September was a bad month for electricity in California. On September 8, a failed high-voltage transmission line from Arizona left nearly 1.4 million without electricity in San Diego County and surrounds for an entire day. That same week, the state's legislature defeated an extension of its existing tax on utility bills, AB724, that had been funneling some $400 million per year into renewable energy projects and technology research. Finally, amid wide reports of questionable federal loan guarantees, the California-based solar photovoltaic company, Solyndra, filed for bankruptcy. It was the third high-profile solar bankruptcy in the country this year.
This volatility is perhaps symptomatic of the new normal for electricity in the United States, one that Americans accustomed to flicking on a switch for cheap power will have to accept as the industry attempts to address the need to reduce carbon emissions, improve efficiency, and adopt clean generation technologies. In California, there has been a rush toward adopting renewable energy projects on commercial and residential scales in response to the state's Renewables Portfolio Standard (RPS), which now requires 33 percent of retail electricity sales from renewable sources by 2020. Nearly 30 other states have similar RPS requirements. Sales of solar photovoltaics, for example, have been on the uptick, buoyed by this demand and lower prices, reflecting China's heavy public investment in its own solar companies. But even as California and others ramp up their mix of renewable generation, there is still some disagreement about how wise it is to pursue a larger portion of renewable energy on the national grid.
Unlike coal-powered or natural-gas-fired electricity generation, wind turbines and PV systems are unpredictable. They depend on nature's whims to supply our erratic demand for electricity. The University of Cambridge physics professor, David Mackay, in his 2009 book, Sustainable Energy—Without the Hot Air, calls these whims "lulls" and "slews." A lull occurs when there is no wind or no sun, while a slew describes a significant fluctuation in demand, such as when the nation begins its morning commute. Critics fear that the investment needed to make a fully renewable-supplied electrical grid reliable in both circumstances would not only be unreasonable, but also overly ambitious technologically. Mackay's basic solution is twofold: investment in energy storage, such as pumping water uphill during peak electricity generation periods; and use-based controls, such as demand response. However, the devil is in the details.
Demand response relies on smart grids that can shed demand on the electricity grid by shutting down many small things like air-conditioning units. The idea is that customers can agree to these load-shedding schemes in advance in return for a discounted rate, and override them if it presents an emergency or an inconvenience. Overriding the scheme would then result in the customer paying a higher price for electricity consumed during a demand-response event. In several studies and trial installations, such systems have been shown to lower average electricity demand by 20 percent or more, but large-scale implementation has been slow, largely due to cost (Illinois' governor vetoed comprehensive smart grid legislation in September).
Adding significant renewable energy generation capacity to the existing grid will require an investment in smart technologies like advanced switching, integrated communications systems, and relays to detect these supply and demand fluctuations. A large portion of the cost of increasing our reliance on renewable energy are mundane things like electrical substation upgrades and smart metering installations, almost akin to replacing the existing grid (one of Mackay's wackier digressions explores building a secondary grid only for renewables). A new study by the Lawrence Berkeley National Laboratory, "Mass Market Demand Response and Variable Generation Integration Issues," published in October, concludes that the "largest variability and uncertainty in variable generation [wind and solar] power production is over time periods of 1 to 12 hours." Ultimately, the study argues, demand-response technologies, especially smart meters, have fallen in price enough to make them viable for addressing these relatively short lulls and slews associated with wind and solar. The case for the large scale must be made at the small scale.
But it may be too soon to worry about solving a large-scale problem if it never emerges. Regardless of the appeal of a 100-percent renewable energy grid, it's not really a question for the United States in the short term. In 2010, all of the wind turbines in the country generated only two percent of our electricity. But official predictions of how this will change nationwide in a "low-carbon" economy are fairly low. The U.S. Energy Information Administration (EIA) projects that renewable-generated electricity will account for 17 percent of total U.S. electricity generation in 2035, up from 10 percent in 2010.
We are far from this number, although Texas broke a record for wind-generated electricity supplied on its grid, hitting 15.2 percent, or 7,400 megawatts, of demand for October 7. Other estimates suggest 20 percent or more much sooner, but nothing over 50 percent. Most of those gains will come from hydropower and wind, but renewable also includes solar photovoltaics, geothermal, landfill gas, biomass, waste-to-energy, and fuel cells. Of course, only 17 percent may seem like small potatoes for a 20-year projection, but it's also an average. Five states alone generate over half of the country's renewable energy—California, New York, Oregon, Texas, and Washington. Even with a bad electricity month like September, these states have the most to gain from the smart expansion of renewables. The rest of the country's states may have their own bad Septembers, perhaps much worse, ahead of them if they don't start diversifying—and smartening up—their grids in the next few years.