Introduction to Water Engineering

Slide 1 Welcome to Module 2 and Pipe flow.

Introduction to Water Engineering

Module 2 Pipe flow

4. Flow Under Varying Head

Dr James Ward
Lecturer
School of Natural and Built Environments

Slide 2 Copyright Notice

Please note
Do not remove this notice.

COMMMONWEALTH OF AUSTRALIA
Copyright Regulations 1969

WARNING
This material has been produced and communicated to you by or on
behalf of the University of South Australia pursuant to Part VB of the
Copyright Act 1968 (the Act).

The material in this communication may be subject to copyright under the

Act. Any further reproduction or communication of this material by you
may be the subject of copyright protection under the Act.

Do not remove this notice.

Slide 3 Today we’re looking at flow under varying head.

2.4 Flow under varying head

DO NOT REMOVE THIS NOTICE. Reproduced and communicated on behalf of the University of South Australia pursuant to Part VB of the copyright Act 1968 (the Act) or with permission of the
copyright owner on (29/3/08) Any further reproduction or communication of this material by you may be the subject of copyright protection under the Act. DO NOT REMOVE THIS NOTICE.

1
Introduction to Water Engineering

Slide 4 The intended learning outcomes from this presentation

Intended Learning Outcomes
are for you to understand how flow changes under
At the end of this section, students will be able to variable head, what this means for calculating the time
understand:-
- Flow characteristics under variable head
to empty a tank, and how flow moves between two
- Time required to empty a tank tanks.
- Flow between two tanks

Slide 5 In chapter five, we looked at relationships between

Recap: discharge-head effects head and flow for different situations. Hopefully you
• Table 5.1 (p. 152)
remember that there was generally a non-linear
• Note non-linear relationship relationship between flow and head.
• Q ∝ H1/2, H3/2 , H5/2 etc.
• e.g.

QA  CD A 2 gH (small orifice) For flow through an orifice the discharge was

proportional to the square root of head, for flow over a
2
3

QA  C D b 2 g H 2  H1
32 32
 (large orifice)
rectangular weir it was H ^ 3/2 and 5/2 for a triangular
weir.

Just looking at a couple of those equations – one for

small orifice

And one for a large orifice – we can see that the flow
isn’t directly proportional to the head, it’s proportional to
some power of the head. This means if our head is
changing – for instance if a tank is draining out of a
hole – the change in discharge isn’t going to be exactly
proportional to the change in head.

Slide 6 Okay, let’s say we’re interested in working out how long
Time required to empty a reservoir it’d take for a reservoir to empty. Well discharge is
• Q = Volume / Time
volume over time
• So T = V / Q ?
– No, because head in reservoir reduces as it empties
 Q also reduces and you could say that rearranging this gives time
equal to volume over the discharge rate.
• If Q ∝ H, we might be able to assume an average Q (at
0.5V),
BUT
• If non-linear relationship between H & Q, then we can’t
But that’s not right because the head goes down as the
assume an average discharge. tank empties out, so the flow goes down.

Now, if we had a proportional relationship between flow

and head, maybe we could assume an average flow
rate and go from there,

But we know we generally don’t have a proportional

head-discharge relationship so that’s not much use to
us.

2
Introduction to Water Engineering

Slide 7 Alright, let’s take this tank as a basic illustration

Uniform Cross Section

• T = Time to empty = ?? AWS = LB It’s got a length and width here

L

B And that gives us the water surface area

So we’re talking about starting out at a particular head

h1 of water
h2

QA Which gives us a high flow rate, which gradually

decreases as the driving head gets smaller and
smaller. What we’re going to need to do is use a bit of
Calculus to estimate the emptying time.

Slide 8 Right, so let’s build up the description of emptying time

Q varies as h decreases using calculus.
• Consider small time increment, δt
• Volume of water drained during time δt  QA δt We start out by considering a miniscule increment of
– This volume also equals -AWS δh
where δh is the reduction in h during time δt time, delta T.
– (minus sign compensates for δh being negative)
 AWSh  QAt
Now, during that little time increment, the tank drained
h2   AWS  T
out at a particular flow rate, and if we multiply that flow

h1

 QA
h   t  T
 0
h2  1  rate by the time increment we get the volume
T   AWS   h
h1
 QA  discharged during delta T.

The other way to work out the volume drained during

the little time interval is to look at the physical change in
volume in the tank, so that means taking the water
surface area and multiplying it by the change in height,
delta H. Area times height equals volume.

