Correct energy metering — what is it really about?

1. Completeness

Metering is meant to reproduce the complete picture of consumption, not the set of meters that happen to be available. It happens that the installation covers only selected areas, omits process equipment, or — something we see surprisingly often — meters are wired incorrectly and measure something other than what their labels say. Data that does not reconcile with the supplier's bill is a signal that part of the picture is missing.

2. Read frequency matched to the significance of the point

From our experience: electricity meters at the grid connection points, and those covering at least 20 percent of consumption, are worth reading at least once a minute, and gas and heat meters of similar significance at least once every five minutes. Other electricity points should come down to five minutes, at worst fifteen; other energy vectors to fifteen minutes, at worst an hour. Reading too rarely loses demand peaks and price windows — precisely the things there is money in.

3. Time harmonisation

A subject that is often ignored, particularly where metering was installed at different times or the meters are read by different concentrators. All readings have to be synchronised, because comparing data from drifting time stamps produces conclusions that look credible and are sometimes false. The time stamps of the operators reading the connection-point meters have to be accounted for as well.

4. Retention

Data is best kept for at least several years. Without history, we cannot calculate seasonality, cannot set a reference point and cannot prove the effect of any change. This is where the IPMVP methodology comes in.

5. Availability

Correct measurement has to transfer without friction into multidimensional analytics — into OLAP cubes in which consumption is sliced by site, energy vector, time and process. Data locked inside a single screen looks attractive but is analytically useless.

6. How many meters? Build measurement in layers

Once we know what the measurement should look like, an apparently simple question arises: how many meters should we actually install? The answer requires reconciling two opposing forces. Too few points leave blank spots in the picture of consumption; too many produce noise, and an excess of information works in practice like disinformation.

So we start from the question, not from the meter. Every measurement point should support a specific decision — cost allocation, anomaly detection, verification of the effect of an upgrade. If nobody will change anything on the basis of the reading, the meter only adds to the cost of maintaining the system and adds one more device that can fail.

Building measurement in layers works well. The foundation is the meters at the connection points, reconciling with supplier invoices. A level below come measurements of the main areas — buildings, switchboards, individual energy vectors — chosen so as to reproduce the structure of consumption rather than the set of places where a meter happens to be easy to fit. We go deeper only where a single load accounts for a significant share of consumption — from a modest threshold of a few percent — or where the process requires separate supervision.

The correctness of a set-up built this way is verified by a simple check. The reading of the meter at the connection point is compared against the sum of the subordinate meters; the difference shows how much consumption remains outside direct measurement. In well-chosen metering, it is small — of the order of a few percent. When it starts to grow, a gap has appeared in the picture and something is consuming energy outside the measurement. Driving it down to zero, however, is not justified — the last few percent require disproportionately many meters and no longer improve the picture.

A foundation, not a formality

A well-chosen number of meters usually turns out to be noticeably smaller than the original plans, and still gives a clear and balanced picture of consumption. Because metering is about the picture, not about a collection of devices — and it is the foundation on which everything that comes later stands or falls: from monitoring, through optimisation, to proving savings.