Electric Energy Storage and the Bulk Power System: An Introduction to the Applications and Impact of Storage

admin
Print Friendly
Share on LinkedIn0Tweet about this on Twitter0Share on Facebook0

______________________________________________________________________________

Eric Selmon Hugh Wynne

Office: +1-646-843-7200 Office: +1-917-999-8556

Email: eselmon@ssrllc.com Email: hwynne@ssrllc.com

SEE LAST PAGE OF THIS REPORT FOR IMPORTANT DISCLOSURES

______________________________________________________________________________

February 15, 2017

Electric Energy Storage and the Bulk Power System:

An Introduction to the Applications and Impact of Storage

We expect storage to be a rapidly expanding part of the electric power system going forward with significant and growing impacts on regulated utilities and power markets. In this note we examine the need for storage on the bulk power system, the most frequent uses of storage and their economics, the impact on competitive generators and the potentially important and under-appreciated role utilities may have to play in driving further adoption of storage.

Portfolio Manager’s Summary

  • The inability to store power economically on the 1,000 GW U.S. power grid has required costly investment in peaking capacity, and, to a lesser extent, in the transmission and distribution network, materially increasing system costs and the price of electricity.
  • Peaking plants that operate less than 15% of the hours of the year comprise 170 GW or 15% of U.S. generation capacity, with an estimated replacement cost of $115 billion (see Exhibit 4).
  • By contrast, there are only 24 GW of electric energy storage capacity on the U.S. power grid, accounting for just 2% of total U.S. generation capacity (see Exhibit 10). Of this existing capacity, 93% is pumped hydro, which faces limited opportunity for future deployment due to the geographical limitations on new installations (see Exhibit 11).
  • However, installations of new storage capacity have surged over the last decade.
    • Deployments of electro-chemical and small scale electro-mechanical energy storage, i.e., batteries and fly wheels, averaged 6 MW annually over 2008-2010, but then jumped ten-fold to an average of 63 MW over 2011-2014. Installations surged again to an average 168 MW annually over 2015-2016 (see Exhibit 15).
    • By far the most widely deployed of the new storage technologies is the lithium ion battery; over 2015-2016, it accounted for 97% of capacity additions (Exhibit 17).
  • Given the high cost of today’s battery technologies, however, the economics of deploying energy storage on the grid remain daunting.
    • In particular, batteries remain uncompetitive with peakers. A regulated utility seeking to add peaking capacity can deploy a gas turbine peaker at a third of the cost of a lithium ion battery (see Exhibit 21). And in competitive markets, the gross margin from purchasing power off-peak and selling it on-peak recovers only a tiny fraction of the cost of a new lithium ion battery (see Exhibit 20).
  • Initially, batteries will be competitive in more limited but higher value uses on the grid. In PJM, market prices for frequency regulation are at or near the level required to recover the cost of new zinc or lithium ion batteries. Even on a national basis, however, this is a tiny market, comprising 2.5 GW of capacity with annual revenue of less than $400 million (Exhibits 23, 24).
  • The bulk of the new energy storage projects being deployed in the U.S. today therefore seek to capitalize on revenue streams from more than one use. Of the ~800 MW of electro-chemical and electro-mechanical energy storage capacity operating in the United States, almost two thirds is deployed in a secondary as well as a primary use, and almost half is deployed in a tertiary use as well (see Exhibit 26). Of the 800 MW, ~70% is providing frequency regulation to the grid; ~30% provides spinning reserves; another 30%, is deployed for electric energy time shifting; and 14% to firm renewable energy capacity (Exhibit 27).
  • While competitive generators face relatively few opportunities to deploy storage on the grid at attractive rates of return, regulated utilities can create significant positive externalities (social benefits) by deploying storage and have been encouraged by regulators in California to do so. By reducing the need for capacity during peak demand hours, storage can reduce the marginal cost of supplying electricity, and therefore its price, during those hours when demand is highest. In regions where generation has been deregulated and utilities procure power for their customers in the wholesale market, suppressing peak hour prices by deploying storage on the grid can create material savings for utility customers.
  • We expect, therefore, that regulators in states with competitive generation will follow California’s example and encourage or require regulated utilities to invest in storage. Utilities will capitalize their investments in energy storage in regulated rate base, while passing through to their customers the benefit of lower wholesale power prices. This may rebound to the benefit of the transmission and distribution utilities in these states, such as EIX and PCG in California, ES and AGR in New England, and AGR and ED in New York.
  • Over time, we expect utility investment in storage to moderate spikes in peak power prices, sharply curtailing the revenue of merchant generating fleets and creating further challenges for CPN, DYN and NRG. And as energy storage facilitates the integration of renewable resources on the grid, round-the-clock wholesale power prices will be eroded as well.
  • Although we do not cover the suppliers of energy storage technology, we have included a table of the leading providers of technology for existing and announced storage projects in Exhibit 30.

Table of Contents

  1. The Bulk Power System and the Need for Electric Energy Storage
  2. Storage Capacity on the Grid Today
  3. Energy Storage Applications
  4. In Which Applications is the Deployment of Energy Storage Economic?
  5. Is Energy Storage More Economic If Deployed by Regulated Utilities?
  6. Industry and Company Impact

Exhibit 1: Heat Map: Preferences Among Utilities, IPP and Clean Technology

Details

We expect storage to be a rapidly expanding part of the electric power system going forward with significant and growing impacts on regulated utilities and power markets. In this note we examine the need for storage on the bulk power system, the most frequent uses of storage and their economics, the impact on competitive generators and the potentially important and under-appreciated role utilities may have to play in driving further adoption of storage.

  1. The Bulk Power System and the Need for Electric Energy Storage

Historically, a critical feature of the power industry, which it shares with barber shops, brothels and amusement parks, is that it provides a service which must be supplied at the same moment it is consumed. Just as it is impossible to store a haircut for when the customer requires it, historically it has not been economic to store electricity for future use. The other industries that share this characteristic manage imbalances in supply and demand by requiring the customer to wait (e.g., lines at your favorite rides). But in developed countries, consumers are not accustomed to wait for electricity; rather, they flip a switch on entering a room and expect it instantly to be lit.

Each such increase in the demand for electricity must be met with an instantaneous increase in supply. Similarly, when demand falls, power supplies must be reduced in parallel. Large imbalances in the supply and demand for power cause surges in voltage frequency that can damage electrical equipment, or dips that render this equipment inoperable. For grid-connected equipment and appliances to operate as designed, power generation must be maintained in continuous and instantaneous balance with power demand. A critical implication is that the capacity of the power grid to generate electricity must equal or exceed the peak demand for power – even if that level of demand is reached only a handful of hours per year, and the capacity required to supply it otherwise remains idle.

As a result, the power generation capacity on the grid is commonly a multiple of the average level of demand, a fact that results in very low capacity utilization rates for generation plant. This is illustrated in Exhibit 2, which presents for the major power markets of the United States the ratio of (i) peak demand for power to (ii) the average level of power demand for the course of the year. As can be seen there, this ratio varies between ~1.6 and 1.8.

These ratios imply that, were it possible to store electricity economically for use during peak hours, the scale of the generating fleet required to meet peak demand could be reduced by some 40%.

In fact, the problem is worse than this, because a grid’s generation capacity is sized not only to meet peak demand but to meet other unforeseen, adverse circumstances as well, including the loss of the largest power plants supplying the grid due to mechanical failure or natural disasters. As illustrated in Exhibit 3, the North American power grid aggregates 1,040 GW of generation capacity to serve only 842 GW of expected peak demand, keeping some 198 GW of capacity, or almost 20% of the total, in reserve.

