correct scale…in reality, it´s very flat!
Returning to the enlarged profile, there are 2 obstacles before reaching the outlet.
Point A because it´s so close to the inlet. Point B is a small obstacle. Point C is the
outlet with 1 bar of pressure.
One way of doing it is to aim straight for point C and check if the HGL gets
dangerously close to the other points:
2. The maximum head loss is: 643m – 618m - 10m = 15m. Seeing you can
lose 15m in 2.8 km, the HGL is: Jmax = 15m / 2.8 km = 5.36 m/km.
You can work out the height of the HGL at the critical points by calculating the grade
line as you move down along the pipe, i.e. multiplying it by the distance at each point:
The pressure at this point will be the total available pressure, less the head
loss, and less the elevation of the terrain at that point: P = 643m - 1.6m -
632m = 9.4m.
Despite being less than 10 meters, this is not a big problem, since the system
can´t build up the necessary pressure straight away: it needs a certain
distance and drop before it can pressurise sufficiently. This means A isn´t
really a critical point.
5. Seeing as there are no obstacles, you can aim straight for the outlet. To find
out which PVC pipes are needed, look for head losses of around 5.36 m/km
for 5 l/s:
J110 = 4 m/km at a flow of 4.97 l/s
J90 = 12 m/km at a flow of 5.057 l/s
110mm pipe is sufficiently close to avoid having to combine pipe sizes. The
pressure at the outlet would be:
1. In this example, it´s clear there´s a critical point at 1,300m. The HGL will also
be close to ground level at 300m. At this point, it´s not possible to have 10
meters of pressure due to the lack of available height.
2. To reach the point at 1300m with 10 meters of pressure, the maximum head
loss is:
3. Using the HDPE tables, J200 = 0.7 m/km for 8.221 l/s. To find the
corresponding value at 8 l/s you can interpolate:
Note that for such low head loss values you don´t gain much in terms of
precision, only 0.007 bar in this case.
4. The pressure at the 1300m point is: P = 281m - 270m- 1.3km * 0.67 m/km =
10.13m.
5. The HGL is almost flat in the first reach, which also means that the first 300
meters have most of the available pressure.
6. The second reach begins at 280m, with 13m to reach 265m (255 + 10m).
The available energy is 15.13m (280.13m – 265m), over a distance of
1,200m:
7. In the HDPE tables for 110mm and 10 bar, at a flow of 8.04 l/s, the head loss
is 12m/km: J110= 12 m/km.
8. As the previous reach left at 280.13m, the pressure at the outlet is:
Ten meters is the minimum pressure and 10.73 isn´t much more than that. In
the real world, it´s better to leave yourself a bigger margin. To do so, would
mean lengthening the 200mm pipe a little. Work out on your own how much
pressure you´d gain if it was 200 meters longer. (Answer: 13m)
There´ll be maximum pressure in the pipe when the water is not flowing and no
pressure is lost due to friction. This means to calculate maximum pressures you don´t
need to deal with flow or head loss. Here the HGL is horizontal (also referred to as a
point of static pressure) and all you need to do is look at the difference in height
between the HGL and the low points.
For example, in exercise 10 the maximum pressure will be in the lowest valley:
The maximum pressure is the difference between the elevations: 281m – 253.5m =
27.5 m. Since 2.75 bar is less than 8 bar, you can install PN10 pipe.
1. The difference between the highest point, 345m, and the lowest, 255m, is
90m. When the water is at rest you’ll have 90m of pressure, which exceeds
by 10 meters the 80% pressure limit for 10 bar pipe. At some points in the
pipeline you´ll need to install PN16 pipe.
2. To work out where, subtract the maximum pressure, 80m, from the highest
elevation:
Any point below 265m means you need to install 16 bar pipe.
PN 16 pipe has a higher frictional loss than PN 10, which is why the tables
have differing values for each pressure rating. Knowing beforehand which
sections require PN 16 will save you a lot of work.
3. Use the same method you´ve been using so far to work out the pipe
diameters, paying special attention to the points where the pipe diameters
change. As there are no obstacles, you can aim straight for the outlet point at
324 meters + the residual pressure.
The outlet is freely flowing into a reservoir tank. To avoid wear of the parts
and minimise repairs and replacements, and also to avoid a situation where
the float valve won´t close due to excessive pressure, it´s best to arrive with
the minimum pressure of 10m.
4. The outlet point is 324m + 10m = 334 m. If the starting point is at 345m,
maximum head loss is 11 metres: 345m – 334m = 11m. The resulting head
loss is:
J= 11m / 3.75 km = 2.93 m/km
PVC degrades in the sun, so you can´t install it without burying it. That
means you have to look at the HDPE tables.
5. You´ll see that at 2.957 l/s, J110= 2 m/km. This value is close enough and
leaves some margin.
8. Of the 11m of head we could lose, and after the head loss in the previous 2
reaches, we are left with: 11m – 2.4m – 1.8m = 6.8m.
Ja*x + Jb(d-x) = D
J90*x + J110(d-x) = D
2. 8 BRANCH LINES
In this section you’ll learn how to design distribution systems with branch lines, which
are the most common. The procedure is exactly the same, you just have to take into
account the flow in each respective pipe and calculate the branch lines one by one.
