Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional

Detalhes bibliográficos
Autor(a) principal: Tavares, Hugo Menezes
Data de Publicação: 2020
Tipo de documento: Dissertação
Idioma: por
Título da fonte: Repositório Institucional da UFS
Texto Completo: https://ri.ufs.br/jspui/handle/riufs/14187
Resumo: Electric power is very important for the economic development of all countries, and its consumption has been growing at an impressive rate, doing so faster than other types of power. Along with the increase in consumption, there are also concerns about environmental sustainability; after all, ensuring access to electric power in a reliable, sustainable, modern, and affordable way for all is one of the objectives of the Sustainable Development Goals, an agenda proposed by the United Nations (UN). In addition to encouraging the usage of renewable and less environmentally impactful energy, there are also concerns to create increasingly more power efficient devices, and to reduce the waste of electric power by seeking alternatives for a more efficient use of it. The active involvement of consumers results, for the most part, in a more efficient use of electric power, which increases interest in the development of technologies that make them aware of their habits. Studies show that, the greater the detail of information about electrical power consumption, the greater the amount of electric power saved by consumers. One of the most used techniques to analyze such details is Non-Intrusive Load Monitoring (NILM), who, by disaggregating the loads, distinguishes between each of the appliances and explores the electric power consumption of each one individually. Therefore, in order to contribute to this technique, and considering the ever-growing progress in the electronic and machine-learning areas, this study proposes a set of training strategies using a deep-learning method for load classification in an embedded system, and therefore, contribute to a more efficient use of electric power. Based on literature and experiments, we adopted the binary image of the voltage-current as the distinguishing feature, as it obtained the best results. In order to classify the devices, we used said images as input to the Convolutional Neural Network (CNN), which was chosen after obtaining the best results in the tests that were performed. After we used the leave-one-out cross-validation method, our CNN model was evaluated using the PLAID dataset and obtained an F-Score macro-average of 74.76% for PLAID1, 56.48% for PLAID2, and 73.97% for PLAID1+2, and those results were very close to literature. The novelty of this study is the quantization of the CNN model using TensorFlow Lite, and its application in a resource-constrained embedded system (ESP32). The accuracy rate achieved in testes performed with all data from the PLAID1+2 dataset was 98.55%, which shows that the embedded device can be used to perform the load classification with a high accuracy rate.
id UFS-2_2d34fbc6da3197357e66fda7863cbce1
oai_identifier_str oai:ufs.br:riufs/14187
network_acronym_str UFS-2
network_name_str Repositório Institucional da UFS
repository_id_str
spelling Tavares, Hugo MenezesPrado, Bruno Otávio PiedadeBispo, Kalil Araujo2021-05-06T16:52:21Z2021-05-06T16:52:21Z2020-11-25TAVARES, Hugo Menezes. Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional. 2020. 