Exhibit 2: Ratio of Peak to Average Load Exhibit 3: North American Generation

by Region (%) Capacity & Forecast 2017 Peak Demand (GW)

Source: Energy Information Administration, SSR analysis Source: North American Electric Reliability Corporation, SSR analysis

These two facts — that peak power demand is 60 to 80% higher than average demand, and that the capacity of the North American power grid exceeds peak power demand by 20% — are reflected in a very low level of capacity utilization for the U.S. power generating fleet. As illustrated in Exhibit 4, the annual power output of the U.S. fleet is only 43% of its potential output or, put another way, 57% of the potential power output of the fleet is not used. That the aggregate cost of the U.S. power generation fleet must be recovered during only 43% of the hours of the year materially increases the per kilowatt hour price of electricity.

Exhibit 4: Average Capacity Factor by Generation Technology

Source: Energy Information Administration, SSR analysis

To make this point more concretely, consider that 170 GW of the installed generation capacity in the United States is used to meet peak demand and supply backup capacity to the grid in the event of a disruption in supply. These peaking plants – comprising primarily gas turbine generators and antiquated oil and gas fired steam turbine generators – operate on average less that 15% of the hours of the year (Exhibit 4). When these plants reach the end of their useful life they will be replaced with the lowest cost peaking capacity available today: an advanced gas turbine generator with an average installed cost of some $680 million per GW. The replacement cost of these rarely utilized but essential assets can thus be estimated at $115 billion – well in excess of the enterprise value of Duke Energy, the nation’s largest utility.

The grid’s need for peaking capacity not only requires significant capital expenditure; it also results in higher operating costs during the hours of peak demand for electricity. To minimize the cost of providing peaking capacity to the grid, utilities will generally select for this use generating plants that are the cheapest to build (e.g., gas turbine generators) but which often are less fuel efficient and consequently more costly to operate during the limited hours they run. Importantly, utilities dispatch their power plants on an economic basis, operating their lowest cost units first and firing up incremental generation in rising order of variable cost to meet system demand. As a result, the cost of generating power during hours of peak demand tends to be materially higher than the cost of off-peak supply. A further inefficiency of a grid that operates without energy storage, therefore, is a higher level of fuel expense than would be required if the grid’s most efficient power plants could be continuously operated and their output stored to meet peak demand. In competitive power markets, where the price of electricity at any moment reflects the variable operating cost of the last unit dispatched to meet demand, the problem is a particularly serious, as inefficient peaking plants set the price of power for all customers precisely when the demand for electricity is at its highest. The differential between peak and off-peak power prices on the nation’s largest wholesale power markets is illustrated in Exhibit 5.

Exhibit 5: The Spread Between Average Peak and Off-Peak Power Prices by Region in 2016 ($/MWh)

Source: SNL Energy and SSR analysis

Exhibit 6: The California Load Curve on a Typical Summer Day, Total and by Segment

Source: Lawrence Berkeley National Laboratory

These are not the only inefficiencies of a power system that cannot store electricity. Smaller but nonetheless costly inefficiencies arise from the need of the power generation fleet to adjust its output to meet the daily swing in power demand from the minimal levels required in the pre-dawn hours to its evening peak. Exhibit 6 illustrates the swing in demand on the California power grid during a typical summer day. Two things are striking about the chart: first, that the peak evening demand for power is twice the minimum level recorded during the pre-dawn hours, and second, the nine hour, 25,000 MW ramp in power production required to bridge this gap. Power plants, like automobiles, run most efficiently when operated continuously; the need to cycle them on a daily basis to supply this swing in demand materially reduces their fuel efficiency and increases their variable maintenance costs.

To be more specific, the process of heating boilers and raising the pressure in steam turbines at an idle nuclear or coal fired power plant takes hours and, due to the extreme changes in temperature involved, places significant stress on critical plant components. As a result, it is generally physically or economically impractical to cycle nuclear and coal fired plants on a daily basis. Gas turbine generators, which are derived from jet engines, can cycle more quickly and easily, although not without increased wear and tear and consequently higher variable maintenance expense. Moreover, as these plants ramp up to full capacity, or ramp down from their maximum output, they are operated below design capacity and consequently at less than optimal fuel efficiency. Least fuel efficient of all are the plants providing “spinning reserves” to the system. These gas-fired units are not connected to the grid but are nonetheless fired up and spinning so that they can be connected and provide power to the grid within ten minutes to meet the daily ramp in the power demand or offset the loss of an operating generation unit.

The daily cycling of gas turbine and combined cycle gas turbine generators, and the cost of holding a portion of this fleet in spinning reserve to meet the daily ramp in power demand, is bad enough when there is single daily cycle from minimum to maximum demand and back. Unfortunately, the introduction of large scale solar generation introduces a second daily cycle: the increase in solar power output as the sun rises to its peak at noon and the subsequent decrease as the sun descends in the afternoon hours. Because the variable cost of generating electricity from solar panels is zero, system operators will use solar power to the maximum degree possible, offsetting its daily rise with a reduction in the output of the grid’s gas turbine and combined cycle power plants. The result is that the gas fired capacity on the California grid is increasingly required to cycle twice each day: up from the pre-dawn hours until sunrise, then down until 12-1pm (solar noon), up again as the sun descends and load rises into the evening peak, and then down again as night falls. This dual cycle is illustrated in Exhibit 7 below, which estimates the need for conventional thermal and hydroelectric generation on the California grid by subtracting from system wide demand the output of the region’s growing solar fleet.

Exhibit 7: CAISO Net Load: Forecast System Wide Demand Less Solar Generation, 2012-2020

Source: California Independent System Operator

Seasonal and daily swings in demand are not the only drivers of the swings in output of the generating fleet. Small variations in power demand on a minute by minute basis can cause volatility in the frequency and voltage of the power on the grid, requiring an immediate response from the system’s generators (Exhibit 8).

Exhibit 8: The Variability in Power Demand over a Three Hour

Period on the California Power Grid (MW)

Source: California Independent System Operator

The causes of these fluctuations are numerous and difficult to predict. On a hot summer day, changes in wind speed or cloud cover can materially affect ambient temperatures, altering the demand for air conditioning. In systems with high penetrations of wind and solar power, changes in wind speed or cloud cover can also have a material impact on the amount of renewable generation supplied to the grid. Random fluctuations in the use of industrial or transportation equipment can materially affect the level of power demand on the grid. Whatever their cause, short term imbalances in the demand and supply of power cause instability in the frequency and voltage of the electric energy on the grid. If prolonged, these fluctuations can cause damage to critical grid components, including power plants, and so must be quickly offset by adjustments in power output.

In the absence of batteries to provide the rapid discharge and storage of electricity, these adjustments in power output are provided by power plants that are connected to the grid and operating below their full capacity, and thus are able immediately to alter the amount of power they supply the grid by ramping their generation up or down. This service, referred to as “regulation”, is expensive for power plants to provide; when deployed in providing regulation service, generators cannot produce at full capacity, must therefore operate at sub-optimal fuel efficiency and are subject to the wear and tear and consequent increase in maintenance costs associated with continually changing levels of power output.

