Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional
Autor(a) principal: | |
---|---|
Data de Publicação: | 2019 |
Tipo de documento: | Dissertação |
Idioma: | por |
Título da fonte: | Manancial - Repositório Digital da UFSM |
Texto Completo: | http://repositorio.ufsm.br/handle/1/19725 |
Resumo: | In the context of the expected advances with the implementation of Smart Grids, enabling the modernization of the electricity billing system and making the management of residential electric power systems attractive, the detailed knowledge of the residential consumption profile gains importance as an element to subsidize the decision making by Home Energy Management Systems. One of the techniques for detailing the residential electrical consumption profile is Non-Intrusive Load Monitoring (NILM), a low cost installation technique that presents its complexity in the development of the disaggregation algorithm. Given this challenge, this dissertation presents a methodology based on low frequency sampling electrical monitoring to perform the extraction of characteristics of the activated and deactivated electric loads during the operation of the electric network. Through these characteristics, an evaluation of the event classification performance is performed using the Convolutional Neural Network - CNN, a type of artificial intelligence specially used for visual pattern recognition. To evaluate the developed method, a case study is performed using monitoring data from the United Kingdom Domestic Appliance Level Electricity database - UK-DALE, a data source widely used in NILM surveys. The performance of the developed method is evaluated using the Precision, Recall, F1-Score and Accuracy metrics as well as the confusion matrix to present the classification errors. To compare the classification performance obtained by the developed method is also modeled classification method called Decision Tree. Through the performance analysis of the developed method it is observed that it presents some restriction to deal with behaviors not foreseen in the training phase, but presents the ability to learn new behaviors through new training phases. Also, it presented good performance in the classification of low power events, when compared to the Decision Tree classification method. In conclusion, the use of convolutional neural network presents positive performance to be used in the event classification in NILM applications. |
id |
UFSM_8ac9f6ee1fcefbaee7f3339d570911e0 |
---|---|
oai_identifier_str |
oai:repositorio.ufsm.br:1/19725 |
network_acronym_str |
UFSM |
network_name_str |
Manancial - Repositório Digital da UFSM |
repository_id_str |
|
spelling |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucionalEvent classification in NILM monitoring of residential electrical loads using convolutional neural networkCNNMonitoramento não intrusivoNILMRede neural convolucionalCNNNon-intrusive load monitoringNILMConvolutional neural networkCNPQ::ENGENHARIAS::ENGENHARIA ELETRICAIn the context of the expected advances with the implementation of Smart Grids, enabling the modernization of the electricity billing system and making the management of residential electric power systems attractive, the detailed knowledge of the residential consumption profile gains importance as an element to subsidize the decision making by Home Energy Management Systems. One of the techniques for detailing the residential electrical consumption profile is Non-Intrusive Load Monitoring (NILM), a low cost installation technique that presents its complexity in the development of the disaggregation algorithm. Given this challenge, this dissertation presents a methodology based on low frequency sampling electrical monitoring to perform the extraction of characteristics of the activated and deactivated electric loads during the operation of the electric network. Through these characteristics, an evaluation of the event classification performance is performed using the Convolutional Neural Network - CNN, a type of artificial intelligence specially used for visual pattern recognition. To evaluate the developed method, a case study is performed using monitoring data from the United Kingdom Domestic Appliance Level Electricity database - UK-DALE, a data source widely used in NILM surveys. The performance of the developed method is evaluated using the Precision, Recall, F1-Score and Accuracy metrics as well as the confusion matrix to present the classification errors. To compare the classification performance obtained by the developed method is also modeled classification method called Decision Tree. Through the performance analysis of the developed method it is observed that it presents some restriction to deal with behaviors