Professional Documents
Culture Documents
Applied Energy
journal homepage: www.elsevier.com/locate/apenergy
h i g h l i g h t s
Novel analysis of unique building with integrated pumped hydro energy storage system.
Full parameterisation of pumped hydro energy storage in buildings.
Feasibility of pumped hydro energy storage in buildings is studied.
Conditions for a better competitiveness of this technology are discussed.
a r t i c l e i n f o a b s t r a c t
Article history: The growing use of variable energy sources is pushing the need for energy storage. With Pumped Hydro
Received 27 January 2016 Energy Storage (PHES) representing most of the world’s energy storage installed capacity and given its
Received in revised form 31 May 2016 maturity and simplicity, the question stands as to whether this technology could be used on a smaller
Accepted 14 July 2016
scale, namely in buildings. In this paper, the feasibility of such an installation is analysed by modelling
Available online 9 August 2016
each one of its components and applying it to several installation scenarios. Proposed and existing instal-
lations are also reviewed, including a first-time analysis of an installation in France, which is presumably
Keywords:
the only existing building with an integrated PHES system. It was found that the economies of scale that
Pumped Hydro Energy Storage (PHES)
Energy storage in buildings
render large PHES installations competitive are not present in small installations. This limitation, associ-
Distributed energy storage ated to other important disadvantages, such as the large volume required, seem to point out PHES as an
Levelised Cost of Energy (LCOE) ill-suited solution for energy storage in buildings, an important finding for building design and energy
policy. Nevertheless, if synergies with existing reservoirs could be found (for example for a building on
a riverside), costs could be significantly lowered. Further research on possible synergies with other build-
ing systems as well as a life-cycle assessment analysis are recommended.
Ó 2016 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.apenergy.2016.07.046
0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.
G. de Oliveira e Silva, P. Hendrick / Applied Energy 179 (2016) 1242–1250 1243
[3]: though water distribution systems can be used for demand- such as CO2 emissions mitigation as an important externality,
response [9] and show good complementarity with PV installations especially for tourism resorts.
[10], the question stands whether synergies could be found For existing PHES systems, the closest that could be found were
between such systems in a building. references to small-capacity installations supplying small
To account for the research gaps described before, proposed and communities. Among these, the only installation small enough to
existing installations found in the literature are reviewed and, for be comparable to an installation in a building is described by
the first time, an existing building with an integrated PHES system Manolakos et al. [17] for a remote village of 13 houses in a Greek
(the only one of its kind, to the authors’ knowledge) is analysed. All island. The installation comprises of a stand-alone system of
components of the PHES system are then parameterised, providing 18 kW photovoltaics coupled to a 100 A h lead-acid battery bank
important data to designers and researchers. A PHES installation in and a PHES system consisting of two 150 m3 water tanks with a
a building is modelled and the technical and economic feasibility of height difference of 100 m from which an energy capacity of about
such installation is discussed. 40 kW h was estimated. Two multistage centrifugal pumps are
Section 2 starts with a description of proposed and existing used: one for pumping while the other is reversed and used as a
installations. In Section 3 the PHES parameterisation and turbine. The pump rotation speed is adjusted according to the
modelling is presented and discussed, followed by the results for available excess power through a frequency regulation device. No
different scenarios and discussion in Section 4 and the conclusions information is provided regarding the economics of the installa-
in Section 5. tion, but the authors point out the additional use of the pumped
water for irrigation and household water supply, as well as the
accessible maintenance, as important advantages. Round-trip
2. Existing and proposed PHES systems in buildings efficiency of the PHES system is estimated at about 30% [17].
For PHES, two reservoirs at different heights are used. To store 2.1. The Goudemand residence
energy, water is pumped from the lower reservoir to the higher
reservoir. To later retrieve that energy, water is transferred from In this section, the Goudemand apartment building complex in
the higher reservoir to the lower reservoir through a turbine [6]. Arras (France) is analysed, where part of the common areas were
The use of small-scale PHES has been studied before but for power rendered grid-independent with the use of wind power, photo-
and energy capacities several times superior to those that could fit voltaics, battery energy storage and an integrated PHES system.
