Loading...
Publication
Sistema de hardware para monitorização não intrusiva de motores elétricos
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 6.5 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Este projeto aborda a manutenção preditiva em contextos industriais, com foco em sistemas de refrigeração e de bombagem. Utilizando uma arquitetura de Internet das Coisas (IoT), o sistema combina hardware personalizado e algoritmos de aprendizagem de máquina para monitorizar o desempenho de equipamentos e detetar anomalias. A inovação do projeto reside na sua facilidade de instalação e natureza não intrusiva, permitindo uma integração eficaz com infraestruturas existentes. O hardware foi testado tanto em ambientes laboratoriais quanto em contextos industriais reais, como estações de tratamento de água e sistemas de ar condicionado em hotéis.
Durante os testes em campo, surgiram desafios na medição de parâmetros como as temperaturas e o consumo energético, embora estes tenham sido bem-sucedidos em ambientes laboratoriais. A falta de conectividade à internet em alguns ambientes industriais foi contornada através de soluções de armazenamento e processamento local de dados, assegurando a continuidade da monitorização, mesmo em condições adversas.
Um ponto forte do sistema é o seu custo económico, oferecendo uma alternativa acessível em comparação a outras soluções mais caras ou específicas a determinadas marcas de equipamentos no mercado. Isso não só aumenta a sua aplicabilidade em diferentes contextos industriais, como também demonstra a sua viabilidade como uma solução de monitorização eficiente e económica.
Este trabalho tem implicações significativas tanto a nível académico como industrial, servindo como um modelo multidisciplinar para a implementação de sistemas IoT inteligentes em cenários industriais. Destaca áreas para melhoria e desenvolvimentos futuros, como a otimização do hardware e a integração com outros sistemas de gestão, solidificando o seu papel como um sistema inovador na manutenção preditiva.
This project addresses predictive maintenance in industrial contexts, focusing on refrigeration and pumping systems. Utilizing an Internet of Things (IoT) architecture, the system combines custom hardware and machine learning algorithms to monitor equipment performance and detect anomalies. The project's innovation lies in its ease of installation and non-intrusive nature, allowing for effective integration with existing infrastructures. The hardware was tested in both laboratory environments and real industrial settings, such as water treatment stations and hotel air conditioning systems. During field tests, challenges arose in measuring parameters such as temperature and energy consumption, although these were successfully measured in laboratory environments. The lack of internet connectivity in some industrial environments was overcome through local data storage and processing solutions, ensuring continued monitoring even in adverse conditions. A strong point of the system is its low economic cost, offering an affordable alternative compared to more expensive solutions or those specific to certain equipment brands on the market. This not only increases its applicability in different industrial contexts but also demonstrates its viability as an efficient and economical monitoring solution. This work has significant implications both academically and industrially, serving as a multidisciplinary model for the implementation of intelligent IoT systems in industrial scenarios. It highlights areas for improvement and future developments, such as hardware optimization and integration with other management systems, solidifying its role as an innovative system in predictive maintenance.
This project addresses predictive maintenance in industrial contexts, focusing on refrigeration and pumping systems. Utilizing an Internet of Things (IoT) architecture, the system combines custom hardware and machine learning algorithms to monitor equipment performance and detect anomalies. The project's innovation lies in its ease of installation and non-intrusive nature, allowing for effective integration with existing infrastructures. The hardware was tested in both laboratory environments and real industrial settings, such as water treatment stations and hotel air conditioning systems. During field tests, challenges arose in measuring parameters such as temperature and energy consumption, although these were successfully measured in laboratory environments. The lack of internet connectivity in some industrial environments was overcome through local data storage and processing solutions, ensuring continued monitoring even in adverse conditions. A strong point of the system is its low economic cost, offering an affordable alternative compared to more expensive solutions or those specific to certain equipment brands on the market. This not only increases its applicability in different industrial contexts but also demonstrates its viability as an efficient and economical monitoring solution. This work has significant implications both academically and industrially, serving as a multidisciplinary model for the implementation of intelligent IoT systems in industrial scenarios. It highlights areas for improvement and future developments, such as hardware optimization and integration with other management systems, solidifying its role as an innovative system in predictive maintenance.
Description
Keywords
Manutenção preditiva Internet das coisas Aprendizagem de máquina Monitorização industrial Sensores Anomalias