Preventive and Predictive Maintenance
From schedule to condition
Preventive maintenance is calendar- or runtime-based: lubrication routes, filter changes, belt inspections. Predictive (condition-based) maintenance acts on measured condition - vibration, thermography, oil analysis, ultrasound, and performance trending. Reliability-centered maintenance asks what function matters and which failure mode threatens it. The P-F interval is the time between a detectable potential failure and functional failure; inspection frequency must be shorter than that interval.
Vibration
Match units to the problem: displacement for low speed, velocity (in/s or mm/s RMS) for the mid band roughly 10–1000 Hz, acceleration for high frequency and early bearing defects. ISO 20816/10816 zones A through D rank severity. Classic signatures: 1× for unbalance, 2× for misalignment, many harmonics for looseness, bearing defect frequencies (BPFO, BPFI, BSF, FTF) with sidebands, and broadband noise for cavitation or rubbing.
Oil analysis
Track particle counts (ISO cleanliness), viscosity, water content, and wear metals such as iron, copper, chromium, lead, and tin against a baseline. A sudden iron rise indicates wear; silicon indicates dirt ingress.
Thermography and ultrasound
Thermal imaging finds hot electrical connections, bearing heat, and failed steam traps. Ultrasound detects leaks, steam, and slow-speed bearing faults.
Program discipline
CMMS work orders with as-found and as-left data, alarm limits built from baseline, and route-based collection at identical points with identical mounting.
Traps
Reading 1× as always unbalance (could be resonance or a bent shaft), ignoring sensor mounting, and setting alarms from a single reading.
Rule of thumb
Trend against baseline and alarm on change - a doubling of vibration velocity is significant regardless of absolute value.