thesis_presentation_28_07_2015

MASTERS THESIS DEFENSE
Adaptation of Water Distribution
Networks for Urban Renewal
BY
MUSTAFA E. ÇELİK
ADVISOR: ASSOC. PROF. DR. NURİ MERZİ
JULY 2015
Civil Engineering Department Middle East Technical University

Outline
 Introduction
 Theoretical Considerations
 Methodology
 Case Study
 Summary and Conclusions

Introduction
 Started in England and France in
19th century due to
industrialization
 Came Turkey’s agenda in 1980s
due to increase in shanties
Dikmen Valley Urban Renewal
Project
 To determine maximum possible
outflow at a single node without
disturbing hydraulic conformity
criteria in a WDN
 Necessary to investigate capacity
of existing system whether it can
respond changes due to demand
increase
History of Urban Renewal Aim of the Study

Theoretical Considerations
Components of WDN
 Distribution reservoirs, pipes, valves, fire hydrants,
pumps

Pump Station Operations

Pump Characteristic Curves

 For the same discharge,
heads are additive
Pumps in Series Pumps in Parallel
 For the same head,
discharges are additive

Outline
Introduction
 Methodology
 Case Study

Methodology
In this study, sensitivity analysis proposed by Bhave
(1991) is used as a main guide.
What is sensitivity analysis?
It can be defined as determining reaction of water
supply system if some parameters like nodal
demands change at a single point.
S.A. establishes upper and lower bounds for input
data. Within boundary, input data can vary without
causing violent changes to the optimal solution.

Methodology
Example (Bhave, 1991)
Using sensitivity analysis
determine the max. possible
outflow at node 3 w/o affecting
the outflows at all other nodes?

Methodology
 To make sensitivity analysis, HGL values
should be known.
Linear Theory using nodal head equations
has been applied in this study to find HGL
values.
Same problem has also been solved by
Watercad for comparison.

Methodology
 In general, node-flow continuity equation
can be written as

Methodology
 Node-flow continuity equation can be
modified as
Node Head Analysis (NHA) by Linear Theory

Methodology
The linearized H equations at nodes are

Methodology
Equations can be rearranged and expressed
as

Methodology
Equations can be written in the matrix form as
Discharges in pipes are assumed to be 0.1 m3/min for the first
iteration.

Methodology
After performing matrix operation, HGL values will be
H2(1) = 99.44 m, H3(1) = 99.13 m, H4(1) = 99.58 m, H5(1) = 99.17 m,
H6(1) = 98.93 m
First and second iteration values are given below

Methodology
 The iterative procedure is stopped when
εmax is less than 0.0001.
εmax is the max ε value at tth iteration.
ε values belong to first iteration are
εmax

Methodology
Final values obtained by Linear Theory are
compared with Watercad results

Methodology
 After Node Head Analysis (NHA), S.A. can
be made now.
When the nodal demand at node 3
increases, HGL values of all nodes decrease
until some point where minimum permissible
HGL value is reached.

Methodology
 The node-flow continuity equations at
nodes are
Sensitivity Analysis

Methodology
Let yk be the parameter which changes and we want the
sensitivity analysis of the network for the change in yk.

Methodology
When derivative of Fj with respect to yk is
taken, the matrix form will be

Methodology
Substituting the values of the partial
derivatives

Methodology
Summary table is given below

Methodology
Working Range of the Technique Proposed by Bhave
A graph of z = x2 + xy + y is
given in Figure. This graph
is given because it gives an
idea about visualization of
simple nonlinear equation
in space
Permissible increase in q3 is
found minimum at node 5

Methodology
Working Range of the Technique Proposed by Bhave
A slice of z = x2 + xy + y at y=1 and ∂z/∂x=3
at (1,1,3)
𝜕H5
𝜕q3
is actually a slope. Straight line coming from
derivative intersects H5 equation like
intersection of z= x2 + xy + y at y=1 and
𝜕z
𝜕x
= 3
As it can be seen from Figure, this method
allows working range named as ∆.
If ∆ value
(H5,initial (obtained by linear theory) − H5,min) is
chosen much like 2, straight line can not fit
nonlinear equation.