The negative sign’s important because delta H is

negative, so we need to put a minus sign in to give the
volume as a positive.

So we put these two expressions of volume together

and we’ve got Aws by delta H equal to Q by delta T.

Now we’re going to divide both sides by QA, and then

integrate. This means taking the H from H1 to H2,
which is going from the upper level H1 down to the
lower level H2, and on the other side we’re just
integrating delta T from the start, which is time equals
zero, to the end, which is T. So that just becomes T.

So for the special case where there’s a constant water

surface, which just means a reservoir with vertical
sides, the Aws term comes out and we’ve got the
emptying time T equal to Aws by the integral from h1 to
h2 of 1 on QA delta H.

3
Introduction to Water Engineering

Slide 9 So then you take that basic integration

Substituting QA
h2  1 
T   AWS   And sub in whatever discharge equation you need – for
h
 QA  example if you’ve got a small orifice, you chuck in the
h1

• Use discharge formula for the corresponding mechanism orifice discharge equation in stead of QA.
e.g. small orifice, weir, etc (Table 5.1)
• Sharp orifice:
QA  CD A 2 gh
Say that’d look like this
 1 
 h
h2
T   AWS 
h1  C A 2 gh 
 D 
AWS h2  1  And the equation for emptying time’d come out like this.
T 
CD A 2 g 
h1

 h
h
I know it looks a bit exciting but actually most of that
stuff inside the integration sign’s just constant

So we can pull it outside like this – constant water

surface area, constant discharge coefficient, constant
orifice area, so in this case all we’re having to integrate
is a simple “1 on root H”. Actually, it’s really not that big
a deal if any of those parameters happened to not be
constant – for instance if the water surface area
changes with water depth because the tank’s got
sloping sides – it just slightly changes what’s inside the
integration sign but at the end of the day we’re not
talking about really tough integration.

Slide 10 Uniform Cross Section: Let’s run through a simple example where we’ve got a
Example 7.1 small orifice and constant water surface area.
AWS h2  1 
T 
CD A 2 g  h1

 h
h
3m In this case the water level’s dropping from 1.5 metres,
2m measured above the centreline of the orifice, all the
way down to zero. So that gives you the values for H1
and H2 in the integration and everything else is given.
1.5m See if you can work it out and head to the textbook if
you need any help.
Diameter = 0.05m
CD = 0.61

Slide 11 Okay, now let’s cut our teeth on a more involved

Example 7.3: Variable x-section example. In this case we’re still considering a small
4m
orifice but we’re going to move away from the constant
water surface area. Now it’s a pyramid shaped tank
Diameter = 0.1m with all four sides sloping out.
CD = 0.60
4.5m
AWS  f (h) So this time we’ve got to take a step back and make
h2  AWS  sure we keep the AWS term inside the integration
T    Q h
h1
 A  because it’s not a constant.
QA  CD A 2 gh
1m
AWS is going to be a function of the height,

getting larger as you go further up the tank.

Luckily we’ve been given a bunch of dimensions we

can use to work out the function AWS – this might
mean going back to basic geometry for you though.

The equation for QA gets plugged in the bottom, same

4
Introduction to Water Engineering

as before, and anything that stays constant can get

taken out of the integration.

Give it a shot. If you need help – especially when it

comes to remembering how to work out a geometric
relationship like this – the textbook goes through it, but
just remember there are different ways you can go to
get the same answer and the way they do it might not
be the way you intuitively get there. See how you go,
anyway.

Slide 12 Alright, it’s all well and good if you’ve got a nice
Reservoirs with varying slope constructed tank where you can either assume the
• Simplified method using slices
sides are vertical or you can derive a neat equation for
(Box 7.2, p. 216) water surface area as a function of depth, but what are
you supposed to do if you’ve got a dam or reservoir in a
more natural landscape where there’s no easy
relationship between the surface area and the depth of
water?

The answer is that you need to simplify it down to slices

that behave in a mathematically predictable way, and
just put up with a pretty inaccurate answer as a result.
What we do in the simplified method is divide the
reservoir into slices that are preferably of equal
thickness, and we assume each one is a big flat thing
with vertical sides.

Slide 13 Providing we’ve got some sort of contour map we can

Reservoirs with varying slope work out the area of each slice in plan view, and
• Simplified method using slices
convert that to a volume by multiplying it by the slice
(Box 7.2, p. 216) thickness. Then we can make a rough assumption that
the time taken for each slice to empty is going to be the
V1
volume divided by an average flow rate, which we work
Ti = Vi / Qi out using whatever discharge equation is appropriate –
V2
in this case let’s assume the reservoir’s emptying
H1

H2

H3 V3
through an orifice so the discharge is proportional to H
Ttotal = ΣTi ^ (1/2).
Q∝ (H321)1/2

So the first slice empties with a discharge proportional

to some sort of average head. Or we might work out
the discharge based on the head at the top and bottom
and take the average.

