1

UNIVERSITY OF MANCHESTER

SCHOOL OF COMPUTER SCIENCE

COMP 60990: Research Methods and Professional Skills

A "traffic light" Decision Support Tool for Water Engineers

Progress Report

Student Name: Muhammad Zuhairi Bin Zainal

Supervisor: Dr. John Brooke

Programme: MSc Advanced Computer Science

2

ABSTRACT

Dependability of Decision Support System (DSS) to the availability of communication

network has occasionally limits the capability of Water Distribution System (WDS)'s

field workforce in decision making. Communication network serves as transportation

mechanism to transfer data between client's device and server systems. As a fact,

some factors have impact on the performance of communication network, for

instance network loads and geographical constraints. Some of the factors are

uncontrollable by human capabilities. Therefore, there is a need to have alternative

to these uncontrollable factors.

Cache technology is considered as one of the alternative solution for communication

network unavailability. Cached data stored in client's local memory will enable the

DSS's application to works under minimal resources and would still enable to

support field workforce in making decision.

3

Table of Contents

ABSTRACT ..................................................................................................................... 2

Table of Contents ............................................................................................................ 3

1. INTRODUCTION ...................................................................................................... 5

1.1. Problem Statement ............................................................................................ 6

1.2. Project Objectives .............................................................................................. 6

1.3. Project Challenges ............................................................................................. 6

1.4. Report Outline .................................................................................................... 8

2. BACKGROUND RESEARCH ................................................................................... 9

2.1. Water Distribution Systems ................................................................................ 9

2.1.1. Overview of Water Distribution System ..................................................... 10

2.1.2. Leakage in WDS network .......................................................................... 12

2.1.3. District Meter Area ..................................................................................... 13

2.2. Decision Support System ................................................................................. 14

2.2.1. Decision Support Tool for WDS Field Operation ....................................... 15

2.2.2. Static Contextual Knowledge ..................................................................... 16

2.2.3. Dynamic Contextual Knowledge ................................................................ 17

2.3. Distributed Systems, Wireless Communications and Cyber-Physical Systems 18

2.3.1. Introduction to Distributed Systems ........................................................... 18

4

2.3.2. Introduction to Wireless Communication ................................................... 21

2.3.3. Introduction to Cyber-Physical System ...................................................... 23

3. METHODOLOGY ................................................................................................... 25

3.1. Study of overall project ..................................................................................... 25

3.2. Software Development and Testing Methodology ............................................ 25

3.3. Updated project plan ........................................................................................ 29

4. RESULT AND DISCUSSION .................................................................................. 30

4.1. Review and study of existing project documentations ...................................... 30

4.2. Server system set-up for testing and development .......................................... 32

4.3. Hands on existing prototype client application ................................................. 33

4.4. Proposed improvements in technology, implementation, functionality and

usability ...................................................................................................................... 34

5. CONCLUSION ........................................................................................................ 36

6. REFERENCES ....................................................................................................... 37

5

1. INTRODUCTION

Operational activities for field water engineer are crucial and require thorough pre-

execution analysis and studies. A trivial mistake at the field during operation may

lead to disastrous consequences to many of components (for example valves and

pipes) involved in the Water Distribution System (WDS) network, Therefore, a

predictive Decision Support Tool (DST) that is able to predict accurate result from

any on-field activities carried out is necessary to the WDS stakeholders to ensure

the stability of WDS.

A solid distributed computing application framework was established in previous

research. Thus the current research will be an extension from the existing

framework. Collaboration with Thames Water Utilities LTD, a UK's largest water and

wastewater provider, was formed to enable research in WDS area to be conducted

with real-world data and scenario.

In current research context, DST is an application that runs on mobile devices,

and equipped with functionalities to assist field workforce in making decision of

proceeding with scheduled planned provided by central control room. The examples

of functionalities are listed in Section 4.3 and further explanation of the usage of the

functionalities can be found in [2]. The DST application in this project offers support

services to components scattered in a WDS network. Main purposes of using DST in

WDS area are [1]:

to provide access to the static and dynamic contextual information, that are

generated from computational model, in order to locate a problematic area

to assist in analyzing the structure and attributes of physical system and

its components

to help in determining the impact of possible decisions on customers,

regulatory requirements and state of the system

6

to assist in rectification plan preparation (based on up-to-date system state),

and scheduling planned decisions in order to coordinate concurrent

engineering decisions, and so forth

1.1. Problem Statement

Availability of communication network between central servers with distributed

DST applications on the field is considered to be a problem to be analysing in

this research. Communication network availability might vary from one location to

another. Consequently, it affects the reliability of DST application to access

loads of data in the server. This issue might cause the field workforce in failing to

retrieve accurate and real-time data and leads to a delay in actions since they

have to search for available network connection and in some worst case, they

might have to go back to the office and reassess the scenario. This can be

considered as time lost in business operational perspective since it involves time

wasted to collaborate between field workforce and remote personnel.

1.2. Project Objectives

To propose a solution that able to support field workforce's decision making

process in any possible condition in the field.

A study about the network communication reliability to cater for different

possibilities of wireless network availability.

1.3. Project Challenges

Real-time aspect has been among recent goal in many Cyber-Physical System

(CPS) research. The emergence of Internet, advancement of communication

network and sophisticated computing devices are some of the factors that boost

the demand for real-time attribute in CPS. In addition to advancement in

technology, the fact that CPS is dealing with physical world adding the demand

for the real-time attribute in CPS implementation.

