No abstract

Summary In distribution power systems, one of the main objectives is to ensure uninterrupted service. Accordingly, the fast identification and isolation of faults is a matter of major interest. This work proposes a fault location method applicable to any power distribution system, which considers a random number of nodes as measurement points. An index value based on the voltage sags generated by different fault types is calculated for each node; then, considering the current measurements on the network and a preexisting database value, it is possible to determine fault type and location. This approach is analyzed by means of IEEE 34 node test feeder, and the results show that the proposed method can efficiently detect fault type and fault location.

No abstract

En esta tesis se presenta una serie de investigaciones y desarrollos enmarcados en dos campos principales. En primer lugar, la administración y control de redes de distribución eléctrica a través de software de gestión; y finalmente, la detección y localización de fallas utilizando métodos basados en el conocimiento y en el modelo. En lo que respecta al software de gestión, se puede observar que los nuevos paradigmas de implementación de las redes de distribución están cambiando la forma de operación de las mismas. Muchas empresas suelen crear varias soluciones personalizadas de sistemas para gestionar la distribución de energía. Sin embargo y a pesar de su relevancia, aquellas empresas que son medianas o pequeñas tienen problemas para instanciar una de estas porque son muy costosas, incompatibles o limitadas por dificultades en el intercambio de datos. Como posible soluciona este problema, en la presente tesis se expone el desarrollo de un Sistema de Gestión de Distribución Avanzado (ADMS) basado en tecnología de código abierto, lo que permite adaptarse con facilidad a las necesidades particulares de cada distribuidora.Por otra parte, debido a la expansión de los sistemas de distribución de energía,garantizar un servicio sin interrupciones es un ́área de gran interés en el estudio de los sistemas el ́eléctricos. Para ello es sumamente importante la identificación y aislamiento de la ocurrencia de una falla tan rápido como sea posible. Referido a este tema, el presente trabajo además propone una investigación y desarrollo de métodos de detección y localización de fallas que puedan ser integrados dentro del software de gestión mencionado en el párrafo anterior. El principal objetivo es la eliminación del problema de la múltiple estimación de falla presente en muchos delos métodos, por ello los desarrollos incluyen técnicas basadas en el conocimiento de la red para mitigar esta problemática. This thesis presents a set of investigations and developments framed in twomain fields. In first instance, the administration and control of electricity distri-bution networks through management software; and finally, the faults detectionand localization using knowledge-based and model-based methods.Regarding with management software, it can be observed that the new imple-mentation paradigms of distribution networks are changing the way of operation.Many companies often create several custom system solutions to manage powerdistribution. However, small and medium companies have problems installing oneof them due to they are very expensive, incompatible or limited by difficulties indata exchange. As a possible solution, the present thesis shows the developmentof an Advanced Distribution Management System (ADMS) based on open sourcetechnology, which allows it to be easily adapted to the particular needs of eachcompany.On the other hand, due to the expansion of power distribution systems, en-suring uninterrupted service is an area of great interest in the study of electricalsystems. This means that it is extremely important to identify and isolate theoccurrence of a fault as quickly as possible. For this reason, this work also propo-ses an investigation and development of fault detection and localization methodsthat can be integrated into the management software mentioned in the previousparagraph. The main objective is the elimination of the multiple fault estima-tion problem present in many of the methods, therefore the developments include knowledge-based techniques to mitigate this problem.

A problem of great interest for power distribution companies is ensuring uninterrupted service in extensive power distribution systems. Thus, the monitoring of networks and identification of system faults become essential. This work focuses on identifying a fault’s occurrence from a small number of low-cost measurements in a power distribution system. The determination of sensor locations is based on the recent feature selection approach LassoNet, where the measurement locations are ranked. It provides the most informative measures during a fault resulting in a shortening data set. It is used as input to a deep neural network without a significant loss in accuracy. We validate our method on the IEEE 13 and 34 node test feeders for distribution systems to conduct the suggested approach’s experimental studies.