In a system with just one branch line and 2 final consumer nodes of 3 and 2 l/s, the
mainline carries the total flow, and then divides into each branch line:
The procedure is the same for adding another branch line. Pipes d, c and e carry the
final consumer demands of 1, 3 and 2 l/s respectively. b carries the flow for d and c:
1+3 l/s = 4 l/s, and a the sum of b and e, 4+2= 6 l/s.
Size the PVC pipes needed for the following system, fed by a
spring at 32m, with 2 public tap stands at 7m and 0m.
2. There is no point in the system where the pressure is more than 8 bar, so all
the pipes can be PN 10.
3. The energy in the branch line must be enough to allow the water to reach
each point. Choosing the pipe is a skill you´ll develop with practice. In the
meantime, use trial and error.
4. Seeing as the highest point is 7m, it makes sense to try and reach the
branch point with 20m. This leaves 3m of energy to burn off in pipe c, if
there´s 1 bar residual pressure: 20m – 7m – 10m = 3m.
J= (32m-20m) / 4 km = 3 m/km
6. Looking at the PVC tables for 5 l/s, J110= 4 m/km and J160= 0,7 m/km. We
can´t install 4km of 110mm otherwise we won´t reach the branch lines with
enough pressure. On the other hand, if we install 4km of 160mm the project
may end up being too expensive.
If you´re at the design stage and there are likely to be system enlargements,
it´s best to use a larger pipe. If the project is tight on funds, you´ll have to
combine pipe sizes, as you´ve seen in the previous exercises.
7. Since funds are almost always in short supply, we combine pipe sizes:
J110*x + J160(d-x) = D
The first reach will be 1.21km of 160mm, followed by 2.79km of 110 mm.
8. In pipe b the flow is 3 l/s over a distance of 1.2km. Starting with 20m of
energy at the branch point, the head loss to arrive with at least 1 bar is:
9. Looking at the tables for 3 l/s, J63 is a value between 20 and 30 m/km (too
much). J90 = 4.75 m/km. Over a distance of 1.2km, the head loss is D = 1.2
km * 4.75 m/km = 5.7m. If we start with 20m, the pressure at the outlet of
pipe b will be:
P= 20m – 0m – 5.7m = 14.3m or 1.43 bar, which give us some margin above
the minimum.
Size the PVC pipes needed for the following system to work, from
a reservoir tank at 102m to 3 further tanks in three villages at
elevations of 71m, 81m, and 12m respectively.
1. Since we´re dealing with reservoir tanks, we´re going to try and arrive with
the minimum pressure, 1 bar, so as to avoid wear in the float valves and
save on piping.
3. The pipe d will need PN16 at some point: 102 – 12 = 90m > 80m. Seeing as
there´s no topographic profile, and to keep things simple in this exercise,
we´ll assume it´s the final 200m that needs PN 16.
5. Pipe e:
J63, 1.5 l/s = 7.5 m/km J90, 1.5 l/s = 1.4 m/km (combination of pipes needed)
7. Pipe c:
8. As you saw in point 3, the pipe d has 2 reaches, one of 1.1km PN 10, and
the other of 0.2km PN 16. Both have to lose head for the pressure to be
between 1 and 3 bar:
Ideally you´d combine pipe sizes in which the first reach is PN10 and the
second is PN16. Over 1.1km, the head loss will be 26.39 m/km * 1.1 km =
29.03m.
The 200m reach needs a head loss which leaves the pressure at a maximum of 3 bar:
Woops, there are no values high enough in the tables! This is a rare
occurrence, but sometimes comes up in very rough terrain. In these cases,
you can calculate the values you´re looking for using the Hazen-Williams
formula, found at the end of Appendix B, using a friction coefficient of C-140
for PVC and HDPE and C-120 for galvanized iron. Be careful with these
values if you´re using pipe of less than 50mm and velocities higher than 3
m/s.
9. The value for a 40mm pipe (careful: interior diameter is 36.2mm) is 252 m/km
according to the Hazen-Williams formula: too much.
If the pipe is small enough, less than 25mm, and there aren´t many points, you could
also install pressure reducing valves (PRV), which are cheap and easy to find at those
diameters:
!
Next you have a confirmation exercise, with fewer explanations and more
realistic topographic profiles. It´s basically like the previous exercise, but with the
profiles. If you follow it through and can do it on your own, congratulatons! The rest
of the book will be all downhill. If that´s not the case, carefully take the time to go
back over the previous pages.
To avoid aiming for very low pressures, take 15m as the minimum design
pressure. If you then get values between 15m and the absolute minimum of
10m, you can decide whether it´s acceptable or not. The idea is to leave more
than the minimum pressure for unknowns. If it is impractical to have it above 15m at
some point, then 10m is acceptable.
REACH A
The pipe a has to transport 9 l/s. At point 1, the elevation is 1,852m, very close to that
of the sources F1 (1,846m) and F3 (1,850m). We need enough residual head for the
branch lines. With 10m we´ll be left short. We can tentatively aim for 25m of pressure,
which allows for a head loss of:
(1,892m-1877m)/0.500km = 30m/km
In the tables, there´s no value close enough to 9 l/s for 90mm pipe. Use lineal
interpolation to find the answer:
J x − J inf Q − Qinf
= x
J sup − J inf Qsup − Qinf
J x − 30 9 − 7.798 Æ J = 39.28m/km
= x
45 − 30 9.74 − 7.798
REACH C
The pipe c has to transport 6 l/s in total, over a distance of 200m. Keeping the
pressure of point 2 at 20m, the water will have enough pressure to climb the branch
line d:
No pipe is close to this value. However, since the distance is small, you don´t have to
get it exactly right on 20m (22m or 24m are equally valid), and we can use 90mm
pipe. For 6.2 l/s, J= 20 m/km. The pressure at point 2 will be:
REACH E
The pipe e must transport 4 l/s in total over a distance of 400m, with a residual head
of 1.5 to 3 bar (15-30m). We need to find a pipe with a head loss between those
values for 4 l/s:
For 4 l/s the J value will be between 45m and 60m, and between 40.9m and 78.4m.
