Duke Energy issued a press release on Friday (2/20/15). The areas that they provide electrical power in much of North and South Carolina (3.9 million customers) set a record for the most electricity consumed in one hour. From 7am to 8am, customers consumed 36,676 megawatt-hours.
The demand for electricity varies constantly. Historically, utilities simply make sure they can supply power on the worst days in summer in winter. Supply is simply ramped up to match demand. For a typical work day, demand looks something like this:
When the majority of the population sleeps, consumption is low. As people wake up and prepare for work, demand ramps up. It surges again later in the day as people go home and use energy-hungry devices like ovens and TVs. This supply-demand matching presents challenges that are beginning to be addressed with devices like batteries and smart meters. Hopefully, these devices will reduce the need for peaking power stations that are used only a handful of days each year. Renewables with their variable but predictable nature will also play a role.
Unfortunately, the circumstances Friday morning illustrate the difficulty of relying largely on renewables like wind and solar. On Friday in central NC, it was around 9 degrees Fahrenheit and sunny (sunrise was at 6:56am). The wind speed was about 7 mph. The power demand was 36.7 GW. With really rough estimates, I hope to show how renewables would cope with this unusual hour. These are rough estimates and the grid would be more resilient with distributed production, but indulge me.
Solar spill or solar trickle?
A typical spot in NC gets about 4.61 kWh per square meter over the course of a day in February. While much of this energy would come down near noon by averaging this output over all the hours in a day, we get 192 watts per square meter. Matching demand for that hour corresponds to the sunlight falling on 74 square miles. This is before expanding to panels or accounting for efficiency of the cells. You’d also have to assume no snow or long shadows have blocked the panels and energy storage could make the sunlight available throughout the day and night. Imagine driving down the interstate and for an hour all you see is a half mile wide blanket of solar panels.
Is the answer blowing in the wind?
Wind turbines are designed to work best at certain wind speeds. If the speed is too slow, the blades won’t turn. If the speed gets too high, a switch is flipped to stop the blades and protect them from damage. A typical curve looks like this:
To estimate the conversion of meters/second to mph just multiply by 2. For example, the 7 mph converts to 3.1 m/s. From this plot, we see that output is at maybe 2% of rated capacity (assuming the 7 mph is constant and not averaged). That means that the current wind fleet for the entire US (65 GW) would provide maybe 3.5% of the Carolina electricity demand. Again this is assuming all the turbines were located in central NC and turning in 7 mph wind.
How about energy storage?
The current US electricity storage capacity is 24.6 GW. Most of it is pumped hydro which means that extra power is used to pump water up a hill and then released through turbines when needed. This would provide 67% of the Carolina’s electricity used during that hour. Of course that’s assuming the amount of stored water wasn’t depleted in previous hours or needed in future hours.
The bottom line
It seems unlikely that renewables, hydro power, and energy storage will be able to provide the requisite peaking demand on very cold and very hot days. We would have to overbuild capacity by two or three fold in anticipation of low production periods and that would cost a lot of money. Another option is that people voluntarily reduce their demand, but Duke Energy claims that they asked their customers to do this and still hit peak demand. And I think nearly everyone would find blackouts unacceptable. It might be unavoidable that we use fossil fuels for these peaking periods. That is just one reason why we should intensely fund research on carbon capture and storage technologies.