In summary, abundant opportunities exist to enhance the efficiency of the bulk power system by deploying technologies to store electric energy. These opportunities include storing cheap off-peak power for later use during hours of peak demand, thus arbitraging between peak and off-peak power prices (referred to as “energy arbitrage” or “energy time shift”); substituting storage for power plants held in inefficient “spinning reserve;” and deploying storage to provide “regulation” by charging or discharging as needed to offset short-term fluctuations in power demand on the grid. In the next section, we will consider how electro-chemical and electro-mechanical energy storage devices can serve these needs.[1]

  1. Storage Capacity on the Grid Today

There are ~24 gigawatts (GW) of electric energy storage capacity on the U.S. power grid, accounting for 2% of total U.S. generation capacity (see Exhibits 9 and 10). Of the electric energy storage capacity currently deployed in the United States, pumped storage hydroelectric facilities account for 23 GW or 93% (Exhibit 11). Thermal energy storage, primarily molten salt systems at solar thermal power plants, accounts for 3% U.S. electric energy storage; electro-chemical systems (batteries) for 3%; and electro-mechanical energy storage, including compressed air energy systems and flywheels, for 1%. In aggregate, the gross electric output of these facilities is equivalent to ~1% of U.S. power output. On a net basis, however, grid connected electric storage consumes some 6 million MWh of electricity, as typically 15% to 30% of the energy stored by these facilities is lost.

Exhibit 9: Breakdown of U.S. Generation by Exhibit 10: Breakdown of U.S. Generation by Source, 2016 (GWh) Capacity by Energy Source, 2015 (GW)

Source: Energy Information Administration Source: Energy Information Administration

Exhibit 11: Breakdown of U.S. Storage Exhibit 12: Breakdown of Electro-chemical

Capacity by Technology, 2016 (%) & Electro-mechanical Storage Capacity (%)

Source: U.S. Department of Energy, SSR analysis Source: U.S. Department of Energy, SSR analysis

Hydroelectric Pumped Storage

From the 1920s to the mid-1990s, 22.5 gigawatts (GW) of hydroelectric pumped storage capacity were built in the United States. Today, the nation’s 38 pumped storage facilities account for 93% of U.S. electric energy storage capacity. Reflecting the scarcity of cost-effective, geologically suitable and environmentally acceptable sites, no major pumped storage facilities have been commissioned in the United States since 1995.

U.S. pumped storage facilities range in scale from 100 MW to 1000 MW, with an average capacity of ~600 MW. Designed to generate electricity for four to twelve consecutive hours, pumped storage facilities are capable of generating throughout the peak demand hours of the day. This long duration of discharge, and the capacity for daily cycling, are necessary to amortize the high cost of the storage reservoirs, hydroelectric dams, and other civil engineering works required by this technology.

 

As illustrated in Exhibit 13 below, pumped storage hydroelectric facilities deploy two reservoirs, one at an elevation substantially higher than the other. Water from the lower reservoir is pumped to the upper reservoir at night, when power demand and electricity prices are low; during the peak demand hours of the day, when electricity prices are at their highest, the water in the upper reservoir is released through hydroelectric turbine generators and discharged into the lower reservoir for re-use.

Exhibit 13: Schematic Diagram of a Pumped Storage Hydroelectric Facility

Source: British Broadcasting Corporation

Thermal Energy Storage

Thermal energy storage is an economically efficient of means of conserving electrical energy for use during hours when its value is higher. There are 820 MW of thermal storage capacity in the United States today, accounting for 3% of U.S. storage capacity. Of these, 540 MW are deployed at four solar power plants using concentrated solar technology. At these facilities, mirrors are used to concentrate the sun’s energy to melt salt; pumped through tubes between parabolic solar collectors, the temperature of the salt is raised to over 1000 degrees Fahrenheit. The molten salt is used to store this heat until early evening, when the sun is low but when power prices reach their maximum levels for the day. The molten salt is then pumped to a steam generator (boiler) to produce superheated steam for use in a steam turbine similar to that found at a conventional coal fired power plant.

Another 200 MW of thermal energy storage has been deployed by commercial and municipal electricity consumers, primarily for building climate control. During off peak hours when electricity prices are low, the chillers of a building’s air conditioning system are used to chill water. The chilled water, which is mixed with ethylene or propylene glycol to prevent it from freezing, is circulated through tanks of water to make ice, which is then melted during the hottest hours of the day. Cooling loops running through the ice extract the cold to provide cooling to the building, decreasing the power required by the air conditioning system’s compressor and chiller motor during the highest priced hours of the day. In addition to reducing on-peak electricity consumption, thermal energy storage can also cut the capital costs of air conditioning systems. Because the system’s chillers are being used not only during the hottest hours of the day, but during off-peak hours as well, their scale can be reduced by approximately half.

Compressed Air Energy Storage

A similar concept is applied by compressed air energy storage systems. These comprise the bulk of electro-mechanical energy storage, which in turn accounts for ~1% of the electrical energy storage in the U.S. today. Compressed air energy systems operate electrically powered condensers at night, when energy prices are low, to compress air and store it in underground caverns. During peak demand hours, when electricity prices are at their highest, the compressed air is released, cooling it due to its expansion, to supply chilled air for use in a conventional gas turbine generator. Approximately two thirds of the energy produced by gas turbines is used to condense and chill air for the turbine’s own combustion chamber, so as to maximize the efficiency of fuel combustion. By substituting compressed air from the underground chamber, the bulk of the turbine’s energy can be dedicated to operating the electric generator, materially increasing the power output of the turbine and thus its maximum generation capacity. In essence, rather than the use the turbine’s own energy to compress air during peak hours, when the value of that energy is highest, the process of air compression is shifted to hours when energy is cheapest. The stored air is then released to maximize the output of the gas turbine when power prices are highest. The consequent increase in the peak hour capacity of the gas turbine provides a further benefit.

Compressed air energy storage systems can be of substantial size, ranging up to 1000 MW, and can provide peak hour electricity for two to thirty hours depending on the scale of the underground chamber. Like pumped storage hydroelectric systems, therefore, they are well suited to supply peak hour power demand. Also like pumped storage, however, their deployment is limited by geology, and in particular tby he need for an underground cavern or man-made equivalent. In the United States today, only three utility scale compressed air storage systems are currently in operation.

Electro-chemical

Finally, electro-chemical technologies to store electric energy – batteries – account for ~3% of the storage capacity in operation in the United States today. While very high cost, batteries can be deployed in ways that hydroelectric pumped storage and compressed air energy storage cannot. As we will discuss below, the small scale and modular nature of battery storage technology permit their use in a range of distributed applications, whether by utilities or electricity consumers, that other storage technologies cannot serve. The speed and precision with which they can be charged and discharged, moreover, renders batteries highly suitable for regulating variations in the frequency of the power on the grid (regulation) and for providing a quick and continuous response to diurnal increases or decreases in power demand (spinning reserves).

To help the reader conceptualize the key differences between (i) utility scale storage systems such as pumped hydro and compressed air energy storage on the one hand, and (ii) electro-chemical and small-scale electro-mechanical energy storage systems on the other, we have reproduced below a useful chart (see Exhibit 14) from the DOE/EPRI Electricity Storage Handbook published by Sandia National Laboratories in February 2015. The storage technologies are plotted on a chart whose horizontal axis shows the typical capacity range in MW of the different storage systems and whose vertical axis shows the discharge time for which they are generally appropriate. Note that both axes use a log scale.

Exhibit 14: Storage Technologies Plotted by Capacity in MW and Discharge Time

(note that both axes use a log scale)

____________________________________________________________________________

Source: Akhil, Huff, Currier, Kaun, Rastler, Chen, Cotter, Bradshaw, and Gauntlett, DOE/EPRI Electricity Storage Handbook in Collaboration with NRECA, Sandia National Laboratories, February 2015, pg. 29, SSR analysis

On its horizontal axis, Exhibit 14 illustrates the difference in the scale of hydroelectric pumped storage and compressed air energy systems versus electro-chemical and small-scale electro-mechanical storage devices. Pumped hydro ranges from 100 MW to 1000 MW (1GW) in scale, and compressed air energy storage from approximately 50 MW to 1000 MW. By contrast, electro-chemical and small-scale electro-mechanical storage devices range from 1 kW and 50 MW. Even if we assume a modular deployment of the smaller devices, so that a series of batteries is deployed to provide 50 MW of storage, the capacity of a hydroelectric pumped storage system would typically range from 2x to 20x times larger.

