Spotting a Failing Asset Before It Fails

Most equipment failures are preceded by weeks of degradation. The information is usually being collected already and nobody is looking at it.

Signals in the temperature log

The most available and least used diagnostic in a restaurant.

A slow warm trend over days. The unit is losing capacity. There is time to schedule rather than to emergency-dispatch.

Slower recovery after door openings. Capacity loss, before the setpoint is missed.

Widening variation between readings. Control or airflow problem.

Holding temperature but running constantly. The unit is compensating. This is a failure in progress that a twice-daily log will never show, which is the argument for continuous monitoring.

Overnight excursions that are recovered by morning and invisible to manual logging.

Signals in the work order history

Increasing frequency. Intervals between calls shortening.

Escalating severity. Adjustments, then component replacements, then major components.

The same part twice.

Multiple technicians unable to resolve it.

Repairs described as temporary with no follow-up.

Technician observations, which are the most valuable field on the work order and the least read. A note saying the condenser is beyond cleaning is a prediction.

Signals in energy

Refrigeration and HVAC consumption rise as equipment degrades. Fouled coils, failing seals and low refrigerant all increase run time.

A site whose consumption rises without a volume change has equipment working harder.

Sub-metering makes this visible per circuit, and few restaurants have it. At site level, the signal is coarse and still real for a significant degradation.

Compare sites of similar size and volume. An outlier is worth investigating.

Signals from staff

Managers and kitchen teams notice things and do not report them because they are not faults yet.

Noise changes. A compressor or fan that sounds different.

"It takes longer to come down after a delivery."

"It has been making that noise for a while."

"We stopped using that one." Equipment quietly taken out of service is a failure nobody logged, and it happens more than operators realise.

Ice appearing where it did not.

Doors needing a push.

Ask explicitly during site visits. "What is annoying you about the equipment?" produces more useful information than any inspection.

The review that uses this

Monthly, for the critical assets:

Temperature trend for each refrigeration unit. Work orders in the last quarter. Anything staff have mentioned. Energy, at site level.

Quarterly, across the estate:

Assets with more than two work orders. Assets approaching expected life. Assets with rising cost. Sites with rising consumption.

Half an hour a month. It converts emergency failures into scheduled repairs, which is the entire economic argument for maintenance measurement.

Acting on a signal

Schedule an inspection, not a repair. The point of early detection is that you have time to diagnose properly.

Ask for a condition assessment, not a fault fix. A technician asked "is this unit failing" gives a different answer from one asked to fix a symptom.

Decide repair or replace with the history in front of you, which is possible now and will not be at 8pm on a Saturday.

Plan the downtime rather than absorbing it.

Order parts ahead where lead times are long.

Where sensors help and where they do not

Temperature monitoring on refrigeration is the one clear case. Affordable, the signal is real, the failure is expensive, and it doubles as compliance evidence.

Energy monitoring at circuit level is useful at scale and rarely justifies itself for a single site.

Vibration, current draw and run-time monitoring exist and are generally aimed at industrial equipment. The restaurant application is thin.

Everything else in this article requires no hardware. The temperature log, the work order history, the energy bill and the conversation with the manager are already available, and they are not being read.

The condition inspection

When a signal appears, the correct response is an inspection with a specific brief, not a repair call.

Ask for a condition assessment, in writing, naming the asset and the symptom.

Specify what you want assessed: remaining life, what is likely to fail next, whether the current symptom is a cause or a consequence, and what it would cost to address properly.

Ask for a recommendation between repair, planned replacement and monitoring, with reasoning.

Give them the history. A technician shown three years of work orders on that unit gives a much better answer than one looking at it cold.

Expect to pay for the visit, and expect it to be worth several times its cost. A planned replacement during a closure is dramatically cheaper than the same replacement during a Saturday service, and this is the visit that makes the difference.