For the next slice, we jump to a different discharge

because we’re at a lower elevation

And likewise for the next slice. For each slice, the flow
rate’s assumed to be constant but it changes from one
slice to the next.

Since we’ve got a reasonable idea of the volume of

each slice, we can work out the approximate emptying

5
Introduction to Water Engineering

time by dividing the volume by the average discharge

rate

And finally by summing together all the emptying times

of all slices, we get an approximate value of total
emptying time for the reservoir.

Slide 14 So here’s an example of the slice method. We’ve been

Example 7.4 given a contour map of the reservoir and someone’s
figured out the area of each contour.
Neglect friction
Entrance loss K = 0.5

The first one’s about 6 hectares, or 60,000 square

2
60,600 m2 VB
Z + entrance loss
2g
22,400 m2 metres
3,200 m2 Diameter = 0.8m

 A  A2   H Z = 40  VB = ?
The next one’s about 2 hectares
Vol1  1
2 Z = 30 etc.

Ti = Voli/Qi
T = ∑Ti
And the last one’s about 3,000 square metres. We’ve
also got a cross-section view down the bottom showing
the elevation of each contour line.

To work out the volume of each of these slices we take

the average area from the upper and lower contours

and multiply by the height difference.

To work out the average flow rate for each slice, this
time we don’t have a simply discharge equation – we
actually need to solve for velocity using the energy
equation and then multiply by the pipe’s cross-sectional
area to get flow rate. So take the elevation at the top
and bottom of each slice and solve for VB.

Repeat this for each elevation to get a value of velocity

at each depth and convert each one into a flow rate
using the continuity equation – we’ve been told it’s a
0.8 metre diameter pipe. Each slice then gets an
average flow rate based on the upper and lower values.

Once you’ve worked out the average flow rate for the
slice you can get the approximate emptying time for
that slice

And sum them together to get the total approximate

emptying time. As always check in the textbook if you
need help with the worked example.

Image source- Les Hamill 2011, Understanding

hydraulics.

6
Introduction to Water Engineering

Slide 15 Just as a point of caution,

Caution: slice method is rough!

• Provides an approximation only

The slice method only provides an approximation,
• Assumes constant (average) Q over the time interval to
empty each slice
• Becomes a question of cost/benefit as it assumes constant discharge over the time interval
i.e. time taken versus necessity of accuracy
to empty each slice. Depending on the thickness of the
slice, this might be relatively accurate or very
inaccurate.

It really becomes a question of cost/benefit, meaning

the time taken to get a precise and accurate answer
versus the benefit that comes from having a high level
of accuracy. If all you need is a rough idea, like,
whether the emptying time for a dam is going to be
closer to three hours versus three days, then the slice
method is good because it gives you a quick answer.
But if you need a very accurate prediction, you’ll need
to move to a more sophisticated model.

Slide 16 By way of an example, let’s take a situation where we

Testing accuracy of slice method know we can get the true answer using integration and
• Example 7.5
then use the rough slice method to see how close we
θ/2 = 30
get.
h2  1 
T   AWS   h 0.8m
h1
Q Here we’ve got water discharging out of a tank via a V-
0.2m notch weir.
8   5
Q CD tan  2 g  H 2
15 2

20m
It’s a pretty big tank, 20 metres long and 6 metres wide
6m

Theta / 2 for the weir’s 30 degrees

And we’re interested in the time taken to get from an

initial height of 0.8 metres

Down to a height of 0.2 metres above the bottom of the

triangular weir.

So since we’ve got vertical sides, we can assume the

water surface area’s a constant and our basic
integration looks like this

And we’re going from H1 = 0.8 metres to H2 = 0.2

metres in the integration.

The final thing to do is chuck in the appropriate

discharge equation, which for a V-notch weir looks like
this. I know it’s a little scary when you first look at it, but
don’t forget you can take anything out that’s a constant,

which basically means all of this – so you’re only left

with a simple H ^ (5/2) term that you have to integrate.

