Embed Size (px)
Citation preview
/1/
Assessment of residential battery systems (RBS): profitability, perceived value
proposition, and potential business models
by Markus Graebig1, Georg Erdmann2, Stefan Röder3
Abstract
The boost of electricity generation from fluctuating renewable sources comes with the challenge of
how to cope with intermittence. Energy storage with batteries could be one solution. The German
government has launched a program to incentivize battery installations in conjunction with roof-top
PV (photovoltaics) for home owners. This segment of “residential battery systems” (RBS), with
capacities typically between 4 and 15 kWh, may grow rapidly over the next few years. This study
investigates the value proposition of RBS to customers. For the case of RBS in Germany, research has
been conducted in three independent parts: model-based analysis of the battery system’s expected
profitability, empirical study of customer expectations and perceived value proposition, and model-
based analysis of technology acceptance. Finally, potential business models around RBS have been
evaluated based on a small-scale field test together with two utilities. A brief taxonomy of potential
international markets for RBS has been also derived. The authors conclude that RBS will not become
cost-efficient in the foreseeable future. Under very favorable circumstances (rising electricity prices,
decreasing feed-in tariffs, strong economy of scale for batteries), RBS may eventually become cost-
efficient. However, from a marketing point of view, RBS appear to be a “win-win” for home owners –
as a potentially profitable way of optimizing residential electricity consumption, and as a tool which
provides the reassurance of being “self-sufficient”. Other markets outside Germany, e.g. off-grid
applications, may be even more promising for RBS.
Acknowledgement
Parts of the research for this paper at TU Berlin have been conducted within the research project
“SW-Agent” which investigates the role of municipal utilities in the German “Energiewende” (energy
turnaround). “SW-Agent” is kindly sponsored by the German Federal Ministry of Education and
Research (BMBF) under the SOEF (social-ecological research) program. We wish to thank our
partners from Stadtwerke Bühl who greatly supported the customer survey among their customers.
We are also indebted to Johannes Herden who was a great help in conducting the online survey.
1 Corresponding author Markus Graebig, Dipl.-Ing. M.Phil., Technische Universität Berlin | Department of Energy Systems,
Einsteinufer 25 (TA 8) | 10587 Berlin (Germany), phone +49 30 314-28163, fax +49 30 314-26908, e-mail [email protected]
2 Prof. Dr. Georg Erdmann, Technische Universität Berlin | Department of Energy Systems (department head), Einsteinufer
25 (TA 8) | 10587 Berlin (Germany), phone +49 30 314-22890, e-mail [email protected] 3 Jun.-Prof. Dr. Stefan Röder, Steinbeis-Hochschule Berlin, Mollerstr. 12 | 20148 Hamburg (Germany), phone +49 40 41 33
086 18, e-mail [email protected]
/2/
1 Introduction: motivation and status quo of residential battery systems (RBS) in Germany
Within the last decade, the share of renewable electricity generation in Germany has increased
tremendously. In 2013, more than 25 % of total electricity generation in Germany came from
renewable sources such as hydro, solar, wind, and biomass (BDEW, 2014) – compared to only 6.8 %
in 2000. Photovoltaics (PV) have seen the strongest growth rates, peaking in 2012. Within only one
year, 7.4 GW of new capacity was installed (BDEW, 2014). The current interest in residential battery
systems (RBS) is very much linked to the economics of PV in Germany. For a long time, the “PV
boom” had been triggered by generous feed-in tariffs – private PV owners, with up to 10 kWp
installed capacity on their roofs, were granted guaranteed feed-in tariffs which were significantly
higher than the average electricity tariff (figure 1) and subsidized though a “renewables surcharge”.
Hence, the business model of roof-top PV was not self-supply but much rather the subsidized
electricity feed-in – even though PV owners instantaneously rebought electricity from the grid at a
lower price. Since feed-in tariffs (only for new-built PV!) and electricity prices have taken strongly
opposing trends, the so-called “grid parity” – PV feed-in tariffs equaling prices for electricity supply –
was reached around 2012. (Note that this “grid parity” only reflects the microeconomic perspective
of the private decision-maker since one side of the equation is still hugely subsidized.) From now on,
self-supply has become the predominant business model for every new PV installation. PV owners
are incentivized to maximize their own consumption since one kilowatt-hour (kWh) from PV is
available for 17 €-Cents (January 2013) while one kilowatt-hour from the grid costs almost 29 €-
Cents. The spread of 12 €-Ct/kWh (and probably around 15 to 17 €-Ct/kWh in 2014) between
electricity supply and feed-in tariff is the economic benefit of a residential battery systems (RBS)
which allows the PV owner to store excess PV electricity during sunny hours for on-site consumption
during evening and night hours. Batteries may also be one approach to cope with intermittence of
electricity generation from fluctuating renewable sources.
