National Grid, the company selling electricity here in eastern Massachusetts, ran a Smart Energy Solutions pilot program in Worcester from 2015-2018. Serving about 15,000 customers, this program gave discounts when customers used energy off-peak.
Here’s how it worked. As of November 2018, they sold energy at three rates:
- Off-peak: $0.10697 per kWh
- On-peak: $0.13100 per kWh
- Peak-event (brownout conditions): $0.66997 per kWh
In other words, pilot program customers got a 2.3¢ per kWh discount when they used power off-peak, and they were fined 54¢ per kWh for power they used when it was in short supply. For comparison, they charged about 13.7¢ per kWh for customers not in the pilot program. Considering that these prices don’t include distribution charges, the customer savings are modest.
But a small pilot program naturally misses a couple of things. A large scale smart grid can, by controlling domestic hot water heaters, level out the electricity load grid-wide. If the fleet of controlled heaters is large enough, the potential generation cost savings are big. The traditional, crude, way to handle times of overload is to fire up the expensive peak-load steam turbines and beg residential customers to turn off unnecessary appliances. But real-time control of load provides a much more useful way to handle overload. For one thing, it’s fast. A radio or internet signal to smart water heaters can turn them off, or on, in less than a second. That allows the people managing the generators to smooth out their load dynamically.
This sort of real-time control of load also allows frequency regulation. When grids are heavily loaded there’s a tendency for the alternating current frequency to fall below the standard 60Hz. Similarly, when the load decreases the frequency can creep up. Either one is bad: For wide-area grids (Regional Transmission Operations) to connect to local distribution grids the frequencies must match. So frequency regulation–called Frequency Ancillary Services by the utilities–is a big deal.
Tesla installed a large scale 120 megawatt battery system in South Australia. Its rapid response to changing grid load proved a great way to provide Frequency Ancillary Services, costing about a tenth as much as using peaking steam turbines for the purpose. That’s an excellent result.
Here’s the thing. Batteries are expensive. The cost of a domestic Tesla Powerwall, installed, is about $10,000, for a 6.7kWh capacity. Controlled hot water heaters can serve almost exactly the same purpose at almost no additional equipment cost, and with the same responsiveness as the battery.
There’s a pilot program for a grid operator to try.