AI for energy and utility cost reduction in small premises

5 days ago
3 min read

Updated: 5 days ago

Introduction

Energy has moved from a background overhead to a line item owners actually look at, and the response is usually to change supplier and turn some lights off. Both help a little. The larger opportunity is almost always in when energy is used rather than what it costs per unit, and that pattern is invisible without data at a finer resolution than a monthly bill.

Half-hourly consumption data, which most business meters can provide, tends to produce one immediate finding: substantial overnight and weekend consumption in a building nobody occupies. Explaining that baseline is frequently worth more than a tariff negotiation, and it requires analysis rather than investment.

1. AI for energy and utility cost reduction starts with the overnight baseline

The clearest finding available.

Consumption between midnight and five in the morning, when the premises are empty, is your baseline. In many small businesses it is a third or more of the daytime rate, and every unit of it is being wasted.

2. Get half-hourly data before doing anything else

Monthly bills cannot show a pattern.

Most commercial meters provide it and most suppliers will supply it on request. Without this resolution you can see the total and nothing about its causes, which is why energy management usually stalls at changing supplier.

3. Explain every part of the baseline

The diagnostic exercise.

Refrigeration, servers, standby loads, heating left on a timer, lighting on a circuit nobody controls, air conditioning fighting a heater. Each has a fix, and several of them are free.

4. Look at the shape of the day, not the total

Where the tariff interacts.

If you are on a time-of-use tariff, moving a load out of the peak period saves money without using less energy. Identifying the shiftable loads — charging, heating water, batch processes — is a scheduling decision.

5. Separate weather-driven consumption from the rest

Otherwise comparisons mislead.

Heating and cooling vary with temperature, so a cold quarter looks like a failure and a mild one looks like a success. Normalising for temperature is what lets you tell whether anything you did actually worked.

6. Sub-meter the largest consumers

Where the payback is quick.

A handful of sub-meters on your biggest loads converts guesswork into fact. Without them you are attributing consumption by inference, and inference tends to blame the wrong equipment.

7. Set an alert on unexpected consumption

The recurring value.

A notification when overnight consumption rises above its normal level catches a failed timer, a door left open on a chiller, or a heater left running, within a day rather than at the next quarterly bill.

8. Do the cheap fixes before the capital projects

Ordering matters.

Timers, controls, thermostat settings, draught proofing, maintenance of seals and filters, and turning things off. These have short paybacks and their effect is measurable, which is what makes the case for the larger projects afterwards.

9. Check the water and waste bills as well

Consistently ignored.

A continuously running toilet or a leaking supply can cost more annually than the lighting. Waste collections are frequently sized for a volume the business no longer produces, and both are corrected by reading the bill carefully once.

Take any efficiency claim from a supplier and test it against your own data after installation. Savings projected from a model and savings visible in half-hourly consumption are different things.

Conclusion

Look at the overnight baseline first, because it is usually the largest and cheapest finding.

Obtain half-hourly consumption data before making decisions, account for every component of the baseline load, examine the shape of the day and shift what you can out of peak periods, normalise for temperature so comparisons mean something, sub-meter your largest consumers, set an alert on abnormal overnight consumption, complete the cheap control and maintenance fixes before considering capital projects, and read the water and waste bills with the same attention.