Reliable operation of main mine ventilation fans is essential for mine safety and production continuity. In many mines, fan condition is still judged against fixed vibration limits that do not account for changes in airflow and pressure, which can lead to frequent false alarms and ambiguous interpretation of mechanical degradation. This paper proposes a maintenance-event-constrained vibration health index (MEC-HI) that combines operating-condition modelling with long-term residual vibration analysis using SCADA-level measurements. A linear regression model is first fitted to relate bearing RMS vibration velocity to airflow, differential pressure and motor electrical quantities during confirmed healthy operating phases. The model is then used to estimate the expected vibration level and to compute condition-normalised residual vibration. Positive residuals exceeding a statistically derived tolerance are smoothed and accumulated over time within segments separated by major maintenance events, and the cumulative index is reset after bearing replacement. Unlike many health-indicator studies that rely on high-frequency waveforms or fault-specific feature engineering, the proposed framework targets practical deployment when only routine RMS and operating tags are archived. The approach is demonstrated using a three-year dataset (24,672 operating hours) from an axial-flow main fan in a large underground copper mine. The case study shows that MEC-HI captures the onset and progression of bearing degradation more clearly than raw RMS trends and suppresses load-driven false alarms, while remaining implementable with routinely available SCADA measurements. The framework can be extended to other ventilation fans and rotating machinery operating under strongly varying loads.
Meng et al. (Thu,) studied this question.