Outline
 Introduction
 Methodology
 Case Study
Computer Science Department Virginia Polytechnic and State University

Case Study
 Study area is N8.3 pressure zone located at Keçiören and partially
at Yenimahalle districts.
N8.3 consists of six DMAs. These are Northern and Southern
Sancaktepe, Şehit Kubilay, Upper and Lower Çiğdemtepe, Yayla.
At entrance of the DMAs, there exists measurement chamber.
Pressures and discharges are measured through measurement
chambers.
Discharges measured through measurement chambers are used to
obtain daily demand curve of each DMA.

Case Study
STUDY AREA

Case Study

Case Study
 Some buildings have already been replaced by
residences due to urban renewal in N8.3 pressure zone.
An example for application of urban renewal in
Çiğdemtepe neighborhood can be seen below.
Residence Single-storey building

Case Study
 Case study is conducted in Yayla DMA.
Within urban renewal, there will be demand increase
due to residence(s).
It can be asked that how much area in terms of hectare
is influenced due to change in discharge.
Answer of this will be given at the end of this section.

Case Study
 Many facilities like high rise buildings, schools,
shopping malls have been building and some shanties
have been pulled down according to the urban renewal
based on the article 73th of Municipalities Law No.5393.
This will cause population increase in the near future for
N8.3 pressure zone among the others.
Generally increase in consumption of potable water
accompanies problems.
For example, if new residence has 30 stories, residents
at the upper stories may not use water due to low
pressure head.

Case Study
 Sensitivity analysis gives an idea about the critical region, which could have a
pressure problem.
The methodology employed here allows the user to overcome following situations:
• In general, the study area (pressure zone) has been assumed to have a pressure
head of 40 m for the existing situation.

Case Study
• The developer is assumed to select the node at which the nodal demand is to be
maximized without disturbing existing nodal pressure heads in the neighborhood
(DMA).
• If at the end of the analysis, permissible increase in discharge at a particular node
without disturbing other regions is higher than existing situation, developer may
want to extract this discharge by adding new stories to high rise building.

Case Study
• However for some cases low pressure head problem can be faced
with and this is developer’s responsibility. Hydrophere can be
used to solve the problem.

Case Study
• How about if one developer has decided to develope the area (red one) around a node with a
"very high" demand so that other developers cannot find a possibility to realize similar projects?
NHA of neighborhood municipality should specify minimum pressure head (40 m in this study)
S.A. allowable nodal demands

Case Study
For operational reasons, the network has been run as DMA.
Insulated system has been studied under five different scenarios
by Şendil (2013). These are:
• INS-1PMP (SM)
• INS-1PMP (LG)
• INS-2PMP (SM+SM)
• INS-2PMP (SM+LG)
• INS-3PMP (SM+SM+LG)
Here SM and LG indicate small and large pump.

Case Study
Scenario codes and their definitions of N8.3 pressure zone (Şendil,2013) are given
below
According to hourly energy prices, there exists two type of energy prices, namely
Constant and Multi Tariff

Case Study
Calculated Energy Costs for CT Calculated Energy Costs for MT
Since minimum energy cost is obtained at INS-3PMP (SM+SM+LG)-MT,
which has energy cost of 123.456 TL/day, results of this have been used
while studying Yayla neighborhood.

Case Study
Junction number is decreased from 147 to 40 by Watercad
Skelebrator for Yayla neighborhood.

Case Study
Before carrying out sensitivity analysis, existing pump combination
graph is drawn to understand capacity of system.
Since pump combination is parallel, discharge values are additive
for the same head.

Case Study
Maximum discharge drawn from N8.3 is 650 m3/hr.
Extra 150 m3/hr can be drawn from the system without adding new
pump to the current system.

Case Study
Twelfth hour is selected while studying Yayla neighborhood.
JC-0702 is selected as a source node because it is close to the outlet
of pumps.

