Energy Storage Deployment and Innovation for the Clean Energy Transition

The University of California, Berkeley is one of CDTM’s partner programs for a semester abroad. Besides taking classes, a research stay is possible as well. Felix, Class of Spring 2014 joined the Renewable & Appropriate Energy Laboratory (RAEL) at UC Berkeley for this master’s thesis.

The research focus was to develop an advanced price forecasting model for lithium-ion batteries. Batteries are one of the key components for the Energiewende as they are required for electric vehicles as well as energy storage to integrate the increase share of intermittent sources of renewable energy into the grid. Comparing historical price forecasts with actual prices showed substantial deviations. However, reliable forecasts are requisite to develop well-functioning policies to promote the energy transition.

Figure 1: Boxplots of forecasted prices versus (actual) historical prices (black dots). Years refer to the date of the predicted price and not to the date when the forecast was published. Each year in the figure can include forecasts of various years. Prices are adjusted to 2015 US dollars.

In order to improve the accuracy of the forecasts, patent data as a proxy for innovation was added to the traditional model of economies of scale (one-factor approach). By adding innovation to the model, the two-factor approach significantly increased accuracy as both economies of scale, and innovation account to price decreases of high-tech products such as lithium-ion batteries.

Figure 2: Development of patents in the area of lithium-ion technology. The figure depicts the development of patents registered at US and the European patent offices, and through the Patent Cooperation Treaty. The last five years have seen a slowdown in patent activity in terms of patents being filed and approved. In particular, European Patent Office’s (EPO) and PCT patents show that a plateau has been reached.

Research Outcomes

The results show that storage prices are falling faster than expected and more rapidly than solar PV or wind technologies. The fall in prices is allowing new combinations of solar, wind, and energy storage to outcompete coal and natural gas plants on cost alone.

Together with Prof. Daniel Kammen and Noah Kittner, Felix found that R&D investments for energy storage projects have been remarkably effective in bringing the cost per kWh of a lithium-ion battery down from $10,000/kWh in the early 1990’s to a trajectory that could reach $100/kWh next year. The pace of innovation is staggering.

Figure 3: Overview of price development of LIB, solar and wind prices. The plot incorporates the fit of traditional one-factor models (economies of scale, experience curve) and our two-factor model to historical prices. LIB forecasts are based on projects for production output, and patent activity on the average of the past five years. Prices for wind display averages of data from Qiu and Anadon until 2007. From 2008 to 2019, prices are interpolated using the 2020 forecast of Lantz, Hand and Wiser. Prices for solar are taken from Zheng and Kammen. All prices are adjusted to 2015 US dollar.

Ordinarily, public research investment and private venture capital money undergo tough scrutiny before money can be spent on research and the results from years of work are not immediately visible. However, this study shows that long-term R&D spending played a critical factor in achieving cost reductions, and a recent lack of investment for basic and applied research may miss the $100/kWh target for cost effective renewable energy projects. Modest future research investment from public and private sectors could go a long way to unlock extremely low-cost, and low-carbon electricity from solar, wind, and storage.

Figure 4: Global corporate and VC investment in the energy storage sector. In comparison to the post-financial crisis years 2009, 2010 and 2011, investment levels dramatically decreased by 2014.

Figure 5: Overview of public R&D spending in the US and worldwide since 1980. Federal R&D spending has declined over the past four decades from about 1.2% to 0.8% of US GDP. In the same timeframe, federal R&D spending on energy-related topics plunged from over 0.3% to 0.013%. The dark green dots show a similar development for the share of energy-related R&D in total R&D spending. In the late seventies, energy R&D accounted for over 10% of total R&D (with over 50% of this allocated to nuclear energy) globally, whereas the international community allocated 3.9% of R&D funds to energy topics in 2013.

As Tesla moves to install a Gigafactory in Nevada and the largest lithium-ion storage facility in the world in southern Australia, new combinations of energy storage in terms of size, scale, and chemistry are emerging quicker than ever.

Tesla’s storage projects are not the only examples. Cities like Berlin have already embraced grid-scale storage. Berlin plans to install a 120 MW flow battery underground to support wind and solar efforts at integrated prices as low at 15 cents/kWh, in line with forecasts made in this paper. California is home to the first energy storage mandate on the grid, requiring utilities procure 1.325 GW of storage by 2020. These innovative policies showcase the range of storage options that may benefit clean energy, from small Powerwall batteries in the home to city-scale storage facilities providing back-up to utility-scale wind and solar farms.

There is an important co-evolution of battery developments for electric vehicle usage, grid-scale storage that supports solar and wind electricity, and other consumer applications for new electronics. To forecast future energy storage prices, Felix compiled a new dataset looking back to prices from the early 1990’s and development of new lithium-ion batteries through international patent databases. The team also looked at how storage co-evolved with solar and wind innovations. They found that for storage technologies, investment in applied research may actually be a more effective in $/kWh cost reduction than pure economies of scale mass production.

This past year (2017) the US reached its goal of $1/W SunShot solar power three years early. However, low-cost solar is usable during the day and experiences intermittency, which causes researchers to question the reliability of solar power. That’s why energy storage makes a big difference.

The study follows a string of research investigating the relationship between research funding and deployment of new technologies for solar panels and wind turbines. The study’s outcome highlights the need for more research in emerging storage technologies, as there is not a clear winner, and a diverse range of options may outlast lithium-ion batteries. There may be room for a number of different battery chemistries that all provide different services on an evolving grid, some providing voltage regulation and frequency control, and others serving long duration outages and providing back-up for buildings and communities.

The essence of the research project was published in Nature Energy and can be cited as:

Kittner, N., Lill, F. & Kammen, D. M. Energy storage deployment and innovation for the clean energy transition. Nature. Energy 2, 17125 (2017). Paper and supplemental data are available online at: https://rael.berkeley.edu/project/innovation-in-energy-storage/