Figure 1: “Grid parity”: Guaranteed feed-in tariff (10 kWp roof-top photovoltaics, new installation) vs. average electricity
tariff for households (typical German household, 3 people, 3,500 kWh/year). [Sources: IWR, BMU, BNetzA, BDEW]
In 2013, the German government launched a 25-million-Euro program to further incentivize RBS
installations in conjunction with roof-top PV for home owners (KfW, 2013). New installations of RBS,
together with small-scale PV, are eligible for a grant of up to 30 % of the RBS’s net price. This
program has probably raised interest in the new technology. Manufacturers as well as utilities are
trying to push RBS into the market4, and they are considering potential new business models around
the battery systems. Little research has been published on the RBS economics and their perceived
value proposition. In our study, we investigated the following questions:
4 Most recently, EWE – one of Germany’s top-5 utilities – launched their RBS “EQOO” (http://eqoo.ewe.de/) which is
advertised as a “carefree package”: 10-year guarantee on the RBS, five-year fixed-price contract for residual electricity supply, individual services, EUR 2,000 bonus (in addition to the government’s KfW program) for the first 100 customers, etc.
/3/
(1) What are the economics of an RBS in conjunction with PV? How much is the value of maximized
own consumption, and does it amortize the total cost of ownership of an RBS?
(2) What is the customers’ perceived value proposition of an RBS, and how high is their willingness
to pay? Which role for the purchasing decision do “soft factors” (self-sufficiency, environmental
concerns, etc.) play?
(3) The study concludes with a brief discussion and an outlook on potential international markets
which might be better suited for the sale of RBS.
Figure 2 gives an overview of RBS options which were available on the German market around mid-
2013 (note that there has been a very dynamic market development, i.e. the figure is unlikely to be
fully up-to-date). The systems are compared by their economic indicators “effective capacity” and
“price”. Also, the systems are distinguished by their underlying battery technology (lead- or lithium-
ion-based). For a more detailed comparison, other indicators (peak power for charging/discharging,
low-load operation, lifespan and number of charging/discharging cycles, energy efficiency, interfaces,
etc.) should be considered as well.
Figure 2: Market overview of selected residential battery systems (RBS) in Germany, 2013. “Pb” denotes lead-based
batteries, “Li” lithium-ion-based batteries. [Sources: providers’ websites, research by André Huschke; Focus magazine, 2013]
2 Functional principle of RBS in conjunction with photovoltaics (PV)
This study does not focus on technical details of residential battery systems (RBS), yet their
functional principal should be briefly discussed. RBS are usually designed as integrated power
management units, consisting of a battery stack (lithium-ion batteries in the advanced systems), an
DC/AC inverter module, power and control electronics, AC and DC interfaces, and internet
connection (figure 3).
Some RBS make separate DC/AC converter modules redundant and can be directly connected to the
DC output of the PV module. The RBS automatically monitors the balance between PV generation
and electricity demand in the house. First priority is always maximum coverage of the instantaneous
electricity demand. Any excess electricity from the on-site PV module will be used to charge the
battery stack in the RBS. Only when the battery has been fully charged, further excess electricity will
be fed into the grid to earn the fixed feed-in tariff. When the PV generation falls below the current
/4/
electricity demand in the house (e.g., during night hours or cloudy weather), the battery will be
discharged. The RBS is equipped with a DC/AC converter module and will route AC electricity (230 V,
50 Hz in Germany) to the on-site consumers. When the battery is fully discharged, the RBS’s power
management will draw power from the grid. Most RBS provide an additional uninterrupted power
supply (UPS) function. Also, RBS are usually equipped with telecommunication interfaces for data
exchange (remote maintenance of the system, monitoring the system status with mobile devices)
and connections to home automation systems. Further solutions could be included, such as a DC/DC
interface for electric vehicles (e-mobility).
Figure 3: Functional principle of an RBS in conjunction with PV, E-mobility is an optional add-on.
3 Methods: economic model, customer survey, structural equation model
3.1 Model-based estimation of RBS’s expected profitability
With a model-based approach, the expected profitability of residential battery systems (RBS) has
been analyzed. The battery helps home owners maximize the share of on-site consumption of
electricity from their own roof-top photovoltaics (PV). The battery operation has been modelled as
outlined in figure 3: excess electricity which is not needed instantaneously is stored for consumption
during evening and night hours, therefore saving money for external electricity supply from the grid.