Case Study
Since pumps are working at 22:00-06:00 and 15:00-16:00, at 12:00
water comes from T53 to the junction JC-0702.
Discharge value drawn from the network will be higher after
carrying out sensitivity analysis.
Therefore HGL value of JC-0702, which is 1153.16 m at 12:00, will
be less.
Any analysis related to tank T53 has not been made within the
scope of this study.

Case Study
Worst scenario is considered for JC-0702. It is assumed that HGL value of JC-0702
equals to 1151.69 m while applying sensitivity analysis.
JC-0702 is replaced by a reservoir, which has a HGL of 1151.69 m.

Case Study
Height of T53 is 6.50 m. At 12:00 water level is 3.53 m, which is higher than half
height of tank.
In this study, it is assumed that water in tank would be enough for the network at
12:00 after completing sensitivity analysis.

Case Study
 NHA has been made with Linear Theory and Watercad to compare HGL values.
 Since results are close to each other, it can be said that Linear Theory is a good
alternative for obtaining HGL values.

Case Study
After obtaining HGL values, S.A. can be made now.
Since Yayla DMA consists of 40 nodes, it is difficult to show location
of all nodes in screenhoot of Watercad.
Instead of this, three nodes are selected.
Two of them are critical because they are at highest and lowest
elevations.
The other is selected in between these two nodes.
 Location of three selected nodes for Yayla DMA will be given in
next slide.

Case Study
Location of Three Selected Nodes for Yayla DMA

Case Study
Before giving results of S.A. of all nodes in Yayla, flowchart is given below.

Case Study
Permissible increase in discharge values of all nodes in Yayla neighborhood are given
below

Case Study
S.A. has been completed now.
In this study, each node represents an (influence) area.
Area value of Yayla neighboorhood was given previously.
Since municipality may want to see how much area will be influenced due to change in nodal
demand, weight of nodes has been calculated to obtain influence area of each node.
Weight of each node is obtained by demand distributor macro (Şendil,2013).

Case Study
Weight results page is given below.

Case Study
Weight of each node is multiplied by area of neighborhood to
obtain influence area of each node.

Outline
 Introduction
Theoretical Considerations
 Methodology
 Case Study

Summary and Conclusions
Shanties and abandoned regions have been replaced by residences;
this brings consumption increase at existing nodes.
This study underlines two important points.
Capacity of existing WDN Minimum pressure head
-max. discharge drawn from existing -minimum pressure
system is 650 m3/hr, this value can be head obtained by Linear Theory is 40.31 m,
increased up to 800 m3/hr without
adding new pump. this value was allowed to decrease 40.00 m
in this study.

To deal with nonlinear equations in WDNs, Linear Theory can be applied even to the
neighborhood with 40 nodes and 58 pipes.
To check accuracy of minimum pressure head value of 40.00 m obtained by sensitivity
analysis method, nodal demands were increased by permissible amount and network
analysis was refreshed 40 times.
Results show that minimum pressure heads range between 39.75 m and 39.90 m after
demand increases.
Bhave (1991) says that “results taken from sensitivity analysis are not exact because they
are derivatives at a point but it is a fast way for comparing possible changes “.
It can be stated that the method proposed by Bhave (1991) is searching a solution near
Node Head Analysis (NHA). If Δ value is chosen much, permissible increase in discharge
value obtained by sensitivity analysis will get away from the real value.

Location of Most Suitable Node (JC-0118) and Most Unsuitable Node (JC-0433) for Yayla

Each node represents an influence area. With the help of Mapinfo and demand
distributor macro (Şendil, 2013), the municipality can see the magnitude of the
influence area for each node within the scope of urban renewal.
This study is a preliminary work for determining maximum discharge that can be
drawn from a single node without affecting other nodes in the network.
All calculations have been made according to the nodal demands.
Any developer can benefit from results of nodes used in this study.
For future studies, 2 or 3 nodes could be selected at the same time and their effects
on system could be investigated.
Calculations could be made according to the parcels instead of nodes to approximate
real life scenario.

Thank you!!!!
QUESTIONS ???

thesis_presentation_28_07_2015

More Related Content

Viewers also liked (8)

Similar to thesis_presentation_28_07_2015 (20)

thesis_presentation_28_07_2015