Desenvolvimento de uma infraestrutura de dados com código e com capacidade analítica para monitoramento de bombas industriais utilizando internet das coisas industrial

• Ano de defesa: 2024
• Autor(a) principal: Marcelo Simões Reis
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ICB - DEPARTAMENTO DE FISIOLOGIA E BIOFÍSICA; Programa de Pós-Graduação em Inovação Tecnológica e Propriedade Intelectual; UFMG
• Palavras-chave em Português: IIOT; IaC; LoRa; Containers; Kubernetes; Inovação tecnológica; Internet das Coisas; Arquitetura de rede de sistemas (SNA); Indústria
• Link de acesso: http://hdl.handle.net/1843/77158

This work presents the development and implementation of a data infrastructure for industrial environments, covering the entire information flow, from data collection from sensors in the field, cloud storage to presentation on dashboards. The hardware development steps, decisions on communication technologies and wireless network protocols, the definition of the architecture, the implementation of firmware and software for cost and performance optimization, and a data structure model are described. Initially, the development is a low-cost physical infrastructure designed to connect islands of information within industrial environments. The objective was to ensure that even the most isolated areas within an industrial plant can communicate effectively with the central system and create a poor database. Next, research focused on creating a scalable and cost-effective logical infras-tructure. This infrastructure was designed to store large volumes of industrial data, with a fo-cus on rapid implementation and traceability of modifications made, using concepts of contai-ners and infrastructure as code. Finally, the capacity of the data infrastructure, both logical and physical, to provi-de information on monitored elements was guaranteed so that third-party systems can con-sume this data for analysis. This ability to provide quality data with good response time is cru-cial to transforming the industry's maintenance profile from a reactive model to a proactive model. The results obtained demonstrate the solutions and effectiveness of the proposed solution, offering a robust and low-cost infrastructure that can be easily integrated and scaled in different industrial environments. The dissertation concludes that the implementation of such systems can bring significant operational and economic improvements to the industry, promoting proactive maintenance and more efficient data management.