A Predictive Maintenance Approach for Centrifugal Pumps: Case Study of Vibration Analysis Using FFT and Phase Difference Diagnostics

Effective condition monitoring is crucial for maintaining industrial rotating machinery. This study investigates a predictive maintenance approach to diagnose and resolve excessive vibration in a refinery’s centrifugal pump. Vibration analysis using Fast Fourier Transform (FFT) and phase difference diagnostics identified a dynamic imbalance caused by contaminant accumulation on the impeller, resulting in vibration spikes of 11.61 mm/s with a dominant FFT spectrum at 1X RPM. To mitigate this issue, dynamic balancing was performed in accordance with ISO 1925 and ISO 1940-1 standards, successfully reducing the vibration to 3.59 mm/s. The results validate vibration-based diagnostics as a proactive maintenance tool, minimizing downtime and improving reliability. This study demonstrates how predictive maintenance strategies can enhance the performance of rotating machinery, with potential applications in other industrial systems.