Correlation Studies of a Multi-Sensor Data for Induction Motor Health Assessment

Abstract

Induction motor failures impose substantial costs on industrial operations, yet conventional maintenance strategies struggle to balance preventive interventions with the risk of unexpected breakdowns. This study investigates the application of statistical analysis and correlation techniques to multi-sensor data for early detection of motor faults. It focuses on how vibration, temperature and current of an induction motor can be monitored under different operating conditions. In this study, a system having a vibration sensor, current sensor and temperature sensor was designed and tested on a 3-phase induction motor. The test was carried out for three days and the results were analyzed using Python. The analysis produced a correlation matrix and a performance graph which forms a basis for the health assessment of the induction motor. The results show that multi-sensor correlation patterns offer superior diagnostic capability compared to single-parameter thresholds. The performance analysis graph shows that the cycle of the motor goes on and off with the peak current around 10 A, the core temperature around 55  to 60 ℃, while the vibration level is around 15 . Also, the correlation matrix shows a high correlation between these measurement criteria.  This ascertains the effectiveness of the designed system and reveals the efficiency achieved in multi sensor data assessment, thereby making them accessible for widespread deployment in condition-based maintenance programs.