7

Reliable real-time data from physical world is expected to be sending to the

servers in cyber world to be compute and process. The data transportation

requires low latency, if not zero, between physical world and cyber server.

However, the availability of communication network itself is considered to be a

pre-requisite to have real-time CPS.

There are many factors have effects on the performance and the availability of a

network, namely signal sharing, network usage and load, local environment

characteristics, and network range and distance between devices [41]. Some of

these factors are beyond human's control. Thus, ensuring the availability of

communication network is one of the challenges in implementing real-time

distributed CPS.

Another identified challenge to this project is data consistency during cache

synchronization with database in case of device re-connected to communication

network. The flow of the cache implementation for this project will be discussed

in Section 4.4. As this research dealing with distributed CPS, there is tendency

that numbers of distributed field workforce make changes or updates to the

earlier identical cached copy stored inside each individual's device. The

complication may arise during data synchronization to the server once their

devices reconnected to communication network and may affect the consistency

of data in database server. In supporting this claim, Tanenbaum et al. 1 agree

that it is hard to ensure consistency in large-scale distributed system.

1 Tanenbaum, Andrew S, and Steen. Distributed Systems: Principles and Paradigms. Upper Saddle River, NJ: Pearson Prentice

Hall, 2007, p273

8

1.4. Report Outline

Chapter 1 provides an overview about the project, problem statement of the

project, objectives to be achieved by performing the research, and finally general

challenges anticipated for the project.

Chapter 2 describes the background research related to the project. This chapter

will cover elaboration on subjects about Water Distribution Systems, the needs

for Decision Support Systems in Water Distribution area, and technological

aspects that contributed to the development of Decision Support System.

Chapter 3 discuss the methodologies involved in this research. Performed and

planned activities such as project timeline, software development, and software

evaluation will be explored in depth.

Chapter 4 describes results and discussion of to-date project's progress, project's

improvement suggestions based on studies and analysis done from Chapter 3.

Chapter 5 and Chapter 6 serve as conclusion of overall project and list of

references respectively.

9

2. BACKGROUND RESEARCH

2.1. Water Distribution Systems

Demand for reliable and high-quality fresh water supply increases from time to

time. Apart from domestic consumption, water managers also have to be able to

handle staggering demand from industries and agricultural sectors. Metals, wood

and paper productions, gasoline and oils are only few numbers of industries that

heavily depending on water [5]. As a result, European Environment Agency

(EEA), an independent agency that studies, evaluates and provides sound

information about environment, anticipated that there will be a struggle to meet

water needs in the future due to appreciation of demand for water and because

of climate change[6].

Awareness about the importance of effective Water Supply System management

has been extensively developed by many parties. Various alternatives has been

discovered and been proposed to ensure the sustainability of water supply. A

team consists of top academics from Newcastle and Oxford University, has

published academic papers related to usage of nuclear power that could result in

increase of tidal and coastal water by almost 400% [7]. They had also studied

about the potential of using gas or other fossil fuels with high levels of carbon

capture and storage (CCS) that could lead to increment of fresh water

consumption by almost 70% [7]. In agricultural sectors, EEA highlighted several

alternatives, ranging from efficient water irrigation, educating farmers on best

practises on efficient water consumption and finally, the possibility of using

treated wastewater to meet agriculture's demand for water [6]. Daniel Loucks et

al [4] conducted a comprehensive study on how to use quantitative analysis, and

in particular computer models, to support and improve water resources planning

and management. Besides that, United Kingdom's government has also played

significant role in committing to secure water supplies, high standards of drinking

water and effective sewerage services by enforcing policy that require water

companies to update their plan every 5 years on [8]:

10

how to deal with future populations needs

climate changes

develop - where needed - new water supply resources such as reservoirs

Water Supply Systems comprise of 3 major components, which are [2] [10]:

raw extraction and transport

water treatment and storage

clear water transport and distribution

This research will be focusing only on third component as Water Distribution

Systems fall into this category.

2.1.1. Overview of Water Distribution System

Water Distribution System, or simply WDS, is a network of water supply

resources such as pumps, pipelines, storage tanks, and other appurtenances

[3][9]. It is a subsystem to a larger system which is Water Supply System. The

definition for Water Supply System is "infrastructure for the collection,

transmission, treatment, storage, and distribution of water for homes,

commercial establishments, industry, and irrigation, as well as for such public

needs as fire fighting and street flushing" [3]. Description about some of

elements inside WDS can be referring to Table 2.1. Main functionality of WDS

is to be able to carry water from treatment plant (or from the source in the

absence of treatment) to the consumer [9]. Optimal water pressure is

necessary to ensure the continuity of operating plumbing fixture and avoid

leaks and pipeline breaks [3].

11

Table 2.1:Basic elements inside WDS

Figure 2.1:Water Supply Distribution System[9]

12

2.1.2. Leakage in WDS network

Water Distribution System is among the largest distributed infrastructures built

in UK and due to its scale and component's age, daily operational problems

has never been easier to the stakeholders. Examples of the issues that have to

be solved are pipes leakages, malfunctioning of valves, and discoloration and

water contamination [17]. Leakage in WDS context is water loss in a WDS

network that may caused by burst pipes, sudden rupture of a joint and fittings,

and overflowing service reservoirs [14]. Two groups of factors identified that

influenced leakage, which are [13][14]:

Technical factors. In this group, qualitative and quantitative technical

aspect of the WDS infrastructure that cause the leakage. For instance,

conditions of the infrastructure such as pressure, soil movement, pipe

condition, poor quality material, fittings and workmanship (e.g. poor

installation activities), traffic loading and also due to leakage control method

itself.