J x − 45 4 − 3.752
= Jx = 50.78m/km
60 − 45 4.396 − 3.752
This value is in between the previous ones. The residual pressure will be:
REACH B
Pipe b must transport 3 l/s over 500m, to an elevation of 1,846m, beginning at
1,852m. The pressure at 1 has been calculated to be 20.36m.
Looking at the tables for 90mm pipe, we get 5.5m/km. Check the maximum pressure
is not exceeded:
In case this is excessive, a combination of pipe sizes would have been installed.
REACH D
The pipe d must transport 2 l/s over a distance of 200m, to an elevation of 1850m,
departing from 1842m. The pressure at 2 has been calculated as 26.36m.
Looking at the tables for 93 mm pipe, we get 15m/km. Check the maximum pressure
is not exceeded:
Note that when there´s no demand, for example, at night, the pressure at
point F2 is 1,892m-1,822m = 70m. This much pressure at a tap outlet is
dangerous and is basically useless for the final users. In the next chapter
you´ll see how to resolve this.
Aside from the user points, it´s best for water systems to work with as little
pressure as possible. More pressure means more problems, burst pipes, higher flow
through leaks, more water wasted at point of use and frustrated users. To reduce
excess pressure there are 2 solutions: pressure reducing valves, or break-pressure
tanks.
The fundamental difference between the two is that you can install a PRV in several
points in a system and adjust the pressure setting you´re looking for. A BPT is fixed
and must be located at a specific elevation above the lowest user point, according to
the maximum pressure limits you are working with.
Calculate the HDPE pipe you would need to feed a tap with a
maximum of 3 bar, assuming you are using a pressure reducing
valve.
(This is the same topographic profile as in the previous exercise, except you´re only
dealing with the mainline).
1. Do the calculations as if there was no PRV. The valve does not affect the
pipe diameters: it simply reduces the pressure during times of low demand.
3. The diameters for a flow of 4 l/s with at least 10m of pressure all along the
line are:
Jmax = (1,892 - 1,851 -10)m / 0.6 km = 51.66 m/km (for the low point)
Jmax total = (1,892 - 1,822 -10)m / 1.1 km = 54.54 m/km
1. The first step is to work out where to place the BPT. If the maximum
pressure is 30m, it should be situated 30m above the tap: 1,822m + 30m=
1,852m. This amounts to 500m downstream in the topographic profile.
Take note: this would not be a good place to locate the tank, because the
terrain over the next 100m is very flat and the water would need to travel
some distance before pressurising again to the minimum of 10m. The best
place would be at an elevation of 1,851m, at 600m.
2. The water entering the BPT should arrive with 10m of pressure, to avoid
excessive wear of the float valve. So the head loss you´re looking for is:
3. For 63mm pipe at 4 l/s, J63 = 45 m/km, close enough to avoid combining pipe
sizes:
P = 1,892m – 1,851m – 0.6 * 45 m/km = 14m
4. For the pipe leaving the BPT it´s enough to find a diameter that leaves the
pressure between 10 and 30m:
2. 10 PRESSURE ZONES
It´s not always easy to supply an entire area so that all users are within the right
pressure limits. People often live on the side of a valley. If the highest users are more
than 20m above the lowest, it´s impossible to supply them all within the 10-30m
range. If the difference is 30m, for example, when the lowest users have 30m of
pressure the highest ones have 0m. If the highest have 10m of pressure, the lowest
have 40m.
You´re dealing with a problem which is similar to sleeping with a blanket that´s too
short: either you´re feet are freezing or your head gets cold!
It´s fairly common that in the zone you´re dealing with, the difference in elevation
between the high and low points is too great. To avoid making too many zones, a
useful rule of thumb is this: make only enough zones to remain within a 35m pressure
range.
Creating separate zones can require quite a lot more pipe. In the case that follows,
you can´t break the pressure on one hillside as you won´t then be able to get up the
one on the opposite side. What you need to do is install piping that feeds the BPT,
together with a separate bypass line which maintains the pressure and allows you to
reach the opposite hillside:
1. The difference between the highest user and the lowest is 81m – 43m =
38m. You´ll have to make two pressure zones, one high and one low.
2. Next, work out where you´ll put the BPT to supply the lower zone. You need
to find a site which:
In this system, the branch line just after the 73m point fits all the
requirements. Below you´ll see where to place the BPT. We´ll also number
the pipes:
4. Pipe 5. From the BPT with 73m of pressure, we´re looking for a low-friction
pipe:
In the next point, we´ve gone straight for a pipe with a head loss of less than
5m/km, without calculating Jmax as before. This will save you time on the
calculations, once you´re more familiar with the pipes.
