טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentLifshitz Ron
SubjectDynamic District Metering Areas Optimization for Water
Distribution Systems
DepartmentDepartment of Civil and Environmental Engineering
Supervisor Professor Avi Ostfeld
Full Thesis textFull thesis text - English Version


Abstract

To date, water distribution systems (WDS) grow more and more complex as they may include up to tens of thousands of connected components with a common goal of delivering good water quality with sufficient amount to consumers. Since any deviation from the required goals in these mega networks can result with massive damage to consumers, the operation of such complex systems require new, innovative methods to address the above challenges.

In recent decades, the technique of partitioning the network into separate, manageable zones has grown popular. Like the ancient romans, with their “divide and rule” technique for controlling vast empires, a methodology often called District Metered Areas (DMAs) generates controlled, separately metered, parts of the network to improve system performance both in quality and quantity. The result of such an activity is a clustered network.

In this research, water quality simulations on clustered networks had been performed to explore new, unique, patterns in infection spread, categorized into distinct behavioral points. This research also focuses on the lack of dynamic methods to operate WDS and emphasizes the benefits trough dynamic, connectivity analysis based, clustering process. The combination of infection patterns and dynamic clustering is used to generate a new methodology for cluster assessment for contamination event detection and response. 

Although many clustering and partitioning techniques are explored, this research work develops a new parameter called Infection Delay Time (IDT) to meet the network operator’s available sources for contamination events in the form of First Response Time (FRT). IDT value, in time units, describes how long it takes for an infection to cross the cluster. If this value is larger than the FRT of the network operator, it can be isolated, located and treated in a known perimeter with minimum damage to consumers over the network. The IDT analysis over the network components can also guide the network operator when investing resources for water quality improvement in the distribution system. These methodologies are described in this research and applied on a few water distribution systems example applications.