Solving the Variability of Renewable Power

An oft-repeated problem with renewable sources such as wind, solar, wave, and tidal is that they are variable. When the wind is not blowing or the sun is not shining then they don't produce power; conversely, there are times where more power could be produced than could be used. This variability, often exacerbated by its unpredictability, has significant implications, especially for utility-scale generation that is connected to the electrical grid.

Current grid management is, in its simplest form, the matching of electrical generation and electrical use, the matching of supply and load. Electrical grid managers are able largely to rely on the load profile, the historical variation of the load over time. There are two basic techniques today to match generation and load: generate additional electricity from various sources when needed (the usual approach) or reduce demand (demand response, much less common.) A detailed explanation can be found here:

The power utilities are able to predict to a reasonable accuracy (generally to within one or two percent) the demand pattern throughout any particular day. This means that the free market in electricity is able to schedule just enough base load in advance. Any remaining imbalance would then be due either to inaccuracies in the prediction, or unscheduled changes in supply (such as a power station fault) and/or demand. Such imbalances are removed by requesting generators to operate in so called frequency response mode (also called frequency control mode), altering their output continuously to keep the frequency near the required value.

The grid frequency is a system-wide indicator of overall power imbalance. For example, it will drop if there is too much demand because generators will start to slow down slightly. A generator in frequency-response mode will, under nominal conditions, run at reduced output in order to maintain a buffer of spare capacity. It will then continually alter its output on a second-to-second basis according to the needs of the grid.

This spinning reserve is a significant expense to the power utilities as often fuel must be burned or potential power sales lost to maintain it. The kind of generation used for fast response is usually fossil fuel powered which produces emissions of between 0.48 and 1.3 tonnes of CO2 equivalent for every megawatt hour (MWh) generated. Thus a significant environmental burden, in the form of increased greenhouse gas emissions, is associated with this imbalance.

Most forms of generation are unsuitable as peaking power plants (peaker plants, spinning reserves) because they cannot be efficiently started/stopped or operated on an intermittent or sudden demand basis. As a practical matter, only natural gas turbine generation can serve as peaker plants. This is the core reason why T. Boone Pickens, Chesapeake Energy and others are so interested in wind power--it will increase demand for natural gas.

Thus the paradox: the desire to add renewable sources of electrical generation is motivated in part by the need to mitigate climate change; however, the addition of variable renewable sources increases the need for spinning reserves, which currently adds to the carbon problem.

What to do? What other than natural gas, with its carbon footprint problems, could serve as a spinning reserve or, more broadly, as a peaking power plant or some kind of load following capability from storage that would enable near-instantaneous supply increases to respond to changes in the electrical demand?

An alternative is grid energy storage. With the growing interest in and development of electric vehicles, especially plug-in hybrid electric vehicles (PHEVs) some have suggested that a growing array of distributed batteries in PHEVs could serve as a source of additional electricity in periods of high demand.

The concept, called vehicle to grid (V2G), is based on the fact that your car is typically not being used 90 percent of the time. "What if it could work for you while it sits there?" said Jeff Stein from the University of Michigan.

The National Science Foundation has granted a research team lead by Stein $2M to explore the possibility of V2G technology using PHEVs. There are many problems to be solved, however. The cars would need to be plugged into a socket not just when being charged, but also so electricity could be drawn back out. How would this be controlled? No PHEV owner will be happy to wake up in the morning and find the battery (half-)drained after being plugged in all night, presumably charging. There are (potentially significant) efficiency losses in charging/discharging batteries, and the life of the batteries would likely be shortened by an arbitrary cycle where complete charge or discharge may not occur. Lastly, there would need to be substantial elements of a future smart grid deployed to even allow this distributed storage to be harnessed in a centralized way. Interestingly, there is already a test of this concept underway at the University of Colorado (Boulder) by Xcel Energy. Other tests are also underway by Southern California Edison, Austin Energy, Duke Energy, Wisconsin Power, Excel Energy, and Pacific Gas & Electric, amongst others.

Hydro is another mostly green approach. Here in Washington state we get about 70% of our electricity from conventional (big dam) hydroelectric power, which has the ability to serve as a peaker plant by letting more or less water flow out of the reservoirs and through the turbines. There is competition for the water, however, especially from irrigation, but also from navigation and fisheries concerns, so the degree to which these dams can serve as peaker plants is somewhat limited.

Pumped storage hydroelectricity is another storage mechanism that might be explored, and may be very well-suited in coastal settings with large amounts of ocean energy generation (offshore wind, wave, etc.) Some of the drawbacks of this form of energy storage would be mitigated by a reservoir built for the purpose, rather than the use of a pre-existing (freshwater) lake.

Storage could also be achieved via flywheel arrays, hydrogen generation, compressed air, or other techniques.

Longer term, an updated, expanded, and smarter electrical grid is necessary. Wind generation is more variable the more local the scale and geographic reach of the turbine array. As more wind generation comes on line in greater density and over a more diverse, interconnected geographic area, local variations even out and become less significant. Offshore wind, despite its higher cost has several significant advantages over onshore wind; a large one is greater wind (power) on a steadier basis. Large coastal arrays (example) would take out some of the variability.