Storage is the ‘Holy Grail’ of the Energy Transition – or is it?

New research raises some crucial questions about the energetic performance of available energy storage options and their ability to support the transition to renewable energy

A global energy transition is underway, which is intended to cure the world from its fossil fuel addiction and to bring it on a more sustainable renewable energy path. Solar and wind energy sources, in particular, are growing fast and are forecasted to provide a rising share of the world’s energy supply in the coming decades. Almost 150 years after photovoltaic cells and wind turbines were invented, they still generate only a small fraction of the world’s primary energy, but they are now growing faster than any other energy source and their falling costs are making them increasingly competitive with fossil fuels. Many now predict that an era of clean, unlimited and cheap energy is possible, or even that it is looming.

However, the expected transition to renewables faces a number of technical and economic difficulties, which are by now well documented. In particular, the ‘clean energy’ obtained from solar panels and wind turbines has, as The Economist put it a few weeks ago, a ‘dirty secret’: it tends to disrupt electricity systems. Solar and wind are indeed ‘intermittent’, ‘variable’ or ‘fluctuating’ energy sources (meaning that they are not continuously available due to some factor outside direct human control), and hence cannot provide ‘dispatchable’ power generation (meaning that they cannot be turned on or off or adjust their power output according to an order). Therefore, they cannot be used to provide the ‘base load’ on a power grid (i.e. to generate the electrical power needed to satisfy the minimum demand), and they need to be ‘backed-up’ by other generating technologies to ensure that power generation can meet demand at any point in time, even when the sun doesn’t shine and the wind doesn’t blow. The profitability of these other generating technologies is however hampered by the need to modulate production to accommodate the fluctuating output of renewables, which sometimes produce power in excess of actual demand.

The more variable renewables are deployed, hence, and the more they push down the price of power from any generating source, depressing revenues and hampering the profitability of many generating technologies that need to keep on being used if the lights are to stay on – and that require increasing government subsidies as a result. The higher the penetration of renewables, the worse these problems get. In Germany, for example, the ambitious and subsidized push for renewables in recent years as part of the so-called Energiewende (‘Energy Transition’) has boosted solar and wind power generation, which represented 18.2% of the gross power generation mix in 2016 (5.9% for solar and 12.3% for wind). It has however caused a steep drop in wholesale power prices and contributed to push massive losses on the country’s largest utility companies, which still generate most of their power from coal and gas and find themselves at risk of falling into what some have called a ‘death spiral’.

Higher penetration levels of variable renewables, even as they push down the costs of electricity generation, thus potentially bring higher system costs, as the overall power system needs to be able to accommodate more intermittent energy sources while also ensuring that it can keep functioning despite the dwindling profits and investment capacities of traditional utilities. According to a recent report by the UK Energy Research Council (UKERC), the additional costs intermittent renewables impose on the electricity system can potentially rise very high and even spiral out of control in some cases. To avoid this, says UKERC, increased technical and economic ‘flexibility’ has to be built into the electricity system: if electricity systems and markets are redesigned and re-optimised to reflect the need for flexible supply and demand, using ‘smart’ new technology to smooth demand and move power around the network more efficiently, the system costs of intermittency can remain contained. ‘Smart grids’ that use digital communications technology to detect and react to local changes in usage, or ‘demand response’ mechanisms and devices that help adjust power demand with supply, are widely seen as instrumental to ensuring that electricity systems can incorporate larger proportions of variable renewable generation with only modest additional system costs.

A new energy revolution?

A full-scale transition to renewables, however, would need more than demand-side management. It would also require supply-side adaptations to ensure effective ‘load balancing’, i.e. the storage of excess electrical power from variable renewables during low demand periods for release as demand rises. Countries or regions that have already achieved a high penetration of variable renewables, such as Denmark and South Australia, have access to a sort of ‘virtual storage’ mechanism in the form of interconnectors to larger grids, through which they can export power when oversupplied and import it when renewables’ output is low. As wind and solar take on a greater role everywhere, however, physical storage mechanisms will be needed. This is why electricity storage is often presented as the ‘holy grail’ of the transition to renewables, ‘the’ solution that will enable renewables to finally take off and start displacing fossil fuels on a significant scale.

