Big data energy cuts would mean big changes by Peter Judge
Big Data is everywhere in the tech media, and is going to be responsible for filling up lots of data centres. What I want to know is, how green is it?
“Big Data” is really just a new name for analytics. An old and esoteric discipline, which used to be provided by adepts to high-paying clients, is now available to the masses through open source software and cloud services, based on ever-cheaper hardware in bigger data centres.
A lot of green initiatives are based on more processing of more data. Smart Cities and Smart Transport infrastructures are supposed to be able to reduce energy use by orchestrating public and private transport so less energy is wasted.
Smart Grids are supposed to even out the fluctuations in electricity usage, and reduce transmission losses, by making sure that local generation produces power that can be used efficiently, and sources of loss (equipment left turned on) are automatically turned off.
Smart Grids are supposed to respond in a micro-management way to waste by individual devices, and at the same time, take account of large changes in power demand.
But most efficiency measures, take a lot of effort and only shave a few percent off the energy use, because underlying behaviour remains the same.
To get more savings, people have to change the way they live. I’ve been told that Big Data might help put mechanisms in place that could enable those changes.
Suppose energy tariffs could change smoothly according to demand, and everyone can supply electricity to the grid as well as draw from it - and suppose the feed-in prices also vary according to demand.
Also suppose that lots of people have electric cars with large batteries, which spend large amounts of time plugged into the grid to change them up. A Smart Grid might allow us to add up all those batteries and treat them as an energy store - a big “virtual battery” for the whole grid - and use them an energy source when demand is high.
With that backup, the grid could use fewer back-up plants - which currently have to burn fossil fuels, as that’s the only source which can respond quickly enough to demand. Less backup generation really would mean less emission.
The drawback to this is that the virtual battery is available at the wrong time. The cars will mostly be plugged in and charged up at night, when demand is low, so the virtual battery will be fully charged when no one needs it.
But suppose Big Data analytics could make calculations, and offer people better tariffs if they changed their behaviour to improve that virtual battery.
For instance, if enough electric-car commuters delayed their journey into work, the grid could use their batteries to handle the breakfast time peak, without switching on extra capacity. The utility could offer them money to shift their work day by a couple of hours.
Night shift workers could even make a net profit. Their cars would be plugged in and charged during the day, and might be available to back up the grid during the highest daytime demand peaks.
The drawback to all this will be in the sums. To make a real difference, would take a large proportion of the population would have to be night shift workers with (currently expensive) electric cars, providing a significant amount of power during the day. That means a lot of higher-paid staff would have to be persuaded, by lower energy bills, to change their life completely.
Big Data offers the kind of detailed calculations that might make that sort of offer. But will society really make that sort of big change? We don’t yet know.