The pathway to renewables and the role of liquid air19 February 2021
To accommodate a growing share of intermittent renewables in the energy mix, while also maintaining grid stability, will require a major expansion of energy storage capacity. Among the range of long duration energy storage options currently available, liquid air offers a number of advantages, including small footprint, gigawatt scalability, great siting flexibility, no cycling capacity degradation, ability to provide synchronous power, and low levelised cost of storage. By Robert Giglio and Stefan Nygard Sumitomo SHI FW
Nearly all forecasts predict that most of the electricity produced in the world will be from intermittent solar and wind resources by 2050. While wind and solar plants currently make up 18% of the world’s installed electric power capacity this number is expected to grow to nearly 70% by 2050 (Figure 1). Although some new solid fuel, hydro and nuclear plants will be built over this same period, an equal amount is expected to be retired, effectively keeping their net contribution to the world’s power capacity at about the same level as today.
Integrating significant amounts of wind and solar power capacity into the power grid will be essential in meeting environmental commitments around the world and in addressing the challenges related to global climate change. But renewable energy brings its own set of challenges.
First, there is the need to address the intermittency of renewables. Because the sun doesn’t always shine and the wind
doesn’t always blow, optimising the use of renewable energy requires more dynamic grid management and some degree of backup power.
Second, wind and solar plants are located where sun and wind resources are optimum, which is typically far from major load areas (cities, manufacturing hubs, industrial parks, commercial centres). This leads to higher T&D losses and increased congestion on electric grids.
Third, wind and solar plants produce power in the form of unidirectional direct current, which needs to be electrically synthesised into bidirectional alternating current before being connected to the grid. This synthesised AC – asynchronous power – lacks important characteristics needed to support grid stability and reliability.
In contrast, AC generators produce synchronous power, at a frequency determined by the speed at which the rotor spins. Electromechanical coupling links the spinning mechanical inertia of the generator rotor to the frequency of the produced electricity, enabling AC generators to passively provide mechanical inertia to the grid, independent of the MWh produced. This inertia provides valuable voltage and frequency stability during unexpected grid upsets.
If there is a disruption in the power balance on the grid, due to a short circuit or a mechanical failure at a power plant, this stored inertia acts instantly to slow the rate of change of grid frequency, giving valuable time for primary controls (fast reserves) to rebalance the grid.
Today, some 82% of the power on the grid is produced by traditional spinning AC generators, resulting in a high level of grid stability and reliability. As we look out to 2050, when the number of conventional generators will be substantially lower, it may become more difficult to maintain grid frequency and voltage during upset events.
Supporting the transition
Energy storage is poised to bridge this gap, supporting the widespread transition to renewable energy without threatening grid stability. The role and importance of energy storage is illustrated in Figure 2, which provides a simple example for a region with high solar installed capacity. A similar example could be illustrated for regions with high wind capacity.
During a sunny day, there is an overabundance of power from solar plants, well beyond that needed to meet load. The first job of an energy storage plant, therefore, is to store the excess solar energy over the 11-hour daytime period and release it later in the evening and in the early morning of the following day. This allows full utilisation of the power produced by the solar plant, helping to smooth out energy prices over the entire 24- hour period.
The energy storage technology needs to do more than smooth the energy supply curve, however; it also must contribute to maintaining and/or restoring grid stability. As noted above, solar and wind plants are asynchronous generators, so most of the power on the grid over the 24-hour period will lack the inertia provided by synchronous power.
If a grid fault were to occur, it would be difficult to control and likely would result in an extended upset, tripping additional plants and loads off the grid as frequency and voltage levels exceed predetermined safe levels. Energy storage can help restore the inertia during such situations.
As a scalable technology option, energy storage can support high levels of renewable energy integration across the entire electricity supply chain. As shown in Figure 3, energy storage can be applied to the supply chain at the point of power generation, at the transmission and distribution level, and at the point of electricity use.
At the generation level, in addition to time- shifting renewable power, energy storage plants can be used to improve the ramping response of less-responsive power plants, thereby balancing the minute-by-minute output fluctuations in from wind and solar plants. Energy storage plants can also provide both real and reactive power for fast acting voltage and frequency control. Moreover, energy storage plants integrated with spinning AC generators can provide inertia for grid stability and for black start services needed when re- energising a grid after a complete black-out.
At the transmission and distribution level, energy storage plants can relieve congestion and avoid curtailment of wind and solar plants by storing electricity during periods when the grid is congested and then releasing it at off- peak times. Energy storage plants can provide reactive power for local voltage control, and those with AC generators can add inertia to improve stability at points across the entire T&D network. In many cases, energy storage plants can be a faster, lower cost and more flexible alternative to traditional T&D solutions such as more cables, towers and substations.
At the point of use, for large power uses such as industrial and manufacturing facilities and commercial hubs, energy storage plants can improve power quality and reliability, provide uninterruptible back-up power, and reduce energy costs by shifting power consumption away from times with high grid congestion and corresponding high peak power rates. Energy storage plants can also enable facilities to install their own wind or solar plants to store and smooth the produced energy for their own use or to export to the grid at peak demand times. Finally, some energy storage technologies, notably liquid air energy storage, can use waste heat and cold produced at the site to further lower overall energy costs.