On its vertical axis, Exhibit 14 illustrates the marked difference in the duration of discharge of the various storage technologies. The chart shows high power supercapacitors, flywheels, nickel-metal hybrid batteries, and nickel cadmium batteries providing short bursts of power measured in seconds or minutes. Lead acid batteries, lithium ion batteries and sodium based batteries can achieve substantially longer duration of discharge. These technologies may be deployed to provide regulation, discharging for up to 30 minutes at a time, or to operate as peakers designed to discharge for up to four consecutive hours. Importantly, however, even four hours of continuous discharge falls short of the requirements that many independent system operators (ISOs) impose on capacity resources. Hydroelectric pumped storage and compressed air energy storage systems, which are intended to provide generation to meet peak load during the highest demand hours of the day, generally are designed to operate between four and twelve continuous hours.

Flow batteries – batteries in which the electrolyte is stored outside of the cell, and is fed into the cell in order to generate electricity – hold out the prospect of materially longer discharge durations. A flow battery’s energy is a function of the volume of electrolyte; depending on the size of its electrolyte storage tanks, the duration of discharge of a flow battery can range from two to ten hours.  However, the power capability of flow batteries, which is determined by the size of the stack of electrochemical cells, remains relatively low, generally 100 MW or less.

The marked differences in power capacity and duration of discharge of pumped hydro and compressed air energy storage on the one hand and batteries and flywheels on the other have important functional implications. First, pumped storage and compressed air energy storage are utility scale systems designed to combine substantial power capacity with hours of continuous discharge. Their role on the grid is to serve as centralized sources of peak hour generation capacity. The smaller scale and, more importantly, the short duration of discharge of batteries and flywheels limit their suitability for this purpose. However, the smaller scale and modular nature of batteries facilitate their deployment on a distributed basis on the grid. Thus batteries may be deployed to supplement the capacity of overloaded transmission lines or distribution circuits during the limited number of hours per year when they are loaded beyond their design capacity, or to decrease the peak demand of large commercial and industrial utility customers, whose bills often include a demand charge calculated on the basis of their peak power consumption for the month. In addition, batteries can be economic sources of energy in certain applications requiring shorter term capacity commitments. Thus batteries that provide fast, precise but short term (less than 30 minute) bursts of power increasingly have found a role providing regulation services to the grid, and may in future supply spinning reserves as well. Finally, as we will see below, batteries deployed in such uses often can also be used to arbitrage energy prices, charging during off-peak hours when power prices are low and discharging on-peak when prices are higher (sometimes referred to as “electric energy time shift”).

While still a tiny portion of the storage capacity on the U.S. power grid, deployments of electro-chemical and small scale electro-mechanical energy storage, i.e., batteries and fly wheels, have surged over the last decade. Capacity additions averaged 6.4 MW annually from 2008 through 2010, but then surged ten-fold to an average of 63.1 MW from 2011-2014. Installations over 2015-2016 have averaged 167.9 MW annually, more than 2.5x the level of the prior four years (see Exhibit 15).

Exhibit 15: Deployments of Electro-chemical and Small Scale Electro-mechanical Energy Storage (Batteries and Flywheels) since 2007 (capacity adds in MW)[2]

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org) , SSR analysis

Among the electro-chemical and small scale electromechanical storage technologies, which have been deployed most widely on the grid to date, and which are most likely to be deployed going forward? Exhibit 16 provides a breakdown by technology of the various electro-chemical and small scale electro-mechanical (i.e., flywheel) storage devices deployed in the United States today. To allow a closer analysis of the contribution of these innovative, small-scale storage technologies, the charts in Exhibit 16 and Exhibit 17 exclude hydroelectric pumped storage, thermal storage and compressed air energy storage.

As these charts illustrate, by far the most widely deployed of the new storage technologies is the lithium ion battery developed for use in consumer electronics and electric vehicles. The dominance of the lithium ion battery, moreover, has increased dramatically in the last two years: over 2011-2014, lithium ion batteries accounted for 53% of the electro-chemical and small-scale electro-mechanical capacity added in the United States; over 2015-2016, they accounted for 97% (Exhibit 16).

Exhibit 16: Breakdown by Technology of Electro-chemical and Small-Scale Electro-mechanical Storage Capacity Deployed in the United States Today (MW) (Excludes Pumped Storage, Thermal Storage and Compressed Air Energy Storage)

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org) , SSR analysis

Exhibit 17: Deployment of Electro-chemical and Small-Scale Electro-mechanical Storage Capacity in the U.S. by Technology, 2010-2016 (MW)

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org) , SSR analysis

Finally, if we analyze planned additions of storage capacity over the remainder of the decade, it appears that lithium ion batteries are poised to contribute a larger share of future additions of storage capacity than more traditional technologies such as pumped storage, compressed air and thermal energy storage. Based on an analysis of all U.S. electric energy storage projects announced in 2015 and 2016, and which are scheduled to come on line over 2017-2020, lithium ion batteries will comprise three quarters of total storage capacity additions through the end of the decade (Exhibit 18).

Exhibit 18: Breakdown by Technology of Planned Deployments of Electric Energy Storage Capacity in the United States, 2017-2020 (MW)

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org) , SSR analysis

3. Energy Storage Applications

As we have seen, utilities have historically sought to deploy large scale, long duration energy storage to meet the peak power demand of the grid, primarily through the construction of massive hydroelectric pumped storage facilities. The average size of the U.S. pumped storage facilities approaches 600 MW, and they are designed to operate for between 4 and 12 hours. They are, in other words, utility scale sources of power during peak demand hours. By contrast, the average capacity of the electro-chemical batteries connected to the grid today is only 2 MW. More importantly, their average duration of discharge is less than an hour. As we discuss in more detail below, three characteristics of batteries — their high cost, limited capacity and short duration of discharge – have to date rendered them unsuitable for use as peaking capacity.

Rather than supply peaking capacity, therefore, electro-chemical and electro-mechanical storage units (e.g., batteries and flywheels) are being deployed in higher value uses where their smaller capacities and short duration of discharge are not a limitation. Exhibit 19 provides a sense of the range of these applications. Of the ~800 MW of electro-chemical and electro-mechanical storage capacity on the system today, 88% are deployed in uses related to the bulk power system. The most important of these is regulation, for which 45 storage projects totaling 314 MW (average capacity of 7.0 MW) have been built. The second most important application is electric energy time shift, for which 36 projects totaling 215 MW (average capacity of 6.0 MW) have been deployed. Together, these two applications account for two thirds of the electro-chemical and small electro-mechanical storage capacity on the grid today. Other bulk power applications include spinning reserve and ramping capacity, to meet the daily swing in load on the system; the supply of electric capacity, often in order to render reliable or “firm” the output of wind or solar power plants; and black start capability, or the supply of electricity during a blackout to restart the generating plants on the system.