For the further discussion, the following terms shall be defined:
EPV = EPV,on-site direct + Eexcess (eq. 1)
Eexcess = ERBS + Efeed-in (eq. 2)
with: EPV ....................... total annual electricity generation from on-site PV
EPV,on-site direct .. annual electricity from PV which is directly used for on-site self-supply
Eexcess ................. annual excess PV generation
ERBS ..................... annual excess PV generation which is stored in the RBS
Efeed-in ................ annual excess PV generation which is fed into the grid
Ehousehold = EPV,on-site direct + Eresidual (eq. 3)
Eresidual = · ERBS + Egrid (eq. 4)
with: Ehousehold ........... the household’s total annual electricity demand
/5/
Eresidual ............... annual residual electricity demand in addition to direct PV on-site use
........................... total efficiency of the RBS
Egrid ..................... annual electricity supply from the grid
Ctotal = Egrid · Pgrid – Efeed-in · Pfeed-in (eq. 5)
with: Ctotal .................... total annual cost of electricity for the household
Pgrid ..................... price of electricity supply from the grid (EUR per kWh)
Pfeed-in ................ feed-in tariff for excess PV generation fed into the grid (EUR per kWh)
Based on the given load and generation profiles, the model calculates the “E” variables in (eq. 1) to
(eq. 4), both for the PV stand-alone case and the PV & RBS case. In the PV stand-alone case (ERBS = 0),
the total annual cost of electricity will be:
Ctotal,PV = (Ehousehold – EPV,on-site direct) · Pgrid – (EPV – EPV,on-site direct) · Pfeed-in (eq. 5-a)
In the PV & RBS case, the total annual cost of electricity will be:
Ctotal,PV&RBS = (Ehousehold – EPV,on-site direct – · ERBS) · Pgrid
– (EPV – EPV,on-site direct – ERBS) · Pfeed-in (eq. 5-b)
With a Pgrid significantly higher than Pfeed-in, the following relation will always be fulfilled, i.e. the
annual electricity bill in the PV & RBS case will be lower than in the PV stand-alone case:
Ctotal,PV&RBS < Ctotal,PV Pgrid >> Pfeed-in (eq. 6)
The annual revenue R from an RBS is the difference between Ctotal,PV and Ctotal,PV&RBS:
R = Ctotal,PV – Ctotal,PV&RBS (eq. 7)
We define the pay-back PB of an RBS as the net present value (NPV) of all annual revenues Rt over
the T years lifespan of the RBS:
∑ (
( ) )
∑ (
( ) ( ))
(eq. 8)
with: T ........................... expected lifespan (years) of the RBS
Rt .......................... revenue from the RBS in year t
q ........................... discount rate (cost of capital)
We define the degree of self-supply as the percentage of the household’s overall electricity demand
which is actually covered by on-site PV generation. Obviously, the degree of self-supply will
increase with the availability of an RBS:
with < 100 % (eq. 9)
Given data availability and time constraints, it was not yet possible to design the model for analyses
of a broader variety of locations, system configurations, load and PV profiles. For the time being, the
existing model serves as one case study for a typical RBS-PV setup in Germany and is based on the
following data and assumptions:
(1) Time interval: the model is based on 15-minute time intervals for all 366 days in 2012 (2012 was
a leap year). The 15-minute interval is in line with most load and generation profiles which are
publicly available – even though it is certainly a substantial simplification. Finer time resolution
/6/
might be needed to better account for very sharp power gradients. Current research5 will shed
more light on the question of what the ideal time interval in models like ours should be.
(2) Load profile: the 15-minute/366-day load profile for the household in our model has been
represented by an H0 standard load profile (SLP) provided by EWE6, a major German grid
operator (DSO). We consider the SLP a suitable proxy for an average household, knowing that
SLPs have substantial limitations. Since an SLP always reflects a statistical average of various
households, it will always flatten sharp power peaks. Our current research aims at developing
high-resolution load profiles for different types of individual households which will then allow for
more realistic stress tests of the RBS.
(3) PV generation profile: Stadtwerke Bühl, our partnering municipal utility in southern Germany
(near Stuttgart), kindly provided their 15-minute/366-day reference profile for electricity
generation from roof-top PV. In the future, the model will be extended by a variety of load
profiles from other locations (a synthetic PV profile, based on geographic coordinates and
irradiation data, will be implemented5).
(4) Feed-in tariffs for new built PV modules were 0.17 €/kWh in January 2013 according to BNetzA7,
with an expected 20 % degression per year: Pfeed-in,2013 = 0.17 €/kWh, Pfeed-in,2014 = 0.136 €/kWh.
