Plataforma computacional baseada em Edge Computing para estimativa de geração e diagnóstico operacional de sistemas fotovoltaicos utilizando ESP32 e dados meteorológicos em tempo real

Abstract

The expansion of solar photovoltaic generation in Brazil has increased the demand for tools capable of monitoring and estimating the energy performance of photovoltaic systems under real operating conditions. However, environmental factors such as elevated photovoltaic cell temperatures and the accumulation of particulate matter on module surfaces can significantly reduce system efficiency. To overcome these limitations, this paper introduces a computational platform leveraging edge computing architecture for the monitoring and operational diagnosis of photovoltaic systems. The proposed solution employs the ESP32 microcontroller as an edge node responsible for the local processing of real-time meteorological data obtained through the Tomorrow.io API. The platform implements analytical models for estimating thermal and soiling losses, enabling the prediction of generated power and energy performance indicators. The results are delivered through a Human-Machine Interface consisting of a responsive web dashboard, historical data storage using the LittleFS embedded file system, and a voice assistant based on the Web Speech API. The validation process was conducted through computational simulations using real meteorological data from the city of Goiânia, Goiás, Brazil. The results demonstrated the platform’s ability to process information locally, estimate operational losses, and provide real-time energy indicators while reducing dependence on external processing resources. It is concluded that the proposed approach presents significant potential for low-cost photovoltaic monitoring applications based on edge computing technologies.