in a building [11,12]. Caralis et al. [13] and Ma et al. [7], for This is probably the only existing building with an integrated PHES
instance, provide estimates for the cost of PHES components in system, and the descriptions given here are based on a visit with
islands. Stenzel and Linssen [3] analysed the use of existing the contractor to the building, in April 2015, followed by e-mail
waterways for energy storage, while Kusakana [14] studied the exchanges. The owner of the building, Pas-de-Calais habitat,
feasibility of small-scale PHES coupled to hydrokinetic power for proved reluctant in providing any information other than the little
electricity supply to isolated areas. Nevertheless, the cost parame- already made public, limiting the analysis that could be done of
terisation done for high-capacity PHES (usually in the form of €/kW such an innovative installation.
or €/kW h cost estimates) is often the one used for smaller-scale The Goudemand residence is managed by Pas-de-Calais habitat,
projects [14], without taking into account the cost differences that a social operator responsible for planning, developing and main-
could result nor the fact that these cost estimates are very site- taining real estate and related infrastructures in the Pas-de-Calais
specific and not enough consistent data is publically available to region in Northeastern France. It is mostly active in rundown areas
extrapolate costs on such terms [15]. where it builds or restores existing buildings, creates new public
Regarding PHES in buildings, there are few references in the lit- services and lays out improved transportation facilities with the
erature and only references to projects could be found, not existing goal of revitalising problematic neighbourhoods. The social opera-
installations. Fonseca and Schlueter [16], for instance, proposed tor currently has about 100,000 tenants in 40,000 dwellings, and
and simulated such a system for an informal community of 3000 builds around 700 more per year according to the needs of its
people occupying an abandoned complex of five unfinished build- inhabitants [18,19]. Among these buildings is the Goudemand
ings in Caracas (Venezuela). The proposed system comprised of a residence, a three building complex built in 1975 with 240 apart-
total storage capacity of 85 kW h using water tanks distributed ments, housing approximately 700 people (Fig. 1).
through different floors due to load floor restrictions. The use of The Nord-Pas-de-Calais region, where the city of Arras is
a PHES system increased the calculated energy cost when com- located, has set an energy development path largely based on Jer-
pared to lead-acid batteries, but the authors defended that impor- emy Rifkin’s third industrial revolution concept presented in his
tant externalities, such as a secure provision of water and 2011 homonymously titled book [21]. By 2050, the region intends
technological simplicity, rendered it as the best choice. Stoppato to approximately halve its energy consumption and have its entire
et al. [1] simulated and optimised the installation of a hybrid energy needs covered by renewable energy sources [21]. This new
cogeneration system in an isolated tourism resort, with a capacity strategy, coupled with the rising strain of energy costs on family
of 170 people, in northern Italy, for the supply of electricity, heat budgets in France [22], has pushed the Pas-de-Calais habitat social
and water. The system consisted of an internal combustion engine operator to test strategies aimed at reducing energy consumption,
coupled to photovoltaics, a boiler, lead-acid batteries, thermal and increasing the share of renewable energy sources and reducing
water storage and a PHES system. Water is stored in concrete energy costs [19]. In 2012, a major renovation took place in the
reservoirs at a height difference of 50 m and a single, reversible Goudemand complex, with the goal of testing different technolo-
pump is used. Although the cost figures used seem optimistic, gies and setting a roadmap for further implementation in other
the optimum required 175 m3 water reservoirs coupled to a buildings. First, the energy consumption of the common areas
7 kW pump (giving an estimated storage capacity of 24 kW h and was strongly reduced with the installation of LED lighting coupled
an energy-to-power ratio of 3.5 h) together with 148 kW h of bat- to presence sensors and energy-efficient intercommunication
tery storage. The use of a PHES system resulted in a reduction of devices (intercoms), allowing for a reduction in the electricity con-
the discharge rate and depth-of-discharge of the batteries, sumption of common areas from 255 kW h to 47 kW h per year (an
effectively increasing their lifetime and decreasing the overall 80% reduction). In one of the buildings, nine 240 W photovoltaic
costs. The authors also discussed the public awareness to issues panels were installed together with two 500 W vertical-axis wind
1244 G. de Oliveira e Silva, P. Hendrick / Applied Energy 179 (2016) 1242–1250
Fig. 2. View of the systems installed on the roof of the easternmost building of the
Goudemand residence complex with the two original wind turbines [19].