150 f. Dissertação (Mestrado em Ciência da Computação) - Universidade Federal de Sergipe, São Cristóvão, Sergipe, 2020.https://ri.ufs.br/jspui/handle/riufs/14187Autorização para publicação no Repositório da Universidade Federal de Sergipe (RI-UFS), concedida pelo autor.Electric power is very important for the economic development of all countries, and its consumption has been growing at an impressive rate, doing so faster than other types of power. Along with the increase in consumption, there are also concerns about environmental sustainability; after all, ensuring access to electric power in a reliable, sustainable, modern, and affordable way for all is one of the objectives of the Sustainable Development Goals, an agenda proposed by the United Nations (UN). In addition to encouraging the usage of renewable and less environmentally impactful energy, there are also concerns to create increasingly more power efficient devices, and to reduce the waste of electric power by seeking alternatives for a more efficient use of it. The active involvement of consumers results, for the most part, in a more efficient use of electric power, which increases interest in the development of technologies that make them aware of their habits. Studies show that, the greater the detail of information about electrical power consumption, the greater the amount of electric power saved by consumers. One of the most used techniques to analyze such details is Non-Intrusive Load Monitoring (NILM), who, by disaggregating the loads, distinguishes between each of the appliances and explores the electric power consumption of each one individually. Therefore, in order to contribute to this technique, and considering the ever-growing progress in the electronic and machine-learning areas, this study proposes a set of training strategies using a deep-learning method for load classification in an embedded system, and therefore, contribute to a more efficient use of electric power. Based on literature and experiments, we adopted the binary image of the voltage-current as the distinguishing feature, as it obtained the best results. In order to classify the devices, we used said images as input to the Convolutional Neural Network (CNN), which was chosen after obtaining the best results in the tests that were performed. After we used the leave-one-out cross-validation method, our CNN model was evaluated using the PLAID dataset and obtained an F-Score macro-average of 74.76% for PLAID1, 56.48% for PLAID2, and 73.97% for PLAID1+2, and those results were very close to literature. The novelty of this study is the quantization of the CNN model using TensorFlow Lite, and its application in a resource-constrained embedded system (ESP32). The accuracy rate achieved in testes performed with all data from the PLAID1+2 dataset was 98.55%, which shows that the embedded device can be used to perform the load classification with a high accuracy rate.A energia elétrica é de grande importância para o desenvolvimento econômico dos países e o seu consumo vem crescendo em um ritmo vertiginoso, mais rápido que os demais modais energéticos. Concomitantemente com o aumento do consumo, surge também a preocupação com o meio ambiente e a sustentabilidade, sendo que assegurar o acesso à energia elétrica de forma confiável, sustentável, moderna e a preço acessível para todos é um dos objetivos da Agenda 2030 proposta pela Organização das Nações Unidas (ONU). Além do incentivo ao uso de energias renováveis e de menor impacto ambiental, há também duas preocupações: criar dispositivos energeticamente cada vez mais eficientes e reduzir o desperdício de energia elétrica, buscando alternativas para um uso mais eficiente desta. O envolvimento ativo dos consumidores resulta, na maioria das vezes, em um uso mais eficiente da energia elétrica, aumentando o interesse no desenvolvimento de tecnologias que os conscientizem quanto aos seus hábitos. Estudos mostram que quanto maior o detalhamento de informações acerca do consumo elétrico, maior a quantidade de energia elétrica economizada pelos consumidores. Uma das técnicas mais utilizadas para esse detalhamento é o Monitoramento Não-Intrusivo de Cargas, que através da desagregação de cargas, faz a distinção entre as cargas elétricas e explora o consumo elétrico de cada uma delas individualmente. A fim de contribuir para essa técnica, e diante do crescente avanço nas áreas de eletrônica e aprendizado de máquina, este estudo propõe realizar a classificação