5. Pipe 6:
Æ Pipe 6 is 63mm.
6. Pipe 7:
7. Pipe 1. This needs to be really efficient, as it starts with 73m and then needs
to get up to 81m.
J110, 4 l/s = 3.5 m/km (This pipe would leave the pressure only slightly above
81m). Any unauthorized connection or increase in flow could cut off the
supply. If the project is very tight on funds and you´ve got no choice in the
matter, 110mm is the answer. Whenever you can though, use a pipe which
is one size up in these kinds of situations:
Æ Pipe 1 is 160 mm
8. Pipe 2. This is still the critical section. You need a more efficient pipe.
10. Pipe 4. For reasons of symmetry and similar elevations you´ve probably
already got an idea which it is:
There´s never one right answer for a system, as you will have gathered
working through the exercises. There are other criteria which will mean you
choose one design over another.
2. 11 MULTIPLE SOURCES
Sometimes you need to use several different sources. Unless you´re incredibly lucky,
they are rarely at the same elevation. In these situations, both sources need to have
the same residual pressure when they meet. When there´s no demand, the higher
source will discharge into the lower one. To avoid this, you can install a no-return
valve in the pipe line leading to the lower source.
(Intentionally blank)
We want the pressure at the junction to be greater than 10m, although the topography
won´t let us pressurise the system much more than that.
1. The pipe with less room to manoeuvre comes from the south source. Let´s
start with this. The maximum head loss is:
For a flow of 2 l/s, the pipe which leaves us enough pressure is 90mm. The
pressure at the junction is:
2. The north source must arrive at the junction with 13.25m too. The required
head loss is:
To reach the junction with precisely the required pressure means combining
pipe sizes, namely 63mm and 90mm. We can interpolate for 63mm:
Note that the order you place the pipes in is very important. If you put the smaller pipe
first, the HGL will go underground:
The graph showing the topography and the HGL looks like this:
The HGL´s of each pipeline meet at the junction with the same pressure: 13.25m.
2. 13 LOOPED SYSTEMS
Bad news!
Even the simplest looped system cannot be calculated the way we´ve been doing up
to now. This is because the water can follow various paths to get to the same point.
To calculate by hand you can use the Hardy-Cross method. It´s more
time consuming than complicated for simple systems, and can be as
exhausting as having to sit through your brother-in-laws holiday snaps.
I would recommend that instead of losing time learning this method, learn how to do it
with a computer from the start. These two books can show you how:
If you want a quick tutorial to get over your fear and apprehensions about
using a computer for the calculations, chapter 6 of the theory book takes you
through an example.
Now that you´ve been warned, this is how to do the calculations by hand, in case you
don´t have access to a computer.
a. In each node the mass is conserved. That means the water coming in is
the same as the water going out, and results in all the flows in a node
equalling 0.
b. In each loop energy is conserved, that means the sum of all head loss in
a loop is 0.
The difference between this and what you´ve seen up to now is that this method
doesn´t select the pipe sizes for you. It only tells you what happens with the pipes
you´ve chosen, i.e. it´s a matter of trial and error. That´s why it becomes a mammoth
task, because you´re not only doing the iterations, but you have to keep going over it
again and again until you find the optimum pipe sizes. One small change somewhere
in the loop and you have to calculate everything over again.
Size the PVC pipes for this system, in which each one is 1km long,
and where the nodes all have the same elevation with an input
pressure of 3 bar.
2. Assign a flow to each pipe such that the law of the conservation of mass is
respected. But try and fine tune how you distribute the flow, as it´ll save you
a lot of calculations. For example:
Pipe 1: - 4 l/s
Pipe 4: - 3 l/s
Pipe 5: - 3 l/s .
4. We can use the tables to find the pipes which have a head loss of around
5m/km for flows of 1, 3 and 4 l/s:
5. Let´s see how the loops perform. For loop 1, beginning at a, we add the head
losses as we move in the direction of the arrows and subtract those which go
the opposite way:
1km * 4.75 m/km + 1km * 3.75 m/km – 1 km * 2.75 m/km = 5.75 m/km
5.75 m/km is a little different to 0, which is what the law of loops demands.
This means the chosen flows are wrong and they need to be adjusted in the
second round.
For loop 2, beginning with a, we add the head losses as we move in the
direction of the arrow and subtract the ones in the opposite direction. I´ve
removed the units here to make it simpler:
∑hi , is the sum of the friction produced in each pipe of the loop,
i.e. the value you calculated in point 5.
Q = flow
When a pipe belongs to two loops, the correction is made for both, i.e. for
pipe 5, the correction will be -0.52 + 0.29 = - 0.23 l/s.
If the flow goes in the direction of the arrow it´s considered positive, otherwise
negative. The variation ∆Q is always added up.
This marks the end of the first iteration; now with these flows you need to
repeat the cycle you began in point 2.
Just let me run off here and give you the values calculated by a computer, in
case you really want to carry on with the calculations by hand. Ignore the
arrows, the computer hasn´t necessarily chosen the directions the way you
have. You can consider the flow values to be acceptable when the variation
is less than 5%, i.e. ∆Q < 0,05*Q.
Once you´re done, to calculate the pressure values, just take any route
towards a node and calculate the head loss along the whole route, as you
have done in the previous exercises. Any route has the same head loss, as
the system is in equilibrium.