Social factors. This factor is perceived to be human incompetency to have

proper infrastructure planning, institutional (e.g. government, national and

local agencies) attitude towards leakage problems, and policies to control

leakage.

Leakage in WDS may affect in several ways [12][14]:

consumer inconveniences

collateral damage to larger WDS network and infrastructure

substantial amount of cost incurred for compensation and repairing

requirement to over-design sewer capacity as a result from increased

loading on infiltrated sewers

error in water bills because of introduction of air in WDS network

13

pollution in pipe network that may also lead to health risks

Given numbers of inconvenience consequences due to leakage in WDS, there

is a need to have leakage monitoring system and rectification plan to overcome

leakage. Basically, three identified strategies in leakage management, which

are [14]:

Passive control. Rectification activities after bursts or drop in pressure

Regular survey. This includes listening to pipe work, and manual inspection

to the distribution system

Leakage monitoring. Monitoring the water flows into zones to measure

leakage and prioritize detection activities

District Meter Area technique falls into third strategy and its introduction has

been considered as one of best leakage monitoring technique that able to

produce the best result [11].

2.1.3. District Meter Area

International Water Association (IWA), an association that gather water

professionals and has ambition to provide innovative solutions for water

challenges, had produced an international standard for terminology of Water

Supply System to avoid misunderstanding about terms in Water Supply System

domain. This initiative is a consequent from the increment of attention and

interest from water experts all around the world that having difficulties to

understand each other because of differences of definitions used by individual

countries [12]. In one of its published documents, John Morisson et al. [13]

define District Meter Area (DMA) as a "discrete area of a distribution system

usually created by the closure of valves or complete disconnection of

pipe work in which the quantities of water entering and leaving the area

are metered."

14

DMA approach has been one of heavily used approach in water distribution

leakage management. The rationale underlying this approach is to divide large

WDS network into smaller and scalable areas or zones, and being able to

regulate the flow measuring system in these zones. In this approach, flow

meters are installed at strategic points throughout the WDS [11][13]. Main

purpose of the flow meters are to record flows of water that flow into a zone.

Figure 2.2 depicts example of DMA installation in WDS network.

Figure 2.2:Complete Water Supply System inputs and outputs [12]

2.2. Decision Support System

Researches in Artificial Intelligence(AI) has contributed to undisputable advanced

technology development to various area in human life, including pharmaceutical,

gas production plant, fermentation process, telecommunication, and

environmental problems managements system [18][19]. Among the main

objectives of AI development is to have capability of computer to function like

15

human brain by being able to gather information, develop logical hypotheses and

thus complement human's decision making ability [20]. Environment Decision

Support Systems (EDSS) are one of the systems that applied the advancement

of AI in environment field, which related to WDS.

At glance, EDSS is an intelligent information software system that integrate

environmental models, databases and assessment tools which are integrated

under a graphical user interface (GUI), often realized by using spatial data

management functionalities provided by geographical information systems (GIS)

[18][21][22][23]. Ideal EDSS tools should incorporate two components as its

source of information, which are [1][2][18]:

Qualitative or analytical component. This includes personal knowledge,

obtained through education and experiences, that produce personal

knowledge

Quantitative component. Mathematical calculations and computer

modelling that produce contextual knowledge, a type of knowledge that is

specific to system involved.

2.2.1. Decision Support Tool for WDS Field Operation

Eventhough there has many DSS and tools develop to solve environmental

problems, the chances of developed DSS failing to meet with real world

problems are reported to be high [24]. Uncertainty and approximate knowledge

are some reasons behind the failing. In some environmental DSS test cases,

data inputs to be process in environmental field are too many, yet with limited

available knowledge that able to digest the data correctly and accurately [23].

In the current research project of WDS domain, field operation decision making

process receives both type of knowledge that was previously mentioned in

Section 2.2, personal and contextual knowledge. Contextual knowledge is

perceived to be complement knowledge for field personnel that possess

16

personal knowledge, in their decision making process. Combination of these is

an ideal approach as to reduce the probability of DSS failing because of

knowledge limitations.

Figure 2.3:Components inside EDSS [23]

2.2.2. Static Contextual Knowledge

Static contextual knowledge is a kind of knowledge that will not vary over time

such as location of pumps, size of pipes, and type of valves. In contrast, a

dynamic contextual knowledge has the tendency to vary from time to time,

depending on many factors, for instance consumption demand from

consumers, and water pressure and flows that differ if some elements in WDS

infrastructure are altered.

