MTTF: Formula, MTTF vs. MTBF & Spare-Part Strategy

What is Mean Time To Failure (MTTF)?

MTTF (Mean Time To Failure) is the average operating time from when a non-repairable component is put into service until it fails. A conveyor belt has an MTTF of 8,000 hours. A proximity sensor has an MTTF of 25,000 hours. A servo motor bearing has an MTTF of 12,000 hours. When the belt breaks, you do not repair it — you replace it.

When the sensor fails, you replace it. When the bearing fails, you press in a new one. MTTF applies to components that are replaced, not repaired. That is the fundamental difference between MTTF and MTBF: MTBF measures the time between failures of a repairable system (the entire machine). MTTF measures the lifespan of a non-repairable component (the part inside the machine). An MES does not calculate MTTF from a formula — it measures it from reality: how long did this type of bearing actually last in this machine under these operating conditions? That measured MTTF drives the spare-part strategy that prevents unplanned downtime.

How do you calculate MTTF?

MTTF = Total Operating Time of All Units / Number of Failures

Where:

• Total Operating Time of All Units = the cumulative operating hours of all identical components in the population. If you have 10 identical proximity sensors, each with different run times, you sum all their hours.
• Number of Failures = the count of components that failed during the observation period.

The critical subtlety: MTTF is a population average. It does not mean every sensor will last 25,000 hours. Some will fail at 15,000. Some will last 35,000. The distribution matters more than the average — and an MES that tracks the actual failure time of every replaced component builds this distribution over time, giving maintenance managers the data to set replacement intervals that balance cost (replacing too early) against risk (replacing too late).

What is the difference between MTTF, MTBF and MTTR?

The relationship: when a bearing (MTTF component) fails inside a press (MTBF system), the press experiences a failure event. The time to replace the bearing contributes to the press's MTTR. The press's time between this failure and the next failure is the press's MTBF. So MTTF of the component drives MTBF of the system. Improving the MTTF of the weakest component is the most direct way to improve the MTBF of the entire machine.

Why does MTTF matter for spare-part strategy?

MTTF is the metric that answers the single most expensive maintenance question: "When should I replace this part before it fails?"

There are only three strategies, and MTTF data determines which one is right:

The mistake most plants make: applying the same strategy to all components. They replace cheap filters on a schedule (wasteful) and run expensive bearings to failure (risky). MTTF data — collected from actual failure history in the MES — enables the right strategy for each component type. At Neoperl, SPS-based alarm correlation identified which component failures caused the most downtime. That data drove targeted spare-part stocking: the components with low MTTF and high downtime impact were stocked on site. The rest were ordered on demand.

How does an MES measure MTTF in practice?

MTTF is not printed on the component's datasheet — or rather, the manufacturer's datasheet number is a theoretical value derived from accelerated life testing in a laboratory. The real MTTF in your plant, with your operating conditions, your ambient temperature, your load profile, is different. Often very different.

The MES measures real MTTF by combining three data streams:

• Alarm data with component-level resolution. When the SYMESTIC alarms module captures alarm #3012 ("hydraulic pressure below 180 bar"), and the maintenance team records that the fix was "replaced hydraulic seal," the MES now has a failure event tagged to a specific component type. Over time, these events build the actual MTTF distribution for hydraulic seals on that machine. At Neoperl, SPS-based alarm capture provided this component-level failure tagging automatically.
• Operating hours per machine. The MES tracks running hours from the PLC signal — not calendar time, but actual production time. A machine that runs 16 hours/day accumulates operating hours twice as fast as one running 8 hours/day. MTTF must be measured in operating hours, not calendar months. The SYMESTIC production metrics module provides this data automatically.
• Maintenance event logging. When a component is replaced (whether preventively or after failure), the replacement is recorded in the MES with a timestamp. The difference between installation timestamp and failure timestamp is the component's actual time-to-failure. Aggregated across all units of the same component type, this becomes the measured MTTF.

The result: after 6–12 months of operation, the MES contains enough failure data to calculate real MTTF values for the 20–30 component types that cause the most downtime. That data replaces the manufacturer's theoretical MTTF with plant-specific, measured reality — and the spare-part strategy shifts from guesswork to evidence.

How does MTTF fit into the bathtub curve?

Components fail in three phases — the classic "bathtub curve" of reliability engineering:

MTTF is meaningful primarily for the wear-out phase. In that phase, failure is age-related and predictable. The MES provides the operating-hour data to detect when a component population enters wear-out — alarm frequency per component type increases, even if each individual alarm clears quickly. That rising trend is the leading indicator that MTTF is being approached.

FAQ

Can MTTF be used for repairable systems?
Technically no — for repairable systems, MTBF is the correct metric. But in practice, many maintenance teams use "MTTF" loosely to mean "how long until the next failure." The distinction matters: MTBF assumes the system is repaired and returned to the same condition. MTTF assumes the item is replaced with a new one. For a bearing inside a machine, MTTF is correct (the bearing is replaced, not repaired). For the machine itself, MTBF is correct (the machine is repaired, not replaced). The MES calculates both: MTBF for the machine, MTTF for the component types within the machine.

Is the manufacturer's MTTF reliable?
It is a starting point, not a fact. Manufacturers derive MTTF from accelerated life tests under controlled conditions — constant temperature, clean environment, rated load. Your plant has temperature swings, dust, vibration, overload conditions, and operators who occasionally bump into sensors with forklifts. The real MTTF in your environment is typically 40–70 % of the manufacturer's stated value. The only way to know the real number is to measure it — and the MES provides the data to do so.

How does MTTF relate to OEE?
MTTF affects OEE Availability indirectly. When a component with low MTTF fails, the machine stops — and that stop appears as unplanned downtime in the OEE Availability calculation. Improving the MTTF of the weakest components (either by sourcing better parts or by replacing them before they fail) directly reduces unplanned downtime and improves OEE. The MES connects the dots: "The 3 failures on press 5 this month were all bearing type SKF-6205. The measured MTTF is 4,200 hours. Replacing at 3,500 hours would have prevented all 3 failures and saved 18 hours of unplanned downtime."

How many failure events do I need to calculate a reliable MTTF?
As a rule of thumb: at least 5–10 failure events for the same component type to get a meaningful average, and 20+ events to understand the variance (distribution shape). With fewer than 5 events, the MTTF is a rough estimate. This is why the MES approach is powerful: by tracking failures across all machines in the plant (or across multiple plants, as at Meleghy with 6 sites or Carcoustics with 500+ machines), the failure population grows much faster than tracking a single machine. A component type installed in 50 machines across 3 plants generates failure data 50× faster than tracking one machine alone.

Related: MTBF (Mean Time Between Failures) · MTTR (Mean Time To Repair) · MTBM (Mean Time Between Maintenance) · TPM · Predictive Maintenance · OEE Explained · SYMESTIC Alarms Module · SYMESTIC Process Data · SYMESTIC Production Metrics · MES: Definition & Functions

About the author

Martin Brandel

MES Consultant at SYMESTIC. Dipl.-Ing. Nachrichtentechnik. Over 30 years in industrial automation. Has replaced enough bearings on enough continents to know that the manufacturer's MTTF and the plant's MTTF are never the same number — and the difference is where the spare-part budget lives. · LinkedIn