The remaining 12% of electro-chemical and electro-mechanical storage capacity on the grid is deployed in a range of uses requiring smaller scale and often distributed energy storage. Among the principal use of small scale, distributed storage units are:

  • To enhance reliability, power quality and grid resilience, as when an industrial, large commercial or institutional consumer of power may deploy a battery on-site to mitigate the consequences of a blackout or brownout.
  • To defer transmission or distribution system upgrades. As power demand grows, high voltage power lines and substation may become over-loaded during periods of peak demand. Replacing these pieces of equipment can be both expensive and risky. Given the long useful life of power lines and substations, they are generally scaled to meet expected demand over the next 30 to 40 years; in designing replacements, therefore, utilities run the risk of miscalculating the scale of capacity required over the coming decades. Utilities may therefore choose temporarily to deploy storage on the downstream side of the bottleneck to supplement the capacity of the power line or substation, especially during the current period of slow-to-no demand growth in many regions, deferring the more substantial capital expenditure and thus alleviating pressure on customer rates.
  • To manage electricity bills. Utilities will often bill their industrial and large commercial customers a combination of (i) a demand charge, expressed in $/MW, calculated on the customer’s peak monthly demand for power from the grid, and (ii) an energy charge, expressed in $/MWh, calculated based upon electricity consumed and its time of use, and tending therefore to be significantly higher during hours of peak demand than during off-peak hours. Such customers have an incentive to deploy small scale distributed storage units at their premises to store electricity during off-peak hours and discharge it during on-peak hours. By discharging the battery during hours of peak demand, the customer can shave its peak load, and thus the demand charge it must pay the utility, as well as reduce its consumption of high priced peak hour energy. Commercial and industrial customers that have their own solar generation capacity can also use batteries to store electricity for their own use, or sale to the utility, when rates are at their highest.
  • Demand response. Some regional transmission organizations (RTOs), such as the California ISO, will allow commercial and industrial consumers of power to offer reductions in their load during hours of peak demand (“demand response”) as a substitute for the RTO acquiring system reserves, and will pay these consumers for their peak shaving commitments.

Exhibit 19: Average Size of U.S. Electro-chemical and Electro-mechanical Storage Projects by Primary Use (MW) (Excludes Pumped Hydro and Thermal Energy Storage)

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org), SSR analysis

  1. In Which Applications is the Deployment of Energy Storage Economic?

Given the high cost of today’s energy storage technologies, rarely will any one of these applications in Exhibit 19 above provide the revenue required to offset the capital and operating costs of deploying energy storage.

Let us begin our cost analysis with the most obvious use of energy storage on the grid: storing cheap off-peak power for later use during hours of peak demand. As noted above, the U.S. power grid includes ~170 GW of peaking capacity whose capacity utilization rate is less than 15%. The low fuel efficiency of these peaking units sustains peak hour power prices well above the level of off-prices (see Exhibit 5). This price differential is insufficient, however, to recover the cost at which energy storage technologies can be deployed today. We illustrate this point in Exhibit 20, where we compare (i) the annual cash flow required to recover the cost of various energy storage technologies[3] with (ii) the net revenue that could be earned by arbitraging between off-peak and on-peak energy prices. To estimate the net revenue from energy arbitrage, we have assumed that an electric energy storage system capable of four hours of discharge, with a cycle efficiency of 85% (power discharged over power used to charge the storage system), is deployed in a daily cycle of purchasing power off peak and selling it on during on peak hours. Our net revenue calculation is based upon the average peak and off-peak power prices prevailing in the seven U.S. RTOs in 2016 (see Exhibit 5). As Exhibit 20 illustrates, the annual net revenue from energy arbitrage never exceeds 5% of the cost of deploying storage.

Exhibit 20: 2016 Annual Net Revenue from Energy Arbitrage in U.S. RTOs Compared to the Cost of Various Storage Technologies in $/kW-year (1)

_______________________________________________________________________________

  1. Annual revenue requirement reflects the low end of the range of estimates in Lazard’s Levelized Cost of Storage – Version 2.0, December 2016

Source: Sandia National Laboratories, DOE/EPRI Electric Storage Handbook, February 2015; Zhi Zhou, Todd Levin, and Guenter Conzelmann, Argonne National Laboratory, Survey of U.S. Ancillary Services Markets, January 2016, pg. 34; Lazard and Enovation Partners, Lazard’s Levelized Cost of Storage – Version 2.0, December 2016; SSR analysis

If the margins from the arbitrage of peak and off-peak power prices in the wholesale power markets are too low to justify the deployment of batteries, could battery storage nonetheless represent an economic alternative to a gas turbine peaker for a regulated utility that must increase the peak capacity on its system? We analyze this question below, and present the results in Exhibit 21.

Lithium ion and zinc batteries are two of the lowest cost forms of electrical energy storage suitable for widespread deployment on the grid as sources of peaking capacity. Their relatively small scale and modular characteristics facilitate their deployment at sites where geologically constrained hydroelectric pumped storage and compressed air energy storage are not feasible. Their small scale also renders them attractive relative to conventional gas turbine peakers in urban areas where power plant sites are not readily available. Finally, both technologies are capable of four hours of continuous discharge, the minimum level a reliability coordinator might deem sufficient to meet demand during peak hours.

At current prices, however, these batteries are completely uncompetitive with a conventional gas turbines as sources of peaking capacity. Exhibit 21 illustrates this point by comparing the all-in cost (i.e., the sum of operating and capital costs) of a gas turbine peaker to that of a lithium ion battery in a peaking application. Our source for the capital cost, fixed and variable operating cost, and heat rate of the gas turbine peaker is the Energy Information Administration’s Capital Cost Estimates for Utility Scale Electricity Generating Plants, published in November 2016. Based on the EIA’s cost assumptions, we have estimated the annual revenue requirement of a new gas turbine peaker over its useful life given the weighted average cost of capital of a typical U.S. regulated utility today. Our source for the cost per MWh of a lithium ion battery in a peaking application is Lazard’s Levelized Cost of Storage – Version 2.0, prepared by Lazard with assistance from Enovation Partners and published in December 2016. As can be seen in the final line of Exhibit 21, the cost per MWh of the lithium ion battery ranges from 2.9x that of a conventional gas turbine, assuming a 15% capacity factor, to 3.7x that of a gas turbine at a 5% capacity factor.

Both in a regulated context and in competitive wholesale markets, therefore, our analysis suggests that batteries cannot yet be economically deployed to provide peaking capacity to the grid.

Exhibit 21: Comparison of the Cost of a Lithium Ion Battery in Peaker Application to That of Gas Turbine Generator

_____________________________________________________________________________________________________________________1. Assumes a pre-tax weighed average cost of capital of 9.9%, reflecting a 15% pre-tax cost of equity (equivalent to 10% after-tax), a 4.5% pre-tax cost of debt, a 50/50 debt/equity mix and a 35% tax rate.

Source: Energy Information Administration, Capital Cost Estimates for Utility Scale Electricity Generating Plants, November, 2016; Lazard and Enovation Partners, Lazard’s Levelized Cost of Storage – Version 2.0, December 2016; SNL and SSR analysis

We have also analyzed the value that can be realized in providing operating reserves to the grid. Operating reserves include spinning reserves, which are available within 10 minutes to meet changes in load, and non-spinning reserves, and which are not synchronized to the grid and take longer to deploy. Many of the regional transmission organizations operating in the United States have created markets for spinning and non-spinning reserves; historical prices for these services are therefore available,[4] allowing the annual revenue that can be realized from the supply of these services to be compared with the cost, expressed in $/MW-year, of the various storage technologies. As can be seen in Exhibit 22, however, the expected revenue from spinning and non-spinning reserves is sufficient to recover only a fraction of the cost of energy storage.

