Management and supply of water in an urban water distribution system is a complex process, which include various complexities like pressure variations across the network depending on topography, demand variations depending on customers’ requirement and unaccounted water etc. Applying automatic control methods to water distribution systems is a way to improve the management of water distribution. There have been some attempts in recent years to develop optimal control algorithms to assist in the operation of complex water distribution systems. The difficulties involved by these hydraulic systems such as non-linearity, and diurnal demand patterns make the choice of a suitable automatic control method a challenge. For this purpose, this study intends to investigate the applicability of different controllers which would be able to meet the targets as quickly as possible and without creating undue transients. As a first step towards application of different controllers, PD and PID linear controllers have been designed for pump control and valve control in water distribution systems. Then a Dynamic Inversion based nonlinear controller has been designed by considering the non-linearities in the system. Here, different cases considering the effects of initial conditions used, linearization methods used, time step used for integration and selection of gains etc., have been studied before arriving at best controller. These controllers have been designed for both the flow control problems and level control problems. It is found that Dynamic Inversion-based nonlinear controller outperforms other controllers. It is well known that the performance of controllers is much dependent on the tuning of the gains (parameters). Thus in this study various alternative techniques such as Ziegler--Nichols rules (ZNPID), Genetic algorithms (GAPID) and fuzzy algorithms (FZPID) have been studied and a comparative study has been made Although with all the three gain tuning methods, required states have reached their target values, but the responses vary much in reaching to final targets. The self-tuned FZPID controller outperforms other two controllers, especially with regard to overshoots and the time taken to tune the gains for each problem. Further, an optimal DI controller is developed for the over determined case with more controls and less targets. Energy loss is considered as an objective function and normal DI controller equations are considered as constraints. Hence, an attempt is made to reduce the energy minimization in water distribution system by formulating an optimal control problem using optimal Dynamic Inversion concept. Finally, leakage reduction model is developed based on excessive pressure minimization problem by locating valves optimally as well as by setting valves optimally. For this purpose, optimization problem is solved using Pattern search algorithms and hydraulic analysis is carried out using EPANET program.

A redução das perdas em sistemas de distribuição de água tem se tornado, além de uma questão ética, um determinante da eficiência e mesmo da sobrevivência das concessionárias destes serviços. A quantificação e determinação da natureza das vazões perdidas, que possuem relação íntima com as pressões de operação, são aspectos que ainda merecem análises mais profundas e práticas. Os parâmetros do modelo pressão-vazamento permitem o detalhamento das perdas físicas e foram obtidos por meio de calibrações para um setor de distribuição de água da cidade de Piracicaba, SP, em padrões hidráulicos distintos - sob hipótese de consumo noturno nulo (vazão mínima noturna), descargas de rede, patamares de pressão e sob consumo diurno dirigido pela pressão através de curvas encontradas na literatura e propostas. Os resultados obtidos indicaram que os valores dos parâmetros do modelo pressão-vazamento não são fixos, mas variam conforme o padrão hidráulico de operação da rede. Estudos acerca do parque de hidrômetros e seus desvios de indicação também foram realizados, em conjunto com uma análise das vazões típicas de abastecimento residencial, e apontaram para fortes erros de medição em razão do uso de reservatórios domiciliares. / Losses reduction on water supply networks came to be, beyond the ethical question, a condition for efficiency and even survival of enterprises which develop those services. Quantification and determination of the nature of leakage, which are closely related to operational pressure, are aspects which still claim for deeper and more practical analysis. Parameters to the pressure-leakage model allows to detail physical losses at one given system, and they were obtained throughout employment of distinct hydraulic patterns - minimal night flow (no usage), network flush situations, pressure levels and day usage driven by pressure from curves found on lectures and author\'s propositions - when proceeding calibration at a water supply section at the Piracicaba city, SP. Achieved results have shown that pressure-leakage model parameters values are not constant, but really change according to network\'s operational hydraulic pattern. Studies concerning section\'s water flow meters and its measuring deviation were also done, and joined to an analysis of typical patterns of home usage flows; they have pointed to huge metering errors because of employment of home reservoirs.

