How to find cold storage performance loss using the records you already have

What data do you need, and what if some is missing?

Most established operations already hold enough evidence to diagnose performance. Collect, per room or per site:

Record	What it shows	If it is missing
Temperature and alarm logs	Excursions, recovery time, defrost behaviour	Install calibrated loggers for 4–8 weeks before deciding
Electricity bills or sub-meter data	Energy trend and peak demand	Use utility monthly totals; mark the result as site-level only
Door / dock activity or shift logs	Infiltration load drivers	Estimate from receiving and dispatch records; label as estimate
Maintenance and breakdown tickets	Recurring faults, downtime hours	Interview technicians; list faults by month
Spoilage, claims and returns	Product outcome	Start recording lot, room, date and cause now
Throughput (tonnes or pallets in/out)	Normalises energy and loss	Use invoices or warehouse system movements

Never fill gaps with assumed values and then treat the result as measured. Mark every estimate.

How do you build a baseline?

1. Choose a 12-month reference period that reflects normal operation.
2. Put all records on one monthly sheet.
3. Calculate kWh per tonne handled = monthly kWh ÷ monthly tonnes.
4. Add the monthly average outdoor temperature from a local weather station.
5. Count excursions (minutes outside the product's specified range), downtime hours and spoilage value per month.

This follows the logic of an energy baseline and energy performance indicators used in ISO 50001 energy management.

Which patterns point to process, equipment or management?

Pattern in your data	Likely category	Check next
Excursions cluster at receiving or loading times	Process	Door discipline, staging, warm product intake
kWh per tonne rises steadily over seasons with similar weather	Equipment	Coil fouling, refrigerant charge, compressor wear, insulation damage
Long recovery after defrost	Equipment / controls	Defrost settings, drain heaters, control logic
Same fault recurs after repairs	Management	Root-cause practice, spare parts, maintenance contract
Spoilage without matching excursions	Process / product	Incoming quality, rotation, sensor placement

Hypothetical worked example

Hypothetical figures, for method only. Assume a room used 42,000 kWh for 600 t in a reference month (70 kWh/t), and 46,800 kWh for 600 t a year later with similar average outdoor temperature (78 kWh/t). The 11% rise per tonne, with no change in weather or throughput, suggests an equipment or control issue worth inspecting — it does not prove the cause or the value of a fix.

Diagnosis checklist

• 12 months of records gathered, gaps marked
• Energy normalised by throughput and weather
• Excursions, downtime and spoilage counted per month and per room
• Patterns classified as process, equipment or management
• Sensor calibration and placement verified
• Questions listed for a refrigeration technician

What should you do with the findings?

Fix low-cost process issues first and keep measuring. Where data points to equipment or capacity limits, turn the baseline into a project brief so suppliers quote against the same facts. Acceptable product temperatures are product-specific and governed by local law; this method does not set them.