GreenFramework: um framework para monitoramento inteligente de recursos energéticos em cidades inteligentes

Abstract

Amid the rising global energy demand and the expansion of smart cities utilizing IoT for integrated management of electricity and water resources, public institutions such as UFMA face a lack of low-cost monitoring tools, leading to waste, high operational costs, and elevated CO2 emissions due to inefficiencies in building load management. This work proposes the GreenMonitoring framework, based on IoT, Artificial Intelligence, and Model-Driven Engineering (MDE), to support intelligent monitoring of electricity and water in buildings. The framework integrates data from collection via sensors to availability through Application Programming Interfaces (APIs) or a web application front-end. The study’s motivation stems from the need to align institutional management with the Sustainable Development Goals (SDG 7 and 11), overcome the scarcity of modular open-source platforms, the predominance of complex and expensive proprietary systems, the limited integration of the water-energy nexus in university campuses, and the lack of intuitive interfaces for non-technical users, based on the hypothesis that an IoT- and machine learning-based framework can identify consumption patterns, detect anomalies, and predict demands with high accuracy, significantly reducing waste of both resources. The GreenFramework is proposed as a modular and scalable architecture for integrated resource management, instantiated in this work as GreenMonitoring. In the case study, GreenMonitoring was deployed to monitor electrical substations and water intake in UFMA buildings. Electricity measurements were conducted in six UFMA buildings, identifying demand peaks between 1 p.m. and 5 p.m., waste during low-occupancy hours, and CO2 emissions calculated based on the monthly MCTI factor, along with predictive models showing excellent performance (R2 up to 0.9725 in the Center for Exact Sciences and Technology building — CCET); for water consumption, only one meter was installed at the Institute of Electrical Energy (IEE)—due to the large physical distance between water and energy measurement points, which hindered installation in multiple buildings—with the equipment functioning correctly and data collection and transmission fully validated, although deeper quantitative analyses and identification of water consumption peaks remain as future expansion prospects.