(5) Electricity prices for private households averaged 0.256 €/kWh in 20138, with an expected
average electricity price increase of 4.7 % per year: Pgrid,2013 = 0.256 €/kWh, Pgrid,2014 =
0.268 €/kWh.
(6) RBS: as a reference technology, we defined a Li-ion-based RBS with an effective capacity of
4.0 kWh. The following additional assumptions have been made: overall efficiency of one
charging/discharging cycle = 95 %; peak power for charging/discharging 2.0 kW; lifespan T =
20 years. In our reference scenario, this RBS is installed in a house with 7.0 kWp PV and an annual
electricity demand of Ehousehold = 5,000 kWh.
(7) Cost of capital is assumed to be q = 3.1 % per year.
Given the current state of the model, assumptions (1) to (3) are invariant. Assumptions (4) to (7) are
variable, as well as various technical parameters9. We are currently refining the model with regard to
assumptions (1) to (3), implementing a much broader range of load and generation profiles which
will allow for analyses of locations outside Germany10. The refined model will also include other
applications besides PV, such as the connection with a domestic micro CHP (combined head and
power) and the provision of primary control energy11.
3.2 Customer survey on purchasing criteria and willingness to pay
In the empirical part of the study, the customer perspective was investigated. 2,134 customers of a
“typical” municipal utility, Stadtwerke Bühl in southern Germany, were invited to take an online
5 Ongoing research by Anselm Eicke and Markus Graebig (TU Berlin).
6 Downloaded from http://www.ewe-netz.de/strom/1988.php on 2013-04-06.
7 BNetzA (Bundesnetzagentur, Federal Network Agency for Electricity, Gas, Telecommunications, Post and Railway),
http://www.bundesnetzagentur.de/cln_1911/DE/Sachgebiete/ElektrizitaetGas/ErneuerbareEnergienGesetz/VerguetungssaetzePVAnlagen/VerguetungssaetzePhotovoltaik_node.html#doc149586bodyText4, accessed on 2013-04-07
8 Stadtwerke Bühl tariffs
9 The following parameters are variable (values for the reference scenario are put in parentheses): the household’s total
annual electricity demand (Ehousehold = 5,000 kWh); PV peak capacity (7.0 kWp); degradation of PV peak capacity (0.5 %
p.a.); RBS effective capacity (4.0 kWh); overall efficiency of one RBS charging/discharging cycle ( = 95 %); peak power for charging/discharging (2.0 kW); RBS lifespan (N = 20 years) which is defined as the time when the effective storage capacity has degraded by 15 %.
10 Ongoing research by Richard Schwausch and Markus Graebig (TU Berlin).
11 Ongoing research by Heiko Birkholz, André Huschke and Markus Graebig (TU Berlin).
/7/
survey about RBS. 360 private customers (16.9 %) finished the entire survey. It covered questions on
customer expectations and concerns, perceived value proposition, willingness to pay, and
sociodemographic data. Also, potential drivers for the “intention of use” were measured in a sub-
study, based on a structural equation model (see chapter 3.3).
3.3 Structural equation modelling for the analysis of RBS’s perceived value proposition
The purpose of this sub-study is to gain an initial insight into both the customer expectations and the
perceived value proposition regarding RBS. Very little research has been published on RBS
acceptance. We suggested that not only financial considerations but also “soft factors”, particularly
the desire of becoming more independent from the electricity supplier (self-sufficiency), may play an
important role for the purchasing decision. This hypothesis was tested in the empirical study by
means of structural equation modelling. The Technology Acceptance Model (TAM) by Davis (1989) is
a leading theoretical foundation for measuring the customer´s intention to use a new technology.
We developed our research design based on Davis’ TAM structure in which a person’s attitude
towards using a technology is directly influenced by the technology’s perceived usefulness and its
perceived ease of use. The attitude towards using a technology drives the intention to use which
appears to be an appropriate indicator for the actual use. Based on expert interviews and our own
experience, the basic model was extended by the following three latent variables.
(1) Personal innovativeness: We derived this construct from Agarwal/Prasad (1998), originally
applied in the information technology (IT) context, and included measures that describe “the
willingness of an individual to try out any new (…) technology”.
(2) Subjective norm: In line with the theory of reasoned action (Fishbein/Aizen 1975),
Venkatesh/Davis (2000) revised the original TAM version and included subjective norm as an
additional construct. It is defined as a “person’s perception that most people who are important
to him think he should or should not perform the behavior in question” (Fishbein and Ajzen 1975,
p. 302). Customers may use an RBS, despite any individual cost-benefit consideration, just
because they believe that this is their contribution to the “Energiewende” (transition from fossil
fuels and nuclear power to renewable energy). Previous research is ambiguous about the effect
of subjective norms on customers’ intentions: Mathieson (1991) found no significant effect of
subjective norms on intention while Taylor and Todd (1995) did find a significant effect.
