A long-term challenge to the widespread adoption of grid-connected PV (photovoltaic) power is managing the instability that can be introduced into the grid because of the intermittent and variable nature of PV power generation. Every type of solar installation—from a collection of residential installations (in the same city, perhaps) to a single utility-scale solar farm—loses power output rapidly when a cloud passes overhead. This common occurrence turns power production off and in the case of residential and commercial rooftop PV installations, immediately turns grid-power producers into power consumers. In the case of utility PV installations, cloud cover diminishes a major power source, disrupting power to potentially thousands of customers. Dramatic and immediate shifts in the nature of the power flowing throughout the grid can disrupt distribution across even larger areas because of how interconnected the system is (consider that the Northeastern Blackout of 2003 was traced back to instability caused by a power plant in Ohio). Large-scale grid-tied energy storage is the most direct solution to managing the fluctuations in power output from PV systems (along with other intermittent power sources such as wind and tidal power). Market-research firm Pike Research estimates that the global grid-energy-storage market could rise from $1.5 billion in 2010 to $35.3 billion in 2020.

Types of Power Generation

In some ways, electricity is a just-in-time manufactured product with zero shelf life. No way to maintain an "inventory" exists, so production capacity must always exceed demand or shortages (in the form of brownouts and blackouts) occur. Because electricity demand fluctuates from day to day and minute to minute, power companies rely on detailed forecasting models that incorporate historical demand, major events, and weather patterns. Even with these tools, power companies must constantly generate electricity in excess of anticipated demand to supply power to all users reliably.

Power plants traditionally fall into three categories: baseload, peaking, and load following. Baseload power plants—including coal, nuclear, and hydroelectric—constantly operate near full capacity (stopping only for maintenance or repair). Peaking power plants operate only during periods of peak demand. Relatively expensive and inefficient to operate, peaking plants include gas-turbine systems that can shut down output during off-peak hours. Load-following power plants—including hydroelectric and steam turbines—sit somewhere in between baseload and peaking plants, and they can adjust their output to follow demand. Essentially, baseload power plants (typically the cheapest and most efficient of the three types) should be designed to meet the lowest level of demand a utility experiences to minimize wasted power, and a mix of load-following and peaking power plants ought to cover the anticipated fluctuations in demand that cycle throughout the year (for example, more electricity consumption during summer days to power air conditioning or more electricity in winter evenings to power lights when the sun goes down earlier) and throughout any given day (typically, more power is consumed during the day than at night).

PV power plants do not fall neatly into any of the three categories of conventional power plant. Although the output from PV power plants is variable like that of a peaking or load-following plant, it is uncontrollable and thus not as easy to incorporate into a conventional grid-power-management scheme. Fortunately, PV generates electricity during the day when demand is often high. Unfortunately, when demand spikes at an inopportune time (at night, for example), PV is not useful. PV can certainly offset some load-following-plant output, but displacing meaningful amounts of load-following-, peaking-, or even baseload-plant capacity requires the coupling of some form of energy storage to the PV plant.

Grid-Tied Energy Storage

Adding large-scale energy storage to the electricity grid will give grid electricity a shelf life that provides utility companies with greater flexibility in dealing with fluctuating demand. Storage systems can collect energy as an intermittent energy source (such as PV) produces it and output a regulated amount of electricity that is not dependent on uncontrollable factors (such as when the sun is shining).

On a small scale, a residential customer could conceivably rely entirely on PV electricity by using a PV system that has several times the power capacity he or she might need at any given moment (for example, 20 kilowatts). This (very expensive) solution places excess power from the daylight hours into storage that the user can draw from at night. On a utility scale, storage allows for large-scale PV power plants to function more like load-following power plants. Such a model requires significant overbuilding of capacity to meet 24-hour demand (as with the residential example) but would make PV far more useful for powering the electric grid. (On a side note, this is another reason that a simple cost-per-watt comparison between types of power plants is not always a fair comparison. To replace a baseload power plant, such as nuclear plant, and provide 24-hour electricity at comparable output levels, a PV power plant with storage capability would require a rated capacity that is between three and nine times higher than that of the baseload plant.) Grid energy storage can currently mitigate wind- and solar-energy variability and provide load-following capability for different types of power plants. Although the cost of storage is currently too high for several other functions, it may eventually be low enough to provide time-shifting capabilities (saving PV electricity for predictable spikes in evening demand) and provide peak shaving (generating and storing power during low demand for use in periods of high demand). Cost-effective energy storage could also enable utilities to provide grid stability without relying on long-distance transmission and distribution or perhaps make it possible for commercial and industrial users to optimize their power consumption by purchasing power when it is cheapest (often at night) and storing it for later use.

