Development of an anomalous data treatment method dedicated to ADMS applications

• Grant number: 22/10271-0
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate
• Start date: December 01, 2022
• End date: November 30, 2023
• Field of knowledge: Engineering - Electrical Engineering - Power Systems
• Principal Investigator: Fernanda Caseño Trindade Arioli
• Grantee: Felipe Bayma Barbosa Rolim
• Supervisor: Bala Venkatesh
• Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Institution abroad: Toronto Metropolitan University (TMU), Canada
• Associated to the scholarship: 20/07103-3 - Anomaly detection in electrical power distribution systems based on sensors and smart meters data analytics, BP.DR

Abstract

The electrical power distribution systems have been subjected to a modernization process characterized, among other factors, by a higher level of monitoring and measurement. As such, the amount of installed smart meters is increasing worldwide. The global smart meter market was evaluated at US$ 21.13 billion in 2019, and, according to Allied Market Research, it is projected to grow at a Compound Annual Growth Rate of 8.8%, characterizing a market value of US$ 39.2 billion by 2027. The smart meters and other meters installed throughout the feeders are part of the Advanced Metering Infrastructure, which provides essential data to improve applications of the Advanced Distribution Management System, such as losses management, outage management, and Volt-var control. However, the improved observability of the systems comes with increased vulnerability of the distribution system operation management due to data anomalies. Different methodologies based on artificial intelligence, statistics, or optimization have been proposed. Nonetheless, methodologies integrating physics-based approaches leading to better performance for anomalous measurement data treatment are still required. Therefore, this project proposes a scalable and automatic anomalous data treatment methodology that continuously detects, filters, and recovers anomalous data into the dataset, resulting in reliable processed input data and a diagnosis report specifying the possible occurred event. Furthermore, applying the processed data, an aggregate model of distributed energy resources will be developed. Real measurement datasets will be applied to carry out studies to test and validate the methodology, exploiting the variety of measurement data. (AU)