We have been discussing off and on how smart power meters (see Smart metering data storage appetite) and intelligent sensors (see the Sensor cloud comes home) such as smart thermostats generate copious amounts of data. Sensors and meters such as these are used in a new power distribution network called the smart grid.
But recently there was a report in Technology Review of how smart thermostats can “communicate” with power companies to determine current energy costs and then alter temperature settings to reduce power use.
Power companies give discounts to those customers who install the smart thermostats in the hope that such devices will lower peak power use. If this happens they will save significant investments in new power generation and power lines to satisfy the ever-growing peak power level. This probably works best in summer with A/C equipment which is a prime user of electricity during peak consumption periods. The story goes on to say that these smart thermostats have been used in a small test bed but are about to be rolled out on a wider basis.
One serious concern brought up in the article is that when such devices are rolled out on a large scale there is a high chance of power thrashing.
As smart thermostats reduce home power consumption in volume, they may end up driving the cost of power down low enough such that the thermostats, on the next cycle, will start to consume more power, leading to power thrashing.
If this happens on a big enough region, such oscillations of power use may lead to an even higher peak than what was in place prior to the smart grid. And if the thermostats have been in place for a while and succeeded in reducing peak power capabilities, such thrashing may lead to a smart grid crash or congestion collapse.
Thrashing prevention on the power grid
There are three ways to prevent power thrashing:
- Increase the amount of peak power to handle the worst case regional energy working set.
- Decrease the power consumption of the largest power users.
- Decrease the number of consumers.
It would seem the first solution defeats the purpose of the smarter grid. I believe the second approach can best be implemented by having even more smart appliances which can reduce power consumption on demand. As for the third approach, this may be infeasible, as you cannot just drop power to consumers without some serious consequences.
Dumb grid solutions
We have for a couple of years now received a discount on our power bill by having a special device attached to our A/C unit that reduces our power consumption on one day a year. It’s not quite as intelligent as the smart thermostats discussed above but it does the job (at least once/year). Such devices, used on a large scale could provide the capabilities of the smart thermostat but eliminate the potential for thrashing and large power oscillations.
So the smart grid is coming but the smarter it gets the more care we need to take in implementing it.