de cargas em sistema embarcado utilizando um método de aprendizado profundo e, dessa forma, contribuir para um consumo mais eficiente de energia elétrica. Baseado na literatura e em experimentos realizados, adotamos como característica de distinção a imagem binária da trajetória tensão-corrente, que foi a que obteve melhores resultados. Para realizar a classificação dos aparelhos, utilizamos essas imagens como entrada para um método de aprendizado profundo, que foi a Rede Neural Convolucional (RNC), escolhido após obter melhores resultados nos testes que foram realizados. A contribuição deste trabalho é a quantização do modelo da RNC usando o TensorFlow Lite e a sua aplicação em um dispositivo embarcado, que foi o ESP32. Usando o método de validação cruzada leave-one-out, nosso modelo da RNC foi avaliado usando o dataset PLAID e obteve uma média macro F-Score de 74,76% para PLAID1, 56,48% para PLAID2 e 73,97% para PLAID1+2, resultados bastante próximos ao da literatura. Em testes realizados com todos os dados do PLAID1+2, a acurácia foi de 98,55% no ESP32, demostrando que o dispositivo embarcado pode ser utilizado para realizar a classificação de cargas com alta taxa de acurácia.Fundação de Apoio a Pesquisa e à Inovação Tecnológica do Estado de Sergipe - FAPITEC/SESão Cristóvão, SEporComputaçãoInteligência artificialRedes neurais (computação)Sistemas embarcados (Computadores)Monitoramento não-intrusivo de cargasRede neural convolucionalNon-Intrusive load monitoringLoad classificationConvolutional neural networkEmbedded systemsCIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAOClassificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucionalinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisPós-Graduação em Ciência da ComputaçãoUniversidade Federal de Sergipereponame:Repositório Institucional da UFSinstname:Universidade Federal de Sergipe (UFS)instacron:UFSinfo:eu-repo/semantics/openAccessLICENSElicense.txtlicense.txttext/plain; charset=utf-81475https://ri.ufs.br/jspui/bitstream/riufs/14187/1/license.txt098cbbf65c2c15e1fb2e49c5d306a44cMD51ORIGINALHUGO_MENEZES_TAVARES.pdfHUGO_MENEZES_TAVARES.pdfapplication/pdf6491620https://ri.ufs.br/jspui/bitstream/riufs/14187/2/HUGO_MENEZES_TAVARES.pdffdef03084f92492c225bba5be693dbcaMD52TEXTHUGO_MENEZES_TAVARES.pdf.txtHUGO_MENEZES_TAVARES.pdf.txtExtracted texttext/plain311691https://ri.ufs.br/jspui/bitstream/riufs/14187/3/HUGO_MENEZES_TAVARES.pdf.txtc86f09c4c24ffd9a57bf6d885cf84eefMD53THUMBNAILHUGO_MENEZES_TAVARES.pdf.jpgHUGO_MENEZES_TAVARES.pdf.jpgGenerated Thumbnailimage/jpeg1439https://ri.ufs.br/jspui/bitstream/riufs/14187/4/HUGO_MENEZES_TAVARES.pdf.jpg8ab48e9b850cb7f9d3f63ec947e073b5MD54riufs/141872021-05-06 13:52:24.327oai:ufs.br:riufs/14187TElDRU7Dh0EgREUgRElTVFJJQlVJw4fDg08gTsODTy1FWENMVVNJVkEKCkNvbSBhIGFwcmVzZW50YcOnw6NvIGRlc3RhIGxpY2Vuw6dhLCB2b2PDqiAobyBhdXRvcihlcykgb3UgbyB0aXR1bGFyIGRvcyBkaXJlaXRvcyBkZSBhdXRvcikgY29uY2VkZSDDoCBVbml2ZXJzaWRhZGUgRmVkZXJhbCBkZSBTZXJnaXBlIG8gZGlyZWl0byBuw6NvLWV4Y2x1c2l2byBkZSByZXByb2R1emlyIHNldSB0cmFiYWxobyBubyBmb3JtYXRvIGVsZXRyw7RuaWNvLCBpbmNsdWluZG8gb3MgZm9ybWF0b3Mgw6F1ZGlvIG91IHbDrWRlby4KClZvY8OqIGNvbmNvcmRhIHF1ZSBhIFVuaXZlcnNpZGFkZSBGZWRlcmFsIGRlIFNlcmdpcGUgcG9kZSwgc2VtIGFsdGVyYXIgbyBjb250ZcO6ZG8sIHRyYW5zcG9yIHNldSB0cmFiYWxobyBwYXJhIHF1YWxxdWVyIG1laW8gb3UgZm9ybWF0byBwYXJhIGZpbnMgZGUgcHJlc2VydmHDp8Ojby4KClZvY8OqIHRhbWLDqW0gY29uY29yZGEgcXVlIGEgVW5pdmVyc2lkYWRlIEZlZGVyYWwgZGUgU2VyZ2lwZSBwb2RlIG1hbnRlciBtYWlzIGRlIHVtYSBjw7NwaWEgZGUgc2V1IHRyYWJhbGhvIHBhcmEgZmlucyBkZSBzZWd1cmFuw6dhLCBiYWNrLXVwIGUgcHJlc2VydmHDp8Ojby4KClZvY8OqIGRlY2xhcmEgcXVlIHNldSB0cmFiYWxobyDDqSBvcmlnaW5hbCBlIHF1ZSB2b2PDqiB0ZW0gbyBwb2RlciBkZSBjb25jZWRlciBvcyBkaXJlaXRvcyBjb250aWRvcyBuZXN0YSBsaWNlbsOnYS4gVm9jw6ogdGFtYsOpbSBkZWNsYXJhIHF1ZSBvIGRlcMOzc2l0bywgcXVlIHNlamEgZGUgc2V1IGNvbmhlY2ltZW50bywgbsOjbyBpbmZyaW5nZSBkaXJlaXRvcyBhdXRvcmFpcyBkZSBuaW5ndcOpbS4KCkNhc28gbyB0cmFiYWxobyBjb250ZW5oYSBtYXRlcmlhbCBxdWUgdm9jw6ogbsOjbyBwb3NzdWkgYSB0aXR1bGFyaWRhZGUgZG9zIGRpcmVpdG9zIGF1dG9yYWlzLCB2b2PDqiBkZWNsYXJhIHF1ZSBvYnRldmUgYSBwZXJtaXNzw6NvIGlycmVzdHJpdGEgZG8gZGV0ZW50b3IgZG9zIGRpcmVpdG9zIGF1dG9yYWlzIHBhcmEgY29uY2VkZXIgw6AgVW5pdmVyc2lkYWRlIEZlZGVyYWwgZGUgU2VyZ2lwZSBvcyBkaXJlaXRvcyBhcHJlc2VudGFkb3MgbmVzdGEgbGljZW7Dp2EsIGUgcXVlIGVzc2UgbWF0ZXJpYWwgZGUgcHJvcHJpZWRhZGUgZGUgdGVyY2Vpcm9zIGVzdMOhIGNsYXJhbWVudGUgaWRlbnRpZmljYWRvIGUgcmVjb25oZWNpZG8gbm8gdGV4dG8gb3Ugbm8gY29udGXDumRvLgoKQSBVbml2ZXJzaWRhZGUgRmVkZXJhbCBkZSBTZXJnaXBlIHNlIGNvbXByb21ldGUgYSBpZGVudGlmaWNhciBjbGFyYW1lbnRlIG8gc2V1IG5vbWUocykgb3UgbyhzKSBub21lKHMpIGRvKHMpIApkZXRlbnRvcihlcykgZG9zIGRpcmVpdG9zIGF1dG9yYWlzIGRvIHRyYWJhbGhvLCBlIG7Do28gZmFyw6EgcXVhbHF1ZXIgYWx0ZXJhw6fDo28sIGFsw6ltIGRhcXVlbGFzIGNvbmNlZGlkYXMgcG9yIGVzdGEgbGljZW7Dp2EuIAo=Repositório InstitucionalPUBhttps://ri.ufs.br/oai/requestrepositorio@academico.ufs.bropendoar:2021-05-06T16:52:24Repositório Institucional da UFS - Universidade Federal de Sergipe (UFS)false
dc.title.pt_BR.fl_str_mv Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
title Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
spellingShingle Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
Tavares, Hugo Menezes
Computação
Inteligência artificial
Redes neurais (computação)
Sistemas embarcados (Computadores)
Monitoramento não-intrusivo de cargas
Rede neural convolucional
Non-Intrusive load monitoring
Load classification
Convolutional neural network
Embedded systems
CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
title_short Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
title_full Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
title_fullStr Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
title_full_unstemmed Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
title_sort Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional
author Tavares, Hugo Menezes
author_facet Tavares, Hugo Menezes
author_role author
dc.contributor.author.fl_str_mv Tavares, Hugo Menezes
dc.contributor.advisor1.fl_str_mv Prado, Bruno Otávio Piedade
dc.contributor.advisor-co1.fl_str_mv Bispo, Kalil Araujo
contributor_str_mv Prado, Bruno Otávio Piedade
Bispo, Kalil Araujo
dc.subject.por.fl_str_mv Computação
Inteligência artificial
Redes neurais (computação)
Sistemas embarcados (Computadores)
Monitoramento não-intrusivo de cargas
Rede neural convolucional
topic Computação
Inteligência artificial
Redes neurais (computação)
Sistemas embarcados (Computadores)
Monitoramento não-intrusivo de cargas
Rede neural convolucional
Non-Intrusive load monitoring
Load classification
Convolutional neural network
Embedded systems
CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
dc.subject.eng.fl_str_mv Non-Intrusive load monitoring
Load classification
Convolutional neural network
Embedded systems
dc.subject.cnpq.fl_str_mv CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
description Electric power is very important for the economic development of all countries, and its consumption has been growing at an impressive rate, doing so faster than other types of power. Along with the increase in consumption, there are also concerns about environmental sustainability; after all, ensuring access to electric power in a reliable, sustainable, modern, and affordable way for all is one of the objectives of the Sustainable Development Goals, an agenda proposed by the United Nations (UN). In addition to encouraging the usage of renewable and less environmentally impactful energy, there are also concerns to create increasingly more power efficient devices, and to reduce the waste of electric power by seeking alternatives for a more efficient use of it. The active involvement of consumers results, for the most part, in a more efficient use of electric power, which increases interest in the development of technologies that make them aware of their habits. Studies show that, the greater the detail of information about electrical power consumption, the