ColdStore Logistics: a 6.8% EnPI gain resting on an uncalibrated meter
ColdStore Logistics operates a 42,000 m² refrigerated warehouse holding chilled and frozen product for grocery clients. Refrigeration is 68% of electricity consumption. Twelve submeters cover compressor room, condenser fans, dock doors, lighting and forklift charging; the site procedure sets a 24-month recalibration interval, but the two compressor-room meters were last calibrated in 2022. A recent spot check with a clamp meter indicates the compressor-room submeter reads about 6% high. The EnPI, kWh per pallet-day, is calculated monthly by the energy team from the submeter data and evaluated once a year at management review. In June 2024, 3,000 m² of chilled space was converted to frozen at −25 °C. 2023 EnPI was 0.44 kWh/pallet-day, 2024 is 0.41; pallet-days rose 14% and absolute refrigeration electricity rose 3%. In August the condenser fans ran about 20% above expectation for three weeks; the BMS alarm was acknowledged and cleared, and nothing was recorded. The energy manager wants to report a 6.8% energy performance improvement to the board.
1. Which requirements of clause 9.1 are not being met at this site?
Read the worked answer
What to monitor and measure. 9.1 requires determining what needs monitoring, including relevant variables affecting energy performance. Pallet-days are used as a denominator, but ambient temperature, door cycles and the chilled/frozen mix are not monitored, so the EnPI cannot be interpreted.
Valid methods. The standard requires methods for monitoring, measurement, analysis and evaluation that ensure valid results. There is no documented EnPI method: no formula, no data sources, no treatment of estimated or missing data, no uncertainty.
When to evaluate. Calculation is monthly but evaluation is annual. Nine months of drift and a three-week excursion go unexamined.
Measurement equipment. Equipment must be calibrated or verified, with documented information retained. Two meters are roughly three years past a 24-month interval and one carries a 6% bias - larger than the trend being claimed.
Significant deviations. 9.1 requires investigation of significant deviations in energy performance. A 20% excursion for three weeks is significant by any threshold, yet the alarm was cleared with no investigation and no record.
Comparison with the EnB. No documented comparison of EnPI values against the baseline and its adjustments.
2. Can ColdStore report a 6.8% energy performance improvement? Assess the evidence and the structural change.
Read the worked answer
The direction is plausible - absolute energy rose 3% while activity rose 14% - but the specific figure is not yet demonstrable, for three reasons.
Instrument bias. The compressor-room submeter reads about 6% high. The claimed improvement is 6.8%. Until the meter is calibrated and the historical data corrected or re-estimated, the error is the same size as the claim. The calibration certificate should show when drift began; if the meter was accurate in 2023 and reads high in 2024, part of the 'improvement' is instrument error.
Structural change. Converting 3,000 m² from chilled to frozen raises energy per pallet-day with no loss of efficiency. Without normalisation for the frozen share, the true improvement is understated - the site may be better than 6.8% - but the reported figure is still not the right number.
Method. The EnPI must have a documented method (6.4), a normalisation approach and an adjusted EnB (6.5) before a public claim.
A defensible route: recalibrate, correct the series, split the EnPI into chilled and frozen pallet-days (or add frozen share as a variable), recompute both years on the same basis, and report the improvement with its uncertainty and the adjustment method stated. If the corrected figure is 2%, report 2%.
3. What should be done about the August deviation and the calibration gap, and how does this connect to corrective action?
Read the worked answer
Quantify first. Estimate the excess consumption of the excursion from logged fan power over the three weeks; if fans draw about 90 kW, a 20% excess is roughly 9,000 kWh for the period. Size matters - a deviation worth 0.2% of annual refrigeration energy is handled differently from one worth 2%.
Investigate, don't just clear. Establish why the fans ran high: fouled condenser, ambient spike, control fault, or a set-point changed during maintenance. The investigation and its conclusion belong in documented information under 9.1.
Corrective action (10.2). If the cause was a control fault or a set-point change with no authorisation, the root cause is likely the alarm management process - alarms that can be acknowledged away without an owner, threshold rationale or follow-up. Fix the process, not just the fan.
Calibration. Recalibrate both meters, record the as-found error, apply an interim correction factor to the data, and review the interval; a 24-month interval is inappropriate if drift occurred earlier.
Connect to the EnMS. Feed the finding into the risks and opportunities register, check whether the energy review missed condenser performance as a relevant factor, and add meter health to the internal audit checklist.