Industrial Electrical Energy Audit: A Practical Guide for UK Manufacturing Sites

A practical assessment should establish where significant demand exists and connect each load to the production outcome it supports. For the major assets, record rated capacity, normal operating demand, duty, hours, control method and the process conditions that govern operation.

It should also distinguish facts from estimates. A reading from a suitable power meter is different evidence from a calculation based on current and assumed power factor. An operator's account of running hours may be the best available starting point, but it should be labelled as such rather than presented as logged data.

The assessment should identify avoidable running, poor control, abnormal conditions and assets that need deeper investigation. It should rank actions by likely energy or cost impact, confidence, operational risk and implementation difficulty. Assumptions and uncertainty belong in the report because they determine how much weight a decision-maker can place on the result.

Begin before the walkround. Gather recent electricity bills, tariff information, half-hourly consumption where available, maximum-demand records, shift patterns, shutdown periods and production schedules. Note seasonal loads and large equipment that operates intermittently. This gives the field survey a scale: a 10 kW observation means something different on a 100 kW site and a 10 MW site.

List significant loads by area and duty. Typical groups include motors, pumps, fans, compressors, refrigeration and chillers, electrical process heating, conveyors, extraction, HVAC, lighting, EV charging, batteries and other production machinery. Use an existing asset register where it is reliable, then verify it in the field.

Rated kW is useful for identifying scale but is not the same as consumed kW. A 75 kW motor may operate near full load, lightly loaded or intermittently. A heater may cycle around a setpoint. A compressor may unload while still consuming substantial power. The important questions are what the equipment actually does, how it is controlled, how often it runs and what prevents a different operating pattern.

Record enough context to recognise the asset later: tag, location, application, control arrangement and normal operating state. Photographing a nameplate can reduce transcription effort, but extracted values must be checked against the plate—especially where it contains several frequency, voltage or connection rows.

Use a proportionate boundary. Start with the loads most likely to explain the site profile, then expand where evidence shows a material gap. Small loads still matter when they operate continuously or in large numbers, but surveying every plug-top appliance before understanding the main process loads rarely improves the first decision.

Use the strongest evidence reasonably available. A suitable power analyser or trusted revenue/submeter provides real power directly. A VSD, SCADA historian or equipment display may provide useful data if the signal definition, scaling and operating state are understood. Current, voltage and operating hours are valuable screening inputs, while nameplate values and operator knowledge add context.

Current alone does not equal true electrical power. A three-phase estimate often uses √3 × voltage × current × power factor, but its reliability depends on representative voltage and current, an appropriate power factor and reasonably balanced conditions. Nameplate power factor describes a rated operating point; it may not represent a lightly loaded motor. The result should therefore be labelled as an estimate.

Take readings in a known production state and date them. Three phase-current readings are more useful than one because they also support an imbalance check. Where duty varies, a single spot measurement may not represent the day: use logging, control-system history or repeated readings where the decision warrants it.

Energy efficiency removes unnecessary consumption while preserving the required outcome. Demand flexibility changes when or how much electricity is used for a limited period. A tank may allow a transfer pump to stop temporarily; chilled-water storage may allow cooling demand to move; EV charging may be delayed provided vehicles are ready when required.

Flexible demand is not spare nameplate capacity. It depends on minimum and maximum useful demand, duration, response speed, recovery effects, process buffers, controls, measurement and whether several loads are available simultaneously. Safety, quality and production constraints take precedence.

The industrial demand flexibility guide explains how to screen turn-down and turn-up separately and why technical potential should be adjusted into a more credible operational estimate.

1. No- and low-cost operational changes: remove unnecessary running and correct schedules or setpoints where the process permits.
2. Validate weak assumptions: measure real kW, log variable duty or confirm hours before committing capital.
3. Control improvements: improve sequencing, interlocks or demand-based control.
4. Maintenance and duty corrections: address abnormal conditions and equipment operating outside intended duty.
5. Retrofit opportunities: assess VSDs, improved controls or system changes with process evidence.
6. Capital replacement: consider efficient equipment at an appropriate replacement point using whole-life cost.

For each action, show the estimated energy and cost impact, evidence confidence, operational constraint, implementation effort and next validation step. Avoid made-up paybacks. If cost, hours or savings are uncertain, give the assumptions rather than hiding them inside a precise-looking number.

Assign an owner and a decision gate. “Measure compressor real power for one production week” is more useful than “investigate compressor”. For operational changes, agree how performance and product impact will be checked. For capital measures, specify the evidence needed for design and quotation. Close out rejected actions with a reason so the same unsupported proposal does not return at every review.

A useful report identifies the site and area, assessment date, assets assessed, measurement conditions and the electricity price and operating hours used. It presents energy and cost estimates, priority actions, supporting evidence, uncertainty and engineering limitations. Readers should be able to trace a headline opportunity back to the relevant asset and assumptions.

This structured field-to-report path is what MotorForge is intended to support: capture motor information and readings, screen indicative opportunities, prioritise review and produce a consistent assessment report. Its calculations remain preliminary engineering estimates and require competent validation.