[ES] Debido a la escasez del recurso hídrico que azota a muchos países del mundo, los gobiernos están implementando políticas públicas para que las entidades gestoras del servicio de agua potable, juntamente con las instituciones públicas y la empresa privada, implementen tecnologías digitales para monitorizar y tener un mejor control de la gestión de las redes hidráulicas. Sin embargo, es posible que las entidades gestoras no posean un modelo matemático que les permita analizar el comportamiento hidráulico de la red ante distintos escenarios como el aumento de la demanda en ciertas zonas, la avería en una tubería principal, la ampliación de la red hidráulica; o tomar decisiones como la ubicación de nuevas fuentes de agua, la ubicación de sensores o elementos que permitan delimitar sectores hidráulicos (caudalímetros y válvulas de corte), entre otras. El objetivo principal de este trabajo de investigación es desarrollar nuevos métodos basados en el modelo matemático y la topología de la red que ayuden a resolver los retos anteriores. Estos métodos, implementados en aplicaciones informáticas, pretenden ser de gran utilidad a las empresas de agua potable para comprobar la conectividad de la red e identificar las subredes con problemas de suministro o sin demandas de caudal asignadas; analizar la fiabilidad de la red bajo diversas circunstancias; ayudar a ubicar válvulas reductoras de presión, sensores y otros elementos de control, validar el proceso de asignación de las demandas, identificar derivaciones que causen un suministro deficiente en determinadas zonas; delimitar sectores de demanda sin afectar la red arterial de transporte y sectorizar una red hidráulica según el aporte de cada fuente al consumo de cada nodo; simplificar modelos de detalle que permitan minimizar los tiempos de cálculo para optimizar los procesos de toma de decisiones como el diseño de ampliaciones, la mejora energética o para dar una respuesta rápida en la operación en tiempo real; y por último, proporcionar métodos basados en la simulación hidráulica y en unas ecuaciones de tipo potencial-logarítmicas para convertir rugosidades absolutas de Darcy-Weisbach y coeficientes de rugosidad de Chezy-Manning en coeficientes de rugosidad de Hazen-Williams para calibrar el modelo, en otras aplicaciones. Todos los métodos han sido implementados en aplicaciones informáticas para automatizar el proceso de cálculo. Para ello, se ha utilizado el entorno de programación de Visual Studio 2019 Community (.NET), la librería de EPANET (v2.2) para ejecutar los cálculos hidráulicos, y la librería shapelib para visualizar los resultados de manera gráfica desde cualquier software de Sistema de Información Geográfica. Los resultados obtenidos han sido verificados con profusión, primero porque cada método se ha desarrollado considerando el modelo de una red real y después se ha validado con otros cinco modelos de redes reales de configuraciones diferentes, tamaños, elementos de regulación y leyes de control. Asimismo, dado que todas las herramientas están publicadas en el portal de investigadores de ResearchGate para su libre acceso, han podido ser testeadas por otros muchos usuarios interesados en su uso para analizar y dar solución a los problemas diversos en la gestión de las redes de abastecimiento de agua. Por último, todos estos algoritmos pueden ser implementados en otras plataformas digitales o entornos SIG desde los cuales se puede mejorar la interacción del usuario con la red, permitiendo al modelador u operador de la red tomar las mejores decisiones. / [CAT] A causa de l'escassetat del recurs hídric que assota a molts països del món, els governs estan implementant polítiques públiques perquè les entitats gestores del servei d'aigua potable, juntament amb les institucions públiques i l'empresa privada, implementen tecnologies digitals per a monitorar i tindre un millor control de la gestió de les xarxes hidràuliques. No obstant això, és possible que les entitats gestores no posseïsquen un model matemàtic que els permeta analitzar el comportament hidràulic de la xarxa davant diferents escenaris com l'augment de la demanda en unes certes zones, l'avaria en una canonada principal, l'ampliació de la xarxa hidràulica; o prendre decisions com la ubicació de noves fonts d'aigua, la ubicació de sensors o elements que permeten delimitar sectors hidràulics (cabalímetres i vàlvules de tall), entre altres. L'objectiu principal d'aquest treball de recerca és desenvolupar nous mètodes basats en el model matemàtic i la topologia de la xarxa que ajuden a resoldre els reptes anteriors. Aquests mètodes, implementats en aplicacions informàtiques, pretenen ser de gran utilitat a les empreses d'aigua potable per a comprovar la connectivitat de la xarxa i identificar les subxarxes amb problemes de subministrament o sense demandes de cabal assignades; analitzar la fiabilitat de la xarxa sota diverses circumstàncies; ajudar a situar vàlvules reductores de pressió, sensors i altres elements de control, validar el procés d'assignació de les demandes, identificar derivacions que causen un subministrament deficient en determinades zones; delimitar sectors de demanda sense afectar la xarxa arterial de transport i sectoritzar una xarxa hidràulica