(3) Being self-sufficient: For various reasons, an increasing number of home owners seem to wish to
(a) actively contribute to an innovative, decentralized energy system, (b) cover their own
electricity needs and ensure for security of supply, and (c) become independent from their utility
company. Our construct “being self-sufficient” includes four items to measure these aspects.
Our theoretical model is shown in figure 4 where each hypothesis is marked with an arrow.
Figure 4: Theoretical research model for structural equation modelling.
/8/
The abovementioned constructs are represented by tested and proven measures. Some items have
been adapted and modified, especially in the “being self-sufficient” construct. “Subjective norm” and
“personal innovativeness” are represented as single measure items. All other latent variables are
measured using reflective multi-item scales. The empirical assessment is based on the answers of
319 private customers who finished this part of our survey (see paragraph 3.2).
For the data analysis, we applied the structural equation analysis (SEA). SEA is a methodology for
describing and exploring complex interrelationships between a large number of manifest and latent
variables. This involves the a priori postulated impact relationships between the variables which are
transferred into a linear equation system and the parameters are then assessed (Weiber/Mühlhaus
2010, 17). One weakness of the classic SEA is the need for extensive previous knowledge of the
interrelationships between the model variables and their respective characteristics (e.g. direction,
course of operation) when formulating the hypotheses (Dorniok 2012, 316 et seq.). This is quite
challenging in explorative studies like ours and should be kept in mind when assessing the results.
Also, potential non-linear relationships will be neglected in simplistic linear models (Dorniok 2012,
317).
The Universal Structural Equation modelling (USM) approach is better suited to cope with these
challenges (Buckler-Hennig-Thurau 2008). The USM was first implemented by Buckler (2001) in the
NEUSREL software. As an advanced, causal-analytical composite methodology, both USM and
NEUSREL are linked to the PLS approach (Buckler 2001). With the structural model assessment, a
universal regression based on artificial neural networks (ANN) is used instead of linear regression.
NEUSREL is based on Bayes’ neural networks (BNN) (Buckler 2001). Depending on the initial
information, ANNs can be structurally altered and are adaptive (Dorniok 2012, 318). Researchers
with limited previous knowledge are now able to discover and analyse a priori unknown
interrelationships, external variable effects and non-linear structures through an automated process
(Buckler 2001, 181 et seq.; Weiber/Mühlhaus 2010, 273). Dorniok (2012, 318) concludes with
reference to Buckler/Hennig-Thurau (2008), Buckler (2010) and Weiber/Mühlhaus (2010) that "up to
now no disadvantages of universal structural equation models compared to conventional structural
equation methods have been ascertained in the literature". In view of our research questions and the
disadvantages of classic SEA approaches, we decided to use USM/NEUSREL to test our hypotheses.
The assessment works in three phases: first, the expected relationships are put into a so-called a
priori matrix. Second, the construct values for all latent endogenous and exogenous variables are
determined (Buckler 2001, 189), using the PLS approach. All interdependencies between manifest
and latent variables are analyzed, using neuronal networks (Buckler 2001; Dorniok 2012, 319).
Finally, at the post-processing level, the results of the BNN are extracted (Weiber/Mühlhaus 2010,
270), including causal directions of action, external variable effects, and functional effect
relationships between the variables (Weiber/Mühlhaus 2010, 270). Linear path coefficients (or LPC
value) and non-linear path coefficients (or NLCP value) are determined. The overall explained
absolute deviation (OEAD) measures the extent of the linear and non-linear influence of an
independent variable on a dependent variable (Buckler 2001; Dorniok 2012, 319). NEUSREL also
assesses the quality of the measurement model and the overall model. The coefficient of
determination R² shows the percentage of an influenced variable that is explained by the causal
variables. The overall “goodness of fit” (GoF) is calculated based on Tenenhaus et al. (2005).
