PhD THESIS 
GEOGRAPHIC INFORMATION SYSTEMS APPLIED TO TECHNICAL 
MANAGEMENT OF DRINKING WATER DISTRIBUTION NETWORK 

Presented BY: 
D. Francisco Javier Martínez Solano 

Directed by: 
Dr. D. Rafael Pérez García 

SUMMARY 
When studying a Water Supply System (WSS), one needs to consider a huge amount of information in order to understand its hydraulic performance or to determine how to manage the existing resources of the network efficiently. There are many types of information, but the information can be classified into three main groups, depending on its nature and its later usage. These groups are: physical features of the network elements, economic information of the water supply system and spatial information about the location of both economic and physical data. 
Traditionally, this information has been saved in different formats. The information about network elements (diameters, lengths, starting date, suppliers, etc.) was saved in work plots or small inventory data bases. The economic information was the most carefully kept data base in the system. This database stored all the consumer data, including customers' addresses, registering dates and other relevant data for correct economic management of the system. Finally, spatial information was usually scattered in various topographic maps where the izolines of the supplied geographic area, the location of the mains and the distribution pipes layout appeared. However, most of the time these were not updated sketches. In any case, a connection rarely existed among the three information systems. 
However, nowadays there is a greater need to link spatial, economic and physic information together. This is now possible thanks to the implementation of a proper geographic information system (GIS). This system allows us not only to link geographic or spatial data with another alphanumeric data, but also to update in a simple way the included data, through an appropriate graphical interface. 
The aim of this thesis consists of supplying an efficient tool for decision making support related to water supply technical management. In order to get this target, some developments have been done: 
. 	Definition of a database structure gathering the whole of the necessary data to create the mathematical model of the water distribution network. This database must collect both cartographic data (such as elevation, roads or parcels) and specific data relating to water supply (consumers and infrastructures). 
. 	Elaboration of a digital terrain model. This model must allow the interpolation of the elevation of every element in the network. In the same way, this interpolation could be applied to the new nodes appeared after the model creation. On the other hand, the model has been completed with an algorithm of tessellation that allows dividing the service area in influence zones of every node in the model. 
. 	Creation of a mathematical model of the network starting from the network data available in the GIS. Since the management tasks in a water supply system are diverse, the model must suit the need it has been created for. So that, the GIS includes a set of routines to simplify the network depending on the use and the level of detail needed. 
. 	The load allocation, this is, the definition of the demand to be assigned to the nodes of the model, depends on the use and the level of detail of the model and the techniques used for model skeletonation. In this case two criteria for load allocation have been proposed. The first criteria responds to the level of detail and purpose of the model. So, there must be defined to kinds of models: strategic and detailed model. Strategic model will be used with average loads while detailed models will work with different situations such as peak demand flows, droughts, faults, etc. The second criteria involves time. The situations may vary from current situation to long term flow estimations. For present day situation, the billing data will be used, but in order to plan the future growth of the network, some tools for estimating future demands have been developed. For projected demands, statistic analysis of demand time series and spatial based regressions have been used. 
. 	Another aim of this PhD Thesis consists of initiating the interoperability among the different geographic information systems. For that, the possibility of sharing information has been taken into account. In order to facilitate the information exchange, a data model has been developed. This data model follows the recommendations made by the Open GIS Consortium. This contribution will allow the use of the tools developed in any system, independently of the used software. 
. 	Lastly, a case study has been included in the Thesis showing the usefulness of applying the methodology proposed to the technical management of a real water supply system. 
The main conclusion that can be extracted from this Thesis is that a powerful tool for making 
decision has been created. A proper use of geographic information system will aid to some 
decisions related to the technical management of a water supply system. These decisions will 
be supported by the results from hydraulic simulation made with a mathematical model of 
the network generated from georeferenced data in a GIS. To get this model a succession of 
problems were solved, which are described along the chapters of the Thesis.