not foreseen in the training phase, but presents the ability to learn new behaviors through new training phases. Also, it presented good performance in the classification of low power events, when compared to the Decision Tree classification method. In conclusion, the use of convolutional neural network presents positive performance to be used in the event classification in NILM applications.No contexto dos avanços esperados com a implantação das redes elétricas inteligentes, viabilizando a modernização do sistema de tarifação de energia elétrica e tornando atrativo o gerenciamento dos sistemas de energia elétrica residencial, o conhecimento detalhado do perfil de consumo residencial ganha importância como elemento para subsidiar a tomada de decisão pelos sistemas de gerenciamento de energia elétrica residencial. Uma das técnicas para realizar o detalhamento do perfil de consumo das cargas elétricas residenciais é o monitoramento não intrusivo (Non-Intrusive Load Monitoring - NILM), uma técnica de baixo custo de instalação que apresenta sua complexidade no desenvolvimento do algoritmo de desagregação. Diante deste desafio, nesta dissertação é apresentada metodologia baseada em monitoramento elétrico em baixa frequência de amostragem para realizar a extração de características das cargas elétricas ativadas e desativadas durante a operação da rede elétrica. Através destas características é realizada avaliação do desempenho da classificação dos eventos utilizando a rede neural convolucional (Convolutional Neural Network - CNN), uma técnica de inteligência artificial especialmente utilizada para reconhecimento de padrões visuais. Para avaliar o método desenvolvido é realizado estudo de caso utilizando dados de monitoramento do banco de dados público de consumo de cargas elétricas domésticas do Reino Unido (United Kingdom Domestic Appliance Level Electricity - UK-DALE), uma fonte de dados amplamente utilizada em pesquisas NILM. O desempenho do método desenvolvido é avaliado através das métricas Precision, Recall, F1-Score e Accuracy assim como é utilizada a matriz de confusão para apresentar os erros de classificação observados. Para comparar o desempenho de classificação obtida pelo método desenvolvido é também modelado o método de classificação denominado árvore de decisões. Através da análise desempenho do método desenvolvido é observado que este apresenta certa restrição para lidar com comportamentos não previstos na fase de treinamento, porém apresenta capacidade de aprender novos comportamentos através de novas fases de treinamento. Também, apresentou bom desempenho na classificação de eventos de baixo valor de potência, se comparado ao método de classificação árvore de decisões. Concluindo-se que a utilização da rede neural convolucional apresenta desempenho positivo para ser utilizada na classificação de eventos em aplicações NILM.Universidade Federal de Santa MariaBrasilEngenharia ElétricaUFSMPrograma de Pós-Graduação em Engenharia ElétricaCentro de TecnologiaCanha, Luciane Neveshttp://lattes.cnpq.br/6991878627141193Sausen, Paulo Sérgiohttp://lattes.cnpq.br/5358851504202129Barriquello, Carlos Henriquehttp://lattes.cnpq.br/4127396473202565Medeiros, Aérton Pedra2020-03-04T15:28:44Z2020-03-04T15:28:44Z2019-10-22info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://repositorio.ufsm.br/handle/1/19725porAttribution-NonCommercial-NoDerivatives 4.0 Internationalhttp://creativecommons.org/licenses/by-nc-nd/4.0/info:eu-repo/semantics/openAccessreponame:Manancial - Repositório Digital da UFSMinstname:Universidade Federal de Santa Maria (UFSM)instacron:UFSM2020-03-05T06:01:31Zoai:repositorio.ufsm.br:1/19725Biblioteca Digital de Teses e Dissertaçõeshttps://repositorio.ufsm.br/ONGhttps://repositorio.ufsm.br/oai/requestatendimento.sib@ufsm.br||tedebc@gmail.comopendoar:2020-03-05T06:01:31Manancial - Repositório Digital da UFSM - Universidade Federal de Santa Maria (UFSM)false |
dc.title.none.fl_str_mv |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional Event classification in NILM monitoring of residential electrical loads using convolutional neural network |
title |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional |
spellingShingle |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional Medeiros, Aérton Pedra CNN Monitoramento não intrusivo NILM Rede neural convolucional CNN Non-intrusive load monitoring NILM Convolutional neural network CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA |
title_short |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional |
title_full |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional |
title_fullStr |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional |
title_full_unstemmed |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional |
title_sort |
Classificação de eventos em monitoramento NILM de cargas elétricas residenciais utilizando rede neural convolucional |
author |
Medeiros, Aérton Pedra |
author_facet |
Medeiros, Aérton Pedra |
author_role |
author |
dc.contributor.none.fl_str_mv |
Canha, Luciane Neves http://lattes.cnpq.br/6991878627141193 Sausen, Paulo Sérgio http://lattes.cnpq.br/5358851504202129 Barriquello, Carlos Henrique http://lattes.cnpq.br/4127396473202565 |
dc.contributor.author.fl_str_mv |
Medeiros, Aérton Pedra |
dc.subject.por.fl_str_mv |
CNN Monitoramento não intrusivo NILM Rede neural convolucional CNN Non-intrusive load monitoring NILM Convolutional neural network CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA |
topic |
CNN Monitoramento não intrusivo NILM Rede neural convolucional CNN Non-intrusive load monitoring NILM Convolutional neural network CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA |
description |
In the context of the expected advances with the implementation of Smart Grids, enabling the modernization of the electricity billing system and making the management of residential electric power systems attractive, the detailed knowledge of the residential consumption profile gains importance as an element to subsidize the decision making by Home Energy Management Systems. One of the techniques for detailing the residential electrical consumption profile is Non-Intrusive Load Monitoring (NILM), a low cost installation technique that presents its complexity in the development of the disaggregation algorithm. Given this challenge, this dissertation presents a methodology based on low frequency sampling electrical monitoring to perform the extraction of characteristics of the activated and deactivated electric loads during the operation of the electric network. Through these characteristics, an evaluation of the event classification performance is performed using the Convolutional Neural Network - CNN, a type of artificial intelligence specially used for visual pattern recognition. To evaluate the developed method, a case study is performed using monitoring data from the United Kingdom Domestic Appliance Level Electricity database - UK-DALE, a data source widely used in NILM surveys. The performance of the developed method is evaluated using the Precision, Recall, F1-Score and Accuracy metrics as well as the confusion matrix to present the classification errors. To compare the classification performance obtained by the developed method is also modeled classification method called Decision Tree. Through the performance analysis of the developed method it is observed that it presents some restriction to deal with behaviors not foreseen in the training phase, but presents the ability to learn new behaviors through new training phases. Also, it presented good performance in the classification of low power events, when compared to the Decision Tree classification method. In conclusion, the use of convolutional neural network presents positive performance to be used in the event classification in NILM applications. |
publishDate |
2019 |
dc.date.none.fl_str_mv |
2019-10-22 2020-03-04T15:28:44Z 2020-03-04T15:28:44Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
format |
masterThesis |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://repositorio.ufsm.br/handle/1/19725 |
url |
http://repositorio.ufsm.br/handle/1/19725 |
dc.language.iso.fl_str_mv |
por |
language |
por |
dc.rights.driver.fl_str_mv |
Attribution-NonCommercial-NoDerivatives 4.0 International http://creativecommons.org/licenses/by-nc-nd/4.0/ info:eu-repo/semantics/openAccess |
rights_invalid_str_mv |
Attribution-NonCommercial-NoDerivatives 4.0 International http://creativecommons.org/licenses/by-nc-nd/4.0/ |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.publisher.none.fl_str_mv |
Universidade Federal de Santa Maria Brasil Engenharia Elétrica UFSM Programa de Pós-Graduação em Engenharia Elétrica Centro de Tecnologia |
publisher.none.fl_str_mv |
Universidade Federal de Santa Maria Brasil Engenharia Elétrica UFSM Programa de Pós-Graduação em Engenharia Elétrica Centro de Tecnologia |
dc.source.none.fl_str_mv |
reponame:Manancial - Repositório Digital da UFSM instname:Universidade Federal de Santa Maria (UFSM) instacron:UFSM |
instname_str |
Universidade Federal de Santa Maria (UFSM) |
instacron_str |
UFSM |
institution |
UFSM |
reponame_str |
Manancial - Repositório Digital da UFSM |
collection |
Manancial - Repositório Digital da UFSM |
repository.name.fl_str_mv |
Manancial - Repositório Digital da UFSM - Universidade Federal de Santa Maria (UFSM) |
repository.mail.fl_str_mv |
atendimento.sib@ufsm.br||tedebc@gmail.com |
_version_ |
1805922053360975872 |