/9/
4 Results
4.1 Model-based results: degree of self-supply, pay-back, expected profitability
All results are calculated for an RBS with an effective capacity of 4.0 kWh. Figure 5 shows the degree
of self-supply (eq. 9) which can be achieved, given the installed photovoltaics capacity (ranging
from 3.0 to 10.0 kWp) and the annual electricity demand Ehousehold (ranging from 3,000 to
10,000 kWh/year). The “degree of self-supply” is the percentage of the household’s overall
electricity demand which is actually covered by on-site PV generation. Even though the total
electricity generation from on-site photovoltaics EPV may well exceed the household’s total electricity
demand Ehousehold, the degree of self-supply will be still significantly below 100 % because not all of
the electricity demand can be instantaneously covered by PV or taken from the battery. Even with a
very large battery in conjunction with a high PV capacity, 100 % self-supply is practically impossible,
given that there will be always some cloudy, dark and cold periods during the year. Therefore, within
reasonable limits of RBS and PV size, the household cannot become fully self-sufficient. Other backup
systems, typically the grid, are necessary. Figure 7 shows a comparison of degrees of self-supply in
the PV stand-alone case and in the PV & RBS case. In the reference case (annual electricity demand
Ehousehold = 5,000 kWh, PV capacity 7.0 kWp), the RBS increases the degree of self-supply from 42.0 %
to 64.2 %.
Figure 5: Degree of self-supply with a 4.0 kWheff RBS.
The expected pay-back PB (eq. 8) of a 4.0 kWheff RBS, depending on the installed PV capacity and the
annual electricity demand, is shown in figure 6. The figure illustrates that the PV and RBS
components should be carefully adjusted to the expected annual electricity demand Ehousehold. For a
given combination of RBS capacity and PV capacity, there is an optimal annual electricity demand
Ehousehold which yields a maximum pay-back. For a 4.0 kWheff RBS and 7.0 kWp PV (figure 7), this
optimum is around Ehousehold = 5,000 kWh with a pay-back of PB EUR 4,000 – calculated under the
optimistic assumption of T = 20 years lifespan. For T = 15 years, the pay-back is PB EUR 2,700.
Current purchasing prices for a 4.0 kWheff RBS (figure 2) are in the range of EUR 6,000 (lead-based
systems) to EUR 12,000 (Li-ion-based systems) – not included any maintenance, repair, or running
expenses of the system. Even under very optimistic assumptions, an RBS is currently not cost-
efficient. Considering a Li-ion-based systems, the pay-back is likely to range between 1/4 and 1/3 of
the purchasing price.
/10/
Figure 6: Pay-back of a 4.0 kWheff RBS over N = 20 years lifespan.
Figure 7: Degree of self-supply and 20-year pay-back of RBS for different electricity demands.
4.2 Empirical results: purchasing criteria and willingness to pay
Customers expressed an extremely positive attitude towards RBS (figure 8). Almost no interviewee
disliked the basic idea of an RBS, 78 % could envision using an RBS, and almost 14 % replied that they
were actually planning on purchasing an RBS. To put these answers into perspective, it should be
noted that the survey was conducted in a comparatively wealthy city with a high share of one-family
houses. Nevertheless, these answers support the hypothesis that in principle, there could be a
significant demand for RBS in Germany.
/11/
Figure 8: Intentions to use an RBS (n = 360).
Regarding the perceived value proposition of RBS (figure 9), financial considerations play the most
important role. As expected, customers wish to save money through higher self-supply, and they are
trying to protect themselves against future electricity price rises. Remarkably, “soft factors” are
similarly important as the financials, with “self-sufficiency” as the biggest goal.
Figure 9: Perceived value proposition of RBS: top-5 reasons for being interested in an RBS (n = 360).
Finally, we tried to understand the customers’ willingness to pay for RBS. The interviewees had the
choice between two modes of payment – purchasing with a one-time payment (either cash or loan)
or leasing with a monthly fee. 50 % preferred the purchasing option, one quarter preferred the
leasing option, and the others did not indicate any preference (figure 10). Arguably, customers in less
wealthy cities might have shown less interest in a one-time cash payment.
Figure 10: Willingness to pay for RBS.
/12/
The interviewees were then asked what they would consider a fair price (one-time purchasing price
or monthly leasing fee, respectively). Assuming a moderate interest rate and a 10-year leasing
contract, EUR 10 monthly leasing fee is equivalent to a EUR 1,000 one-time purchasing price. If
customers were purely financially driven, they should not have offered more than the expected pay-
back, i.e. no more than EUR 4,000 (see paragraph 4.1). This threshold is indicated with the dotted red
line in figure 10. A surprisingly large number of customers seem to be prepared to pay a price
premium beyond the EUR 4,000 threshold, especially those customers who prefer the leasing model.
Those customers offering more than EUR 10,000 (one-time purchasing) or EUR 100 (monthly leasing
fee) could be the early adopters – their willingness to pay is sufficiently high to afford the RBS
technology under current circumstances. Note that the interviewees did not know our financial
analysis outlined in paragraph 4.1.
4.3 Model-based results: perceived value proposition
Overall, the measurement model validation process yields reasonable results (table 1). However,
with regard to selected Cronbach Alpha values, future studies should pay special attention to the
construct’s items in order to enhance the validation results.