GIS is a computer system that enable spatial data handling, map creation,

geographical modelling and simulation that can be use to assist in

understanding some complex interrelationships of natural resources, human

population and their ecological interactions [26]. GIS has been among the

biggest contributors in EDSS field and continuous research of implementation

17

of GIS in EDSS has started even before 1980's. Examples of GIS application in

EDSS field are study of prevention of groundwater deterioration, visualization

and mathematical calculation of the distribution of predicted environment

concentration (PEC) of down-the-drain chemicals in European water surface on

a river, and a study of GIS/remote-sensing techniques to estimate oil erosion

losses from intensive agricultural activities in a watershed in Sao Paulo

State [27]. In this research, GIS application is one of static contextual

knowledge provider as it provides the built application with the capability to

display and zooming geographic map, and functionalities to create simulation

using static data of water network [2].

2.2.3. Dynamic Contextual Knowledge

According to [1], one of the functional requirements of the currently developed

DST prototype for this thesis is to be able to run water quality modelling, a

computer modelling applications that consist of two types of simulations, which

are hydraulic and water quality simulation. The result from running hydraulic

simulation is to be used to predict spatial and temporal behaviour of water

quality, while running the water quality simulation provides insight on the

criticality of water properties such as concentration of chemical in the water and

water age [2]. The simulations are using dynamic data that may change from

any point of times because of many factors. Table 2.2 shows the identified

sources for dynamic data and its usages:

Data Sources Usage in Hydraulic simulation

Supervisory Control And Data

Acquisition (SCADA) system

Set status & speed of pumps

Set status & settings of valves

Set water level in tanks

Wireless Sensor Networks (WSN) Set pressures at junctions

Set flow rates in pipes

18

Field workforce Set status of pipes

Set status & speed of pumps

Set status & setting of valves

Set water level in tank

Table 2.2:Dynamic data sources and their usages in currently developed system [1]

2.3. Distributed Systems, Wireless Communications and Cyber-Physical

Systems

2.3.1. Introduction to Distributed Systems

Recently, development of distributed system has been in hasty pace had

resulted many advancement in technological area. Ranging from high-

performance mainframe computers to small nodes in sensors network,

distributed system has been one of the backbone behind their operations.

Although there were many definition about Distributed System defined by

scholars, but the most famous is [35]:

A distributed system is a collection of independent computers that appears to

its users as a single coherent system.

The underlying concepts behind the definition of distributed system are very

interesting to be elaborated. Two main aspects from the definition that worth

mentioning are [35]:

computers as components that are autonomous

abstraction of distributed system's components from user's perspective as

user perceived that they only deal with single system instead of with

various components inside the system

Typically, three common layers used in describing components in distributed

systems which are:

19

outer layer that handle interaction with user

middle layer, that sometimes called middleware

lower layer that constitutes operating system and networks

Figure below depicts an example of distributed system, sometimes known as

middleware.

Figure 2.4:Four networked computers, with three applications. Application B is

distributed between two computers [35]

Optimization of distributed system may benefit an organization in several ways.

For instance, installation of network printers in an organization enable

operational cost reduction as the network printers are shared among users in

an organization. Another benefit of optimized distributed system is the systems

have the capability to handle concurrent request from multiple clients, which

lead to increase in organizational productivity and operational efficiency. The

existence of Internet has helped to extend distributed system's capability by

providing mechanism for users to communicate, to collaborate and to exchange

files.

Current developed DST for WDS area has embraced the distributed system

approach by having structured components' architecture, that separating client-

end application and server-side systems. As the end-user, field workforce used

their handheld devices to retrieve information without knowing about the

20

computational operation handled by servers that are located in remote area and

not inside their handheld devices. Figure 2.5 illustrates the proposed project

system's architecture that embodies distributed system.

Figure 2.5: Proposed high-level system architecture for WDS field workforce[1]

21

2.3.2. Introduction to Wireless Communication

Two form of communications between computers, servers, printers and nodes

in a network are wired and wireless. Wired communication imposed limitations

to the development and implantation in distributed system economically,

practicality and feasibility. These limitations hinder wired communication to be

used in large-scale distributed system. As opposed to wired communication's

limitation, wireless communication offers lower implementation cost and most

importantly scalable for large system.

In distributed system environment, selecting suitable type of wireless

communication is essential. Latency, data rate, resilience and security are

some example of the criteria in choosing suitable type of wireless

communication. Table 2.3 depicts matrix of some of the communication and

their criteria.

Table 2.3: Type of wireless communications and their criteria weightage [36]

Above matrix table also applicable for DST used in WDS area. As referring to

Figure 2.5, the DST application lies as intermediate layer between end-user

and system-core layer. Criteria such as data rate, resilience, distance and

22

scalability are considered as crucial for the field workforce that used DST in

executing their task. Data rate or bandwidth criterion defines information

transfer capacity of the channel [36]. Current DST prototype possesses the

ability to load GIS map and huge amount of data to-and-from server side. Thus,

bandwidths required are expected to be medium to high, since lower bandwidth

may result in high latency. Distance coverage is also expected to be high since

field workforce will be anywhere around the globe and should still be able to

communicate with centralized server side. The scalability criterion of whole

DSS determines the capability of the system to be expanded in future to cater

for wider coverage expansions. Other criteria are equally important; however

there are still room to tolerate if they are not being met.

Field workforces are expecting information from DST to be available in almost

real-time. However, the accurate output from server-side computations and

availability of information are more importance. Therefore, there is trade-off

between latency and accuracy since sometime servers might require longer

processing time in order to produce accurate output. Security criterion has

relationship with data integrity. Data tempering and data exploit are examples

of security concerns that might occur in wireless communication area. A DSS

require high-level of support commitment from wireless service provider to

provide support in ensuring high level of communication's security between

DST in a field with server-side environment. The least expected criteria for DST

in WDS environment to possess is resilience.