Exhibit 22: 2014 Average Revenue for Ancillary Services Compared to Annual Revenue Requirement of Various Storage Technologies, in $/kW-year (1)

_______________________________________________

  1. Annual revenue requirement reflects the low end of the range of estimates in Lazard’s Levelized Cost of Storage – Version 2.0, December 2016.

Source: Sandia National Laboratories, DOE/EPRI Electric Storage Handbook, February 2015; Zhi Zhou, Todd Levin, and Guenter Conzelmann, Argonne National Laboratory, Survey of U.S. Ancillary Services Markets, January 2016, pg. 34; Lazard and Enovation Partners, Lazard’s Levelized Cost of Storage – Version 2.0, December 2016, SSR analysis

Our analysis suggests, however, that certain battery technologies can be profitably deployed to provide regulation, charging or discharging as needed to offset instantaneous imbalances in the demand and supply of power on the grid and thus maintaining the grid’s frequency within requisite limits. In particular, prices for regulation on the PJM Interconnection[5] are at or near the level required to recover the cost of zinc and lithium ion batteries, and nearly high enough to recover the cost of sodium based batteries (Exhibit 22 and Exhibit 23).

However, the scale of the markets for frequency regulation, spinning reserves and non-spinning reserves is tiny, particularly in comparison to the potential market for storing power for on-peak use. We pointed out above that 170 GW of U.S. generation capacity comprises peaking plants whose capacity factors are below 15%. By comparison, across the seven U.S. regional transmission organizations, only 2.5 GW is procured annually for frequency regulation, 5.4 GW for spinning reserves and 8.7 GW for non-spinning reserves. The dollar value of the services procured in these markets is correspondingly small, totaling less than $400 million annually across the seven RTOs for frequency regulation, ~$150 million annually for spinning reserves and $60 million annually for non-spinning reserves (see Exhibits 24 and 25). While the need for these resources will likely grow with increasing renewables penetration on the grid, we do not see a meaningful change in the magnitude of the ancillary services markets.

Exhibit 23: Cost of Various Storage Technologies Compared to 2014 Market Revenue for Regulation in PJM and Across All U.S. RTOs in $/MW-year (1)

  1. Annual revenue requirement reflects the low end of the range of cost estimates in Lazard’s Levelized Cost of Storage – Version 2.0, December 2016.

Source: Sandia National Laboratories, DOE/EPRI Electric Storage Handbook, February 2015; Zhi Zhou, Todd Levin, and Guenter Conzelmann, Argonne National Laboratory, Survey of U.S. Ancillary Services Markets, January 2016, pg. 34; Lazard and Enovation Partners, Lazard’s Levelized Cost of Storage – Version 2.0, December 2016, SSR analysis

Exhibit 24: Average Prices and Annual Market Revenue for Ancillary Services, by RTO

__________________________________________________________________

Source: Zhi Zhou, Todd Levin, and Guenter Conzelmann, Argonne National Laboratory, Survey of U.S. Ancillary Services Markets, January 2016, pg. 3, SSR analysis

Exhibit 25: Generation Capacity Dedicated to Various Ancillary Services and Peaking Capacity

% of Capacity in the Seven U.S. RTOs

GW of Capacity in the Seven U.S. RTOs

GW of Capacity in the United States (Estimated Based Upon Seven RTOs)

Source: Energy Information Administration; Zhi Zhou, Todd Levin, and Guenter Conzelmann, Argonne National Laboratory, Survey of U.S. Ancillary Services Markets, January 2016, pg. 34; SSR analysis and estimates

Given these limited revenue available from providing spinning and non-spinning reserves, regulation and energy arbitrage, it should not be surprising that the bulk of the new energy storage projects being deployed in the United States today seek to capitalize on revenue streams from more than one use. If we exclude conventional hydroelectric pumped storage and thermal storage, and focus on the newer electro-chemical (battery) and electro-mechanical sources of energy storage (flywheels and compressed air energy storage), there are ~800 MW of energy storage capacity operating in the U.S. today, of which 628 MW are electro-chemical and 171 MW electro-mechanical. The most commonly cited commercial applications of these projects is regulation (39% of the 800 MW installed) and electric energy time shift (27%). However, some 510 MW of this capacity, or 64% of the 800 MW of electro-chemical and electro-mechanical storage capacity deployed, also operates in a secondary use, most frequently to supply spinning reserves (36% of the 510 MW) and regulation (21%). Finally, 390 MW of the storage capacity in operation today (49% of the 800 MW installed) are deployed in a tertiary use as well, most commonly regulation (36% of the 390 MW) and the firming of renewable energy capacity (20%) (Exhibit 26).

Exhibit 26: Primary, Secondary and Tertiary Uses of Electro-Chemical and Electro-Mechanical Storage Capacity

Primary Use Case (100% of Projects):

Secondary Use Case (64% of Projects):

Tertiary Use Case (49% of Projects):

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org), SSR analysis

Adding the capacity of those projects that list frequency regulation as their primary, secondary or tertiary use, it would appear that 565 MW, or 71% of the 800 MW of U.S. electro-chemical and electro-mechanical storage capacity, is providing frequency regulation to the grid. Some 252 MW, or 31% of the total, provides spinning reserves; another 236 MW, or 30%, is deployed for electric energy time shifting; 109 MW or 14% is deployed for the firming of renewable energy capacity; and 95 MW, or 12% for ramping (i.e., enhancing the ramp capabilities of generation resources deployed to follow load) (Exhibit 27).

Exhibit 27: Stated Applications of U.S. Electro-chemical and Electro-mechanical Storage Projects (% of Installed Capacity)

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org) , SSR analysis

  1. Is Energy Storage More Economic If Deployed by Regulated Utilities?

In the preceding sections, we have compared the cost of electric energy storage to the revenues such facilities can realize through the provision of various services, including regulation, electric energy time shift and spinning reserves, given the market prices of those services. The presumption underlying this analysis is that the economic value of the electric energy storage facility should be assessed based upon the private benefit to its owner, i.e., by calculating the potential revenues of the facility and comparing these to the total costs of building, operating and maintaining the facility. An alternative approach, which may give a very different answer, is to assume that the economic value of the facility reflects the private benefit to its owner plus the wider benefit of the facility to society at large. An example of a social benefit of an electric energy storage facility that cannot be captured by its owner is the impact of energy storage on the price of electricity, which would rebound to the benefit of all the electricity consumers on the system. In fact, as we illustrate with an arithmetical example below, only a small fraction of the social value of storage can be realized by the owner of the energy storage facility.

During the summer months, demand on the California ISO tends to peak at ~50,000 MW between 2:00 and 6:00 pm. Minimum demand runs at approximately half this level, between 2:00 and 6:00 am (see Exhibit 6). Reflecting actual market conditions in the CAISO in 2016, we will assume in this example that, due to the marked difference in demand between peak and off-peak hours, the annual average price of power during peak hours is ~$30/MWh and the annual average price during off-peak hours is ~$25/MWh. What is the economic value-added of an electric energy storage designed to arbitrage this difference by purchasing and storing power during off-peak hours when the price is low and discharging and selling power during peak hours when the price is high?

For purposes of this example, let us assume that a private firm proposes to deploy 5,000 MW of electrical storage capacity on the CAISO system. Comprised of a series of lithium ion batteries deployed at various points on the high-voltage power grid, the proposed system would be capable of cycling 365 days per year, discharging for four hours during each cycle, and doing so with a round-

trip efficiency of 90% (i.e., the ability to discharge at peak 90% of the electricity stored during off-peak hours). Given an average peak power price of $30/MWh, the revenue that the system could realize per MW installed would be some $43,800 annually ($30/MWh x 4 hours of discharge per day x 365 days). The cost of charging storage system would be ~$40,600 per MW per year ($25/MWh x 4 hours per day x 365 days/90% to reflect the round-trip efficiency of the battery). The annual gross margin of the facility would thus be $3,200 per MW per year, or some $16 million annually for the 5,000 MW system.

