Energy management in manufacturing: turning a utility bill into a controllable production cost

Split the bill into what production drives and what runs regardless

Plot site consumption against output over a run of weeks and the relationship separates into two parts: a slope that represents energy per unit made, and an intercept that represents everything drawing power whether or not the plant is producing. That intercept — compressors serving leaks, lighting, ventilation, chillers, ovens held hot, control panels — is often startlingly large relative to the production-linked part, and it is usually the easier of the two to reduce. It also needs no process change and no capital, only somebody with authority deciding what may run outside production hours and someone verifying that it does not.

Meter at the process or the analysis stops at the gate

A single incoming meter supports billing and nothing else. Submetering the major consumers — the compressor house, the ovens and furnaces, the chillers, the biggest motors, each production area — is what lets consumption be attributed to a department, a line and eventually a product. Start with the loads that plausibly account for most of the bill rather than instrumenting everything; a small number of meters usually covers the large majority of consumption. Metering hardware and data collection belong to the automation discipline, but deciding what to meter and who receives the readings is an operations decision and should not be delegated to the installer.

Energy per unit, normalised, is the only comparable figure

Absolute consumption falls when the plant is quiet and rises when it is busy, which makes month-on-month comparison worthless. Specific consumption — energy per unit, per tonne, per square metre of product — is comparable, provided it is normalised for the things that legitimately move it: product mix, ambient temperature for heating and cooling loads, and batch size where startup is a large share. Without normalisation, a mild season or a favourable mix will be reported as an improvement, the initiative will be declared successful, and consumption will rise again the following year with no explanation.

When you run costs money as well as how much you use

Tariffs commonly charge for peak demand as well as for units consumed, and a single coincident start of several large loads can set a charge that persists for months. Sequencing startups, staggering compressors, avoiding simultaneous furnace ramps and shifting discretionary loads out of expensive periods reduces cost without reducing production. Where the plant has flexibility in when it runs, that flexibility is a commercial asset and increasingly a saleable one in markets with demand-response arrangements. The prerequisite is knowing the load profile at fine time resolution, which is the second reason to meter properly.

Making it survive the person who started it

Energy savings decay. Setpoints get adjusted for a hot week and stay adjusted, a compressor left on manual keeps running, a shutdown routine lapses when the person who wrote it leaves. A management-system approach — an energy policy, a named owner, a defined review cycle and documented operating practice, of the kind the ISO 50001 standard addresses — is what turns a project into a retained result. Certification is optional; the underlying structure of ownership, measurement and periodic review is what actually holds the gains, and a plant can adopt it without ever inviting an auditor.

Frequently asked questions

Where should a plant start if it has never measured energy by area?

With a walk during a non-production period, clipboard in hand, listing everything still running and asking why. That single exercise routinely finds compressors serving nothing, extraction fans on lines idle for months, ovens held at temperature over a weekend and lighting in empty areas. None of it needs metering to fix. Do that first, then meter the largest consumers so the next round of decisions has data behind it rather than opinion about which machine is thirsty.

Is compressed air really worth treating as an energy issue?

Yes, because it is electricity converted at poor efficiency and then frequently leaked away. Most of the energy going into a compressor ends up as heat rather than useful air, so every leak is expensive out of proportion to its size, and a plant with an unmanaged system can be losing a large fraction of its compressor output through leaks and inappropriate uses such as blowing down parts or cooling panels. It is usually the single best return available on a modest engineering effort.

How do we justify energy projects against production investments?

Express the saving as an annual cost reduction with the measurement method stated, and be conservative about persistence — assume savings decay unless a control or a procedure locks them in. Energy projects lose internal arguments because their benefits are diffuse while a new machine promises visible capacity. The counter is that energy savings are usually low risk and do not depend on winning more orders, which is a genuinely different risk profile and worth stating explicitly to a finance committee.