Resilience criterion, as according to [36] is ability to resist interferences and

recover during catastrophic circumstances such as earthquake. This criterion is

considered as low for DST in workforce area since a DST is handheld device

and assumption of the ability of field workforce to avoid performing task during

catastrophic events. Table 2.4 summarizes level of importance of several

communication criteria from an expected DSS in WDS area.

Criteria Weightage

23

Latency Medium

Bandwidth High

Resilience Low

Security Medium

Distance High

Scalability High

Table 2.4: Expected criteria from a DSS in WDS area and its weightage

It can be inferred from information in Table 2.3 and Table 2.4, suitable wireless

communication to be used are either WiMax or cellular network (2G, 3G and

4G).

2.3.3. Introduction to Cyber-Physical System

Cyber-Physical System (CPS) is defined as integrations of computational and

physical process that able to interact with humans through many modalities

[37][38]. From the definition, there are three identified aspects inside CPS,

which are:

Computational capability expected from technology

Physical process in real world

Human involvement as source of input

Uncertainty and unpredictability of physical processes in real world has caused

tension to engineers in finding predictable solutions. Generic medical

prescription to obese patient will not really help without really knowing factors

24

(e.g. family traits, diet and lifestyle) that lead to the obesity. These factors vary

from every human. Earthquake, volcano eruption and tsunami are example of

unpredictable natural disasters. Behaviours of these disasters have becoming

interesting subjects to be studied.

Edward Lee [38] believes that CPS is not optimized if it is operated in controlled

environment as the environment does not provide robust and unpredictable

conditions. Advancement in technologies is expected to be able to support

computations for solution for complex physical problem in real world. Human

interaction with system is essential as source of faithful inputs in order to

produce accurate computations output. With all the stated aspects inside CPS,

it came down to the question on how to design CPS that able to handle

unpredictability in physical process problems? Current research's CPS design

can be used as an example because the design met some criteria of good CPS

design. The design will be elaborated briefly in Section 4.

25

3. METHODOLOGY

3.1. Study of overall project

Thorough study about the overall existing system was conducted in order to

understand underlying technical concept behind the project and the system

architecture. Some of the activities including:

Reading publications and thesis of previous researchers that related to

current project. This is to ensure:

o comprehensive understanding about WDS domain concept, EDSS and

technical aspect of system architecture is build

o that the suitable objectives, and proposed solutions are selected for

current system prototype

Revision and discussion with other researcher to understand technical

aspects especially about server-side component of the prototype system.

Read-then-run the existing system from source code. This helps to

understand system and the syntax.

3.2. Software Development and Testing Methodology

Current research will adopt iterative and incremental software development

methodology together with several Agile practises. This methodology was

chosen on the basis that current research is expected to provide additional

functionalities to the existing developed DST for WDS. Briefly, iterative and

incremental software development methodology emphasis on functionality's

gradual enhancement, with cyclical release to the system. Basically, there are

four major phases involved [32]:

Inception phase. Define scope, general requirement and risk

associated.

26

Elaboration phase. Working unit that fulfil requirement and minimize

risk associated from inception phase.

Construction phase. Incrementally integrate working unit into existing

overall system

Transition. Deploy system to production environment

Figure 3.1: Iterative development model [33]

In proposing client-side caching technique (will be further discussed in Section

4.4) as a solution to less reliable network communication between DST and

central server in WDS environment, two round of iterations are identified as

necessary. Table 3.1 below describe the planned iterations and phases involved.

Inception Elaboration Construction Transition

Scope Requirement Risk Estimated

Completion

Time (weeks)

Estimated

Completion

Time (weeks)

Estimated

Completion

Time(weeks)

Iteration 1 Cache DST able to Loads of 4 1 1

27

development in

client device

detect network

connectivity

Latest cached

data is stored

Ability to re-

run simulation

manager

data will affect

DST

performance

especially

handphone

devices

Iteration 2 Data

synchronization

between DST and

central server

database

Correct data

updated to

central server

Incorrect

data updated

to database

(data

consistency)

1 1 1

Table 3.1: Software development iteration

Test-Driven Development (TDD) practise will be used in each of iteration

regularly. TDD is an evolutionary approach in agile software practise that

emphasize on creating test script or test code before creating the actual

implementation code. The development of the implementation code will have to

ensure that the code will passed in all test scripts. Main advantages of using this

approach are:

implementation code fulfilled required functionalities

as growing implementation code expanded, it does not lost its

previous tested functionalities

reduce efforts to document the functionalities as test cases can be

considered as documented functionality

Thus, it can be expected that at the end of the research, many test cases will be

shown as one type of proof of system functionalities.

28

Another agile software development practise that will be used is time boxing.

This is a time management approach is crucial to avoid uncertainty related to

time. All tasks, regardless of its scale, will be identified and will be assigned with

weightage and anticipated time required to accomplish each tasks.