greater the amount of electric power saved by consumers. One of the most used techniques to analyze such details is Non-Intrusive Load Monitoring (NILM), who, by disaggregating the loads, distinguishes between each of the appliances and explores the electric power consumption of each one individually. Therefore, in order to contribute to this technique, and considering the ever-growing progress in the electronic and machine-learning areas, this study proposes a set of training strategies using a deep-learning method for load classification in an embedded system, and therefore, contribute to a more efficient use of electric power. Based on literature and experiments, we adopted the binary image of the voltage-current as the distinguishing feature, as it obtained the best results. In order to classify the devices, we used said images as input to the Convolutional Neural Network (CNN), which was chosen after obtaining the best results in the tests that were performed. After we used the leave-one-out cross-validation method, our CNN model was evaluated using the PLAID dataset and obtained an F-Score macro-average of 74.76% for PLAID1, 56.48% for PLAID2, and 73.97% for PLAID1+2, and those results were very close to literature. The novelty of this study is the quantization of the CNN model using TensorFlow Lite, and its application in a resource-constrained embedded system (ESP32). The accuracy rate achieved in testes performed with all data from the PLAID1+2 dataset was 98.55%, which shows that the embedded device can be used to perform the load classification with a high accuracy rate.
publishDate 2020
dc.date.issued.fl_str_mv 2020-11-25
dc.date.accessioned.fl_str_mv 2021-05-06T16:52:21Z
dc.date.available.fl_str_mv 2021-05-06T16:52:21Z
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.citation.fl_str_mv TAVARES, Hugo Menezes. Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional. 2020. 150 f. Dissertação (Mestrado em Ciência da Computação) - Universidade Federal de Sergipe, São Cristóvão, Sergipe, 2020.
dc.identifier.uri.fl_str_mv https://ri.ufs.br/jspui/handle/riufs/14187
dc.identifier.license.pt_BR.fl_str_mv Autorização para publicação no Repositório da Universidade Federal de Sergipe (RI-UFS), concedida pelo autor.
identifier_str_mv TAVARES, Hugo Menezes. Classificação para o monitoramento não-intrusivo de cargas em sistema embarcado com rede neural convolucional. 2020. 150 f. Dissertação (Mestrado em Ciência da Computação) - Universidade Federal de Sergipe, São Cristóvão, Sergipe, 2020.
Autorização para publicação no Repositório da Universidade Federal de Sergipe (RI-UFS), concedida pelo autor.
url https://ri.ufs.br/jspui/handle/riufs/14187
dc.language.iso.fl_str_mv por
language por
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.publisher.program.fl_str_mv Pós-Graduação em Ciência da Computação
dc.publisher.initials.fl_str_mv Universidade Federal de Sergipe
dc.source.none.fl_str_mv reponame:Repositório Institucional da UFS
instname:Universidade Federal de Sergipe (UFS)
instacron:UFS
instname_str Universidade Federal de Sergipe (UFS)
instacron_str UFS
institution UFS
reponame_str Repositório Institucional da UFS
collection Repositório Institucional da UFS
bitstream.url.fl_str_mv https://ri.ufs.br/jspui/bitstream/riufs/14187/1/license.txt
https://ri.ufs.br/jspui/bitstream/riufs/14187/2/HUGO_MENEZES_TAVARES.pdf
https://ri.ufs.br/jspui/bitstream/riufs/14187/3/HUGO_MENEZES_TAVARES.pdf.txt
https://ri.ufs.br/jspui/bitstream/riufs/14187/4/HUGO_MENEZES_TAVARES.pdf.jpg
bitstream.checksum.fl_str_mv 098cbbf65c2c15e1fb2e49c5d306a44c
fdef03084f92492c225bba5be693dbca
c86f09c4c24ffd9a57bf6d885cf84eef
8ab48e9b850cb7f9d3f63ec947e073b5
bitstream.checksumAlgorithm.fl_str_mv MD5
MD5
MD5
MD5
repository.name.fl_str_mv Repositório Institucional da UFS - Universidade Federal de Sergipe (UFS)
repository.mail.fl_str_mv repositorio@academico.ufs.br
_version_ 1802110658134671360