Many predict that the ‘age of storage’ is already dawning thanks to the progress of existing lithium-ion battery technologies that are getting better and cheaper and that are increasingly being deployed at utility and commercial level (with battery ‘megafactories or even ‘gigafactories’) or for home use (with home batteries such as the famous Tesla Powerwall). Moreover, media reports abound about the development of ‘breakthrough’ storage technologies that will remedy the limitations of lithium-ion batteries (boundaries of energy density and power density, risks of battery fire, possible resource constraints, etc.). In 2016 for instance, the Advanced Research Projects Agency-Energy (Arpa-E), a branch of the U.S. Department of Energy, said it was supporting the development of new designs for batteries, and new chemistries, that would enable the deployment of larger-scale energy storage systems and transform the U.S. electrical grid within five to ten years. Research projects in Europe and China, among others, are also reporting rapid progress towards next-generation battery technologies. In early 2017, John Goodenough, a British scientist who in 1980 identified the cathode material that enabled development of the rechargeable lithium-ion battery, published together with his team at the University of Texas a paper claiming a breakthrough in the development of a low-cost all-solid-state battery capable of solving many of the problems that are inherent in today’s batteries (cost, safety, energy density, rates of charge and discharge and cycle life). Breakthroughs are also regularly claimed in storage technologies alternative to batteries, such as supercapacitators or power-to-gas in combination with gas-to-power technology. These various advances, according to some, are about to usher in a new energy revolution that will reduce the need for base load provision and enable rapid decarbonisation of energy systems.

However, determining whether storage can indeed solve the problem of intermittency of wind or solar requires assessing the ‘energetic productivity’ of using energy storage in combination with variable renewables. Gathering energy resources from the environment – including from the wind and the sun’s rays – and making it available for various uses in society indeed requires ‘investing’ some amounts of energy, meaning that only part of the energy obtained is effectively available to do other things than extracting and producing energy. The ‘energy return on (energy) investment’ (abbreviated EROI or EROEI), i.e. the ratio between the total amount of energy delivered by an energy resource during its working lifetime and the amount of energy invested in obtaining it, is the best-known ratio used to measure the ‘energetic productivity’ of energy resources. Resources yielding a large amount of energy to society for each unit of energy expended can be considered as energetically productive, while resources that require large amounts of energy to be invested and yield only a limited amount of ‘net energy’ are considered less productive.

Most EROI analyses show that variable renewable energy sources have lower net energy ratios than fossil fuels, but many suggest that these ratios tend to rise as technological progress improves the performance of solar panels and wind turbines. According to a recent meta-assessment of 29 scientific studies performed by Rembrandt Koppelaar from Imperial College London, rapid technological improvements mean that the EROI of solar-photovoltaic (PV) energy, in particular, is probably already significantly higher than usually estimated.

How, however, does storage affect the EROI ratios of wind and solar power, and of the overall energy system that uses them? Does it increase or rather decrease them? Renowned energy researcher Robert Ayres recently stated that energy storage is “the key to increasing the EROI for intermittent systems”. The EROI for wind and solar PV, he wrote, “may increase radically in the future, as new energy storage technology is implemented”. Most EROI analyses, however, consider that the need for storage always increases the energy burden of electricity supply systems. Compared with a situation with no storage, it indeed requires additional energy to be ‘invested’ but does not, in itself, ‘return’ any additional energy. Therefore it always lowers the EROI of solar- and wind-generated electricity, even if maybe maybe less so in some cases than their ‘curtailement’ (i.e. a reduction in the output of generators from what they could potentially produce given available resources, due to transmission congestion or lack of transmission access, excess generation during low load periods, voltage, or interconnection issues).

This leads to two fundamental questions:

• Can we really afford storage to support a large-scale transition to renewables? Or in other words, can wind and solar PV deliver enough net energy to ‘pay’ the energetic cost of storage deployment?
• Can variable renewables with storage deliver enough net energy to really replace fossil fuels?

Incorporating EROI into electrical storage

There have been a number of attempts in recent years to evaluate the EROI of energy storage, but most of them have so far isolated storage devices to determine a device-specific EROI. This probably stretches the EROI concept too far, as this concept is in principle meant for energy sources and not for energy conversion or storage functions or devices. A long-run electricity transition, however, requires being able to compare the efficacy of new generation technologies in substituting previous ones for capacity, without degrading system reliability. And this requires taking into account the energy storage function, which is to some extent ‘embedded’ in ‘conventional’ generation technologies as it is built into the energy sources themselves (i.e. fossil fuels, and uranium for nuclear power). With conventional generation technologies, discrete storage devices are only needed to a limited extent, to increase operational flexibility. With variable renewables, on the other hand, the energy storage function is not available in the energy source itself and has to be built into discrete storage devices on a wide scale in order to boost availability. Therefore, a dynamic function for incorporating electrical storage into EROI calculations is needed in order to derive useful or predictive information concerning the potential of renewables to substitute conventional generation for capacity.

A new paper by energy researcher Graham Palmer from the University of Melbourne in Australia, published in the new peer-reviewed journal dedicated to biophysical economics, BioPhysical Economics and Resource Quality (BERQ), seeks to address this need. This new study proposes a framework for incorporating storage into net energy calculations by measuring the EROI of variable renewable energy (VRE) and storage as a subsystem, rather than trying to evaluate EROI for storage as a stand-alone unit.