Functionality Level of difficulty (0-9)

*0=easiest,9=hardest

Estimated time (working days including

weekend) *will be based on Elaboration

phase + Construction phase

Iteration 1

(40days)

Detect network

unavailability

1 2

Store page cache

functionality

6 10

Retrieve page cache

functionality

3 4

Store data cache

functionality

6 10

Retrieve data cache

functionality

3 4

Perform simulation on

unavailable network

5 7

Iteration 2

(20days)

Detect network

availability(reuse function in

Iteration 1)

1 1

Update database data

correctly

9 19

Table 3.2: Initial timeboxing chart for this research

29

3.3. Updated project plan

30

4. RESULT AND DISCUSSION

This section describes activities executed as of until this report was written and

related issues since the research started.

4.1. Review and study of existing project documentations

As a result from reading related project materials, it is understandable that this

research will be focusing on the DST for field workforce of WDS. With respect to

system architecture, the DST is considered as one of subsystem inside a CPS

for WDS. Previously, Section 2.3.3 had briefly described about CPS and the

three aspects that may affect a CPS. In current research context, the DST acts

as interaction mechanism between the CPS's third aspects (i.e. human

involvement as source of input) with backend computations server.

Designing robust CPS requires abstraction of process from user's perspective,

distributed computations and networked control, and also validation and

verification supports [37]. Therefore, it is believed that current proposed CPS

design for decision support process for field workforce in WDS has taken into

account the previous mentioned criteria. Figure 4.1 shows the overall CPS

architecture of the current prototype system.

31

Figure 4.1: Proposed CPS of the monitoring and controlling the WDS

Generally, the CPS design has decoupled client-server architecture. The red

rectangle area, which is also the research area, is the client-side from the whole

system, while the remaining areas are the middleware and backend server.

From the selected red area, three resources used as control mechanisms of DSS

in the WDS field area, which are:

Communication networks

Field workforce engineering expertise

Handheld tools or supporting devices for field workforce

These three resources are essential and dependable on each others. Field

workforce engineering expertise is considered to be the major personal

knowledge provider to this CPS and the existence of handheld DST is to provide

reliable and faithful static and contextual information to the field workforce. In

32

order to retrieve contextual information from remote server to handheld devices,

communication network serves as the intermediary.

As mentioned in the Error! Reference source not found. section, this research

will propose an alternative solution for less reliable communication network in

client-side of the system. Setup for server-side for the system has completed and

will be discuss in Section 4.2. Progress work about the implementation in client-

side will be discussed in Section 4.4.

4.2. Server system set-up for testing and development

This activity is needed in order to prepare a Linux environment, Web application

server (Apache Tomcat) and JAVA development environment. In addition, VPN

software (Shrewsoft) was also installed to enable communication with database

servers and computational servers that located inside University of Manchester's

secured network zone. VPN is needed if the development will be done from

outside of University Of Manchester's wireless area.

The unfamiliarity with Linux operating system and tweaking computer's Operating

System (OS) had caused in longer time consumed since substantial research

through Internet was required. Several approaches tried including:

Setup Ubuntu in virtual machine hosted by virtual server (VMware)

Install Ubuntu OS with Windows 8. OS dual booting technique with

Windows 8 required.

Second option was chosen since according observation made found that it is

much faster when running the DST prototype program. Besides OS, the

Integrated Development Environment (IDE) installed was Netbeans Eclipse

(Kepler). All Eclipse's required plugins and extensions were installed including:

Subversion (SVN). This plugin is required to access to centralized

version controlling server

33

Google Web Toolkit. This plugin is to translate Java code to

JavaScript during compilation mode

The client application will be running from local machine and hosted in the same

server as web application server.

4.3. Hands on existing prototype client application

The source code was downloaded into development area and was run in debug

mode, to enable clearer understanding of system flow by adding breakpoint

remarks at several lines of codes.

The prototype system is exercising Model-View-Presenter (MVP) pattern, a

software design pattern that isolates user-interface from business logic. This

pattern is said to be originated from Model-View-Controller concept [28]. MVP is

suitable to be used for web development area since it requires substantial effort

of graphical user interface (GUI) development. Briefly, Model component involve

in structuring Data of the system, View component is used to provide user

interaction mechanism to system backend and finally Presenter component is

used as computation unit and as bridge between Model and View.

By studying the developed prototype MVP architecture in research [2] and

running the application using Eclipse IDE, several knowledge obtained, which

are:

Understand the flow between MVP components' classes.

Understand communication between client and server.

Understand data transferred between client and server.

Understand on how to operate the developed DST prototype

Moreover, this activity helps to understand functionalities and features of current

system prototype. Features identified are [2]:

34

DMA network model selector and loader

Display network map and user’s geolocation

Display WDS components and their attributes

WDS component searcher

Run simulations

Simulation status notification

Results summary

Visualization of results on network map

Results graph creator

Alerts notification

Automatic simulation re-run

4.4. Proposed improvements in technology, implementation, functionality

and usability

As mentioned in Section 1.2, one of the project objectives is to propose solution

to support decision making process for field workforce, particularly in any

wireless network communication conditions. As described in Section1.1,

dependency of current DST prototype to the availability of communication

network may limit the capability of decision making for workforce engineer while

in the operation field.

Therefore, one alternative proposed to encounter this limitation is to have client-

side caching technique. Briefly, cache technology is high speed buffer memory-

based storage used to store temporary portions of contents of computer's main

memory [30]. In current project, cached data and web pages are expected to be

stored in client's (in this research context, the field workforce) device memory.

35

Identified format of data received from server side is in eXtensible Markup

Language (XML) and web pages are in Hypertext Markup Language (HTML)

format.

