Desenvolvimento de um sistema IOT de baixo custo para monitoramento de grandezas elétricas e sua aplicação em usinas fotovoltaicas

Abstract

This work presents the development and validation of an Internet of Things (IoT)-based monitoring system applied to a medium-sized photovoltaic power plant. The proposed solution aims to provide a low-cost, flexible, and independent alternative to the native monitoring platforms embedded in inverters, which often limit data access, customization of indicators, and the depth of performance analysis. The system was implemented using an ESP32 microcontroller as the acquisition device, performing readings through the Modbus TCP protocol and transmitting the collected measurements to the cloud via MQTT. The data were stored in a database and visualized through a dashboard developed in Power BI. To validate the system, a case study was conducted at the UFV 05 – Solar Power – Oeste plant, located in Teresópolis, Goiás, composed of three Sungrow SG333HX inverters and 1,944 photovoltaic modules. The practical application demonstrated that the system was able to continuously collect, process, and transmit operational data, enabling instant, daily, and monthly generation analyses. The results obtained were consistent with the expected performance of the plant, confirming the accuracy of the readings and the stability of the solution under real operating conditions. In addition to the technical validation, an economic feasibility analysis was performed, highlighting the low cost of the proposed system — approximately R$ 30.00 initial investment plus R$ 30.00 per month for infrastructure — compared to commercial monitoring solutions, which can exceed R$ 250.00 per asset per month or R$ 1,500.00 per month in specialized platforms. This cost difference reinforces the potential of the solution to expand monitoring capabilities in small and medium-sized photovoltaic plants, democratizing access to advanced supervision tools. The results demonstrate that the developed IoT solution is technically feasible, economically advantageous, and suitable for real distributed generation contexts, offering a robust and scalable alternative for monitoring the performance of photovoltaic power plants.