Predictive analysis of photovoltaic generation and liquidity price in the energy market via artificial intelligence techniques and exploratory data analysis

• Grant number: 21/08298-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): October 01, 2021
• Effective date (End): September 30, 2022
• Field of knowledge: Engineering - Electrical Engineering - Power Systems
• Principal Investigator: Wallace Correa de Oliveira Casaca
• Grantee: Leonardo Fernando Fini
• Host Institution: Faculdade de Engenharia e Ciências (FEC). Universidade Estadual Paulista (UNESP). Campus de Rosana. Rosana , SP, Brazil

Abstract

In the speculative and competitive environment such as the energy market, predicting electricity generation, as well as its corresponding price, are very challenging tasks that can support decision-making for the market agents. In addition to wind and hydraulic generation, photovoltaic energy has attracted a lot of contracts signed in this market. Therefore, this project aims to study the problem of forecasting the energy generated at the Pirapora solar plant, the second-largest solar complex in Brazil, as well as predicting its corresponding price in the national energy market. For this purpose, we will study Exploratory Data Analysis (AED) tools and Artificial Intelligence (AI) models, including Random Forest, Support Vector Machines, and Gradient Boosting. The AI models will be prototyped and validated by combining different data collections, which gather several energy-related features such as solar energy generation, climate data, and natural affluent energy and liquidity prices, thus enabling the development of new strategies and computational apparatus for agents in the Brazilian electricity market. (AU)