Solar Eclipses and the Electric Grid: Markets and Automation

Yesterday’s solar eclipse across the US amplified a dominant issue in electricity policy discussions over the past couple of years — does increasing use of distributed energy resources like solar photovoltaics make the grid more resilient, or does it lead to imbalance and inadequacy? In California during the eclipse (Financial Times), solar generation dropped compared to the day before, while system operators implemented algorithms to bring on more hydroelectric generation to meet demand:

As solar production declined, the California grid manager tapped more supplies from natural gas-fired power plants and hydroelectric dams, executing plans they have been crafting for months. It also imported power from neighbouring states.

Solar output then roared back as the eclipse rolled eastward at midday, when the sun’s radiation is strongest.

The event was a kind of dress rehearsal for a future in which intermittent solar sources have a bigger role in the power supply. California aims to generate half its power from renewable sources by the year 2030.

“We were really pleased with how smoothly everything went,” an ISO spokeswoman said. “It bodes wells for renewable energy on the grid during an event like this.”

Automation and smart grid technologies are already making this kind of distributed resource interconnection both feasible and valuable. The combination of market institutions and digital technologies at the consumer retail level would amplify that capability and that potential value creation. A key factor here is dynamic pricing, a topic that I covered a lot several years ago.

In a recent Forbes post, Koichiro Ito of the University of Chicago connects these ideas and reports some findings from his research on consumer response to time-of-use (not dynamic) pricing in Japan. Ito and his co-authors tested two hypotheses: the information given to consumers about how much money they are likely to save on TOU pricing changes their behavior, and a cash payment to enroll in the program changes their behavior.

We studied more than 2,000 households in Japan who could join a new time-based program that offered a low electricity rate during the morning and night, and a higher rate in the afternoon when electricity use was at its peak. Our goal was to measure what role information and incentives play in the customers’ decision-making. Some customers received very little information about the programs. Another group of customers was given information on how much their electricity bills would increase or decrease under the new plan. And a final group was given the same information plus a $60 bonus for switching.

Some of their results were expected (cash payments induced switching), while some were unexpected. In particular, the consumers who actually changed their electricity consumption the most were those who did not receive either additional information or a cash incentive — whether through intrinsic motivation to reduce energy consumption or some other motive, the most responsive consumers changed their behavior without information or external incentive.

The behavioral aspects of this result have important implications for policy: information and cash incentives are not necessarily sufficient to induce widespread reductions in electricity consumption, even in response to a high peak price in a TOU contract. But what about technologies that make it easier and cheaper for non-intrinsically-motivated consumers to change their behavior?

I’d like to see the research take the next step to transactive energy, to test the hypothesis that consumers with information and with transactive technology they can use to automate their response to price changes will change their behavior as much as the baseline intrinsically motivated consumers. In an experiment with TOU pricing such transactive technology is not necessary because the pricing is not dynamic, but imagine an experiment with real-time pricing, transactive consumer technology, and this treatment design.