Photovoltaic Power Forecasting using Machine Learning Methods

• Grant number: 23/00297-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2023
• End date: March 31, 2024
• Field of knowledge: Engineering - Electrical Engineering - Power Systems
• Principal Investigator: Marcos Julio Rider Flores
• Grantee: Bruno Vitor Barros Bandeira
• Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated research grant: 21/11380-5 - CPTEn - São Paulo Center for the Study of Energy Transition, AP.CCD

Abstract

Photovoltaic (PV) power plants require accurate forecasting to achieve satisfactory insertion in electricity grids. In other words, a precise PV forecast (PPF) helps grid operators optimize dispatch to balance generation and demand; and, in this way, make better use of this renewable resource. Thus, this Scientific Initiation research project proposes applying two machine learning methods: Long-Short Term Memory (LSTM) and Temporal Convolutional Networks (TCN), for the day-ahead PPF. For both machine learning methods, the day-ahead PPF is obtained with time intervals of fifteen (15) minutes. The proposed methodologies will be programmed in Python, using the TensorFlow open-source libraries for machine learning applicable to various tasks and Pandas for data management, manipulation, and analysis. The methodologies will be validated using a dataset corresponding to a photovoltaic plant located at the State University of Campinas (UNICAMP) in Brazil, with an installed capacity of 336 kWp. This research is part of the Microgrids for Efficient, Reliable and Greener Energy (MERGE) project.