Fig. 1. Southern view of the Goudemand residence complex before the installation
of the PHES system [20].
Element Capital cost (VAT incl.) There are different methods to evaluate the economic value of
Total €100,000 an energy project. The LCOE is a method that allows obtaining
Energy use reduction €27,000 the price per energy unit that balances out all the costs of the pro-
LED fixtures (60)a €5000 ject. It is obtained by adding up all the costs and dividing this sum
Intercoms (240)a €20,000 by all the energy provided. The values must be discounted to take
Presence sensors (30)a €2000
Energy production and storage €73,000
into account the time-varying value of money. The LCOE (€/kW h)
Photovoltaicsb €6000 is given by
Wind turbinesa €24,000 PN n
Batteriesa €5000 ðCn =ð1 þ dÞ Þ
PHES €38,000
LCOE ¼ PNn¼0 n
ð3Þ
Water tanks (lower reservoir)a €17,000 n¼0 ðEout =ð1 þ dÞ Þ
Pumpa €4000
where Cn is the cost in year n (€), Eout is the energy output in year n
Turbinea €2000
Pipinga €2000 (kW h), d is the real yearly discount rate (%) and N is the project
Roofing (upper reservoir)c €1000 duration (year) [24]. This formula can be simplified. It can be
Electronics and controla €7000 considered that the cost Cn consists only of the cost of electricity
Othera €5000
purchase at a price Pin (€/kW h) and the capital cost in year 0 C0
a
Estimate from contractors and equipment suppliers contacted. Includes all the (€). The total efficiency of the system gtotal and the yearly energy
necessary equipment and installation. output Eout (kW h) can also be considered to remain constant. In
b
A specific cost of €3000/kW was considered for an installation capacity of 2 kW.
c
this case, the LCOE (€/kW h) becomes
Total roofing cost was of €10,000 but the contractor estimated that the changes
required for storing water represented about €700. C0 Pin
LCOE ¼ þ ð4Þ
Eout ð1þdÞN 1 gtotal
dð1þdÞN
by Pas-de-Calais habitat is that they expect an immediate gain of
€50 per year per building by avoiding the grid connection fees For a real yearly discount rate d of 5% and a project duration N
for the common areas [19]. Taking into account that the grid of 25 years, the LCOE (€/kW h) becomes
connection will still be needed (as for example for the lifts), and C0 Pin
given the difference in order of magnitude of the given values, such LCOE þ ð5Þ
14 Eout gtotal
justification seems far-fetched. Despite the lack of detailed cost
information, an estimation of disaggregated cost components is This simplified formula intuitively shows the effect of two fac-
presented in Table 1. Values are separated into two main cate- tors on the LCOE for such a project. In the first factor, the energy
gories: energy use reduction (for the three buildings) and local output Eout must be maximised to dilute the capital cost of the
energy production and storage (installed in only one of the build- installation and reduce the LCOE. This means having the system
ings). Most values are obtained from discussions with contractors operate on a high capacity factor, i.e. fully charging/discharging
and equipment suppliers. the maximum number of days per year [25]. The dimensioning of
the system, not only in terms of energy capacity but also in terms
of power capacity will play a significant role on Eout and should be
3. PHES modelling optimised by simulating the system for different capacities using
real consumption data [4]. An over dimensioned system will be
For PHES, two reservoirs at different heights are used. To store underutilised while an under dimensioned system will not be able
energy, water is pumped from the lower to the higher reservoir. to make full use of the scale economies present in such projects,
Later, that energy can be retrieved by passing the water from the increasing the LCOE [7]. In the second part, the energy buy price