After you’ve integrated that and found the true value of

the emptying time, work out the approximate emptying
time using the slice method. Go with three slices, each
one 0.2 metres thick – so you’ll have four elevations to
consider – 0.8, 0.6, 0.4 and 0.2 metres. You can build
up your areas, volumes, flow rates and discharge times
for each slice just the same as last time and work out
whether there’s any significant difference between the
7
Introduction to Water Engineering

true method by integration versus the approximate

method. Check the textbook if you need help.

Slide 17 So far so good I hope! Now let’s step it up and look at

Flow between two tanks how you work out the flow between two tanks. If we
• Example 7.7
start with a tank like this, discharging to the
atmosphere, it’s just like the other examples we’ve
looked at.
4.5m

4m
4m

Slide 18 But what happens if instead of discharging to the

Flow between two tanks atmosphere, it’s actually draining into another tank? In
• The head difference
driving the flow is hD,
this example it’s draining into a narrower tank, with a
which changes smaller water surface area – this means that for a given
constantly
water level drop in the big tank, it’ll rise a lot more in the
hD
smaller tank.

The driving head causing the flow is the difference in

0.5m
water level between the two tanks – as the water level
2m
drops in the big tank and rises in the small tank, the
2m
overall head difference reduces

and eventually as the water levels equalise, the driving

head becomes zero and there’s no longer any flow
between the tanks.

8
Introduction to Water Engineering

Slide 19 Right, so in case you hadn’t picked it, the way to work
A little integration  this out comes back to integration. The thing that
Consider small time interval δt makes this one a bit trickier to work out than the first lot
A2
δV = QA.δt
A1 of integration we did is that in this case, we don’t
=A1.δx actually know what the final water level’s going to be.
=A2. δy
What we do know, though, is that the end point is when
δy = δx.(A1/A2) the driving head HD drops to zero. So we need to set
δhD = δx+δy
QA up the integration with that end point in mind.
= δx.(1+(A1/A2))

Let’s kick things off by considering a tiny increment of

time, delta T.

Now, just like the first lot of integration, we’ve got a

change in volume equal to the flow rate times delta T.
And obviously we’ve got the same problem as always,
which is that the flow rate changes because of the
change in driving head.

So just like before, we can also express the change in

volume as the water surface area times the change in
height. In this case we’ll call the change in height of the
large tank “delta X”

And because the volume leaving tank 1 is entering tank

2, we can also express the same volume as the area of
water surface in tank 2 times the rise in water level,
delta Y.

Rearranging these gives us delta Y in terms of delta X

and the two water surface areas.

Now the overall change in the driving head, delta HD, is

equal to the sum of the change in water level in both
tanks.

We can use the relationship with the surface areas to

get this all in terms of delta X.

Image source- Les Hamill 2011, Understanding

hydraulics.

Slide 20 Alright, so going back to the first part where we equate

Integration the two different expressions for the change in volume
QAt  A1x QAt 
A1
hD
 A  And using our expression for delta HD
 1  1 
A1 
hD  x1    A 2 

 A2 
A1 1
t  hD
x 
hD  A1  QA Rearranging to give delta X instead
 A1  1  A 
 2 
1  A 
 2 
T A1 0 1 We can substitute this into the equation
 0
t 
  hD
A  H QA
1  1 
 A2 
Substitute in appropriate equation for
That gives us an expression like this,
submerged orifice flow

and now it’s just a matter of rearranging

into something we can integrate. So we integrate the

time side which gives us the time taken to equilibrate
the two tank levels, and on the right hand side we end
up with a whole lot of constant areas and an integral of
1 on Q by delta HD.

9
Introduction to Water Engineering

The important thing here’s the interval over which we’re

integrating – remember we’ve reorganised this to be in
terms of the head difference, HD, which is what’s
driving the flow. So the integration goes from the initial
head difference, which we could call H, to the final
head difference, which for equilibration’s just going to
be zero.

As with the other examples, you substitute the

appropriate equation for the type of discharge you’ve
got – bearing in mind in this case it’s going to be an
equation for submerged flow, so it needs to be
expressed as a function of the head difference HD.
Check out pages 137 to 138 back in chapter 4 of the
textbook for a recap on submerged orifices. Meanwhile
go through example 7.7 carefully in the textbook to get
your head around the method we’ve just been through.

Slide 21 So, in summary, we’ve looked at flow characteristics

Summary under variable head, in particular focusing on how this
affects our calculation of the time required to empty a
• Flow characteristics under variable head tank or for water levels to equalise between two tanks.
• Time required to empty a tank
• Flow between two tanks

Slide 22 If you’ve got any questions or need further clarification,

please post a question or comment on the Discussion
Forum.

Thank you

10