Comparative Analysis Among Soft Computing Techniques for Non-Technical Losses Detection in Power Distribution Systems

• Grant number: 23/14980-9
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: February 01, 2024
• End date: April 30, 2024
• Field of knowledge: Engineering - Electrical Engineering - Power Systems
• Principal Investigator: Lucas Teles de Faria
• Grantee: Luiz Paulo Barbosa do Nascimento Filho
• Supervisor: Tiago Manuel Campelos Ferreira Pinto
• Host Institution: Faculdade de Engenharia e Ciências (FEC). Universidade Estadual Paulista (UNESP). Campus de Rosana. Rosana , SP, Brazil
• Institution abroad: Instituto de Engenharia de Sistemas e Computadores - Tecnologia e Ciência (INESC TEC), Portugal
• Associated to the scholarship: 23/03151-1 - Non-Technical Losses Estimation: A Top-Down Approach, BP.IC

Abstract

Non-technical losses or commercial losses cause financial losses to power utilities and weaknesses in the distribution power grids with damage to power quality (with an increase in blackouts) and the power grid reliability due to undue changes in the grid topology. These losses are commonly caused by theft (clandestine connections) and energy meter fraud. There are significant technological advances with the advent of smart grids and phasor measurement units (PMUs) at strategic points of distribution power grids. However, these advances are not available on a large scale in developing countries, particularly those with the most significant losses. Therefore, it is necessary to study methodologies for preventing and combating non-technical losses in conventional distribution power grids. In this sense, this project aims to implement a comparative analysis among soft computing techniques to identify irregular consumer units such as ARTMAP Fuzzy neural network and support vector machines (SVMs). Field inspections are time-consuming and have a high person-hour cost. In this way, identifying the most appropriate technique for detecting irregular consumer units minimizes false positives - unnecessary inspections in regular consumers.