Storage technology choices
No single technology can provide all the needed services shown in Figure 3. A portfolio of energy storage technologies placed at optimal locations across the grid or behind-the-meter is needed to achieve the most reliable, flexible and economic solution. Optimum spots for energy storage plants will depend on grid design and the distribution of generating plants and loads unique to each grid.
Figure 4 shows the services, range of capacity and energy discharge duration for energy storage technologies available today, both from an economic and functional capability aspect.
Table 1 summarises the benefits that each of the larger scale technologies offer to the energy storage market.
Flywheels and supercapacitors
Flywheel and supercapacitor technologies are best suited for small scale (up to 1 MW) energy storage applications requiring instantaneous response. Size, cost and safety issues grow along with the scale of these technologies, making them unsuitable for larger scale applications.
Lithium-ion battery technologies have become the default choice for energy storage plants in the small to medium size range (up to 100 MW, under 4 hours of duration), driven mainly by their high siting flexibility, high efficiency, and growing supply chain from the electric vehicle sector. Beyond this scale, other options such as pumped hydro, compressed air energy storage, and liquid air energy storage are typically more competitive.
The main drawback of battery technologies is that they lose capacity each time they charge and discharge. This imposes a significant limitation on the revenues that a Li-ion plant can realise in peak shifting and energy arbitrage markets. It also drives up life-cycle costs since the battery modules, which make up about 85% of the total plant cost, need to be replaced every 8-12 years and to be oversized to account for capacity decay (4-8%/year). Over a 30-year plant life, these factors can amount to more than 250% of the plant’s initial cost.
Lithium ion battery technologies face other challenges as well:
- Grid support. Li-ion batteries don’t provide the stabilising quality of synchronous power to the grid and are unable to provide inertia and black start services.
- Thermal instability. Manufacturers and system integrators have not been able to solve the instability problem and instead are focusing on improving thermal monitoring of battery cells, improving cooling system redundancy, and the effectiveness of fire suppression systems for large systems.
- Environmental concerns. The need to dispose of or recycle spent modules is potentially the ultimate limiting factor for the growth of batteries in both the energy storage and electric vehicle sectors. At large scales, a gigaton stream of waste batteries may become an insolvable problem both economically and ecologically.
A revealing statistic from the US Department of Energy is that only 5% of all Li-ion batteries produced up to now have been recycled. Unless this number can be dramatically increased, future growth in battery manufacturing will be directly linked to future growth in the mining of nickel, cobalt, lithium, phosphorous and other metals. Metal mining can be destructive environmentally and ecologically, and these impacts could contribute to higher costs for Li-ion batteries.
At the largest scale for energy storage, pumped hydro has been the dominant choice, representing 93% of energy storage operating in the world today on a capacity (MW) basis and 99% on a power generation (MWh) basis. At this scale, response time is not as important (5 minutes or longer), since the common application for these large plants is shifting energy demand/supply by a period ranging from a few hours to days, providing capacity and operating reserve margin to the grid.
Even though pumped hydro has been the long-standing dominant choice for energy storage plants, future growth is limited due to siting concerns. Disruption of both wildlife and human habitat make permitting of new projects difficult and goes against the trend towards a more distributed system of energy storage across the grid to support high levels of wind and solar power.
Compressed air energy storage (CAES) employing underground caverns, is a proven energy storage technology option, but it too has limited locational flexibility. Only two plants have been built and operated worldwide since 1978.
Cavern evaluation and permitting can be expensive and time consuming, exposing developers to significant investment risk well before the viability of the project can be determined.
Environmentally, standard CAES plant configurations cannot be considered “clean” energy resources since the pressurised air is typically heated using fossil fuel combustion before expansion in a gas turbine. Newer designs are being developed that promise zero emissions by replacing the combustion process with heat stored from compressing the air, but these systems have not yet been fully demonstrated.
Liquid air energy storage (LAES) can be seen as a clean, compact version of CAES. Through liquefaction, an LAES plant reduces the volume of air by a factor of 700, as compared to a 40-70 reduction factor for CAES. This means an LAES plant can be very compact, storing air at low pressure cryogenic conditions in insulated tanks instead of as compressed air in large caverns. By using excess electricity from renewable energy plants to cool the air, LAES plants are emissions-free, unlike conventional CAES.
An LAES plant has the gigawatt scalability of pumped hydro, with the advantage of being flexibly deployable across the grid, with the only site requirement being a grid connection.
Liquid air plants also avoid the expensive, long and uncertain permitting and construction process associated with pumped hydro and CAES. Highview Power, the leading developer of liquid air technology likes to say “liquid air is pumped hydro in a suitcase”, due to its very compact footprint.
A liquid air plant can maintain its original energy storage capacity and round-trip efficiency over a 30-year plant life without
the need to replace major plant components. And at the end of life, recycling of LAES plant components and restoring or repurposing the site does not pose long-lasting environmental or site cost liabilities.
