In their testing, the researchers found that fast-switching devices distort the normally smooth 50 Hertz sine wave of the power supply. Instead, the wave takes an erratic shape and that's something electronic meter designers have not made allowances for.
Older, electro-magnetic meters aren't affected by the problem. Meters using current transformer or shunt resistor sensors provide fairly accurate readings even with energy-saving devices.
However, smart and electronic meters using Rogowski Coil sensors provided very high and inaccurate readings. On the other hand, electronic meters with Hall sensors read almost a third too low, and the researchers cheekily suggested these are the best for consumers.
Rogowski Coil and Hall effect sensors are both commonly used in energy meters, and a quick Google search showed that they can be found in New Zealand.
Those type of sensors are popular as they're relatively inexpensive, and used for "sub metering" of power in large buildings, to get a better picture of where energy is used.
Which meter has what type of sensor then?
Unfortunately, Frank Leferink who sent me the paper said the researchers would not reveal the make and model of the meters tested, as they weren't interested in naming and shaming and only wanted to present their findings on a technological issue.
I contacted AMS that supplies meters for Vector, and WEL who installed the Landis+Gyr E350 device on my house, but received no response. The Landis+Gyr meter I have is modular, and it's not clear which type of sensor WEL is using for it. Fingers crossed, it's not Rogowski Coil sensor because I do use energy-saving lights, and there's a dimmer in the house too.
A friend forwarded the research to Genesis, and asked how confident the electricity retailer is that the readings from smart meters are accurate.
Genesis responded by saying the research was conducted in the Netherlands "which would have different regulations regarding their electricity supply and certification of electricity meters." Perhaps, but that ignores the technical issue the researchers found.
Either way, Genesis will reimburse customers whose meters read inaccurately. The meters will be tested by independent contractors if customers suspect they read wrongly, and replaced, Genesis said.
Kudos to Metrix which supplies meters to Mercury for giving a full and detailed answer, however.
Metrix said that the majority of their meters use Current Transformers (CT) or Shunt Resistor sensors. Only a small proportion, less the one-and-a-half percent, of Metrix deployed meters use Hall sensors, and a company spokesperson said no Rogowski Coil meters have been installed by them.
That electronic meters can misread wildly took at least one New Zealand electricity supplier's chief executive, who shall remain anonymous, that I spoke to with surprise.
Which is fair enough, meters are tested stringently before being certified; the effects on meters of customers using energy-saving lights and dimmers was perhaps something the power industry and meter makers didn't think of?
It's an interesting issue though, and suggests meter makers need to incorporate realistic load scenarios with lights and devices that electricity customers have in their homes from now in their designs.
Either way, if your power bill hits the moon after a smart electronic meter is installed, it might be worth pointing your electricity company to the Dutch research
- VAMS responded after the publication of the column, saying their meters do not use Rogowski coil or Hall effect sensors - only current transformers as well as shunt resistor sensors, with digital signal processors (DSPs).
These, VAMS said, are designed and tested to meet or exceed the metering accuracy standards, even when measuring harmonically distorted energy as per above.