higher reservoir to the lower reservoir through a turbine. The Pin is adjusted according to the efficiency gtotal of the energy stor-
amount of energy stored E (J) is given by age system to reflect the cost of energy losses. This allows a fair
comparison of different energy storage technologies with varying
E ¼ qV g h ð1Þ efficiencies. For instance, for a given energy buy price Pin, an energy
storage system with an efficiency gtotal = 100% would have no
where q is the density of water (kg/m3), V is the volume of water at losses and, consequently, the cost of the energy supplied is still
the higher reservoir (m3), g is the gravitational acceleration (m/s2) Pin. On the other hand, for an energy storage system with an effi-
and h is the height difference between reservoirs (m) considering ciency gtotal = 50%, half of the energy input would be lost, effec-
that the water level variation on the reservoirs is negligible when tively increasing its cost to 2 Pin.
compared to their height difference. Considering a constant water
density of 1000 kg/m3 and a gravitational acceleration of 9.81 m/ 3.2. Water reservoir
s2, the amount of energy stored E (kW h) is approximately given by
In a building, to increase the stored energy, the reservoirs
E V h=367 ð2Þ should be placed at the maximum height difference. Unconven-
tional configurations such as underground lower reservoirs
where V is the volume of water at the higher reservoir (m3) and h is [26,27] have been suggested, but the risk-aversion to such novel
the height difference between the reservoirs (m). This gives PHES a configurations would probably favour an installation at the ground
volumetric energy density of 0.056 W h/l for a 20 m height. Other floor and at the roof. There is a limit on the maximum load appli-
energy storage technologies, such as Li-ion batteries for instance, cable on the roof with today’s regulation demanding a minimum
can offer much higher energy densities up to 600 W h/l [6]. Despite load capacity of 1500 kg/m2, which roughly corresponds to
having a low energy density, the low specific cost of large scale 1.5 m3/m2 of water [28]. If placed outside, reservoirs would be
PHES installations in favourable sites has allowed it to become subjected to radiation which could limit the lifetime of some
the most used energy storage technology worldwide. plastic water tanks. Low temperatures could also freeze the water,
1246 G. de Oliveira e Silva, P. Hendrick / Applied Energy 179 (2016) 1242–1250
rendering the system inoperable while the resulting density varia- Horizontal water tanks are available to a higher range but are also
tion of water freezing could also damage the reservoirs. Both reser- limited by the need of a base structure to compensate the base wall
voirs could be open, possibly resembling a decorative lake or a stresses. They are mainly adapted to vehicles. Vertical water tanks
varying height pool. In buildings next to rivers or canals, these are available up to a higher volume only surpassed by flexible
could be used as the lower reservoir. An open top reservoir could water tanks which require a larger area, but given the low thick-
also be partly refilled with rainwater but even considering a yearly ness of the plastic wall and production ease, are relatively inexpen-
rainfall of 4 m3/m2 (the average upper limit in Europe [29]), this sive. They are mainly used for agricultural applications.
would only represent 44 kW h for a 200 m2 roof area, 20 m height Several manufacturers were contacted and a database of plastic
building (218 W h/m2 year). Open reservoirs would also be sub- water tanks was built. All water tanks considered are made of a
jected to debris, requiring a filtering system while losses through single wall of plastic, are UV resistant and the price includes VAT
evaporation would also reduce the system efficiency. Given water’s and delivery to Belgium (Fig. 5).
