1. Definitions: Repairable vs. Non-Repairable Systems
Before computing metrics, engineers must distinguish between items that are discarded upon failure and systems that are restored to operating condition:
- MTBF (Mean Time Between Failures): Strictly applies to repairable systems (such as pumps, compressors, conveyers, and circuit breakers). It measures the average elapsed operating time between consecutive failure events.
- MTTF (Mean Time to Failure): Strictly applies to non-repairable items (such as incandescent light bulbs, disposable mechanical seals, or single-use fuses). Once they fail, their operating life is permanently terminated.
- MTTR (Mean Time to Repair): The average corrective maintenance downtime required to diagnose, disassemble, replace parts, reassemble, and recommission an asset back to full production capability.
2. Mathematical Formulas
For an observation window covering total calendar duration, the asset alternates between operating uptime and repair downtime:
MTBF = Total Operating Time / Number of Failures (N)
MTTR = Total Repair Downtime / Number of Failures (N)
Failure Rate (λ) = 1 / MTBF (failures per operating hour)
Calculating Inherent vs. Operational Availability
Availability represents the mathematical probability that equipment will operate satisfactorily at any given moment under specified conditions:
Operational Availability (A_o) = Mean Time Between Outages / (MTBO + Mean Downtime)
An industrial centrifugal slurry pump is monitored continuously over a 90-day quarterly period (2,160 calendar hours). Maintenance logs record 4 unplanned breakdown events resulting in a combined downtime of 72 hours:
- Step 1: Calculate Total Operating Time
Operating Hours = 2,160 total hours - 72 downtime hours = 2,088 operating hours - Step 2: Calculate MTBF
MTBF = 2,088 hours / 4 failures = 522.0 operating hours/failure - Step 3: Calculate MTTR
MTTR = 72 hours / 4 repairs = 18.0 hours/repair - Step 4: Calculate Inherent Availability (A_i)
A_i = 522.0 / (522.0 + 18.0) = 522.0 / 540.0 = 96.67% - Step 5: Calculate Instantaneous Failure Rate (λ)
λ = 1 / 522.0 = 0.001916 failures per hour (or 1.916 failures per 1,000 running hours)
3. Common Pitfalls and Statistical Limitations
While MTBF is widely reported in CMMS dashboards, reliability engineers must be aware of its mathematical assumptions:
- The Constant Failure Rate Fallacy: MTBF assumes an exponential distribution with a constant hazard rate (β = 1 in Weibull analysis). For mechanical equipment experiencing wear-out (β > 1), MTBF changes over time and can be misleading.
- Including Planned Maintenance Downtime: Scheduled PM inspections, oil sampling, and planned shutdowns should NOT be counted in MTTR; MTTR measures purely corrective repair velocity.
- Using Calendar Hours Instead of Running Hours: If a standby pump sits idle for 1,000 hours, counting those idle hours artificially inflates MTBF without reflecting true mechanical reliability under load.
Frequently Asked Questions
Does a higher MTBF always mean a better machine?
A higher MTBF indicates longer continuous uptime between failures. However, if that machine has an exceptionally long MTTR (e.g. 2 weeks to source rare imported parts), its total availability may still be lower than a machine with a lower MTBF that can be repaired in 20 minutes.
Should scheduled preventive maintenance (PM) be included in MTBF?
No. MTBF measures unplanned failures. Planned preventative activities are part of scheduled downtime and must be tracked separately in Planned Maintenance Percentage (PMP) or Total Downtime.
What is a world-class target for Inherent Availability?
For critical continuous process equipment (refineries, power generation, chemical plants), world-class inherent availability typically exceeds 98.5% to 99.0%.