Método para detecção, identificação e classificação de erros em redes de distribuição por meio do estimador de estados e índices de desvio

• Autor(a) principal: Delavechia, Robson Porsch
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Manancial - Repositório Digital da UFSM
• Texto Completo: http://repositorio.ufsm.br/handle/1/22439
• topic: Detecção; Identificação; Classificação; Estimador de estados; Índices de desvio; Detection; Identification; Classification; State estimator; Deviation indices; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA

The development of new technologies, needs and concepts, have led to numerous contributions to several fields of knowledge, one of which is the Electric Power Systems. Contained in this area, the Electricity Distribution Systems stand out, which have migrated from a traditional model to a more modern and contemporary structure, giving rise to the term smart grids. Smart meters, more dynamic monitoring mechanisms, extensive means of communication, storage, and data processing, among a vast number of variables, constitute the forming elements of Intelligent Electric Grids. Even with the evolution provided by all these factors, the errors inherent in this sector continue to be part of the operational reality and in the scope of planning. Within this context, the present work seeks to carry out the detection, identification, and classification of errors, taking as a scenario of study, analysis and implementation of the methodology, a real distribution network. Starting from the data provided by smart meters allocated throughout the network, we seek to acquire sufficient information to achieve observability. Through servers located in certain points of the base scenario, the aim is to store and process the collected measurements. After filtering the information, it is inserted in a computer platform that will be used for the physical implementation of the network under study, in which simulations, analyzes and diagnoses of the errors found will be carried out. The method used for the detection and identification of errors is at the discretion of state estimation that obtains a real-time model of the network, in view of the merging of remotely obtained measurements and pseudo-measures, derived from the calculation of the Power Flow. The State Estimator presents itself as a solution for the evaluation of errors associated with the measures, because, in addition to the steps of surveying the topology and observability analysis, it performs the processing of incorrect data, through the detection and identification of irregularities. Subsequently, this dissertation proposes two indices, one for individual deviation of meters and another for global deviation of the network, which will be responsible for the classification of errors found by the State Estimator, in pre-defined groups of possible sources that cause deviations. These indexes were developed for this work, aiming to attribute possibilities of the irregularities pointed out. Thus, the energy concessionaires, in addition to detecting and identifying inconsistencies in the network, can send teams previously instructed on the possible type of error to be minimized and/or eliminated, favoring the distributors and the final consumer.