The resulting EROI(VRE+storage) is equal to the ratio of the gross lifetime energy supplied by the VRE-storage subsystem and the sum of its lifetime ‘embodied’ energy (i.e. the energy consumed by all of the processes associated with its production and operation), relative to the displaced capacity of conventional generation (thermal generation or hydro). The benefit of applying this methodology is that it makes it possible to meaningfully compare the EROI of substitutes to conventional generation with the EROI of the generation capacity they replace, at any point along an energy transition pathway. Hence, it makes it possible to trade-off the energetic costs of storage with the value that storage provides, with different grid mixes, VRE penetration, and geographic regions.

The study only considers the role of electrical storage, and more particularly pumped hydro storage (PHS), the dominant form of electrical storage globally, and lithium-ion batteries, the fastest growing one. It leaves aside possible substitutes through sector coupling (e.g. power-to-heat, power-to-gas, power-to-desalination, vehicle-to-grid), and other forms of storage (e.g. thermal storage in concentrated solar thermal), though they could be assessed with the proposed framework at a later stage. For the purpose of the study, the framework is applied to a series of renewable simulations for the Texas ERCOT network. ERCOT (the Electric Reliability Council Of Texas) is one of nine independent system operators (ISO) in the US. It operates the electric grid and manages the deregulated market for 75% of the state of Texas, and supplies electric power to around 24 million customers in a region that has abundant wind and solar resources. The simulations provide a base case with zero VRE and storage, two intermediate steps, and a 100% VRE generation hypothesis.

The most important outcome of this study is that, in energy terms, rising VRE and storage exhibit marked diminishing returns: the first units of storage are the least energetically expensive, while as the penetration of VRE rises the displacement of conventional capacity becomes significantly more energetically expensive. In other words, the marginal EROI of the VRE-storage subsystem (i.e. the EROI of the additional VRE-storage required to provide an additional gigawatt of displaced conventional capacity) goes down as the penetration of VRE increases, not so much because the energetic productivity of wind turbines and solar panels themselves is affected but because the marginal embodied energy of the required additional storage capacity rises due to technical and resource constraints. This suggests that it may be energetically preferable in high VRE penetration scenarios to curtail renewable energy at times, rather than to store energy.

A second outcome is that the rate of these diminishing returns vary depending on the storage technology used. Application of the framework to the ERCOT case shows that the marginal embodied energy per unit of displaced capacity is around four times higher at near-100% VRE versus low penetration VRE for PHS, but around 41-fold for lithium-ion batteries. This implies that it is 4 to 41 times (PHS vs. lithium-ion) more energetically expensive to displace a gigawatt of conventional capacity at near-100% VRE than at low penetration VRE. In case storage is provided for in the from of lithium-ion batteries, the EROI of the VRE-storage subsystem at near-100% VRE penetration, which by hypothesis would then become the total aggregate EROI of the energy system, would probably fall too low for the energy system to be energetically viable. Society would then risk falling off the “net energy cliff”, beyond which prosperity is burdened by excessive direct and indirect requirements of the energy sector.

Key takeaways for policy makers

The results of Graham Palmer’s work now need to be confirmed by other studies, and his proposed framework for incorporating EROI into electrical storage refined. In particular, further simulations based on different regional grids and assumptions would be needed to capture the benefits of geographic diversity. Palmer’s findings however already raise three essential points that need to be taken into account by policy makers when designing energy transition strategies:

1. A shift from an electrical system based mostly on energy stocks (with built-in energy storage function) to one based mostly on natural flows (with the construction of storage devices required to ensure large-scale availability) will probably be constrained by the energetic demands of the VRE-storage subsystem. Or in other words, high penetration of VRE will require the large-scale deployment of storage solutions, but there might be biophysical limits to how much storage can be deployed if the energy system is to remain viable.
2. Lithium-ion batteries, which are the fastest growing form of electrical storage today and are increasingly being touted as capable of supporting the energy transition to renewables, could probably only usefully contribute a short-term role to buffering VRE. The energetic productivity/EROI of an energy system reliant on lithium-ion batteries (and other similar electro-chemical storage devices) would indeed rapidly fall below the minimum useful EROI for society. The energetic requirements of pumped hydro storage, on the other hand, are sufficiently low to enable a greater displacement of conventional generation capacity and penetration of VRE, but wide scale deployment is dependent upon regional topography and water availability.
3. Storage technologies that would enable a full displacement of conventional generation capacity and 100% penetration of VRE at the current system reliability level are, as of today, unavailable. New storage solutions may emerge as a result of current and future research activities, but in order to assess their potential it will be necessary evaluate their energetic performances within the VRE-storage subsystem, all along the energy transition pathway. Only if these performances are markedly superior to existing technologies will storage potentially constitute the ‘holy grail’ of the energy transition that many expect.