Notice how the pressures are 23.99m, 26.28m y 26.8m, which are
acceptable. All these calculations go to show that the initial chosen
diameters work. This doesn´t mean they are necessarily the optimum choice.
2. 12 SPREADSHEETS
All the calculations we´ve done so far in this book are made much easier with the use
of a spreadsheet like Excel, which also allows you to draw graphs. Since it´s the most
commonly used, the explanations here will be with Excel 2007 for Windows, although
there are plenty of free open-source alternatives like OpenOffice Calc:
http://www.openoffice.org/
If you´ve got a computer to hand, and you think that in the future you´re likely
to design more systems, your best bet is to learn how to use a modelling
program like Epanet.
Once you´ve entered the data into the spreadsheet, you can create graphs similar to
the one below, where the pressure, HGL and topography are displayed.
Size the pipes for the mainline supplying the refugee camp Mtabila
II using HDPE with a flow of 9.8 l/s:
2. On the page heading, enter a title which makes sense and include
information regarding the date, who prepared the document, the version
number, and probably some kind of logo of the organisation you´re working
for. The idea is that the document be clearly identifiable to third persons.
3. Save the document with the same name as the title you´ve used:
You´re not going to spend the rest of your life working on this same
system: there will be a moment when other people inherit the project.
There´s nothing more frustrating than trying to fight your way through a
million files called “water 1,” “Peter,” “Dongwe test” and so on...In the case
of version numbers, it´s vital to be really organised so that no one else
starts using a document that´s been only half worked on.
5. Enter the Distance and Elevation values from the topographic survey. You
don´t need to copy the Chainage as it´ll be calculated automatically.
To avoid having to manually enter values or trying to work them out from
the graph, download this file in which they´ve already been entered:
www.arnalich.com/dwnl/19topoeng.xls
© Santiago Arnalich Arnalich. Water and habitat www.arnalich.com
64 CHAPTER 2. Piping
The elevations are negative because the highest point has been arbitrarily
taken to be at an elevation of 0m.
6. Calculate the chainage. To do this, enter the formula “=B12+A13” into cell
B13. Don´t forget the “=”! This adds the new Distance value with the
chainage each time.
7. To complete the rest of the cells, write 0 in B12 (chainage at the beginning is
0). Then place the cursor over the bottom right hand corner of the cell and
note how it changes from a white cross to black one. This means you can
drag and copy the formula down the column into the cells bellow.
8. Click on the corner and without letting go, drag the cross down to the last
cell. All the cells you´re copying the formula to, will be surrounded by a grey
dotted box. When you let go, the formula will calculate the chainage in each
cell. The final value should be 5,911.93m. The pipe line will be almost 6km
long.
If you click on the cell B14 you´ll see the formula changes, rising by 1 each
time from B12+A13 to B13+A14 and so on, in all the cells below.
9. In the tables find the head loss for 9.8 l/s for different diameters. Create an
excel table:
10. You can find the head loss in each reach, J, by multiplying the Distance by
the head loss of the pipe you´ve selected in the table you´ve just made. To
do this, enter the following formula into cell E12: “=A12*$F$5/1,000”.
The $ symbol ensures that the reference always points to the same cell, F5,
even if you copy it down into other cells. Dividing by 1,000 is required as the
distances are in meters but the head loss is in kilometres.
11. Copy the formula until you´ve got the values for each reach. You´ll probably
find it annoying to have more than two decimal points, so right click
anywhere on the screen and select Cell Format. Once you´re there, select
the Number tab, then the Number category and write 2 in the decimals box:
12. To find the cumulative J, do the same as in point 6 and drag down:
13. Now the pressure column. The pressure is the initial elevation, less the
elevation at one particular point, less the head loss. Enter this into the
formula:
The initial pressure is 0. $C$12 is the initial elevation, and the $ symbol is needed
so it points to the same cell when you drag and copy. C13 is the elevation at that
point, and F13 the Cum. J, or cumulative head loss.
14. Drag and copy the formula down and read the pressure values. You´ll notice
that they are all negative values. The 90mm pipe you chose is too small:
To be able to see the results more easily, it´s best to make a graph.
15. The line you´ll pay closest attention to in the graph is the HGL. Seeing as it
doesn´t exist yet, create an HGL column next to the pressure column. The
value is the initial elevation less the cumulative head loss.
Drag and copy the formula to get the remaining values. As the initial
elevation is zero, the values are the same as cumulative J but with the sign
changed.
16. Select the data in the Elevation and chainage column. Click on the centre of
cell with the Elevation title, and without letting go, drag down. You´ll see the
cells are selected, and should be shaded:
17. In the upper menu bar of Excel, go to Insert, then Dispersion, and then the
graph type number 3:
19. The HGL still needs to be plotted. Right click on any blank area on the graph
and choose Select data. You´ll see a box like the one below, where you have
to select Add:
21. Select the cells which have the chainage values and click on the symbol to
accept:
The graph should look like this. You´ll recognise it from previous exercises:
You can make a series of improvements, until you get the result we´re using
here. But since this is not an Excel manual, I´ll let you work out how to do it
on your own:
If you got lost at any point, you can download the Excel document, with
what we´ve done up to now. It can also be useful as a template:
www.arnalich.com/dwnl/gravitytemplate.xls
Now select the pipes until the HGL is 10m above the terrain at all points, with a
pressure of between 10 and 30m.
