California sweltered under record temperatures last week. On Tuesday, Sacramento reached its highest ever recorded temperature of 116°F, breaking a record from 1925. That's 46.7°C for those on the metric system. This was part of a sequence of six days over 106°F and 10 days over 100°F. 

Normally, this is the point where I would complain about how unpleasant it was to live in that heat all week. I'm told it was unpleasant. Fortunately for me, I was on the other side of the country participating in a fantastic conference on the social cost of water pollution.

Agricultural workers faced brutal conditions, even if they worked in the early morning before the temperature peaked. The heat likely damaged some crops, although for many crops the heat wave came late enough in the season not to do too much damage.

 

Yet, much of the focus in the state this week was on the electricity grid. Many observers saw irony in state officials recommending that electric vehicle owners avoid charging their vehicles between 4pm and 9pm just after the state formalized its plan to phase out the sale of new fossil fuel vehicles.

That plan, which I wrote about two years ago when Governor Newsom first announced it, faces many obstacles to success. One challenge is to expand the electric grid and make it resilient enough to charge all those electric cars amid increasing temperatures while simultaneously eliminating carbon emissions from electricity generation.

They have a plan for that; it involves more wind and solar generation and lots of batteries.  It will include more solar generation on farmland.

This week provided a big test of the grid. At 4:55pm on September 6, electricity demand reached its highest ever level in the state. This is the hottest part of the day in the Central Valley, so it is when air conditioning use peaks.  It is also when the sun gets low in the sky and solar generation drops off. 

With solar power fading and demand peaking, could the state keep the lights on?

 

At noon that day, about a third of California electricity was being generated from renewable sources (wind and solar). As peak demand approached and solar started to drop off, the state increased hydro generation, natural gas generation, and imports from other states. The grid also received a significant contribution from batteries, which had been charged earlier in the day when supply was plentiful.  

 

These supplies weren't enough to meet expected demand, which left two options. The first was to plead with customers to reduce consumption during peak hours. The state asked people to raise their thermostat setting and avoid running appliances and chargers during the peak-demand period.

The pleas were successful. On September 6, consumers reduced demand by about 2000-5000 MW during the peak. This was similar to the amount contributed by batteries.

(I obtained this number by projecting the load shape we observed on August 23 onto September 6. Load peaks at 6pm. I assumed load reductions began at 3:30pm and ended at 10:00pm. This is a simple and rough approximation that could be done much better with more time and data. See the orange line in the figure.)

 

The second option was to institute rotating power outages. Thankfully, we never got to that point. Some cities lost power due to a "communication snafu", and most of the city of Davis lost power due to overheating transformers, but there were no rotating outages on September 6 or any other day last week.

This event shows that people are amenable to adjusting the electricity use when they understand the usefulness of doing so, but we cannot rely on them to do so continually out of the goodness of their hearts. My UC Davis colleagues Katrina Jessoe and Dave Rapson showed in a 2014 paper that combining information with a price incentive causes substantial reductions in critical peak electricity use. Such critical peak price incentives are an important tool for the future. We can also use automated demand side management to adjust electricity use in critical peak periods (e.g., automatically slow down car charging).

The other important tool will be batteries. The price of lithium-ion batteries has declined 97% in the past 30 years. This is why we have affordable smart phones and electric cars. As battery prices continue to decline, it will be why we have more battery storage to add resilience to the electric grid.

We kept the lights on this time, but there is work to do.

 

I made the graphs in this article using this R code.