In order to have cached copy inside DST device, field workforce has to run DST

application at least once while there is communication network before going to

the assigned field location. Replication process of server's data will produce a

copy of server's data and then stored inside field workforce's device memory

during the running. While at the operation field, during network unavailability

circumstances, DST will inquire user before using stored cached data. In other

words, field workforce has the option in making decision by using the cached

data or not. However, as discussed in Section 1.3, there will be a challenge to

perform data synchronization. This research will have thorough look in this

matter.

Java Caching System (JCS) will be used in the cache development of this

research. JCS provide various features including, memory management, disk

overflow (and defragmentation), thread pool controls, element grouping, minimal

dependencies, data expiration (idle time and max life), and fully configurable

runtime parameters [39]. These features will be explored to produce optimized

cache memory management in client's devices. Testing of the development will

be carried out using laptop and possibly Android device. Laptop will be used to

test implementation using Wireless Fidelity (WiFi) network and Android device

will be used to test cellular network.

36

5. CONCLUSION

This study has clearly shown the needs to have reliable decision support tool in

assisting field workforce of Water Distribution System to make sound and accurate

decision. One identified improvement in current developed DST prototype is to have

an alternative solution for communication network dependency. As a result from

detailed reading and discussion, cache solution has been identified to be

implemented in the research.

Research will involve programming activities, testing and reflection on the stated

activities. These activities will be divided into two iterations. First iteration is to

implement the cache technology using Java Caching System into the existing DST

prototype. Second iteration will be focusing on updating cached data in the

37

6. REFERENCES

[1] Kashif Khan, "A Distributed Computing Architecture to Enable Advances in Field

Operations and Management of Distributed Infrastructure," Doctoral Thesis, The

University of Manchester, 2012.

[2] Pedro F. Jaramillo Hernandez, "Decision Support Tool for Field Water Engineers",

Master Thesis, University of Manchester, 2013.

[3] “Water Supply System :: Effluent Disposal.” Encyclopedia Britannica. Accessed

March 31, 2014. http://www.britannica.com/EBchecked/topic/637296/water-supply-

system/286138/Effluent-disposal.

[4] Loucks, Daniel P, Eelco van Beek, Jery R Stedinger, Jozef P. M Dijkman, and

Monique T Villars. "Water Resources Systems Planning and Management: An

Introduction to Methods, Models and Applications." Paris: UNESCO, 2005.

[5] “Industrial Water Use, the USGS Water Science School.” Accessed April 13, 2014.

http://water.usgs.gov/edu/wuin.html.

[6] “Water for Agriculture — European Environment Agency (EEA).” Article. Accessed

April 13, 2014. http://www.eea.europa.eu/articles/water-for-agriculture.

[7] Macalister, Terry. “Water Shortages Could Disrupt Britain’s Electricity Supply,

Researchers Warn.” The Guardian, February 16, 2014, sec. Environment.

http://www.theguardian.com/environment/2014/feb/16/water-shortages-electricity-

supply-climate-change.

[8] “Water Resource Management (current System) - Maintaining Secure Water

Supplies, High Standards of Drinking Water and Effective Sewerage Services -

Policies - GOV.UK.”Accessed April 14, 2014.

https://www.gov.uk/government/policies/maintaining-secure-water-supplies-high-

standards-of-drinking-water-and-effective-sewerage-services/supporting-

pages/water-resource-management.

38

[9] US EPA, OW. “Distribution System.” Accessed April 14, 2014.

http://water.epa.gov/lawsregs/rulesregs/sdwa/tcr/distributionsystems.cfm.

[10] Trifunovic, Nemanja. Introduction to Urban Water Distribution: Unesco-IHE Lecture

Note Series. CRC Press, 2006.

[11] Di Nardo, Armando, and Michele Di Natale. “A Heuristic Design Support

Methodology Based on Graph Theory for District Metering of Water Supply

Networks.” Engineering Optimization 43, no. 2 (February 2011): 193–211.

doi:10.1080/03052151003789858.

[12] Lambert, A., and W Hirner. “Losses from Water Supply Systems: Standard

Terminology and Recommended Performance Measures.” International Water

Association,

http://www.iwahq.org/contentsuite/upload/iwa/Document/Losses_from_Water_Sup

ply_Systems_2000.pdf.

[13] Morisson, John, Stephen Tooms, and Dewi Rogers. “District Metered Areas

Guidance Notes.” IWA Publishing, February 2007.

[14] Farley, Malcolm. “Leakage Management and Control. A Best Practice Training

Manual.” World Health Organization, 2001.

http://whqlibdoc.who.int/hq/2001/WHO_SDE_WSH_01.1_pp1-98.pdf.

[15] Khan, Kashif, Robert Haines, and John M. Brooke. “A Distributed Computing

Architecture to Support Field Engineering in Networked Systems.” 54–61. IEEE,

2010. doi:10.1109/CISIS.2010.179.

[16] “Water Distribution System Analysis Field Studies Modeling and Management

Reference Guide for Utilities.” National Environmental Publications Information

System, 2006.

http://nepis.epa.gov/Exe/ZyPDF.cgi/2000D2C2.PDF?Dockey=2000D2C2.PDF.

