Arquitetura computacional para controle de monitoramento de qualidade de energia utilizando ferramentas de inteligência artificial focado na Indústria 4.0 com integração de energias renováveis

Abstract

Electric power quality (QEE) is crucial for the efficiency of industrial systems in Industry 4.0. This dissertation proposes an innovative computational architecture to control and monitor QEE, integrating renewable energy sources and artificial intelligence (AI). Key metrics like power factor (PF) and total harmonic distortion (THD) directly influence operational costs and equipment lifespan. AI-based predictive strategies dynamically manage capacitor banks and harmonic filters, mitigating issues such as inadequate PF and high THD. Reactive power compensation (CPR) is central to improving PF, reducing system losses, and enhancing energy efficiency. The system also integrates solar energy management, optimizing energy savings and sustainability. Data is collected using IMS Smart Cap 485 devices via the Modbus-RTU protocol, measuring voltage, current, active/reactive power, and THD. Stored in a MySQL database, the data is analyzed in real time with deep learning (LSTM) and optimization (AG) algorithms. Python-based dashboards visualize data, predict network issues, and support strategic actions, such as activating capacitors and THD filters. This approach highlights AI's role in energy efficiency and reducing reliance on conventional sources. Gaps in literature include the lack of interoperability standards, the need for explainable AI algorithms, and limited longitudinal studies on real-world applications. This research addresses these challenges, contributing to Industry 4.0's sustainability and efficiency goals.