Storage Technologies

A wide variety of grid-tied storage options exist, many of which are better suited to specific energy-generation technologies. For example, some hydroelectric plants use excess power to pump water up from lower to higher reservoirs to provide greater flexibility in varying the plant's output to meet demand. So-called pumped hydro accounts for 95% of the grid storage in use today. Although it is a relatively cost-effective means of energy storage, pumped hydro is limited by geography, requiring large reservoirs and inclines that are not available in all locations.

Other technologies include compressed-air energy storage (where air is pumped into geological formations and stored for release to turn a turbine at a later date), flow batteries (which use large reserves of electrolyte solutions in storage tanks to increase energy-storage capacity), mechanical flywheels, ultracapacitors, superconducting magnetic-energy storage, and conventional battery technologies, including lead-acid and lithium batteries. Because of its scale and power characteristics, battery storage is the leading near-term choice for providing storage specifically to PV systems. Different battery technologies offer different virtues, each striking a different balance between features, performance, and price.

In September 2011, S&C Electric Company (Chicago, Illinois) announced deployment of what it claims to be the first fully integrated storage system for utility-scale PV. Installed in Albuquerque, New Mexico, the sodium-sulfur-battery-based PureWave Storage Management System can produce 500 kilowatts of power. Lead-acid batteries are the lowest-cost battery technology but have lower energy density and cycle life than other batteries have. Lead-acid batteries are in use in a few large-scale energy-management applications, including a 40-megawatt-hour system in Chino, California. Hybrid systems that combine technologies are also in development. The Australia Commonwealth Scientific and Industrial Research Organisation (CSIRO; Clayton, Australia) has developed the UltraBattery, which uses a lead-acid battery in parallel with an ultracapacitor. The ultracapacitor acts as a buffer during charging and discharging to provide significantly (fourfold) longer life and higher power output (50%) than lead-acid batteries alone provide, at much lower cost. CSIRO has licensed its technology to the Furukawa Battery Company for use in Japan and Thailand and to the East Penn Manufacturing Co. Inc. (Lyon Station, Pennsylvania) for use in North America.

Rechargeable Li-ion batteries and other advanced batteries offer high energy density and efficiency but are too expensive for utility-grid storage. Manufacturers are increasing the power envelope of Li-ion batteries with the development of improved nanobased electrodes and have also made capacity and safety improvements. Auto-industry manufacturers are now scaling up Li-ion batteries for use in hybrid and electric vehicles (EVs), and economies of scale for manufacturing will help to lower costs. For utilities, large numbers of plug-in EVs could also represent a significant source of distributed energy storage in a scheme called vehicle-to-grid (V2G) storage. A report form the US Department of Energy's Pacific Northwest National Laboratory estimates that 2.1 million EVs with a 33-mile electric range (not unlike that of the Chevrolet Volt) could provide enough grid-stabilizing capacity for 10 gigawatts of wind energy in the northwest of the United States (assuming the cars were equipped with the necessary V2G electronics).

A new type of sodium-ion battery for large-scale electricity storage, which should cost significantly less than Li-ion batteries, is also in development by Jay Whitacre at Carnegie Mellon University. Whitacre's start-up company, Aquion Energy (Pittsburgh, Pennsylvania), received $30 million in venture capital in 2011. Because sodium is more abundant than lithium, it has the potential to be far less expensive. The company claims that the battery material is also nontoxic and completely recyclable.

Looking Forward

The utility-energy-storage market is growing as some regional energy policies are requiring increased grid-tied storage. Market-intelligence firm IHS iSuppli estimates that the market for grid-tied lithium batteries will reach nearly $6 billion by 2020. This demand is tied to so-called smart-grid-systems-deployment requirements over the next decade. Unfortunately, the capital expense of most viable energy-storage technologies is just too high to be offset by the long-term cost and energy savings that such technologies provide. However, as prices for batteries decline, they will become more appealing to power companies looking to stabilize their output and incorporate larger amounts of PV capacity (and other intermittent renewable energy sources). Until then, policy mandates will likely support this growing market.