Latent Variable Cronbach´s Alpha (> 0.7)
Average explained variance (≥ 0.5)
Composite reliability (≥ 0.6)
Perceived usefulness 0.83 0.55 0.73
Perceived ease of use 0.75 0.52 0.68
Attitude towards use 0.79 0.63 0.81
Being self-sufficient 0.69 0.52 0.69
Intention to use 0.64 0.73 0.91 Table 1: Validation of measurement models
The structural model yields a reasonable GoF of 0.67, i.e. the empirically gathered data is adequately
reflected by the model estimate (Dorniok 2012, 323). The R2 values (perceived usefulness R2 = 0.59,
attitude towards use R2 = 0.70 and intention to use R2 = 0.65) show robust results. Figure 11 shows
the findings of the USM-/NEUSREL-based causal analysis. The path from “personal innovativeness” to
“attitude towards use” is not significant, and the path from “subjective norm” to “intention to use” is
below the cut-off value of 0.1 (Lohmöller, 1989). All other paths are above the cut-off value and
statistically significant. All paths from “being self-sufficient” are relevant and comparatively strong.
Figure 11: Tested research model with path coefficients (cut-off value: > 0.1) and t-values (cut-off value: > 1.98).
Finally, we analyzed the effect strength using the overall explained absolute deviation (OEAD). We
found a very strong link between attitude and intention to use (0.52) as well as a medium effect
strength between perceived usefulness and attitude towards using (0.27), being self-sufficient and
/13/
perceived usefulness (0.39) and attitude toward using (0.31). Hence, “being self-sufficient” is a
relevant factor which influences a person´s attitude towards RBS.
5 Conclusions: cost-efficiency, value proposition
As a result from the model-based approach, a typical German home owner (5,000 kWh annual
electricity demand, 7 kWp roof-top PV, RBS with 4 kWh effective capacity, current feed-in tariffs for
“green electricity” and expected price trajectory for electricity supply) in the very optimistic
reference scenario may yield a total benefit (NPV) between EUR 2,700 and 4,000 from the RBS over
its expected lifetime (15 or 20 years, respectively). The residential self-supply with electricity would
increase from 42 % (stand-alone PV) to 64 % (PV & RBS). Given current prices for state-of-the-art Li-
ion-based RBS (Focus, 2013), this equals a pay-back in the order of one third of the RBS’s purchasing
price – not considering any maintenance or running expenses. Even the cheapest available lead-
based RBS will not be amortized. In the medium term, pay-back of Li-ion-based RBS is unlikely to
exceed 50 %. However, under very favorable circumstances (rising electricity prices, decreasing feed-
in tariffs, strong economy of scale for batteries), RBS may eventually become cost-efficient.
In the online survey, the majority of customers expressed strong interest in RBS. Top-5 reasons for
customers to be interested in RBS are:
“saving money through higher PV self-supply” (80 % of all customers interested in RBS),
“avoiding future electricity price rises” (70 %),
“independent electricity supply for my home (self-sufficiency)” (69 %),
environmental concerns (66 %), and
the contribution to the “Energiewende” (61 %).
The structural equation model shows that a somewhat vague concept of “self-sufficiency” is the
most important monetary and nonmonetary driver for customers to be interested in RBS. While
“cutting electricity bills” is stated as the most important purchasing criteria, it appears that many
customers do not have a clear concept of the RBS economy (participants of the survey did not know
about the model-based results stated in paragraph 4.1). More than 50 % of the interviewees with a
general interest in RBS indicated a willingness to pay which exceeded the RBS’s realistic pay-back.
6 Discussion and outlook
Li-ion based residential battery systems (RBS) will not become cost-efficient for German users in the
foreseeable future. Eventually, the system might become cost-efficient, provided that electricity
prices continue to rise dynamically, feed-in tariffs for “green electricity” are cut, and RBS benefits
from a substantial economy of scale, especially with regard to Li-ion batteries. However, customers
indicate a substantial willingness to pay for RBS beyond its expected pay-back, which is mostly
justified with soft factors such as the desire to increase “self-sufficiency”.
In any case, Germany is probably not the ideal market for RBS, even though the technology is being
stimulated with government subsidies as one element of decentralized energy systems under the
“Energiewende” (energy turnaround) regime. Since the German electricity grid provides excellent
security of supply, other incentives for using an RBS – such as the need to cope with outages – are
lacking. On a global scale, other markets might be much more promising:
(1) Off-grid applications, e.g. rural areas without any permanent electricity supply. PV and RBS might
be the first and only option for electrification, not only powering villages but also infrastructure
such as cell phone towers.