high heat capacity, a thermal storage system could also be envi- The specific cost of water tanks is dependent on each type and
sioned or the water available could be used to increase the build- volume. Flexible water tanks present an average price of about
ing’s thermal mass. €70/m3 (including a plastic protection layer) while vertical water
Synergies could be found with the building’s existing water tanks stand at about €300/m3. The installation costs must be added
infrastructure, as suggested by Fonseca and Schlueter [16]. to these prices, including all the required accessories and labour
Nevertheless, the high volumes required for a meaningful storage which are very site-specific.
capacity, and, consequently, the large water flows, are in disagree-
ment with a typical potable water network. Also, the installation
3.3. Piping and pressure losses
would have to respect potable water norms, affecting the choice
of the required equipment (such as the water tanks and
Several types of pipes can be used. Several suppliers provided
turbine/pump). On the other hand, in the case of mandatory rain-
the price of grey, rigid PVC pipes with different diameters, with-
water collection (as currently happens in Brussels), the large
standing 4 bar of pressure (Fig. 6).
required volumes and laxer water use norms seem a better match
The average cost of the considered pipe for an area superior to
for PHES. Given the countless possibilities, the detailed analysis is
0.01 m2 is about €500/m2/m. Accessories for fitting the pipes and
focused only on plastic water tanks.
fixing them should increase by 2–3 times the cost plus the labour
There are four main types of plastic water tanks (Fig. 4). Rectan-
required for the installation, which is strongly site-dependent. A
gular water tanks are available in smaller dimensions, since their
linear installation in an open, easily accessible space would be
higher ratio of area to volume and uneven wall stress distribution
much cheaper than one running through the interior of a building,
requires a sturdier construction, which consequently renders them
eventually requiring the opening of passages for the pipe. The
costly. Since conventional buildings usually contain many corners,
the advantage of rectangular water tanks lies in their shape, as it
allows for an efficient placement without creating dead spaces.
Fig. 5. Capital cost per unit of volume for different types of water tanks. Fig. 6. Capital cost per unit of area for 1 m PVC pipe.
G. de Oliveira e Silva, P. Hendrick / Applied Energy 179 (2016) 1242–1250 1247
3.4. Pump/turbine
50 cm, yielding a total capacity of 100 m3 from which only 80 m3 competitive with grid use, it is important to compare it to other
will be used. Equivalent head losses are considered to be 3 m, energy storage technologies. For a fair comparison, the LCOE is
yielding a useful head of 17 m for a system capacity of 3.7 kW h. calculated under the same conditions as those used for PHES:
In the first case study, rectangular plastic tanks with a total real discount rate d of 5%, electricity price Pin of €0.12/kW h
capacity of 80 m3 are used in the basement as the lower reservoir. and a full daily charge/discharge cycle for an equivalent of about
In the second case, the lower reservoir is considered to already 9 months per year, yielding an energy output Eout of about
exist (a canal or a pond for example). The pump/turbine is 270 kW h per kW h of useful storage capacity. For lead-acid bat-
dimensioned in order to respect an energy-to-power ratio of teries with a total efficiency gtotal of about 80%, a capital cost C0
2.5 h since this should increase the system use (Eout) and lower of €600/kW h and a lifetime N of four years [6], a LCOE of €0.78/
the pump/turbine specific price, reducing the LCOE. From Fig. 7, kW h is obtained. For lithium batteries, considering a total effi-
for a 1.5 kW pump/turbine a specific cost of about €1.8/W and an ciency of 90%, a capital cost of €1000/kW h and a lifetime of
efficiency close to 60% are estimated. However, as explained 18 years [6], the LCOE is €0.45/kW h. As seen from the previous
before, a correct dimensioning of the pump/turbine would require sensitivity analysis, these LCOE values below €0.80/kW h are
an analysis of the local conditions of the installation. unreachable with PHES in buildings. Also, there are important
The LCOE obtained for PHES in buildings varies greatly accord- differences between a PHES system and a battery storage system
ing to the type of installation used. Nevertheless, the water tanks that may render the LCOE misleading. It would be much easier,
have the biggest impact on the capital cost of the system and for example, to increase the storage capacity of a battery storage
should be the main point addressed for cost reduction. This can system than that of a PHES system. Also, the cost of electrochem-
be achieved through synergies with other building systems, ical batteries continues to decrease, which should strongly affect
namely water supply [1] as it is already done for large scale PHES the obtained values in the near future, especially for newer che-
installations [2,3]. Another important parameter is the dimension- mistries, while PHES is dependent on mature technologies (like
ing of the pump/turbine since it affects not only the capital cost of hydraulic pumps) whose costs are not expected to strongly
the installation but also its efficiency (gtotal) and use (Eout), though, decrease [14,35,36]. All these externalities should be taken into
as previously mentioned, this dimensioning must be done specifi- account in the final decision.
