Measuring your own consumption without fooling yourself

Three instruments, three answers

A household that starts measuring quickly ends up with numbers that do not agree. The utility meter says one thing. The clamp sensor on the incoming supply says something slightly different. The inverter reports a house consumption figure that matches neither, and the plug-in monitor on the fridge implies a total that cannot be right.

The instinct is to decide which one is broken. Usually none of them is. They are measuring at different points, over different intervals, with different methods and different accuracy classes, and small differences between them are expected rather than symptomatic.

It is worth knowing what each is doing before drawing conclusions, because a measurement misunderstood produces worse decisions than no measurement at all.

The clamp measures current, and current is not power

A clamp sensor works by detecting the magnetic field around a conductor, which gives it the current flowing. To get power it needs voltage as well, and many inexpensive monitors do not measure voltage at all. They assume a nominal figure and multiply.

That assumption introduces an error whenever the actual supply voltage differs from the nominal, which it does routinely and by a few per cent in either direction. It is not a large error and it is a systematic one, so it tends to bias a whole month’s figures in one direction.

The bigger issue is power factor. In alternating current, current and voltage are not necessarily in step, and for loads with motors or certain power supplies they are not. The product of current and voltage gives apparent power; the actual energy consumed is real power, which is smaller by the power factor. A monitor that multiplies current by an assumed voltage reports apparent power, and it will overstate consumption for exactly the loads people most want to measure.

What the utility meter counts

A domestic meter measures real power properly, integrating it over time, and it is the reference because it is the thing being billed. It is also tested and certified to an accuracy class, which no household instrument is.

That does not make it a good diagnostic tool. It reports a single total for the whole property and, unless it provides interval data, nothing about when. It cannot separate the freezer from the immersion. Its value is as a truth check on everything else: if a monitoring setup implies a monthly total that differs substantially from the meter, the monitoring is wrong.

Where an array is installed, the meter may be recording import and export separately, and the sum of those is not household consumption. Consumption is generation plus import minus export, and confusing any two of those four quantities is the most common error in domestic energy measurement.

Plug monitors are precise and easy to over-generalise from

A monitor between the socket and the appliance measures real power at that socket, and it does so reasonably well. Its weakness is not accuracy but sampling. An hour with the kettle plugged into it tells you almost nothing, because appliance consumption is dominated by duty cycle rather than by rated power.

A refrigerator draws its rated power only while the compressor runs, and how often that happens depends on the room temperature, how full it is and how often the door opens. A single day’s measurement in one season, extrapolated to a year, can be substantially wrong. Leave it in place for a week; better, for a week in summer and a week in winter.

Standby is where these devices genuinely earn their place, because standby draws are small, continuous and impossible to estimate. They are also where the cheap ones struggle, since accuracy at very low power is the hardest part of the measurement and inexpensive monitors often round small draws to zero or report them poorly.

Measuring for a decision rather than for a dashboard

The point of measurement is to change something, and a surprising amount of domestic monitoring never gets that far. Before installing anything, it helps to name the question: is the evening peak large enough to justify storage, is there a standing load nobody can identify, does the array cover the daytime baseline.

For most of those questions, a fortnight of interval data and a household meter reading at each end is enough, and permanent monitoring adds little beyond a habit of glancing at a graph. Real-time displays produce an initial burst of behaviour change that decays quickly, which is a well-observed pattern rather than a criticism of anyone.

And where a measurement leads nowhere, it is fair to stop. A household that has already identified its large loads, shifted what can be shifted and fixed what can be fixed does not need to keep watching. Measurement is a tool for finding things, not a permanent obligation, and there is no prize for the most instrumented house on the street.