Skip to content
allanninal.dev/build
Collaborate

Part 5 of 7 · Utility meter reader series ~5 min read

How a finding becomes a saving

The gap between an energy saving on paper and one on the meter is large, well-documented, and mostly closed by the cheapest possible intervention: measuring afterwards.

Key takeaways

  • Measure for two weeks before and two weeks after, at the same resolution.
  • Compare like periods: same days of week, weather-normalised where relevant.
  • A fix that shows nothing on the meter did not work, whatever the specification says.
  • Common causes: it was reversed, it was partial, or something else grew.
  • Report verified savings only, and report the unverified ones as unverified.

Before and after

How an energy saving is verified against the meterA vertical chain of five steps entered by a box labelled A change is planned, whether a timer, a control or a unit. Step one measures before, for two weeks at the same resolution. Step two records the date, exactly when it changed. Step three waits two weeks to settle, then measures. Step four asks whether the comparison is like for like on days, weather and occupancy; if not it exits to Adjust or wait, noting that a bank holiday ruins a fortnight. Step five produces a measured saving, or the conclusion that it did not work. A note says the second box is the one that gets forgotten and it makes the rest impossible.AWS ACCOUNTA change is planneda timer, a control, a unitMeasure beforetwo weeks, same resolutionRecord the dateexactly when it changedWait two weekssettle, then measureLike for like?days, weather, occupancyAdjust or waita bank holiday ruins a fortnightnoSaving, measuredor: it did not workThe second box is the one that gets forgotten and it makes the rest impossible.
Fig 1. The verification sequence. Recording the exact change date is trivial and is the step whose absence makes a before-and-after comparison unanswerable.
  • Compute
  • App integration
  • Machine learning
  • Management

The date matters more than it seems

“We changed the timers some time in April” makes verification impossible, because the before and after periods cannot be defined. A recorded date, to the day, costs nothing and is the difference between a measurable change and an anecdote.

It is worth keeping a simple log of every change made to the building’s energy-consuming equipment, with dates. Over a couple of years that log explains most of the movement in the data and it is the first thing anybody wants when a number moves unexpectedly.

When it did not work

Why a fix shows nothing on the meter

  • It was reversed. Somebody found the compressor off, assumed it was a fault, and put the timer back. This is the most common one.
  • It was partial. Three of five units were changed, and the specification assumed five.
  • Something else grew at the same time and absorbed the saving.
  • The load moved. Turning off a heater means a different heater runs longer.
  • The saving was never there. The specification was a manufacturer’s figure under ideal conditions.
  • All five are worth knowing about, and none of them are visible without measuring.

The first is worth designing against. A change that somebody can reverse without knowing why it was made will eventually be reversed, so a label on the equipment saying what was changed, when and who to ask is a genuinely effective intervention.

Reporting savings honestly

Claimed against verified annual energy savingsA bar chart with three bars in pounds. Two series: verified annual saving in green, and claimed but not visible on the meter in grey. Claimed on paper: fourteen thousand two hundred. Verified on meter: eight thousand six hundred. Unverified: five thousand six hundred. A note says the grey bar is not a failure, it is the part nobody had measured.£0£5000£10000£15000£20000~£14200Claimed on paper~£8600Verified on meter~£5600UnverifiedVerified annual saving, £Claimed but not visible on the meter, £The grey bar is not a failure. It is the part nobody had measured.
Fig 2. A year of energy work reported honestly. The gap between claimed and verified is normal and reporting it is what makes the verified figure believable.

Reporting the gap rather than hiding it has a useful effect: it makes the verified number credible. An energy report claiming fourteen thousand pounds of savings against a bill that did not move by fourteen thousand pounds is a report nobody believes the next time.

The annual reconciliation

Once a year, compare the sum of verified savings against the actual change in the weather-normalised bill. They will not match exactly and the difference is informative: consistently larger actual reductions mean something good is happening that nobody has attributed, and consistently smaller means savings are being counted that are not there.

That reconciliation is the honest close of the loop, and it is the number worth putting at the top of an annual energy report: not what was claimed, but what the meter says the building did, adjusted for the weather.

Next: what all of this costs to run.

All posts