23. Notice that there´s a critical point 2,241m from the origin. Up until that point
we want the HGL to be as flat as possible, so we can maintain 10m of head
and pressurise the pipe as fast as possible. Modify the formulas from
distance 0 up to 2,241m, so that they point at the 200mm pipe. This means
changing $F$6 for $F$8 in the formula, then drag and copy down to the
distance of 2,241m (row 38).
24. To check the pressure at the 2,241m point, read off the value in the Pressure
column: 9.70m. It´s not quite 10m but it´s close enough. The problem is the
topography, which doesn´t allow for any more. Increasing the pipe size will
increase costs for very little result. Check out how the graph has changed:
25. For the next reach, notice how the difference in height is very pronounced
and is going to require PN16 at the bottom of the valley. This means the next
section will leave from 2,241m until the pressure reaches 80m. Starting from
26. Choose the pipe for the reach which goes from 2,241m to 3,912m. To do
this, multiply the first cell after the 2,241m point by the head loss for the pipe
you want to try. In this case, let´s try 160mm (cell $F$7):
27. Copy and drag the cells to 3,912m and check the result:
It looks pretty good! If you lengthened the reach you´ve just done, you´d get
to the end of the line with sufficient pressure.
28. From 3,912m to 4,459m, where the pressure falls back below the 80m mark
(elevation of -77.54m), you´ll need to install PN16 pipe. Seeing as the pipe
you used in the previous reach looks good, try the same diameter here but
with PN16, and see what happens. To do this, go back to the tables, find the
160mm value for PN16, and incorporate it into your Excel table:
29. Multiply the first cell after 3,912m by the head loss for the pipe you´ve just
introduced (in cell $H$7). Drag and copy the cells up to 4,459m and look at
the result:
30. Do the final reach with 160mm PN10, and read off the pressure at the
entrance to the tank: 31.28m.
31. This is too much pressure. You´ll also save money using smaller pipe. To
reduce the pressure, go to the last cell (where the water enters the tank) and
make it point to 110mm pipe ($F$6). Then copy upwards and note what
happens with the pressure at the end point and in between.
32. You can get to approximately 4,800m. The graph looks like this:
33. Congratulations, the calculation´s done! Now use the notes column to
include information such as “Start of 160mm pipe,” or “End of 160mm PN16
pipe,” and so on. Note that I´ve put a colour scale in the friction loss table for
each pipe. You can indicate which reaches use which kind of pipe by
shading the background of the table with these colours.
You can see the result by clicking on this link (click on the image itself to
enlarge): www.arnalich.com/dwnl/19zheng.png
www.arnalich.com/dwnl/19eng.xls
It may seem kind of laborious to use a spreadsheet if you´ve never done so before.
Just remember it has some of the following advantages:
• The height of buildings. The pressure at the tap should be between 10 and
30m, which won´t be at street level. This means that if you’ve got a building
that´s 3 stories high, you´ll have to take into account the building height, and
provide 10m of pressure on the top floor: 10m + 3* 2.5m = 17.5m above
street level. Follow the same reasoning for buildings up to 5 stories high.
Anything more than that, and the buildings themselves will have to take care
of assigning a pressure group for the users on the highest floors.
• Leaks and maintenance. The flow from a leak increases exponentially with
the pressure. So do the chances of a pipe breaking. That means it´s best to
keep the pressure as low as it can be for the system to function properly.
• Head loss in accessories. There are not only frictional losses in pipes, but
in elbows, T´s, valves and other accessories, causing a loss of pressure,
known as minor losses. They are only relevant for high velocities, and
generally don´t have much effect on traditional gravity flow systems.
You can read more about this in section 5.5 of the theory book.
You can read more about this in section 7.4 of the theory book.
In all the exercises up until now, you´ve been given a specific flow to work with. In this
chapter you´ll learn how to work out and establish a flow, with which to begin
designing your system.
3. 1 BASE DEMAND
This is the quantity of water which the popluation will consume for all kinds of different
uses: cooking, washing, drinking, work activities and so on...There is no quick fix to
working out the demand of the specific population. As a general pointer, these are
some minimum figures with which to work:
In practice, the tendency is to provide the maximum amount of water such that:
Work out the minimum base demand for a town of 1,300 people,
where the average family has 5 members, 2 goats and one cow.
1. To work out the number of animals, you need to work out the approximate
number of families in the town:
2. The flow would be 4,150 litres in 24 hours. Coverting the units to litres per
second:
3. It´s always a good idea to check the demand estimates. In the case of a
birth, we can´t be sure how much water will be consumed. The onsite doctor
can give us an idea: 50 litres/birth.
Always check that the figures you are quoted make sense. A doctor in
Mauritania may claim to use 5 litres, but this is really incompatible with
decent hygiene during childbirth. It doesn´t provide enough water to wash
the mother, the baby, the equipment, and the delivery room.
3. 2 FUTURE PROJECTIONS
Populations grow. If you design a system without taking this into account, it´ll soon be
overtaken by rising demand. Expanding a system is much more expensive than
building it with sufficient capacity from the start. In general, 30 years is the standard
design period, although this can change depending on the circumstances.
To try and work out the future projection, there are two basic methods:
Geometric projection
Usually census data is used together with a growth rate. The projected population in
the future is:
Pf , future population
P0 , current population
i , growth rate in %
t , time in years
You then work just with the future population figure rather than with the current, so as
to take into account it´s growth.