/14/
(2) Unstable grids. Even in some countries of the western world, certain regions are prone to
blackouts. Customers in these regions should be interested in RBS as a means of uninterrupted
power supply (USP).
(3) Life-style markets. The German case stands as an example for a market where the use of RBS is
neither cost-efficient nor physically needed. However, some customers are able and willing to
pay for soft factors such as self-sufficiency. Similar to the German case, there might be other
wealthy regions where customers are responsive to such “life-style” products.
The general results of this study appear to be robust, even though they are based on a simplified
model which should be extended and refined. We are currently working on a more detailed model.
As outlined in paragraph 3.1, we will test the time interval (currently 15 minutes) for the load and
generation profiles and implement a broader range of high-resolution load and generation profiles.
The model will then allow for more realistic stress tests of the RBS and for the analysis of regions
outside Germany.
During a field test with two German utilities, it was suggested that RBS may offer additional value
beyond linking with PV systems. These are starting points for future studies: first, and very similar to
the roof-top PV, the RBS could be connected to a micro-CHP (combined heat and power) unit, a
residential heating unit with co-generation of electricity. The micro-CHP produces electricity as an
inexpensive side-product that can be stored for consumption as and when required. Second, a large
number of residential battery systems may be interconnected to a “battery cloud” which then acts
like one big battery, ready to sell system services such as balancing power to the TSO (transmission
system operator). Third, RBS could help utilities keep their “foot in the customer’s door” while
traditional electricity sales are shrinking. If the utility positions itself as a preferred provider of RBS,
this may serve as the anchor point for a variety of new business models or simply as a first
component of a modular “smart home”.
/15/
7 References
Agarwal, R./Prasad, J. (1998), A conceptual and operational definition of personal innovativeness in
the domain of information technology. Information Systems Research, 9 (2), 204 – 215.
AGEB (2000), German Energy Balance 2000, last accessed on 2014-03-27 from http://www.ag-
energiebilanzen.de/EN/downloads/energy-balance-2000-to-2011/energy-balance-2000-to-
2011.html.
BDEW (2014), Energie-Info und Foliensatz – Erneuerbare Energien und das EEG: Zahlen, Fakten,
Grafiken, BDEW (German Association of Energy and Water Industries) newsletter from 2014-02-25.
Buckler, F. (2001), NEUSREL – ein neuer Kausalanalyseansatz auf Basis neuronaler Netze als
Instrument der Marketingforschung, Göttingen.
Buckler, F./Hennig-Thurau, T. (2008), Identifying Hidden Structures in Marketing’s Structural Models
Through Universal Structure Modeling. An Explorative Bayesian Neural Network Complement to
LISREL and PLS, Marketing, Journal of Research and Management, 4 (2), pp. 47–66.
Davis, F. D. (1989), Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information
Technology, MIS Quarterly, 13 (3), pp. 319–340.
Dorniok, D. (2012), Die Einschätzung der Differenz von Wissen und Nichtwissen bei
Unternehmensberatern. Ergebnisse eines universellen Strukturgleichungsmodells, ZfbF, 64, 308–340.
Fishbein, M./Aizen, I. B. (1975), Attitude Intention & Behavior an Introduction theory and Research
Reading, Massachusetts.
Focus (2013-06-08), Marktüberblick Solarstrom-Speicher, last accessed on 2014-01-08 from
http://www.focus.de/ immobilien/energiesparen/solarenergie/tid-31894/strom-im-eigenen-keller-
speichern_aid_1017784.html.
KfW (2013), Erneuerbare Energien – Speicher, Kredit 275, last accessed on 2014-01-08 from
https://www.kfw.de/inlandsfoerderung/Privatpersonen/Bestandsimmobilie/F%C3%B6rderprodukte/
Erneuerbare-Energien-%E2%80%93-Speicher-(275)/#1.
Lohmöller, J. B. (1989), Latent Variable Path Modeling With Partial Least Squares, Heidelberg.
Mathieson, K. (1991), Predicting user intentions: Comparing the technology acceptance model with
the theory of planned behavior, Information Systems Research, 2 (3), 173–191.
Taylor, S./Todd, P. A. (1995), Understanding information technology usage: a test of competing
models, Information Systems Research, 6 (2), 144–176.
Weiber, R./Mühlhaus, D. (2010), Strukturgleichungsmodellierung. Eine anwendungsorientierte
Einführung in die Kausalanalyse mit Hilfe von AMOS, SmartPLS und SPSS, Berlin/Heidelberg.
Venkatesh, V./Davis, F. D. (2000), A Theoretical Extension of the Technology Acceptance Model: Four
Longitudinal Field Studies, Management Science, 46 (2), 186–204.