The developer of the project would compare this estimated annual gross margin with the all-in cost (the sum of all capital and operating and maintenance costs) of the system to determine its economic viability. In the December 2016 study by Lazard and Enovation Partners, Lazard’s Levelized Cost of Storage – Version 2 ($399/kW-year), the annual revenue required to recover the all-in cost of lithium ion batteries is estimated at $399/kW-year. For a 5,000 MW (5,000,000 kW) system, this works out to $2.0 billion per year. The project would clearly be uneconomic.

In addition to this private benefit, however, a 5,000 MW system to store electrical energy on a power grid with peak demand of ~50,000 MW could generate much larger social benefits that the owner the facility would not be able to capture. Let us assume that by reducing peak demand by 10%, the storage system eliminates the need to operate the least efficient power plants on the grid — oil fired steam turbine peakers – that previously had operated and set the price of electricity some 5% of the hours of the year. Such plants, many now 50 years old, frequently have heat rates (ratios of fuel consumption to power output) of 15 million Btus/MWh. As the cost of industrial fuel oil exceeds $10/MMBtu, the fuel cost of these facilities can be estimated at $150/MWh. If we further assume that, once demand is reduced by 10%, the least efficient units dispatched to meet demand are gas fired combustion turbines with a heat rate of 12 million Btus/MWh, and that the price of gas is $3.00/MMBtu, then the price setting units on the system during the 5% of hours when demand is highest would now have a fuel cost $36/MWh. The availability of 5,000 MW of storage would thus reduce the marginal cost of supplying electricity during the peak hours of the year by ~$114/MWh. In a highly competitive wholesale power market like CAISO, this reduction in the marginal cost of supply would be reflected in a similar reduction in the wholesale price of power. The cost to California’s utilities of procuring electricity to supply their customers would fall commensurately.

Critically, this reduction in price would occur during 5% of the hours of the year (438 hours annually) when power demand would be running at 50,000 MW or higher. A total of 21.9 million MWh of power purchases (50,000 MW of demand for 438 hours) would be affected. The savings to electricity consumers across the grid would thus be $114/MWh x 21.9 million MWh or ~$2.5 billion per year. Given the annual revenue requirement for lithium ion batteries of some $399/kW-year ($399,000/MW-year), the savings to electricity consumers would be significantly in excess of the revenue requirement of 5,000 MW storage system ($2.0 billion annually).

A project with an expected annual gross margin of $16 million and an annual revenue requirement of $2.0 billion would be an economic non-starter for a competitive generator; but the $2.5 billion social benefit of the project would offer savings to electricity consumers well in excess of the project’s cost. The attractiveness of deploying energy storage on the bulk power grid, when assessed

from a social perspective, suggests to us that storage is first likely to deployed at scale by regulated utilities rather than competitive generators.[6]

In fact, taking the public benefits of storage into account, the California Public Utility Commission in 2011 set a target for the state’s regulated utilities to procure 1,325-MW of electric energy storage by 2020. California’s regulated utilities are pleased to comply, as a portion of the capital cost of deploying storage on the grid will be added to the utilities’ regulated asset base, on which the companies are allowed to earn a regulated return. The reduction in the prices paid by their customers for peak hour electricity, and the consequent savings on customer bills, will also rebound to the utilities’ benefit; having reduced customer bills through the deployment of storage, utilities may find themselves better positioned with customers and regulators to advance efforts to accelerate rate base growth or raise shareholder returns. Given these advantages to the utilities, and the compelling social benefits of energy storage, we would expect the deployment of storage by regulated utilities to become more common in other states as well.

If so, this pattern may offer a hint as to the potential winners and losers from the eventual roll-out of electrical energy storage at scale. We expect that in the long run regulated utilities will own a larger share of the nation’s energy storage capacity than the competitive generators. And as this storage capacity is rolled out, independent power producers will likely suffer on two counts: from the early retirement of peaking plants in response to the reduction in peak hour demand and, more importantly, from the erosion of peak hour prices, and thus revenues and gross margin, that will be suffered by the remaining plants in their fleet. Much like the growth of renewable energy in response to state renewable portfolio standards, the deployment of storage by regulated utilities will in the long run suppress the power output, prices and revenues of the nation’s competitive generators.

  1. Industry and Company Impact

In this final section we will estimate the size of the market for energy storage in the United States; second, assess the impact of energy storage on individual power companies; and third, identify leading public and private suppliers of energy storage systems and components.

To size the market for energy storage, we start with the premise that energy storage technologies will replace other technologies on the grid, including generation, transmission and distribution, and in doing so will reduce the cost of operating the grid. One way to estimate the potential scale of the investment in storage, therefore, is by reference to the largest part of the U.S. electric power grid that storage will supplant, peakers.

We estimated earlier in this note that there are 170 GW of generation that operate less than 15% of the time. If enough storage were installed to replace this capacity it would comprise 20% of peak demand. Storage capacity on this scale should have the flexibility to provide nearly all of the ancillary services currently required by the grid.

Given this framework, the maximum potential size of the investment in U.S. energy storage is the cost of supplying 680 GWh of storage, or 170GW for 4 hours. While we are not identifying any storage technology as a winner at this time, we will use the price lithium-ion battery costs to estimate the scale of the associated investment, as lithium ion technology has dominated recently installed and planned storage projects. One approach is to assume that rollout of 680 GWh of storage is achieved at current prices for lithium-ion batteries. A second approach is to assume that these prices fall materially over time. The second approach is more realistic: as we discussed above in our assessment of the relative costs of lithium ion batteries and a conventional gas turbine peaker, a marked decline in the price of batteries would be required to render such an investment economic.

Assuming a total installed cost of $475/kWh,[7] the total cost of 680 GWh of storage capacity would come to $323 billion. If this level of storage is ever to be reached, however, it will take many years to be deployed and will probably cost much less as costs decline over time. If the average capital cost of utility-scale storage declines to $200/kWh – approximately the price at which we estimate lithium ion batteries would become competitive with gas turbine peakers – the scale of the necessary investment falls to $136 billion. Assuming a 15-year life for lithium ion batteries is achieved (~5000 cycles), this would give rise to an annual replacement investment of some $9 billion.

We expect investment in electric energy storage to be strongest in those regions of the country where generation has been deregulated and utilities and retail electricity providers procure power for their customers in competitive wholesale markets. As discussed above, it is in markets such as these that the social benefits of storage – the suppression of peak power prices and the consequent reduction in power procurement costs – will be greatest.

To date, the largest deployments of storage capacity have occurred in the California ISO, which saw 264 MW installed over the period 2011-2016, and PJM, with 260 MW (see Exhibit 28), followed by ERCOT and MISO, with 48 and 24 MW, respectively. By far the largest planned deployment of energy storage is also in CAISO, with 483 MW of capacity additions announced and a total of 1,325 MW of new storage capacity required by 2020 by the California Public Utility Commission for California’s investor owned utilities. Other regions where we would expect to see regulated storage investment by T&D utilities are those where state regulators are most consumer friendly or have developed plans to upgrade the grid through the integration of renewable energy, distributed generation and energy storage capability. These regions include ISO-NE (New England), the NYISO (New York) and PJM (Mid-Atlantic and Midwest). We also expect storage deployment in ERCOT (Texas), MISO (Midwest) and SPP (Midwest and Plains states) for the purpose of integrating intermittent wind power onto the grid.