Saturation
Imagine a beach early in the morning. As the day goes by, it starts filling up with
people. There comes a certain point where new people arriving decide to go
somewhere else when they see the beach is too full. This is how a balance is
established in relation to peak density, for example, 1 person per meter squared. If
the beach is 1,000 square meters, when it fills up in the future, it will have 1,000
people.
This focus is very useful in some cases, such as urban populations with rapid and
unpredictable settlements.
Notice how there are many assumptions here. Future projections are not a
precise science, they are an approximated prediction, and predictions are
sometimes...well...have a look at this:
2
An area of 1.3km in a city is being rapidly populated with
immigration from rural areas. Before conditions get worse, basic
services need to be provided. In other areas of the same city
which have been populated by rural immigration, each family
2
occupies approximately 260m and averages 7 people. What is the
population to be supplied?
2 2 2 2
7 people / 260 m * 1,000,000 m / 1 km = 26,923 people / km
2 2
2. The population will be: 1.3 km * 26,923 people / km = 35,000 people
1. All taps open. A tap is assigned with a specific number of users, and the
flow is designed to supply all taps simultaneously. This is the approach taken
in emergencies, refugee camps, and similar situations. For example, one tap
is assigned 250 people with a flow of 0.2 l/s.
2. Temporal variations. Even more important than knowing how much water is
consumed, is knowing when it will be consumed. People use more water at
certain times of the day, on certain days of the week, and sometimes during
certain months of the year. To take this into account, a multiplier is used to
increase the average flow by a certain factor.
The idea is to design the system for the worst case scenario: “the hour of
highest demand, the day of highest demand, and the month of highest
demand.” This is the approach used most often for larger populations. For a
gravity flow water project calculated by hand, multiply the average demand
by a factor of between 3.5 and 4.5 (3.5 if the demand is more regular and the
hot season is less pronounced etc).
It´s a good idea to read section 2.5 of the theory book to get a better
understanding of the nature of these variations.
3. Simultaneity. Use this approach for systems and parts of systems which
have less than 250 connections. Nevertheless, for large systems it´s good
working practice to establish a minimum pipe diameter (generally 75mm to
100mm). This diameter is sufficient to not have to take into account
simultaniety, as it can supply more than 250 connections. A minimum
diameter is established to make enlargements easier, avoid blockages and
provide a certain amount of protection against fires.
Avoid installing pipes which are too small, especially when the distances
are large. Putting in a 40mm pipe over 2km for example, is a really bad
idea. It´ll get blocked, and it then becomes very difficult to work out where
the blockage is. Be careful with saving money here! It´ll soon end up
becoming very expensive for the local population.
To work out the design flow for simultaneity, use this graph to find the
multiplier for the average demand (Arizmendi 1,991):
Most of the time, choosing the method for calculating the corrected demand comes
down to 3 basic questions, which you can see in the flow diagram below:
Since the quantity of water is not a limitation, we´re only concerned with 3
standards:
9 The maximum distance between any home and the nearest water
supply point is 500 meters.
2. The first standard requires that each tap be at a maximum distance of 700
2
meters from one another if the layout is regular. One way of doing it would
be this:
2
In a regular layout, the simplest is to draw out boxes inside circles with a radius of 500m.
Applying Pythagoras´ theorem, each side comes out as 2r/√2=707.1m. Don´t worry if you don´t
understand the reasoning for now, it´s not really that important in the context of this book.
4. To meet the third criteria, you have to work out how many people there are in
the distribution area of each public tap stand:
If one tap for every 250 people is needed, a minimum of 2 taps will be
necessary. This means the flow at each supply point will be 3 * 0.2 l/s = 0.6
l/s.
5. To calculate the design flow you need to look at the context. As it´s a refugee
camp:
Now you´re ready to do the calculation. Note that the system is symmetrical,
which means you only have to size the pipes which supply nodes 1, 2, 3, 4 y
5. The pipes that go from 1’, 2’, 3’ y 4’ will be the same as the ones that go
from 1, 2, 3, 4.
6. Determine the flows each pipe will carry if the supply points have a demand
of 0.6 l/s. Remember the 0 node doesn´t affect the calculation because you
are working with a supply line which has no flow limitations:
7. The system is flat and begins with 3 bar of pressure. The longest route,
which goes from 0 to 3 has 3 * 0.7 km = 2.1 km. If you aim to get there with
15m of pressure, the head loss is:
The material to use is HDPE, as it´s quick to assemble (it comes in rolls),
it´s strong, and because PVC is damaged when exposed to the sun (in
an emergency it´s best not to wait for the pipes to be buried).
The pipes which are closest to the required head loss for 2.4 l/s, 1.2 l/s and
0.6 l/s are:
P0 = 30 m
P1 = 30 m – 0.7 km * 3.5 m/km = 27.55m
P2 = 27.55m – 0.7 km * 6 m/km = 23.35m
P3 = 23.35m – 0.7 km * 1.7 m/km = 22.16m
Pf = 23,300 people.
2. As there are more than 200 connections, the temporal variations approach is
used. If the inhabitants have similar habits they´ll use water at similar times.
This puts pressure on the system as demand will be concentrated during
these times. We´ll use the multiplier of 4.2, between the range of 3.5 and 4.5
(that´s right, an educated guess! The data for more precise calculations
aren´t available.)