[17] Haines, R., K. Khan, and J. Brooke. “Bringing Simulation to Engineers in the Field:

A Web 2.0 Approach.” Philosophical Transactions of the Royal Society A:

39

Mathematical, Physical and Engineering Sciences 367, no. 1898 (July 13, 2009):

2635–44. doi:10.1098/rsta.2009.0047.

[18] Cortés, Ulises, Miquel Sànchez-Marrè, Luigi Ceccaroni, Ignasi R-Roda, and Manel

Poch. “Artificial Intelligence and Environmental Decision Support Systems.”

Applied Intelligence 13, no. 1 (2000): 77–91.

[19] Burras, Simon. “A Real-Time Expert System for Industrial Environments.” In

Industrial Applications of AI (Artificial Intelligence), IEE Colloquium on (Digest No.

014), 2–1. IET, 1992.

http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=167698.

[20] “Cognitive Systems Redefine Business Potential. Adaptive Intelligence Can Help

To Optimize Decision Making.” Ventana Research, 2012.

[21] Matthies, Michael, Carlo Giupponi, and Bertram Ostendorf. “Environmental

Decision Support Systems: Current Issues, Methods and Tools.” Environmental

Modelling & Software 22, no. 2 (February 2007): 123–27.

doi:10.1016/j.envsoft.2005.09.005.

[22] A.E., Rizzoli, and Young W.J. “Delivering Environmental Decision Support

Systems: Software Tools and Techniques” 12 (1997): 237–49.

doi:10.1016/S1364-8152(97)00016-9.

[23] Poch, Manel, Joaquim Comas, Ignasi Rodríguez-Roda, Miquel Sànchez-Marrè,

and Ulises Cortés. “Designing and Building Real Environmental Decision Support

Systems.” Environmental Modelling & Software 19, no. 9 (September 2004): 857–

73. doi:10.1016/j.envsoft.2003.03.007.

[24] Giupponi, Carlo. “Decision Support Systems for Implementing the European Water

Framework Directive: The MULINO Approach.” Environmental Modelling &

Software 22, no. 2 (February 2007): 248–58. doi:10.1016/j.envsoft.2005.07.024.

[25] Goodchild, Micheal F. “Geographic Information System,” 1991, 194–200.

40

[26] “Geographic Information Systems (GIS) | How We Use Data in the Mid-Atlantic

Region | US EPA.” Accessed April 19, 2014.

http://www.epa.gov/reg3esd1/data/gis.htm.

[27] Sweeney, Michael W. “Geographic Information Systems.” Water Environment

Research, 1999, 551–56.

[28] “Using the Model-View-Presenter (MVP) Design Pattern to Enable Presentational

Interoperability and Increased Testability - Dot Net Miscellany - Site Home - MSDN

Blogs.” Accessed April 28, 2014.

http://blogs.msdn.com/b/jowardel/archive/2008/09/09/using-the-model-view-

presenter-mvp-design-pattern-to-enable-presentational-interoperability-and-

increased-testability.aspx.

[29] “Cache.” Wikipedia, the Free Encyclopedia, April 21, 2014.

http://en.wikipedia.org/w/index.php?title=Cache&oldid=605177092.

[30] Smith, Alan Jay. “Cache Memories.” ACM Computing Surveys (CSUR) 14, no. 3

(1982): 473–530.

[31] Singhal, Vivek, Ian Emmons, and Richard Jensen. “Dynamic Web Page Cache,”

August 22, 2006.

[32] “What Is Iterative and Incremental Development? - Definition from Techopedia.”

Techopedia.com. Accessed April 30, 2014.

http://www.techopedia.com/definition/25895/iterative-and-incremental-

development.

[33] “Iterative and Incremental Development.” Wikipedia, the Free Encyclopedia, April

18, 2014.

http://en.wikipedia.org/w/index.php?title=Iterative_and_incremental_development&

oldid=591868914.

[34] “Introduction to Test Driven Development (TDD).” Accessed April 30, 2014.

http://www.agiledata.org/essays/tdd.html.

41

[35] Tanenbaum, Andrew S, and Steen. Distributed Systems: Principles and

Paradigms. Upper Saddle River, NJ: Pearson Prentice Hall, 2007.

[36] Akyol, B. A., Harold Kirkham, S. Clements, and M. Hadley. “A Survey of Wireless

Communications for the Electric Power System.” Prepared for the US Department

of Energy, 2010.

https://www.pnnl.gov/nationalsecurity/technical/secure_cyber_systems/pdf/power_

grid_wireless.pdf.

[37] Baheti, Radhakisan, and Helen Gill. “Cyber-Physical Systems.” The Impact of

Control Technology, 2011, 161–66.

[38] Lee, Edward A. “Cyber Physical Systems: Design Challenges.” 363–69. IEEE,

2008. doi:10.1109/ISORC.2008.25.

[39] “JCS - Java Caching System.” Accessed May 8, 2014.

http://commons.apache.org/proper/commons-jcs/.

[40] Sha, Lui, Sathish Gopalakrishnan, Xue Liu, and Qixin Wang. Machine Learning in Cyber Trust. Boston, MA: Springer US, 2009. http://www.springerlink.com/index/10.1007/978-0-387-88735-7.

[41] “Factors Affecting Wireless Networking Performance - 4Gon.” Accessed May 8,

2014.

http://www.4gon.co.uk/solutions/technical_factors_affecting_wireless_performance

.php.