Today, solar and wind are relatively low cost, and prices will likely fall further. But they are not like fossil fuels—they are what’s known as variable renewable energy (VRE)—meaning they only produce electricity when the sun shines or the wind blows.
Sometimes this variation is predictable. For example, solar doesn’t generate power at night, and generates less power in the winter. But some variation is unpredictable; cloud cover or wind droughts can last weeks at a time. These “VRE deficits” are not a problem as long as wind and solar are a small percentage of electricity generation. But as they become dominant, how do we fill the gaps when VRE supply is low?
Electric utilities are currently adding lots of short duration (4-hour) lithium-ion batteries to address daily and hourly variation, providing enough power in the evening when demand is high. DOE expects that trend to accelerate, but while these batteries are reliable and can be turned on with the flip of a switch, they only solve VRE deficits for a few hours at a time.
Longer duration VRE deficits are another matter. What happens when deficits last for weeks or months? How do we keep the lights on? Detailed studies—for example, from the Royal Society in the UK—show that that there can be sustained periods of poor weather, leading to annual shortfalls as high as 20 percent of total grid production (in a completely decarbonized grid).
Right now, when demand spikes or supply is disrupted, we turn mainly to “peaker plants”—gas plants that run on a single cycle to power turbines for electricity. These plants are less efficient than combined cycle gas turbine plants (CCGTs), and their energy is more expensive, so they are kept in reserve and are paid to provide spare capacity, ensuring that sufficient emergency power is always available. Typically, grid operators pay for about 15 percent of peak energy demand as spare capacity. Any less, and there might be blackouts. Any more, and money is wasted. While VRE use is relatively modest, peaker plants can cover VRE deficits.
But as VRE becomes a larger and larger share of the overall grid, the impact of adverse weather events will become more pressing, and supply will become more variable. A grid with 40 percent VRE will need much more insurance to guarantee the lights stay on.
There are alternatives to gas-fired energy, technologies that can store energy for very long periods and provide it on demand at massive scale—notably hydrogen, hydropower, and compressed air. Hydropower is cheap to run, a mature and well-established technology, and its generators last for decades. But we are removing more dams than we build, and there is no U.S. appetite for the dozens of huge new dams that would be needed. Conversely, green hydrogen suffers from very low efficiency—it takes 100KWh of green electricity to provide 40KWh of stored dispatchable electricity—and it is very expensive to make and store, while transportation is also expensive and can cause serious emissions. Moreover, hydrogen will not suddenly become cheaper quickly, like wind and solar. Compressed air currently operates at about 50 percent efficiency. Substantially improvement will require cooling and then re-heating air, which is costly when the air is stored for long durations. And the electro-chemical batteries now being added to the grid won’t provide the combination of long duration, huge scalability, and low costs that are needed.
This leaves us with gas.
To make a VRE-dominated grid possible, natural gas must be used as insurance against VRE deficits. That’s why DOE projects that peaker plant capacity will actually grow during the clean energy transition, and why it projects no significant decline in CCGT capacity either.
Eventually, we can hope to build a green grid with no emissions. In that world, gas emissions will eventually be phased out altogether. Perhas gas will be replaced with new fuel sources like geothermal or even fusion, carbon capture might eliminate emissions from gas power plants, or scalable storage technologies could finally mature so they can become the energy insurance of tomorrow. Overall though, it’s clear that we don’t have all the technology we need, and that we need to push much harder to improve those critical technologies, many of which are still not ready for full commercial deployment.