Diesel Generator Reliability Benchmarking Report | MCIM
Reliability Benchmarking

Reliability Benchmarking Report/Diesel Generators

The observed reliability of five generator platforms across a live, multi-operator mission-critical portfolio. Built from real operating and failure records, not manufacturer-published figures.

3,072
Units in scope
2,289
Asset-years observed
373
Measured failures

What 2,289 asset-years of data show

Below are four findings that stood out. The full analysis includes confidence ranges, model family detail, failure-mode profiles, and the method behind every figure.

Calendar-hour MTBF with 95% confidence range
Hours between failures, by manufacturer
0 20k 40k 60k 85k hrs Caterpillar 59,473 – 79,211 Cummins 37,143 – 52,103 Kohler 12,332 – 23,707 Whole fleet 48,600 – 59,700
The dot is the point estimate and the bar is the 95% range. Where two bars overlap, those platforms can't be separated on this measure. Kohler's figure is driven largely by one group of newly installed units, explained alongside.
  • 68,456hrs MTBF

    Caterpillar leads on every reliability measure

    Longest MTBF, lowest failure rate at 0.128 per asset-year, fastest resolution at 188 hours, and the lowest rate of planned maintenance finding a defect.

  • No overlap

    The Cummins gap is real, not sampling noise

    Its MTBF range of 37,143 to 52,103 hours doesn't overlap Caterpillar's. Failures concentrate in cooling and fuel, two systems that respond strongly to maintenance discipline.

  • 2.4×

    Resolution time varies more than the machines do

    MTTR runs from 188 to 445 hours across manufacturers. For standby plant, a generator waiting on a part won't answer the next utility event.

  • 23 out of 39

    Brand-new units were behind most of Kohler's 39 failures

    Thirteen newly installed Kohler units, all in their first five months of service, account for 23 of Kohler's 39 measured failures.

Figures reflect recorded operating experience across multiple operators and aren't statements about product quality in general. This report is independently prepared by MCIM and isn't sponsored by, endorsed by, or affiliated with any manufacturer named in it.

What it means for your next purchase

The report closes with four conclusions a buyer can act on, ordered by how strongly the data supports them.

  1. 1

    Weigh resolution time alongside failure frequency

    Ask suppliers for local parts inventory and guaranteed response times, and treat the answer as a reliability specification.

  2. 2

    Put your reliability effort at the front of asset life

    Commissioning rigor is the highest-leverage intervention available. The report lists what to specify before handover.

  3. 3

    Scrutinize the systems around the engine

    Every manufacturer's dominant failure mode sits outside the prime mover: electrical output, cooling, fuel, or controls.

  4. 4

    Specify consumables and verify them

    Off-spec coolant, loose terminals, and unsecured filters showed up repeatedly. It's common and entirely preventable.

From the maintenance recordA planned maintenance finding quoted in the report, with identifying detail removed.
“The coolant is watered down, almost looks like pure water, incorrect color for proper cat coolant. Needs to be flushed and the system filled with cat specific coolant.”

This report covers 3,072 generators. Reliability Benchmarking compares your fleet against more than one million data center assets worldwide.

The same analysis, run continuously against your own estate, so you can see where your assets sit relative to the portfolio and act before the next event.

Questions it answers for your fleet
  • Which of your units are inside their infant-risk window?
  • Which failure modes is your regimen catching?
  • How do your resolution times compare to the rest of the portfolio?
  • Which assets are operating outside benchmark range right now?
  • Where should your next procurement concentrate its acceptance testing?
Pages from the MCIM Reliability Benchmarking Report on diesel generators

Specify the service network and the commissioning regimen as carefully as you specify the machine.

That's where the observed reliability difference actually lives. Get the full report with rankings, model family detail, failure analysis, and method.

Download the report