segons l'aportació de cada font al consum de cada node; simplificar models de detall que permeten minimitzar els temps de càlcul per a optimitzar els processos de presa de decisions com el disseny d'ampliacions, la millora energètica o per a donar una resposta ràpida en l'operació en temps real; i finalment, proporcionar mètodes basats en la simulació hidràulica i en unes equacions de tipus potencial-logarítmiques per a convertir rugositats absolutes de Darcy-Weisbach i coeficients de rugositat de Chezy-Manning en coeficients de rugositat de Hazen-Williams per a calibrar el model, en altres aplicacions. Tots els mètodes han sigut implementats en aplicacions informàtiques per a automatitzar el procés de càlcul. Per a això, s'ha utilitzat l'entorn de programació de Visual Studio 2019 Community (.NET), la llibreria de EPANET (v2.2) per a executar els càlculs hidràulics, i la llibreria shapelib per a visualitzar els resultats de manera gràfica des de qualsevol programari de Sistema d'Informació Geogràfica. Els resultats obtinguts han sigut verificats amb profusió, primer perquè cada mètode s'ha desenvolupat considerant el model d'una xarxa real i després s'ha validat amb altres cinc models de xarxes reals de configuracions diferents, grandàries, elements de regulació i lleis de control. Així mateix, atés que totes les eines estan publicades en el portal d'investigadors de ResearchGate per al seu lliure accés, han pogut ser testades per molts altres usuaris interessats en el seu ús per a analitzar i donar solució als problemes diversos en la gestió de les xarxes de proveïment d'aigua. Finalment, tots aquests algorismes poden ser implementats en altres plataformes digitals o entorns SIG des dels quals es pot millorar la interacció de l'usuari amb la xarxa, permetent al modelador o operador de la xarxa prendre les millors decisions. / [EN] Due to the scarcity of water resources that plagues many countries in the world, governments are implementing public policies so that water management entities, together with public institutions and private companies, implement digital technologies to monitor and have a better control of the management of water networks. However, it is possible that the managing entities do not have a mathematical model that allows them to analyze the hydraulic behavior of the network in different scenarios such as increased demand in certain areas, the failure of a main pipe, the expansion of the hydraulic network; or to make decisions such as the location of new water sources, the location of sensors or elements that allow delimiting hydraulic sectors (flow meters and shut-off valves), among others. The main objective of this research work is to develop new methods based on the mathematical model and network topology that help to solve the above challenges. These methods, implemented in computer applications, are intended to be of great use to drinking water utilities to check network connectivity and identify subnetworks with supply problems or without assigned flow demands; analyze network reliability under various circumstances; help locate pressure reducing valves, sensors and other control elements; validate the demand allocation process; identify bypasses that cause poor supply in certain areas; delimit demand sectors without affecting the arterial transport network; and sectorize a water network according to the contribution of each source to the consumption of each node; simplify detailed models to minimize calculation times to optimize decision-making processes such as the design of extensions, energy improvement or to provide a quick response in real time operation; and finally, provide methods based on hydraulic simulation and potential-logarithmic equations to convert absolute Darcy-Weisbach roughness and Chezy-Manning roughness coefficients into Hazen-Williams roughness coefficients to calibrate the model in other applications. All methods have been implemented in software applications to automate the calculation process. For this purpose, the Visual Studio 2019 Community (.NET) programming environment has been used, the EPANET library (v2.2) to execute the hydraulic calculations, and the shapelib library to visualize the results graphically from any Geographic Information System software. The results obtained have been extensively verified, first because each method has been developed considering the model of a real network and then validated with five other real network models of different configurations, sizes, regulation elements and control laws. Also, since all the tools are published in the ResearchGate researcher portal for free access, they have been tested by many other users interested in their use to analyze and provide solutions to various problems in the management of water supply networks. Finally, all these algorithms can be implemented in other digital platforms or GIS environments from which the user's interaction with the network can be improved, allowing the modeler or network operator to make the best decisions. / Vegas Niño, ÓT. (2023). Nuevas herramientas para la gestión técnica de redes de distribución de agua basadas en el modelo matemático y la topología de la red [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/192474