cally for each case. Typical grid prices in Brussels, for instance, vary
between €0.18 and €0.23/kW h and none of the cases presents a
LCOE competitive to grid use. Grid parity of a PHES system coupled 5. Conclusion
to PV could eventually be achieved by reducing the LCOE by about
85%, which is implausible especially given the maturity of the tech- PHES in buildings is technically feasible. However, when associ-
nology involved [14,35,36]. In the future, provision of services to ated to local sources of energy in a building, it is still far from
the grid can offer better business cases for storage [4]. A sensitivity reaching grid parity. Furthermore, it is not economically competi-
analysis allows for a detailed study of the parameters affecting the tive when compared to other small-scale energy storage systems
LCOE for case study 2 (Fig. 9). that can be incorporated into a building, given the inexistent scale
As expected, the LCOE can be reduced through the decrease in the economies that render high capacity PHES installations competi-
capital cost C0, discount rate d and energy input price Pin or through tive: a fundamental finding for consideration in building design
the increase in efficiency gtotal, lifetime N or energy output Eout. In and energy policy. The inexistence of these scale economies can
any case, even varying these parameters by 30%, the LCOE is always be explained by the low energy capacity of the system (limited
above €1.20/kW h. The cumulated variation of several parameters by the maximum height of the installation and floor load restric-
could still lower this value but to not less than approximately €1/kW h. tions), the high specific cost and low efficiency of some of the
equipment used and the importance of fixed costs. Cost for PHES
4.2. Comparison with battery storage in buildings can be strongly decreased through synergies for the
reservoirs as it is done for large scale PHES through the use of geo-
There are several technologies available for energy storage. graphically suited sites (close-by canals or decorative ponds could
Although PHES in buildings coupled to PV is far from being be used as a lower reservoir for example). Also, the installation
should be large enough to dilute fixed costs and allow the use of
a higher capacity pump/turbine, with lower specific cost and
higher efficiency. The dimensioning of the pump/turbine, which
is very site-specific, would also need to be optimised since it affects
not only the capital cost but also its efficiency and use limits,
strongly affecting the LCOE. Nevertheless, given that PHES is a
mature technology, its cost is not expected to lower significantly,
while other energy storage technologies are expected to enjoy
strong cost reductions [35]. Several externalities are considered
to justify the use of PHES in buildings, namely the easier mainte-
nance due to its simplicity and the synergies with water storage
[16,17]. However, the existing strict potable water norms coupled
with the differences in the dimensioning of PHES and typical water
systems in buildings suggest that such synergies would not be able
to overcome the high LCOE of PHES in buildings, effectively render-
ing it as an ill-suited solution for energy storage in buildings.
The results here obtained can be the basis for a life-cycle assess-
ment of PHES in buildings, allowing for a fairer comparison to
other energy storage technologies. They can also help enhance
research on sustainable construction namely through the analysis
of synergies between PHES in buildings and other building
Fig. 9. Sensitivity analysis for the LCOE for case study 2. systems.
1250 G. de Oliveira e Silva, P. Hendrick / Applied Energy 179 (2016) 1242–1250