Scenario
You've taken over a piece of equipment, and its PM plan calls for maintenance every 2 weeks. You reviewed the past 6 months of breakdown records and found that this machine has never failed within 2 weeks after maintenance. Instead, all breakdowns occurred between the 8th and 10th week.
This PM plan involved unnecessary labor and expense, and maintenance was not performed at the right time.
The Bathtub Curve: Understanding Equipment Life Cycles
Equipment failure rates over time exhibit three phases, resembling a bathtub:
Early Failure Period (Infant Mortality)
- Higher failure rate after new equipment or parts installation
- Causes: Manufacturing defects, improper installation, material flaws
- Countermeasures: Run-in tests after installation, IQ/OQ/PQ validation
Random Failure Period (Useful Life)
- Failure rate stabilizes and occurs randomly
- Causes: External impacts, operational errors, environmental factors
- Countermeasures: SOP standardization, operator training, condition monitoring
Wear-out Failure Period (Aging)
- Failure rate rises rapidly as equipment approaches end of life
- Causes: Wear, fatigue, corrosion
- Countermeasures: Preventive Maintenance (PM), predictive maintenance, component replacement
PM Frequency Setting Methods
Method One: Time-Based PM (Based on MTBF)
Principle: Perform PM at 60-80% of MTBF
Example: A critical component has an MTBF = 1000 hours
Suggested PM interval = 700-800 hours
Logic: Perform maintenance before the probability of failure significantly increases, avoiding waiting for it to break down before replacement.
Method Two: Condition-Based Maintenance (CBM)
Not time-based, but based on equipment condition indicators:
- Vibration value exceeds threshold → triggers maintenance
- Oil quality deteriorates → oil change
- Rising temperature trend → pre-emptive inspection
Condition-based PM is more precise but requires sensor and monitoring system support.
PM Effectiveness Verification
Performing PM does not automatically mean it's effective; tracking is required:
- MTBF before PM vs MTBF after PM: Does MTBF recover after PM completion?
- PM cost vs Breakdown cost: PM labor + parts cost < loss from breakdown downtime is considered reasonable
- Over-maintenance indicator: Most parts replaced during PM are still usable = PM is too frequent
Common Mistake: Over-Maintenance
A common issue in Taiwanese factories: "Maintenance can't hurt anyway."
In reality:
- Every disassembly and reassembly carries the risk of introducing new problems (installation errors, contamination)
- Wastes spare parts and labor costs
- Equipment lifespan is shortened due to frequent disassembly and reassembly
Golden Quote
"The goal of preventive maintenance is not to do as much as possible, but to do the right things at the right time. A PM plan is a living document that must be continuously calibrated with equipment data, rather than remaining unchanged after being set five years ago."