Energy storage offers huge potential for helping renewables meet periods of peak power demand, but the financial and policy cases must keep up with the technology, explain Ralph Masiello, Rick Fioravanti and Jillis Raadschelders.
Our power system is embarking on significant changes that may require storage to be as necessary in the electricity industry as it is in every other type of industry operating today. One area where energy storage may become a necessity is with the integration of renewables into the electricity grid. As more systems look to adopt renewable generation into the electricity system, the network can become more difficult to balance. Before the advent of renewables, variability was evident only in the load. Now, as increasing amounts of renewable generation are added to the grid, variability is introduced to the supply side as well, making the system more unpredictable.
In the long term, the implications of widespread, mass deployment of electricity storage across the power system are profound. Integrating storage into the electric grid can add the flexibility required for the system to adjust to future increases in renewable energy generation penetration by acting as a bridge, a buffer and a reliability component. Harnessing today’s rapidly advancing energy storage technology for the sustainable and distributed grid of tomorrow requires keen insight into storage applications, the financial implications, and the public policies needed to encourage its use.
Renewable resources, especially wind generation, are often characterized as intermittent or variable in nature. This variability can impact the electricity grid in multiple areas. For operators, as renewable deployments hit penetration rates of 20% or more, sudden drops in output can create scenarios where the grid may not be able to respond in the available time. For renewable operators, this intermittency can create problems in trying to meet their forecast output. While it is well within the state-of-the-art to forecast that tomorrow will be a windy day, it is difficult to forecast the actual wind speed accurately hour-by-hour a day ahead. Wind power production forecast-versus-actual typically show significant variations, especially in the timing of wind fall-off and ramp-up.
Though this second-by-second variation in renewables is not generally an operational problem, the variability in how the diurnal cycle occurs certainly can be. This becomes apparent when examining systems that have a high wind and solar penetration. Highly predictable renewable patterns, such as the production from a concentrating thermal solar plant, may exhibit ramp rates that are very high compared to normal load ramping or conventional generation ramping. The potential of diurnal wind production falling off in the morning, while shortly afterwards the concentrating thermal solar plant is ramping up, can create major difficulties. Trying to balance this with conventional generation resources is impractical if not infeasible.
Storage can assist with renewable integration in the ways outlined below:
Evening out Fluctuations, Moderating Ramping
Levelizing these kinds of fluctuations as well as moderating ramping – up or down – is a perfect application for large-scale storage. Today, however, there are no financial incentives for renewable developers to install storage and there is no mechanism to allocate the costs elsewhere. Instead, the prospect is that the market will socialize these costs via increased procurement of regulation and balancing energy services. As regulation is paid for whether or not it is used, the implication is that the market operator must procure enough regulation services to cover the worst case, even if that only occurs on one day in every 10.
Another important aspect of renewable balancing has to do with transmission curtailments that may be imposed on remote wind farm production. In fact, if transmission is built on sound economic principles, this concept is almost a given. The capacity factor of a given land-based wind farm is typically 25%–30%, meaning that the average output of the wind farm is 25%–30% of its peak capacity. It may not be economic at all to build transmission with that kind of capacity factor in mind. But if the transmission is built to some level less than peak capacity, there will be occasions when the wind farm has to be curtailed. This is another perfect use for storage. When wind production is at peak and transmission curtailments are in effect, you can store the energy, then discharge it when wind production falls off and transmission capacity is available. The difficulty with realizing such a solution today has both economic and policy roots. The cost of storage needs to align with the benefit of the opportunity. Storage technologies with lower costs, such as compressed air storage, may work – provided a convenient cavern is located nearby – but the cycle efficiencies are an issue. However, this market requires that storage comes down in cost before it can be realized.
Diurnal and Longer Time Shifting
Because we don’t have direct control over when renewable systems produce power, production shifting is a key area of interest for storage applications. For wind, in some areas production peaks tend to be at night, when prices are low, while production drops during the hottest days, when prices are high and the energy is needed most. Though large solar does align with mid-day peaks, utilities are seeing peaks extend into early evenings, when people return from work and turn on appliances, after the sun has set. For renewables, production shifting with storage has tremendous potential value.
Adjusting to Forecasting Error
Because our ability to forecast output of renewable applications such as wind and solar is limited by our ability to predict the weather (or better stated, the exact timing of expected weather changes), storage can act as a buffer to allow the renewable generation output to more accurately match predicted outputs.
Pumped hydro storage has long been used to levelize energy demand versus production over daily and weekly cycles. As such, it is a great benefit to generation fleets that have a preponderance of lower cost baseload (nuclear, coal, run of river hydro) units that are less flexible in varying their output. Unfortunately, pumped hydro is inherently limited by siting issues. A convenient valley that can be dammed without too much adverse environmental and public consequence is needed, and the reservoir needs to be near adequate water supply – meaning a river, typically.
While many large reservoirs behind major dams also serve as an economic engine for recreational purposes, the typical pumped hydro facility does not because the level fluctuates severely during daily and weekly cycles. This is also a siting issue on the downstream side – the river, lake, or other body of water on the low side has to be able to tolerate widely fluctuating inflow/demand for water as well.
Below: Storage allows wind production to be used when transmission is available
The Financial Case for Storage
Technically, storage provides a viable solution to the renewable intermittency problem. However, putting a price on renewable balancing is not straightforward, in part because the benefit falls on many parties. Opinions vary on how much storage would be needed at a particular wind penetration level. There are a few alternative routes one could take in analyzing the value of storage in this application.