Exhibit 28: Historical (2011-2016) and Announced (2017-2020) Additions of Storage Capacity in MW, by Regional Transmission Organization

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org)

The advance of energy storage should create opportunities for growth in rate base and regulated earnings among transmission and distribution (T&D) utilities, as energy storage is deployed on the grid as an alternative to peaking capacity, spinning reserves and frequency regulation currently provided by conventional generators. T&D utilities would also face some risk, however, of reduced rate base growth if storage is used to defer new transmission and distribution investment. Among the potential beneficiaries of the build-out of storage could be T&D utilities in California (EIX, PCG), New England (ES, AGR) and New York (AGR, ED).[8]

Outside these regions, generation tends to remain subject to cost of service regulation and vertically integrated utilities predominate. The argument for deploying storage — to secure the social benefit of lower wholesale prices and thus power procurement costs – is thus less compelling in these areas. Moreover, while vertically integrated utilities, with transmission, distribution and generation, could see an increase in capital expenditures on storage in their T&D segment, this would come at the risk of cannibalizing investment in their generation segment.

Hybrid electric utilities, with unregulated generation assets and regulated T&D utilities, could see the benefits from increased regulated investment in storage more than offset by the impact of storage on the revenues and generation gross margins of their competitive generation fleets. However, at this time it is too early to predict which impact would be greater.

Finally, among power companies, independent power producers will be the most negatively affected. Almost all of their generation is concentrated in the unregulated regions where storage is likely to be deployed, either by regulated utilities or for renewable integration (see Exhibit 29). In competitive power markets, large scale deployment of storage will almost always act to suppress the power price, as for most applications the need for discharge is during peak demand periods. While the impact today is minimal, we believe that over the next 5-10 years, the increased deployment of storage will contribute to continued weak, and likely declining, power prices and generating margins, particularly in an environment of rising renewable generation and weak demand growth.

Exhibit 29: Breakdown of IPP Generation Fleets by ISO (excluding wind and solar)

Source: Company reports, SSR analysis

Energy Storage Suppliers

As we noted above, the potential market size for energy storage equipment is large, potentially worth a few hundred billion dollars. While we cover almost none of the equipment suppliers, we wanted to highlight the leading companies, both for existing installations and of announced deployments of energy storage. We only highlight these companies in order to provide investors with a list of firms involved in grid scale energy storage and not to provide an opinion on or analysis of any of these companies.

In Exhibit 30 we list the top 10 providers of storage for existing and announced energy storage projects, excluding pumped hydro. As can be seen there, BYD has been the largest supplier of energy storage technology due to a number of large battery installations for renewable integration. Siemens has also been a leader due to its compressed air storage system. Although most of the top suppliers of storage technology are large companies, Younicos, a German startup that acquired Xtreme Power, another storage startup, stands out. Younicos has also been the largest systems integrator for energy storage projects in the US, as well as one of the largest providers of power electronics.

However, looking at announced energy storage projects, only a few of the leading technology providers for past projects are in the top 10 for future projects. Among the largest suppliers of announced projects are AES, Stem and Tesla. All of the capacity to be added by AES and Stem is under contract with Southern California Edison as part of California’s 1,325 MW storage expansion program, as is half of Tesla’s planned capacity.

Exhibit 30: Top 10 Suppliers of Storage Technology for Existing US and Planned Energy Storage Systems (Excluding Pumped Hydro)

Source: U.S. Department of Energy, Energy Storage Database (www.energystorageexchange.org)

Among power electronics providers, there are a number of large public companies, as well as a few startups. Parker Hannifin, however, stands out as the market leader. It provided 30% of the power electronics for existing storage projects and is listed as the supplier to nearly 75% of the announced projects for which there is information on the power electronics provider.

There are few opportunities to invest in publicly traded companies that have a majority of their business in energy storage. Tesla’s exposure is growing due to its “Gigafactory,” the large battery factory it is currently constructing. Among publicly traded battery companies, only Leclanche, a small cap Swiss Li-ion battery manufacturer, appears to have a business focused on large scale battery systems, including grid scale energy storage and batteries for mass transportation and industrial machinery.

©2017, SSR LLC, 225 High Ridge Road, Stamford, CT 06905. All rights reserved. The information contained in this report has been obtained from sources believed to be reliable, and its accuracy and completeness is not guaranteed. No representation or warranty, express or implied, is made as to the fairness, accuracy, completeness or correctness of the information and opinions contained herein.  The views and other information provided are subject to change without notice.  This report is issued without regard to the specific investment objectives, financial situation or particular needs of any specific recipient and is not construed as a solicitation or an offer to buy or sell any securities or related financial instruments. Past performance is not necessarily a guide to future results.

  1. We are not identifying which technology will succeed, but are limiting our focus to technology that has already been successfully deployed at utility-scale in the United States. There are promising new pumped hydro and thermal storage technologies, including using storage tanks for a closed loop pumped hydro system and liquid air energy storage, but as of now they have not been deployed commercially in the U.S.
  2. 2016 data is likely not complete, due to delays between the time a project is completed and the time it is included in the database as operational. Thus 2016 capacity additions could well be higher.
  3. The annual revenue requirements used in this analysis for the various storage technologies reflect the lower end of the range of estimates provided in the December 2016 study prepared by Lazard and Enovation Partners, entitled Lazard’s Levelized Cost of Storage – Version 2.0 and available at https://www.lazard.com/perspective/levelized-cost-of-storage-analysis-20/
  4. See Zhi Zhou, Todd Levin, and Guenter Conzelmann, Argonne National Laboratory, Survey of U.S. Ancillary Services Markets, January 2016, pg. 34.
  5. PJM Interconnection is a regional transmission organization linking 185 GW of generation capacity and serving over 60 million people in a region stretching from northern Illinois in the west to New Jersey in the east, and from Ohio and Pennsylvania in the north to West Virginia and Virginia in the south (see www.pjm.com).
  6. It is important to note that storage assets that are built in order to capture these “public” benefits would probably be unable to sell their power into the wholesale markets due to concerns about distortion of wholesale market prices. Instead these storage assets would either have to be built for other purposes, such as transmission asset deferral or reserves, or would be part of the distribution system. Under current FERC policy, when built to provide other services to the RTO, the additional ability of storage providers to arbitrage the daily spread of power prices must be secondary to the primary services provided and the revenues would likely have to be credited back to the RTOs customers. When built as part of the distribution system, in order to avoid triggering FERC regulation, time arbitrage of power would probably need to be provided as part of a cost based service with no market based revenues. The comparable service would be a retail demand response program that is not bid into the wholesale capacity or power markets. The service would be a reduction of peak demand on the distribution grid both to reduce the congestion and maintenance needs of the system and to benefit retail consumers. Some regulators, for example, in Texas, may be more concerned about maintaining attractive wholesale power prices for generators and would be less likely to allow energy storage included in regulated rate base and receiving cost-of-service revenues to participate in the wholesale markets or reduce demand on the distribution grid. However, there are a number of state regulators who could be interested in these benefits, particularly in PJM and New England, and be willing to incorporate the costs of energy storage in distribution rate base and regulated operating expenses if the benefits are large enough.
  7. This is higher than prices quoted in the press for lithium-ion batteries because the non-battery cost of utility-scale grid storage can be equal to or greater than the cost of the battery itself, reflecting the cost of the storage module, fire suppression and cooling systems, power electronics, and other components.
  8. PJM would also be attractive, but there are no pure play PJM T&D utilities. FE would be the first if it exits its unregulated generation business.
Print Friendly