The emergence of cost effective battery storage

Energy storage will be key to overcoming the intermittency and variability of renewable energy sources. Here, we propose a metric for the cost of energy storage and for identifying optimally sized storage systems. The levelized cost of energy storage is the minimum price per kWh that a potential investor requires in order to break even over the entire lifetime of the storage facility. We forecast the dynamics of this cost metric in the context of lithium-ion batteries and demonstrate its usefulness in identifying an optimally sized battery charged by an incumbent solar PV system. Applying the model to residential solar customers in Germany, we find that behind-the-meter storage is economically viable because of the large difference between retail rates and current feed-in tariffs. In contrast, investment incentives for battery systems in California derive principally from a state-level subsidy program.


Supplementary
Supplementary  The past, current and projected unit power and energy component costs, v p and v e respectively, are estimated based on a blend of academic and industry publications. Our work uses the so-defined technology scope of electrical energy storage technologies, which clearly defines energy and power components of a (lithium-ion) battery storage system, as defined in prior work. 1 Specifically, this approach separates the battery pack from the ex-works system (inverter, container, battery management system) and the system (transport, installation, commissioning) on a technology and cost basis. These can be found in Table 2 and Figure   1 of the Supplementary Information accompanying Schmidt et al. (2017). Note that their approach to separating battery storage system into its components is consistent with other approaches in the literature. [2][3][4][5] In the present work, the energy components refer to the storage modules, which them-  Figure 1 of the main text.
Supplementary Table 1  This yields an annualized average rate of cost reduction of 9.3% over the entire period 3 . The second resource, denoted as (2) and (3) installations. 16 Applying this logic to residential storage installation costs as defined, it is estimated that the fixed cost of installation, the part that is independent of the size of the battery, is $300 (approximately e 260) per system in Germany.

Supplementary Note 4: Unit Cost of Power Components
The range of available estimates for the acquisition unit cost power components, v p , is limited. We therefore rely on project-level data provided by the California Public Utility  Table 1 in the main text, along with calculations for optimal battery energy and power components are based on the values provided in the Source Data file (i.e. those tabs with marked as Munich within the title).
The represented German insolation profile uses publicly available data based on satellite observations. 19,20 To determine seasonal generation profiles the national average installed capacity of 6 kW is used for calculations. 21 Summer and winter are defined as before. It is also presumed that installed solar systems are subject to losses including system losses (86%), inverter efficiency (96%) and temperature losses (97%). Results and Table 1   both are assumed to be subject to the same degradation rate of 1% (with a corresponding degradation factor of 99%). The deflator is calculated as: This specification reflects that the planning horizon in our model is Finally, the analysis for Germany does not consider income taxes, as there are no tax consequences for a homeowner due to savings related to battery storage investment.
Using k e = 4 and k p = 1 as an example, and Equation (1)

Generation and Other Parameters
Raw data for the demand profile in California is provided by the OpenEI dataset. 30,31 Input data is converted into annual representative demand profiles by location. Data is provided in 1-hour intervals over a one year time frame, which is converted into two representative daily demand profiles -summer and winter -using a simple, purpose-built spreadsheet based tool. The summer demand profile is based on data from June -September inclusive; winter is October -May inclusive. Representative demand profiles are expanded to 15-minute intervals through linear interpolation of hourly results. Results and Table 2 in the main text, along with calculations for optimal battery energy and power components are based on the values provided in the Source Data file (i.e. those tabs with the term LA within the title).
For California, raw data for insolation profiles -specifically plane of array irradiance (W per m 2 ) and cell temperature ( • C) -is used. 32 Installed capacity is determined to be that which provides generation equal in magnitude to the average annual demand. This presumption is based on the predominance of net-metering regulations that generally restrict installed system capacity to no greater than 100% of annual demand. The resulting installed capacity is 4.85 kW for California (Los Angeles). It is also assumed that installed solar systems are subject to multiple losses, including system losses (86%), inverter efficiency (96%) and temperature losses (97%). Results and Table 2 in the main text, along with calculations for optimal battery energy and power components are based on the values provided in the Source Data file (i.e. those tabs with the term LA within the title). A summary of input variables for the Los Angeles, California case are provided in Supplementary Table 5.
The deflator is as defined in previously within Supplementary Note 7. The applicable ITC and SGIP rules are described in the Methods section of main article. Using k e = 4 and k p = 1 as an example yields LCOEC of -$0.053 per kWh and a LCOPC of $0.2046 per kW.