One way is to examine the ability of storage to lower the cost of wind integration. Wind integration costs represent the combined impact of incorporating variable or as-available wind power into the grid. A number of recent wind integration studies have been conducted and US estimates range from negligible to around US$5/MWh for wind penetration levels of up to 30%. However, resource plans in the US have quoted numbers as high as $10/MWh. This value could be viewed as an upper boundary for the cost savings captured by storage. According to the American Wind Energy Association, wind generated 52 TWh of energy in 2008. A conservative integration cost of $2/MWh suggests up to $104 million could be captured by storage. However, in such specific cases, it is less clear who the beneficiary of these savings will be, particularly when a storage device is likely to serve more than one purpose.
Another approach to estimating the value of renewable balancing is to look at the annual carrying cost of a conventional generation plant used to firm up the wind production. Financial benefit is the cost to own and operate such a plant. For instance, The California Independent System Operator (CAISO) estimates for 2008 include: combined cycle power plant at $132.6 / €76.0 per kW year; and a combustion turbine at $162.1 / €118.6 per kW year.
This could again be viewed as an upper boundary for what value can be captured by storage. However, if there is already spare capacity, this benefit is significantly reduced. The degree to which the variable wind power would have to be levelized also matters. While yielding benefits from renewable levelizing, storage devices could simultaneously provide returns from energy arbitrage. For example, a battery could cycle continuously throughout the day, levelizing wind. By storing the energy and releasing it at key times, storage would allow producers to take advantage of price differences over time. The maximum benefit would shift energy from off-peak to on-peak price periods. However, the amount of time between such periods would require storage durations most likely applicable for diurnal load shifting, noted opposite on page 89.
In order to calculate shorter-term arbitrage benefit, a detailed model would need to be developed, including price curves, battery operation characteristics, forecast wind and curtailment patterns. KEMA has conducted many studies on this topic.
Diurnal Load Shifting is simply using energy storage for energy arbitrage. This flexibility provided by storage can help increase the value of the energy produced or help lower barriers to investment. This may be the case for a wind farm with high winds at night, when energy prices are low, or worse, the load on the system is too low to accommodate the full wind power production.
While solar energy tends to provide around 80% of its nameplate rating during peak demand periods, wind farms have a much lower value for coincident peak production, often estimated at around 20%–30%. Similarly, a remote wind farm may be limited by transmission capacity during times of maximum production. Since the average power production from a wind farm is around 25%–30% of its nameplate rating (i.e. capacity factor), transmission is rarely built to nameplate capacity. In order to avoid curtailment and to take advantage of on-peak prices, nighttime production can be stored on-site and then discharged during the day, when demand is typically high or during periods when adequate transmission volumes are available.
A storage technology capable of up to six hours of discharge is well suited for this task. The price of curtailment is simply lost production, something a new wind farm developer is eager to avoid as it is otherwise hard to make the economics for wind farms favourable. Assuming, however, that nighttime production is not at risk of curtailment and noting that nighttime prices are generally lower than daytime prices, we can look at arbitrage possibilities.
Make It So. Policy to Encourage Storage
So why is it good public policy to encourage electricity storage at all? As more technologies and renewable generation are added to the electricity grid, storage can help ensure its safe, reliable operation. More importantly for society, the devices can potentially provide a better solution to future grid issues coupled with lower costs.
Electricity grids are periodically subjected to unexpected peaks or ‘super peak’ demands. The current solution to meeting these is to deploy more generation, transmission, and distribution resources. However, society cannot afford to continue to add infrastructure designed for long-term operation to meet a short-term need, as it will eventually lead to significant rate increases for customers.
Another example can be seen with renewable portfolio standards. In this case, public policy is encouraging the use of generation that may necessitate the need for infrastructure to control the variable nature of the renewable generation being encouraged. Today, the solution is similar to the above peak demand scenario, where older generation plants will be used to control the inevitable variability in supply. Society will need to carry the burden of the increased costs and emissions of keeping these older plants operational.
An alternative to building infrastructure exists in the form of demand response. Tools such as demand response and dynamic pricing are widely seen as key policy elements in managing our future energy and infrastructure costs. But even with aggressive and successful demand response, there will be limits on how much improvement we can obtain. In addition, there are unexplored limits on public acceptance and the viability of extreme demand response measures, which may be just as unacceptable to consumers as additional costs. Most consumers do not want to have their lifestyles inconvenienced and will push for alternative solutions.
Above: Pumped hydro offers storage, but is inherently limited by siting issues
Policymakers realize there are potential problems coupled with solutions that are not very acceptable to the public. Storage is an alternative both to building to match peak load and to managing demand. Indeed, it fills this role, as well as providing a backstop against supply failures in every other commodity. Pumped hydro aside, for the first time in more than a century we have the potential to use electricity storage on a large scale. Our task is to decide how best to do so in conjunction with judicious infrastructure expansion, increased demand management, and improved energy efficiency.
Renewables is an area that is being encouraged in many countries, yet the burden to integrate this additional capacity is falling on the grid operators. Storage can be a solution to assist with this integration, but similar issues arise. Policies are mandating that renewables be integrated, but this means that independent power producers are creating renewables, and in some areas regulated utilities are forced to incorporate the generation or import the generation. Paths that policymakers could take to address this problem include investment tax credits, socializing costs and merchant storage.