Figure 5 compares levelised cost of storage (LCOS) for LAES and Li-ion.
Compared to Li-ion battery plants, for the same storage capacity, an LAES plant can be built on a site 5-10 times smaller and can achieve a much lower levelised cost of storage (LCOS) due to its full discharge flexibility, long plant life and low O&M cost, as shown in Figure 5.
The need to oversize the battery modules to offset their capacity degradation (4-8%/year) significantly drives up the cost of the Li-ion plant, as reflected in the ‘debt P&I’ (principle & interest) segments of Figure 5.
The cost of replacing battery modules every 8-12 years (depending on the number of charge/discharge cycles) further drives up the LCOS of the Li-ion plant, as indicated by the ‘major maintenance’ segments of Figure 5.
Taken together, these two factors result in a 37% higher LCOS for the Li-ion plant as compared to the LAES plant on a monthly fixed cost basis ($/month per kW of plant capacity).
On a discharge energy basis ($ per MWh of energy discharged by the plant), the LCOS of the Li-ion plant is 69% more than that of the LAES plant. This is because the Li-ion plant cannot fully discharge its capacity each day, falling short of capturing the maximum revenue available from the fixed $50/MWh energy arbitrage assumed in the analysis. The Li-ion plant’s daily charge/discharge cycles were limited to a 50% depth of charge (200 MWh of total 400 MWh capacity) to maintain a battery module replacement frequency of ten years.
If the Li-ion plant is fully discharged each day to capture the full energy arbitrage revenue, the increase in battery module replacement cost would result in a higher LCOS. In contrast, the LAES plant fully discharges daily and captures the full energy arbitrage without a significant impact on plant maintenance cost or life.
Note that the cost of disposing of or recycling the spent Li-ion battery modules was not included in this analysis as part of the battery replacement cost. The LCOS for the Li-ion plant, therefore, could be significantly higher than that shown in Figure 5. The costs of large-scale Li-ion battery recycling and disposal are not well known today since these activities are at a very early stage.
The round trip (charge-store-discharge) efficiency of a Li-ion battery is assumed to be 85% and that of an LAES plant 55% (with the potential for significant increase, in particular when external sources of heat and cold are available).
Over a 30-year period, the much higher plant round trip efficiency (RTE) of the Li-ion plant coupled with a future decline in Li-ion battery cost (both accounted for in the analysis) are not great enough to offset the short-life and cycling limitations of the Li-ion battery.
Plant round-trip efficiency has a strong financial impact when charging (buying) and discharging (selling) power but seldom is a significant factor in energy storage project viability. The revenue that an energy storage plant can realise from energy price arbitrage (high-low energy price range) falls well short of that needed to support the capital investment and operating cost of an energy storage plant. To be economically feasible today, energy storage plants need revenue from non-energy ancillary services in addition to energy-related revenue.
Figure 6 shows revenue that an energy storage plant can earn from non-energy ancillary services in the Southern California Edison CAISO market today.
Even though the Li-ion plant has a much higher RTE than the LAES plant, it is unable to generate enough revenue from the market to provide a 12% return to investors, as shown by the empty, ‘missing money’, segment. In contrast, the LAES plant can achieve the 12% investor return due to its much lower revenue need and ability to fully discharge its energy each day. These factors have a much greater impact on project economics than round-trip efficiency.
Liquid air technology development
Highview Power has been developing its patented CRYOBattery LAES technology (Figures 7 and 8) since 2005. Based on experience from a 2.5 MWh pilot plant and from a 15 MWh commercial plant in the United Kingdom – which has been providing services to the UK grid for several years – Highview Power has been able to optimise the technical and economic viability of the CRYOBattery technology. It has recently broken ground on a 50MW/250MWh facility in the UK, to be constructed just outside Manchester in a joint venture with Carlton Power.
The round-trip efficiency of the Highview technology has been increased from 25% initially to 55% currently. The primary contributor to this increase is the improved capture and use of waste heat and cold from between the liquefaction and air reheating processes. The design is flexible enough to allow integration of nearby external waste heat and cold sources, such as peaker gas turbine hot exhaust gas, LNG regasification waste cold, and steel furnace hot gas. Further advances in waste heat capture could boost RTE to as much as 80%.
As shown in Figure 7, the LAES plant is charged by using electricity to power an air liquefier. The liquid air is then stored at cryogenic conditions in insulated tanks for hours or days until the energy is needed.
To recover the stored energy, the liquid air is pumped to high pressure, reheated and expanded through a turbine to power an AC generator.
Liquid air may be a newcomer to the energy storage market, but the processes and all the components are well proven within the power, LNG and industrial gas sectors (Figure 9).
Sumitomo Heavy Industries Ltd (SHI) and Sumitomo SHI FW (SFW) have invested in Highview Power to support the delivery of the CRYOBattery technology to markets around the world. As part of the partnership, SHI invested $46 million into HVP.
Highview Power will continue optimising the technology, and SHI and SFW will provide a range of project scope options, from design and engineering to full plant EPC delivery.