• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.3.2022.tde-23032023-080022
Document
Author
Full name
Jônatas Pulz
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Paulo, 2022
Supervisor
Committee
Almeida, Carlos Frederico Meschini (President)
Meffe, André
Rosa, Luiz Henrique Leite
Title in Portuguese
Abordagem alternativa para cálculo regulatório de perdas não-técnicas do sistema de distribuição e técnicas de machine learning para detecção de fraude na baixa tensão.
Keywords in Portuguese
Aprendizado computacional
Energia elétrica
Furto
Abstract in Portuguese
As perdas não técnicas são um problema significativo em países subdesenvolvidos, decorrente, principalmente, de fraudes em medidores e furtos de energia. Para mitigá-las, as distribuidoras realizam inspeções em unidades consumidoras suspeitas. O custo operacional para se realizar essas inspeções é alto e só pode ser justificado por um retorno através da descoberta de fraudes. Para aumentar a precisão na descoberta de fraudes, modelos de machine learning podem ser utilizados. Este trabalho propõe modelos de detecção de fraudes utilizando os tipos de modelos mais atuais e que vem se destacando como bons classificadores. Além disso, este trabalho propõe uma metodologia de cálculo regulatório de perdas mais realista que leve em consideração esse rico banco de dados criado através das inspeções realizadas pelas distribuidoras e o compara com a metodologia regulatória atual numa área piloto da distribuidora Enel de São Paulo.
Title in English
Alternative approach for regulatory non-technical losses calculation of distribution system and machine learning techniques for fraud detection at low voltage.
Keywords in English
Analytics
Electricity theft
Energy losses
Fraud detection
Machine learning
Non-technical losses
Regulatory losses
Technical losses
Abstract in English
Non-technical losses are a significative problem in developing countries, mainly due to tampering in meters and energy theft. To mitigate them, utilities carry out inspections in suspected consumers. The operational cost to carry out these inspections is high and can only be justified by a return through fraud discovery. To increase the precision of fraud detection, machine learning models can be used. This work presents fraud detection models using the most recent types of models that have been showing good results as classifiers. In addition, this work proposes a more realistic regulatory loss calculation methodology that takes into account this rich database created through inspections carried out by utilities and compares it with the current regulatory methodology in a pilot area of the utility Enel of São Paulo.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
JonatasPulzCorr22.pdf (6.01 Mbytes)
Publishing Date
2023-03-27
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.