IMPLEMENTATION OF A PREDICTIVE MAINTENANCE SYSTEM WITH ONLINE MONITORING OF ELECTRIC MOTORS, BASED ON THE USE OF ARTIFICIAL INTELLIGENCE

AN EXPERIENCE REPORT FROM THE INDUSTRIAL ELECTRICAL INSTALLATIONS CLASS

Abstract

Abstract: This work addresses the implementation of a predictive maintenance system with online monitoring in high-voltage (6.6 kV) electric motors. The initiative responds to recurring failures due to thermal degradation of insulation, caused by excessive starts in short intervals, a problem that traditional methods, such as the Baker Test, failed to anticipate. To mitigate these risks and increase the reliability and availability of assets, the project selected, via the INO.VC ecosystem, the technological solution from the startup 2Neuron, composed of the Ultronline and Ultronlink devices. This technology uses artificial intelligence to perform continuous monitoring (24/7) of electrical signals, allowing for the early detection of both electrical and mechanical anomalies, with automatic diagnoses that evolve through machine learning. The detailed planning includes a schedule, definition of roles between contractor and client, a test plan (PIT), and targets for reducing unscheduled downtime, optimizing industrial maintenance management under the Industry 4.0 paradigm.

Keywords: Predictive maintenance, Online monitoring, Artificial intelligence, Electric motors, Industrial reliability.