On-Site Electric Motor Health Monitoring: from Condition Assessment to Predictive Maintenance
Part Of
IFEEC 2025 - Proceedings of the 2025 International Future Energy Electronics Conference
Start Page
125
End Page
129
ISBN (of the container)
9798331549473
ISBN
9798331549473
Date Issued
2025
Author(s)
Abstract
This paper proposes an innovative predictive maintenance (PdM) framework integrating sensor design with state estimation techniques to effectively monitor condition changes in AC motors during operation. In this research, the data acquisition and state observer are conducted using the data acquisition box (AQbox), a low-cost external data acquisition device developed by the NTU Motor Tech Lab. To verify the practical applicability of this framework in industrial environments, four AQboxes have been installed on cooling circulation AC pump motors at Pacific Electric Wire & Cable Co. (PEWC) to test the feasibility of state estimation within the PdM system. This paper presents and analyzes torque anomaly data observed during monitoring. Based on these data characteristics, a robust two-step triggering recursive least squares (RLS) method is developed to predict optimal maintenance timing, effectively addressing data growth and instability while lack of system degradation data. The AQbox will continue to collect field data to form a database for further developing neural network-based diagnostic methods using these electrical signal waveform anomalies.
Event(s)
2025 International Future Energy Electronics Conference, IFEEC 2025, 19 November 2025 - 21 November 2025, Bali
Subjects
condition monitoring
predictive maintenance
state observer
Publisher
Institute of Electrical and Electronics Engineers Inc.
Type
conference paper