Comparative Analysis and Improvement of Predictive Maintenance Techniques in Industry 4.0 using Deep Neural Networks

• Grant number: 23/10429-6
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date until: January 01, 2024
• End date until: December 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Lilian Berton
• Grantee: Lucas Molinari
• Host Institution: Instituto de Ciência e Tecnologia (ICT). Universidade Federal de São Paulo (UNIFESP). Campus São José dos Campos. São José dos Campos , SP, Brazil

Abstract

Predictive maintenance is the periodic monitoring of equipment or machines, through data collected in field measurements such as temperature, vibration, thermography, etc. This maintenance predicts the lifetime of machine components, preventing failures by monitoring equipment in operation. With this, we have several benefits, such as predicting the need for maintenance of equipment, reducing emergency stops, and increasing the useful life and time of equipment availability. Machine learning algorithms can learn new information and detect patterns over time. From sensors collecting real-time data and historical datasets of monitored assets, machine learning algorithms can predict what will go wrong and when. The present work aims to study this topic and apply it to predictive maintenance using machine learning techniques, especially deep neural networks comparing the approaches. We will also develop the front end for the equipment maintenance system, with the aim of providing an efficient and intuitive solution that assists companies in the management and maintenance of their equipment, providing accurate and real-time information on the health of the machines.