4. Let´s look at the tables. Seeing as in the book we´re short of tables, let´s look
at the online version:
www.arnalich.com/dwnl/headloss.zip
HDPE pipes larger than 8” require expensive welding equipment. If there are
none available in the region, it may not make sense to invest in one, as the
project may end up being too costly.
Large diameter pipes are very expensive and it´s worth adjusting the
diameters precisely to reduce costs.
4.1x + 15.2 = 29
1. If they don´t grow, there´s no need to project into the future. On the other
hand, the system is small and the distances are short. There´ll be no
minimum diameter, and seeing as the connection groups are small, the
simultaneity approach will be used.
2. Just as the flow was divided up earlier, it´s best to divide up the connections
for each pipeline:
3. The corrected flows of each pipe are calculated, multiplying the average flow
by the multiplier. The average flow for each connection is:
4. The diameters needed are calculated with the adjusted flows. The longest
route, pipes 1, 3 and 4, is 1700m. Aiming to get there with 10m:
As there´s not much of a pressure margin, it´s best to choose pipes that have
a head loss equal to or less than 4.7 m/km between them all, without much
need of adjustment. In HDPE:
I´ll leave it to you to check that the pressures are greater than 10m in all the
points.
3. 4 UNACCOUNTED-FOR WATER
Unaccounted-for water (UFW) is a little box of disasters in which you´ll find system
leaks, unauthorized connections, authorized but unpaid connections…In some
systems it´s less a question of unmeasured as “unconsumed demand.” For example,
in an emergency system with public tap stands where the users are going to consume
a minimum of 15 litres/day, you have to take into account that some of it will be
wasted. If to this you add the standard losses from a new system of around 20%,
you´ll probably end up with around 10 litres. With such small amounts of water, the
difference between 10 and 15 litres is a world of difference.
Take something in the region of 20% more to take this into account, which means if
the base demand was 10 l/s, the corrected demand will be 10* 1.2 = 12 l/s (adding
20% or multiplying by 1.2 is the same thing. 35% would mean multiplying by 1.35).
The order in which you do the multiplications doesn`t affect the result (adding 20% at
the beginning or the end makes no difference).
Santiago Arnalich
At 26 years old, he began as the coordinator of the Kabul Project CAWWS Water
Supply, providing water to 565,000 people, probably the most important water supply
project to date. Since then, he has designed improvements for more than a million
people, including refugee camps in Tanzania, the city of Meulaboh following the
Tsunami, and the poor neighbourhoods of Santa Cruz, Bolivia.
Currently he is founder and coordinator of Arnalich, Water and Habitat, a private entity
with a strong social commitment dedicated to promoting the impact of humanitarian
organisations through training and technical assistance in the fields of drinking water
supply and environmental engineering.
Bibliography
1. Arizmendi, L. (1,991). Instalaciones Urbanas, Infraestructura y Planeamiento. Librería
Editorial Bellisco.
2. Arnalich, S. (2,008). Gravity Flow Water Supply. Arnalich, water and habitat.
www.arnalich.com/en/libros.html
5. Department of Lands, Valuation and Water (1,983). Gravity Fed Rural Piped Water
Schemes. Republic of Malawi.
9. Santosh Kumar Garg (2,003). Water Supply Engineering. 14º ed. Khanna Publishers.
10. Stephenson, D. (1,981). “Pipeline Design for Water Engineers”. Ed. Elsevier.
11. Walski, T. M. y otros (2,003). Advanced Water Distribution Modeling and Management.
Haestad Press, USA. Haestad methods.
http://www.haestad.com/library/books/awdm/online/wwhelp/wwhimpl/js/html/wwhelp.htm
Below you will find the friction loss tables for the most commonly used pipes. Due to
space limitations not all pipes are listed. If you´re looking for data that isn´t available
here, go to www.arnalich.com/dwnl/headloss.zip.
To use the tables you need to know what material you´re working with, the maximum
pressure, and the type of water you´ll be transporting (clean/dirty). For a flow of 0.02
l/s, an HDPE of 25mm at 16 bar, carrying clean water (k=0.01), has a head loss of 0.6
m/km.
Metal pipe is specified with the internal diameter. Plastic pipe with the external
diameter. This table shows the approximate internal diameters (ID) for plastic pipe:
Approximate values for head loss in m/km of galvanized iron pipe, calculated using
the Hazen-Williams formula for middle-aged pipe.
than 3 m/s). Coefficient C-110 used for pipe of less than 3” diameter, and
Values calculated without velocity adjustments (valid for velocities of less
C-120 for 3” and over, for middle aged pipe with neutral water (Langelier
0.3 344.19 33.60 3.41 1.15
0.35 457.92 44.71 4.53 1.53
0.4 586.40 57.25 5.80 1.96
0.45 729.33 71.20 7.22 2.43
0.5 886.48 86.55 8.77 2.96
0.6 121.31 12.30 4.15
0.7 161.39 16.36 5.52
0.8 206.67 20.95 7.07
0.9 257.05 26.06 8.79
1 312.43 31.67 10.68 0.92
1.1 372.75 37.79 12.75 1.10
1.2 437.93 44.40 14.98 1.29
Index value of ± 0.5).
To calculate intermediate values, you can use the Hazen-Williams formula, taking into
account the warnings and values detailed in the box: