Content uploaded by Martin Bloemendal
Author content
All content in this area was uploaded by Martin Bloemendal on Nov 13, 2016
Content may be subject to copyright.
European Geothermal Congress 2016
Strasbourg, France, 19-24 Sept 2016
1
AFTER THE BOOM; EVALUATION OF DUTCH ATES-SYSTEMS FOR ENERGY EFFICIENCY
Martin Bloemendal1,2, Niels Hartog2,3
1 Delft University Of Technology, Delft
2 KWR watercycle research Institute, Nieuwegein
3 Utrecht University, Utrecht
j.m.bloemendal@tudelft.nl
Keywords: Aquifer Thermal Energy Storage,
Evaluation of ATES, Well design
ABSTRACT
Aquifer thermal energy storage (ATES) is a
technology to sustainably provide space heating and
cooling. Particularly in The Netherlands the number
of ATES systems has grown rapidly in the past
decade, often with the (re)development of urban areas.
To meet objectives for greenhouse gas emission
reduction the number of ATES systems is expected
and required to further rise in future both in The
Netherlands and elsewhere. To evaluate the lessons
learned and the role of practical aspects in the Dutch
development of ATES systems, in this study the
geohydrological conditions and well characteristics for
331 (~15% of total) Dutch ATES systems are
evaluated with respect to optimal well design for
maximal thermal energy recovery. The study shows
that well design of most (70%) ATES systems is
suboptimal. The well design criteria that have been
used thus far in practice, focus on allowing maximum
flow/capacity, disregarding the effect of groundwater
flow on efficiency and the effect of well design on
subsurface space use. Instead, well design should be
based on a more representative value for the storage
volume that takes into account . Based on monitoring
data and analysis of variations and uncertainties of the
actual storage volume, a guideline is defined to reflect
these in the storage volume used for design. Also a
guideline for well design is introduced that accounts
for both conduction and dispersion losses as well as
advection losses in case of high ambient groundwater
flow.
1. INTRODUCTION
Globally, there is a strong drive to meet energy
demand sustainably. Seasonal Aquifer Thermal
Energy Storage (ATES) systems provide sustainable
heating to and cooling to buildings. Although the
potential for using ATES systems depends both on
climatic and hydrogeological conditions, the
application of ATES has potential in many areas
worldwide (Bloemendal et al., 2015) and is therefore
expected to rise in the future. Although the potential
of ATES systems is largely not deployed in many
parts of the world, practical experience with ATES
systems has been developed in several European
countries and elsewhere (Blum et al., 2010; Eugster
and Sanner, 2007; Fry, 2009; Verbong et al., 2001).
Particularly in The Netherlands the number of ATES
systems has grown rapidly in the past decade, often
with the (re)development of urban areas. For an
optimal development of ATES systems, maximizing
the thermal recovery efficiency is crucial as well as
minimizing the required subsurface space
(Bloemendal et al., 2014; Willemsen, 2016). This
depends on hydrogeological conditions, design aspects
as well as operational aspects. Although, operational
aspects are difficult to predict in detail, typical
characteristics for ATES operation should be taken
into account in the design and installation phase of a
new ATES project As after installation it is relatively
costly and complex to change the ATES well design,
ATES wells should a-priori consider local
hydrogeological conditions and characteristic ATES
operational aspects to allow maximizing recovery
efficiency and minimizing subsurface space use. The
experience with the rapid development of ATES
systems so far, may support optimal further
development and use of ATES systems for sustainable
heating and cooling in the future, both in The
Netherlands and elsewhere.
2. MATERIALS AND METHODS
2.1 Theory of heat transport and storage
Thermal energy (cooling or heating capacity) in
infiltrated water in the subsurface is subject to several
processes which cause loss of the stored energy. The
Bloemendal and Hartog
2
processes are diffusion
1
, advection, conduction and
dispersion.
Energy losses due to mechanical dispersion and conduction
Water infiltrated by a an ATES well in an
homogeneous aquifer occupies a cylindrical shaped
volume in the aquifer. Rather than a sharp thermal
interface between the infiltrated water and ambient
groundwater, mechanical dispersion and heat
conduction spread the heat over the boundary of the
cold and warm water bodies around the ATES wells.
Losses due to mechanical dispersion and conduction
occur at the boundary of the stored body of thermal
energy. So to minimize these losses the surface area of
the circumference and the cap and bottom of the
thermal cylinder can be optimized by identifying an
appropriate filter screen according to storage volume
and local conditions. (Caljé, 2010; Gelhar et al., 1992)
Energy losses due to advection
Advection contributes to losses as when injected water
is displaced with the natural groundwater flow, it can
only partially be recovered. The thermal energy
within the injected water volume moves at
approximately half the speed of the water as a
consequence of thermal retardation. The higher
groundwater flow velocity relative to the thermal
radius, the more significant the losses to the ATES
system will be. To minimize these losses the thermal
radius can be optimized by identifying an appropriate
filter screen according to storage volume and
hydrogeological conditions.
Reducing losses
To recover as much of the stored thermal energy as
possible, the ratio between extracted and infiltrated
energy per well (Equation 1a) is a measure for the
thermal efficiency (ηth) of a well. The loss that occurs
depends on the geometric shape of the thermal body of
ground & groundwater, in this study simplified as a
cylinder. The size of the thermal cylinder depends on
the storage volume, filter screen length, water and
aquifer heat capacity (Figure 1 and Equation 1b). The
footprint of an ATES system is the surface area of the
top of the thermally influenced cylinder around the
well, described by the thermal radius (Rth); Equation 1.
()
out out out
th
in in in
E T V a
E T V
1
Diffusion losses are negligible and therefore not discussed in this
paper
R ( )
w in
th aq
cV b
cL
()
R 0,6 (d)
in
h
w
th h h
aq
V
Rc
nL
nc RR
c
Equation 1, Thermal efficiency (a), thermal radius (b)
the relation between thermal and hydraulic radius (c,d).
Rth=Thermal radius [m], V=Storage volume
groundwater [m3], ηth=Thermal efficiency [-],T
=Temperature [°K], cw =Specific heat capacity of water
4,2.106 [J/kg/K], caq=Specific heat capacity of saturated
porous medium 2,8.106 [J/kg/K], n=porosity [-],L=Filter
screen length [m]
Rh
Rth
Filter screen Length
Topview = footprint
Well
Thermal and hydrological
cylinder in subsurface
Figure 1: Schematic presentation of footprint and
subsurface space use of thermal and hydrological
cylinder
2.2 Data used
Permit data from Provinces
The data on the characteristics of ATES systems in
The Netherlands used in this study, was obtained from
provincial databases. Provinces are the local
authorities with the task of permitting and enforcing
ATES systems, they keep a database with
characteristics of the ATES systems for which they
issued a permit. Not all provinces register the same
characteristics in their databases, and out of the twelve
Dutch provinces only five (Gelderland, North-
Brabant, North-Holland, Utrecht, Drenthe) keep data
on the location, permitted yearly storage volume and
filter screen length, resulting in a total of 331 systems
suitable for evaluation.
Operational data
At an aggregated level, operational data of ATES
systems has been used in regional and national studies
and evaluations (CBS, 2005; SIKB, 2015; Willemsen,
2016) all showing that ATES systems yearly use 40-
Bloemendal and Hartog
3
60% of their permitted capacity. Local authorities
keep a record of the yearly pumped groundwater, but
cannot share that detailed information due to privacy
regulations.
Geohydrological Data
Local geohydrological conditions affect the applied
design ATES wells. For instance; when an aquifer has
a limited thickness it is not possible to install a longer
filter screen, or when the groundwater velocity is high,
it may be more beneficial to have shorter filter screen
lengths. Therefore the applied well design is evaluated
with respect to the local geohydrological situation; the
groundwater flow velocity, horizontal conductivity of
aquifer and the aquifer thickness. This data is not
available together with the characteristics of ATES
systems in the provincial databases and collected
separately from the Dutch Geologic databases (TNO,
2002a, b, c) based on the ATES locations. For a
geographically representative subset of 204 ATES
systems it was possible to retrieve local
hydrogeological data for all ATES systems.
For the following hydrogeological parameters data
was abstracted and processed for the aquifer
regionally targeted for ATES systems:
- Hydraulic conductivity. (TNO, 2002a, c)
Hydraulic conductivity values are for each
location provided as a range defined by a
minimum and maximum value. The average
of both extremes was used.
- Groundwater head gradient. (TNO, 2002b)
- Aquifer thickness. (TNO, 2002a, c)
The aquifer thickness is used to identify how
much filter screen length can reasonably be
expected to installed for each ATES system.
2.3 Numerical modeling tools
To realistically simulate subsurface groundwater flow
and heat transport, a geohydrological model was
developed using MODFLOW (USGS, 2000) and
MT3DMS (Zheng and Wang, 1999) (Hecht-Mendez et
al., 2010). MODFLOW and MT3DMS are finite-
difference element packages and well-established
models, widely used for the simulation of groundwater
flow and transport.
3. RESULTS AND DISCUSSION
3.1.2 Size and design of ATES systems
The permitted capacity of the ATES systems ranges
up to 5000,000 m3/year but most (~70%) are smaller
than 500.000 m3/year (Figure 2).
The regional differences in ATES system
characteristics are limited (Table 1), only Drenthe has
relatively small systems with limited variation. The
standard deviation of the other permit capacity varies
between 80% and 95% of average capacity. The
installed filter screen lengths are again similar again
with Drenthe a bit off, as a consequence of the
relatively small systems there. Noord-Holland shows a
bit larger installed filter screens, which may be caused
by the relatively large systems in combination with the
known thick aquifers which are present there.
Table 1, ATES system and geohydrological characteristics in provincial datasets selected for this study
Province
Number of
ATES
systems
Average
permit
capacity
St. deviation
of Per.
capacity
Average of
L installed
Average
Aqiufer
Thickness
Average
Hydraulic
conductivity
Average
Groundwater flow
unit
[-]
[m3/y]
[m3/y]
[m]
[m]
[m/d]
[m/y]
Drenthe
11
87.627
49.340
18
144
20
24
Gelderland
28
197.982
167.715
28
79
42
49
N-Brabant
172
210.754
199.244
28
60
28
23
N-Holland
95
282.946
228.893
43
144
33
6
Utrecht
25
349.620
296.958
33
121
21
10
Total
331
236.790
216.988
32
96
30
18
Bloemendal and Hartog
4
Figure 2, Frequency distribution of selected dataset
according to yearly storage volume
3.1.2 local conditions
ATES systems are spread over the whole of The
Netherlands, but are concentrated in urban areas.
Table 1shows the geohydrological characteristics of the
ATES systems location. Both hydraulic conductivity
and groundwater flow velocity vary little, only the
groundwater flow in Gelderland is higher as a
consequence of pushed/inclined aquifers. The
variation is larger for the aquifer thickness, caused by
local differences in aquifer thickness.
3.1.3 Practical considerations; consequences of dynamic
pumping regimes
To make a thorough assessment of the well design of
the ATES systems in the selected data it should be
evaluated based on the actual storage volume. The
storage volume of groundwater for each well depends
on the energy demand of the building over time,
which in turn depends on use, type and quality of
building and weather conditions. To anticipate on
climate changes, extreme seasons and allow future
growth in future the permitted capacities are generally
larger than the actual stored capacities during
operation .
Table 1, ATES system and geohydrological characteristics in provincial datasets selected for this study
Province
Number of
ATES
systems
Average
permit
capacity
St. deviation
of Per.
capacity
Average of
L installed
Average
Aqiufer
Thickness
Average
Hydraulic
conductivity
Average
Groundwater flow
unit
[-]
[m3/y]
[m3/y]
[m]
[m]
[m/d]
[m/y]
Drenthe
11
87.627
49.340
18
144
20
24
Gelderland
28
197.982
167.715
28
79
42
49
N-Brabant
172
210.754
199.244
28
60
28
23
N-Holland
95
282.946
228.893
43
144
33
6
Utrecht
25
349.620
296.958
33
121
21
10
Total
331
236.790
216.988
32
96
30
18
Several evaluations of ATES systems at an aggregated
level, show that ATES systems use 40-60% of their
permitted capacity (CBS, 2005; SIKB, 2015;
Willemsen, 2016). What further reduces the total
maximum stored volume during the year is that the
storage volume is not injected in once. Particularly in
spring and fall an ATES system may operate
alternating in heating and cooling mode. However
small, this also has a reducing effect on the maximum
stored volume during a year. In contrast, the permitted
stored volume may be incidentally exceeded due to
seasonal extremes which may cause temporal
imbalances. Demand for heating and cooling does not
balance every year, e.g. excess heat may accumulate
in warm wells during a couple of warm winters until a
very cold winter depletes the warm well. The effect of
these aspects is illustrated by different scenarios for
the cumulative build-up of injected volume for a
warm well of a fictitious ATES system and
monitoring data of several ATES systems;
1. All in once pattern. This energy demand
profile is often used to assess ATES-systems;
the total yearly storage volume is infiltrated
and extracted during a relatively short period,
with a period of rest in between.
2. Gradual pattern. The yearly storage volume is
infiltrated and extracted gradually over the
year, during spring and fall infiltration and
extraction alternate.
3. Weather dependent demand pattern based on
the storage volume variation expected based
on the monitored outside air temperature
(2020-2010) of the weather station of De Bilt
in The Netherlands (KNMI, 2013). The
energy demand pattern is derived from the
relative deviation of the daily temperature
from the average outside air temperature of
Bloemendal and Hartog
5
the evaluation period. So at the end of the
evaluation period there is energy balance, but
due to seasonal variations, imbalances occur
over the years.
The effect of these patterns on the storage volume
over time is shown in Figure 3, and shows that, for the
different demand patterns, the maximum storage
volume of weather dependent energy demand profile
uses 70% of the permit capacity. This is confirmed by
Willemsen (2016), who also looked at imbalances and
found that the standard deviation of imbalances over 5
year periods is around 30%.Thus, to make a fair
comparison, the well design will be evaluated based
on the expected maximum storage volume; which is
approximately 75% of the permitted capacity, or
around 150% of the expected yearly average storage
volume.
Figure 3, Volume in storage of well for different energy
demand patterns
3.2 Analytical evaluation of ATES
3.2.1 Loss of thermal energy due to dispersion and
conduction
Relation between storage volume and optimal filter
screen length
Since heat dispersion and conduction occur at the
boundary of the thermal cylinder (Figure 1),
minimizing its total surface area (A) should improve
the recovery efficiency. Figure 4 shows the relative
contribution of the circumference and cap and bottom
to the total surface area of the thermal cylinder in the
aquifer. This reveals that the surface area has a flat
minimum around L/Rth=2. Because dispersion
dominates around the circumference while conduction
dominates at the “cap & bottom” of the cylinder
(section 2), optimizing well design requires to
distinguish between the two to account for the reduced
conduction losses to confining layers after several
storage cycles (Doughty et al., 1982). Doughty et
al.(1982) showed that efficiency increases with the
number of storage cycle to an equilibrium, they found
that the optimal ratio between filter length and thermal
radius (L/Rth) has a flat optimum around 1,5. The
optimal L/Rth-ratio is lower because over multiple
cycles, the conduction losses to “cap & bottom”
reduces. Applying this rule to larger storage volumes
increases the overall efficiency because the surface
area of the “thermal cylinder” relative to the storage
volume decreases with increasing storage volume.
Figure 4: Relation between surface area of cap & bottom
and circumference area of thermal cylinder for different
filter screen lengths (1-25m and a storage volume of 500
m3
Substituting the expression for the thermal radius (Rth)
in the optimal relation of L/Rth=1,5 gives the optimal
filter screen length (L) as a function of storage volume
(V), Equation 2 (a-c). Substituting the expression for
thermal radius in the formula for the surface area of
the thermal cylinder (Figure 4), and equating its
derivative to zero results in a similar expression for
optimal filter screen length according to Doughty et al.
Equation 2 (d-f) shows that the solution for the filter
screen length results in the same third root of the
storage volume, only with the constant 1,23 instead of
1 for (Doughty’s) optimal solution. From the relation
between surface area of circumference and cap &
bottom (Figure 4) can be seen that this effect implies
that shorter filter screens are more beneficial than
simply minimizing the thermal cylinders’ surface area.
3
3
3
2.25
( )
( )
2.25 1 .02 ( )
w
a
doughty
w
doughty a
cV
La
c
L c V b
c
cc
c
2
3
2 2 ( )
2
1(e)
( )
ww
aa
ww
aa
analytical
c V c V
A L d
c L c L
c V c V
Ac L c L
L c V f
Bloemendal and Hartog
6
2 1 .23 ( )
w
analytical
w
a
a
c
cg
c
cc
Equation 2, (a-c) Optimal filter screen length as a
function of storage volume (Doughty et al., 1982). (d-f)
filter screen length for minimizing the surface area of the
thermal cylinder. L=Filter screen length [m], V=Storage
volume groundwater [m3], cw =Specific heat capacity of
water 4,2.106 [J/kg/K], caq=Specific heat capacity of
saturated porous medium 2,8.106 [J/kg/K]
The Dutch guidelines for design of ATES wells do not
give a clear guideline or formula to determine the
filter screen length with respect to storage volume
(NVOE, 2006). In the guidelines determination of
filter screen length is mainly based on maximum
desired flow rate. The relation between filter screen
length, storage volume and thermal losses is briefly
discussed and concluded with the advice to choose a
filter screen length which creates a relatively “flat
cylinder”. From this guideline we conclude that Dutch
ATES systems are supposed to have a filter screen
length equal or shorter than the optimal filter screen
length with respect to the expression for filter screen
length given in Equation 2 (a-c). Although no formula
is given, this approach corresponds with Doughty’s
rule.
Evaluation of the installed filter screen lengths
Equation 2 (l) is now used to assess the installed filter
screen lengths of the ATES systems in the dataset.
From the results of the analysis in Table 2 can be seen
that on average filter screen lengths are designed too
short, the average value for L/Rth of the installed
systems is 74% of what they should be according to
Equation 2; 1,1 instead of 1,5. When the optimal and
installed filter screen lengths are plotted with respect
to storage capacity (Figure 5 ) it becomes clear that
most systems (~76%) have a too short filter screen. As
is shown in Figure 4, also Doughty found a flat
optimum for L/Rth-value, thus it can also be accepted
when the L/Rth-value is between 1 and 4 (Doughty et
al., 1982). In that case 53% of the systems has a too
short filter screen and three systems have a too long
filter screen, Figure 5.
Effect of geohydrological conditions on well design
The design and practical aspects discussed above were
used to compare the applied filter screen length with
thickness available in the aquifer. After analysis of the
local aquifer thickness it appears that 40% of the
ATES systems with a too short filter screen have
space available to make it longer, of which 82% have
space available to meet the optimal length. So in total
about one third of the ATES systems has a too short
filter screen but with enough space available to make
it longer.
Figure 5, L/Rth relative to storage volume
The aquifer thickness found in the data was corrected
to have sufficient clearance between filter screen and
confining aquitards and to take account for variations
in aquifer thickness, considering that the source data
only gives a rough indication of aquifer thickness.
Legal boundaries were also included, for instance in
Noord-Brabant it is not allowed to install ATES
systems deeper than 80 m below surface level, so any
aquifer available below 80 is disregarded in the
evaluation. As a result of this correction the space
available for filter screen length used for evaluation
may in some cases be underestimated.
Table 2, ATES system filters screen length in practice compared to optimal design, L optimal is Doughty
Capacity / well
L installed
L optimal
L installed / L optimal
L/Rth
[m3/y]
[m]
[m]
[m]
[-]
10th percentile
16.000
12
25
0,4
0,4
average
80.000
33
43
0,74
1,1
90th percentile
200.000
56
59
1,26
2,1
Bloemendal and Hartog
7
3.2.2 The effect of ambient groundwater flow on recovery
efficiency
Relation between groundwater flow and energy losses
Additional to the thermal losses that occur through
conduction and dispersion, ambient groundwater flow
may increase thermal energy losses significantly, as it
displaces the stored volume before recovery. Under
these conditions, a body of water in a flowing aquifer
can only be partly extracted by the well which was
used for infiltrating that water body (Bear and Jacobs,
1965). The overlapping surface area of the thermal
footprints before and after the volume of thermal
energy has moved with the groundwater flow is
equivalent to the storage efficiency relative to
groundwater flow, Figure 6.
Figure 6, calculating the overlapping surface area of 2
identical cylinders.
To obtain maximum efficiency the overlapping area of
the thermal footprint must be maximized, in areas with
high groundwater flow velocity this can be achieved
by increasing the thermal radius; thus reducing the
filter screen length. This simple approach is used to
assess well design of ATES systems in areas with
ambient groundwater flow. So for any groundwater
flow velocity it is required to identify a minimal
thermal radius to obtain a sufficient recovery
efficiency during operation of an ATES system in that
specific aquifer. Goniometric rules allow to express
thermal radius as a function of groundwater flow
velocity, substituting a desired minimum efficiency
condition results in a design condition dependent on
flow velocity (u) and the thermal radius; Equation 3.
The velocity of the thermal front (u* ) is QO in Figure
6. Equation 3 shows that the relation between
groundwater flow and efficiency only depends on
thermal radius, so for any storage volume and filter
screen length there is a single Rth/u-value indicating
the expected losses through groundwater flow.
Therefore the Rth/u-value is used to evaluate the ATES
systems design.
Equation 3 shows that for each desired efficiency (ηth)
there is a minimum value for the ratio of Rth and u.
This relation is plotted in Figure 7 and can be used to
identify minimum desired thermal radius (i.e.
maximum desired filter screen length for a given
storage volume) at a location with a given
groundwater flow velocity.
int
2
int
2 2 2
*
*
22
**
*
2
1
2 cos 24
21
cos 24
overlap th footpr
footpr th
overlap th th
th
th th
th th
AA
AR
u
A R a u R u
R
uu
a R u
RR
Equation 3, Calculating the overlapping surface area of 2
cylinders. A =Surface area [-],ηth=Thermal efficiency [-],
Rth=Thermal radius [m], u*=Velocity of the thermal
front [m/y]
To verify this approach numerical MODFLOW
simulations were used to reproduce the relation of
thermal radius, groundwater flow velocity and
efficiency. For different sizes of systems with
different groundwater flow velocities the recovery
efficiency was calculated. The numerical simulation
results are also plotted in Figure 7, which shows that
the analytical relation over-estimates the efficiency
significantly. This makes sense because the numerical
model also includes losses due to dispersion and
conduction which are not taken into account in the
analytical approach to evaluate losses due to
groundwater flow. To take account for this effect the
numerical results were normalized to obtain the
efficiency loss as a consequence of the groundwater
velocity only. This was done by relating the efficiency
of the simulation with groundwater flow to the
associated simulation without groundwater flow (e.g.
normalized result for u= 5 m/y; Ŋ5=η0/η5). The
normalized efficiencies show a better resemblance
with the analytical relation; RMSE=0,14. The
difference is caused by dynamical aspects; the
analytical solution evaluates the advection of an
completely filled storage well, while in practice and in
the numerical model the losses already start to occur at
first injection of (warm/cold) water.
The relations in Figure 7 show that for high flow
velocity and/or small thermal radius (Rth/u < 2) losses
through background groundwater flow are dominant.
While at low velocity and or large thermal radius
(Rth/u >4) conduction and dispersion is dominant;
efficiency is constant. In between (2<Rth/u<4) both are
important.
Bloemendal and Hartog
8
Figure 7, Relation between thermal radius and
groundwater flow velocity for different desired
efficiencies
Evaluation of the installed filter screen lengths
For each of the ATES systems in the data the Rth/u–
value was determined, the relation given in Figure 7
and Equation 3 are used to indicate lines of expected
thermal efficiency, Figure 8. From this can be seen that
many systems (44%) have an expected efficiency
lower than 80% (Rth/u<2,3) only taking into account
losses due to ambient groundwater flow. In addition,
depending on the optimality of L/Rth the actual
efficiency is further reduced (Figure 7).
Figure 8, Rth/u-values for ATES systems in the dataset
with thresholds for different efficiencies
Losses incurred by ambient groundwater flow are in
addition to those by conduction and dispersion. There
is no guideline (NVOE, 2006) or method available to
take account for these losses in well design. Defining a
minimum acceptable efficiency allows to find an
appropriate (maximum) filter screen length, Equation 3.
From simulations and monitoring data we know that
thermal efficiency from ATES well ranges from 70-
90% (Figure 7, Willemsen, 2016, Sommer, 2015,
Caljé, 2010, NVOE, 2006).These efficiencies also
include losses due to groundwater flow velocity,
therefore an acceptable thermal efficiency due to
groundwater flow is assumed to be in the same order
of magnitude; 80%. To identify the minimum thermal
radius a 20% loss due to groundwater flow velocity is
used as threshold.
The analysis shows that 66%of the systems has an
appropriate filter screen length. Table 3 shows the
systems characteristics and groundwater flow velocity
of the systems which do and do not meet the desired
size of the thermal radius. This shows that
groundwater flow velocities around 29 m/y start to
cause problems and mostly smaller systems suffer
from losses due to ambient groundwater flow. The
results from the analysis confirm what logically
follows from Figure 7, smaller thermal radii (i.e.
smaller ATES systems) are most vulnerable for
significant losses as a consequence from ambient
groundwater flow.
3.2.3 Combined results loss of thermal energy by, advection,
conduction and dispersion
For a particular storage volume, increasing the thermal
radius (decreasing filter screen length) will lead to
reduced losses by ambient flow. However, at Rth/u >4
the benefit of increasing Rth decreases and care should
be taken not to decrease L/R below 1-2 (Figure 4) as
this would result in a strong decrease in the loss by
conduction (and dispersion). Assessing the ATES
systems to both relations, 6 types of systems can be
identified as shown in Table 4 and Figure 9. From this
can be seen that 27% of the systems (types C, E and F)
have a too long filter screen mainly because of high
groundwater flow velocity, in Figure 9 can be seen that
these are mainly small systems. Of the 24% of the
systems which need a longer filter screen (type B),
68% has space available to do so (17% in total). Type
D systems meet both requirements. The most
challenging systems are the type A systems, which
should have a longer filter screen to minimize
conduction and dispersion losses, while the
groundwater flow velocity would require a shorter
filter screen.
Table 3, Results of analysis of filter screen length with
respect to groundwater flow velocity
average u
average V
[m/y]
[m3/y]
η < 80%
5
154.307
η > 80%
29
62.617
From this can be seen that depending on the size of
ATES system and groundwater flow velocity,
efficiency of ATES wells is dominated either by
conduction and dispersion, advective transport due to
ambient groundwater flow or a combination of the
two. To get grip on the thresholds and transition area
from one rule to another, both rules can be combined.
Bloemendal and Hartog
9
Figure 10 shows the relations for optimal filter screen
length for Doughty and groundwater flow velocity
combined and plotted together with the ATES systems
characteristics associated with the required L/Rth -value
for different ambient groundwater flow velocities. The
obtained relations are a weighted average of the two
rules with the ambient groundwater flow velocity as
weighing factor; because the higher the groundwater
flow velocity, the higher its impact on the desired
L/Rth-value.
Table 4, Results of combined requirements for optimal
filter screen length
Doughty
condition
Ground-
water flow
condition
η < 80%
η > 80%
Condition
L is too long
L is ok
L/Rth < 1
L is too
short
A = 18%
L is..
#unknown#
B = 24%
L is too short
1 < L/Rth <
4
L is ok
C = 26%
L is too long
D = 29%
L = ok
L/Rth > 4
L is too
long
E = 0%
L is too long
F = 1%
L is too long
Figure 9, Different types of ATES systems with respect
to requirements for optimal filter screen length
3.2.4 Conclusions from analytical analysis
In this analysis, analytical solutions were used and
combined to assess the ATES well design, therefore
the ATES storage was simplified as a cylinder during
operation. Rules and relations available in literature
were used and where necessary new rules were
derived. In at least 52% of the cases the filter screen
length is not optimal, for another 18% it is not clear.
For only 29% of the assessed ATES system it is safe
to assume that based on the expected storage volume
the installed filter screen is optimal. Incorporating the
(thermodynamic) processes which occur in the aquifer
in more detail, may give a better insight in the aspects
influencing thermal efficiency and how to deal with
the type A, B,C and F system types. This however is
future research.
Figure 10, Optimal L/Rth for Doughty and groundwater
flow combined
3.5. Discussion
In practice however, more complex hydrological and
thermodynamic processes occur which are not taken
into account in this analytical analysis. To verify the
validity of the (combined) analytical rules and the
conclusions drawn from them in this work, it is
required to incorporate the operational aspects like
uncertainty and variations in seasons and assess the
effect of well design on efficiency accordingly.
Therefore next steps in this research is to carry out a
Monte-Carlo analysis using multiple secnario’s to
simulate ATES efficiency with a numerical
geohydrological model.
Storage volume as a cylinder
In this research the thermal energy storage in the
subsurface was simplified as a thermal cylinder.
However in practice ATES wells may have a more
ellipsoidal shaped footprint instead of circular as a
consequence of ambient groundwater flow and/or
neighboring systems. The effect of this on the method
followed in this research is limited because the losses
due to groundwater flow are taken into account.
Also the effect of neighboring wells is limited because
of the reciprocal principle; in one season a
neighboring ATES well may cause increased losses,
but the next season it will push back the lost water
because it will then also load its well again with
thermal energy. This is under the assumption that both
systems have a more or less energy balance, which is a
Dutch legislative requirement for ATES systems.
ATES systems in aquifer with high groundwater flow
velocity
Where groundwater velocity is high, filter screen
lengths should be shorter to limit losses due to
advection. This simultaneously results in larger
thermal radii. It might be a better strategy to identify
how two warm and two cold wells can be used to
optimize the overall efficiency by infiltrating in an
upstream and extracting from a downstream well
Bloemendal and Hartog
10
(Groot, 2014). In many areas however this might not
be possible or desirable because of other buildings in
the close vicinity who also have or want to install an
ATES system. In such areas it makes sense to use
planning and organizational procedures to optimize
ATES well positions, to prevent negative interaction,
which is likely to result in the fact that filter screens
can be longer or a vertical separation of filter screens
over the depth of the aquifer.
ATES systems in densely built areas
Planning of subsurface space occurs based on the
thermal footprint (Figure 1) of an ATES well projected
at surface level. As a consequence, the subsurface
space use depends on the presence of neighboring
systems, storage volume (operational aspect) and filter
screen length (design aspect). In areas with many
ATES systems mutual interaction is likely to occur,
and an integrated approach like was proposed by
Bloemendal et al. (2014) or masterplans (Arcadis et
al., 2011; Li, 2014) are a more appropriate way to
organize optimal use of the subsurface. However, also
in these situations the recommendations from this
study will be useful; in such areas it is very wise to
make optimal use of the available aquifer thickness
and reduce thermal radii, which requires longer filter
screens.
Because of accumulation of ATES systems in urban
areas, scarcity of space in urban aquifers is occurring
(Bloemendal et al., 2014; Hoekstra et al., 2015).
Recently it was shown that scarcity of space for ATES
is expected to occur in the near future in many cities in
Asia and the United States, among others (Bloemendal
et al., 2015). Several studies showed that there is a
tradeoff between individual well efficiency and
overall greenhouse gas emission savings in an area
densely populated with ATES systems (Jaxa-Rozen et
al., 2015; Li, 2014; Sommer, 2015). With that in mind,
the question arises to what extent subsurface space
designated for ATES systems is optimally taken
advantage of, in current ATES planning and operation
practice, which is focused on protecting existing
permitted ATES systems (Schultz van Haegen, 2013).
The facts that ATES systems use only 75% of the
permitted volumes, the safety margin around the wells
and that in many cases the filter screens are shorter
than optimal as shown in this study, results in a
underutilization of roughly 30% of the available
subsurface space in urban areas with many ATES
systems. These observations indicate that subsurface
space use (i.e. the projected thermal footprint at
surface level) of ATES systems is much bigger that
would be the case when taking into account optimal
storage volume and filter screen length.
Practical aspects
- Longer filter screens may have another advantage
worth mentioning; a longer filter screen results in
lower groundwater flow velocity around the well.
This reduces the mobilization of particles in the
aquifer and with that risk of clogging of the well
(Beek, 2010; NVOE, 2006). This will have a
positive effect on the life time and maintenance
requirement for the wells.
- In tube wells often the infiltrated and extracted
water is not evenly distributed over the filter
screen (Houben, 2006; Korom, 2003; Sommer,
2015). When relying on longer filter screens for
efficiency or planning purposes, practical
operation must ensure even distribution over the
filter screen otherwise this effect may frustrate the
ATES well efficiency and/or subsurface space use.
Ensuring evenly employment of the filter screen
may be ensured by using multi partially
penetrating screens, special filter screens or pump
inflow tubes at different depths.
6. CONCLUSION
Well design
Thus far, well design is mainly based on the tradeoff
between maximum capacity (flow rate) of the wells
and drilling cost. This research provided simple
methods to design wells taking into the wells thermal
efficiency. This research also showed that with respect
to the recovery efficiency, the optimal filter screen
length has a flat optimum which limits this problem.
Because of the flat optimum and the effect of short
filter screens on the thermal footprint of the ATES
system it is recommended to make them longer in
areas with low groundwater flow velocity and/or
scarcity of space in the aquifer.
Ambient groundwater flow
In case of high groundwater flow velocity it is
recommended to apply the analytical rule for well
design derived in this paper. Groundwater flow is
summarily taken into account while designing ATES
wells in the Netherlands because design guidelines
were not available. This lack on insight is reflected in
the ATES well design of installed systems in areas
with groundwater flow, in most cases the well design
is not optimal.
Storage volume
The estimated storage volume which is used as a basis
for well design is of crucial importance. Variation in
yearly storage volume, groundwater flow, conduction
and dispersion need to be taken into account. Climate
data and aggregated monitoring data indicate that a
proper yearly storage volume to base well design on,
Bloemendal and Hartog
11
is 75% of the permitted value. Using the permitted
volumes as a basis for well design would not result in
the best/highest efficiencies and would also lead to too
big spatial claims.
REFERENCES
Arcadis, TTE, Bos, W., 2011. Handreiking masterplannen
bodemenergie. SKB, Gouda.
Bear, J., Jacobs, M., 1965. On the movement of water bodies
injected into aquifers. Journal of Hydrology 3, 37-57.
Beek, C.G.E.M.v., 2010. Cause and prevention of clogging
of wells abstracting groundwater from unconsolidated
aquifers, geo-sciences. VU-Amsterdam, Amsterdam.
Bloemendal, M., Olsthoorn, T., Boons, F., 2014. How to
achieve optimal and sustainable use of the subsurface for
Aquifer Thermal Energy Storage. Energy Policy 66, 104-
114.
Bloemendal, M., Olsthoorn, T., van de Ven, F., 2015.
Combining climatic and geo-hydrological preconditions as a
method to determine world potential for aquifer thermal
energy storage. Science of the Total Environment 538 621-
633.
Blum, P., Campillo, G., Munch, W., Kolbel, T., 2010. CO2
saving of ground source heat pump systems – a regional
analysis. Renewable energy 35, 122-127.
Caljé, R., 2010. Future use of aquifer thermal energy storage
inbelow the historic centre of Amsterdam, Hydrology. Delft
University of Technology, Delft.
CBS, 2005. Hernieuwbare energie in Nederland 2004, in:
CBS (Ed.). Central authority for statistics in NL, Den Haag.
Doughty, C., Hellstrom, G., Tsang, C.F., 1982. A
dimensionless approach to the Thermal behaviour of an
Aquifer Thermal Energy Storage System. Water Resources
Research 18, 571-587.
Eugster, W.J., Sanner, B., 2007. Technological status of
shallow geothermal energy in Europe, European geothermal
congress, Unterhaching, Germany.
Fry, V.A., 2009. Lessons from London: regulation of open-
loop ground source heat pumps in central London.
Geological society of London 42, 325-334.
Gelhar, L.W., Welty, C., Rehfeldt, K.R., 1992. A critical
Review of Data on Field-Scale Dispersion in Aquifers.
Water Resources Research, 1955-1974.
Groot, J., 2014. Optimizing energy storage and reproduction
for Aquifer Thermal Energy Storage, Geosciences.
University of Utercht.
Hecht-Mendez, J., Molina-Giraldo, N., Blum, P., Bayer, P.,
2010. Evaluating MT3DMS for Heat Transport Simulation
of Closed Geothermal Systems. Ground water 48, 741-756.
Hoekstra, N., Slenders, H., van de Mark, B., Smit, M.,
Bloemendal, M., Van de Ven, F., Andreu, A., Sani, D.,
Simmons, N., 2015. Europe-wide Use of Sustainable Energy
from Aquifers, Complete report. Climate-KIC.
Houben, G.J., 2006. The influence of well hydraulics on the
spatial distribution of well incrustations. Ground water 44,
668-675.
Jaxa-Rozen, M., Kwakkel, J.H., Bloemendal, M., 2015. The
Adoption and Diffusion of Common-Pool Resource-
Dependent Technologies: The Case of Aquifer Thermal
Energy Storage Systems, PICMET, Portland.
KNMI, 2013. Hourly data 1980-2010, climatic projections
KNMI W+ scenario 2010-2045 for Weather station 260, de
Bilt.
Korom, S.F.K., F, Bekker Helweg O.J., 2003. Influence of
pump intake location on well efficiency. Journal of
hydrologic engineering 8, 197-203.
Li, Q., 2014. Optimal use of the subsurface for ATES
systems in busy areas, Hydrology. Delft University of
Technology, 2014.
NVOE, 2006. Richtlijnen Ondergrondse Energieopslag,
Design guidelines of Dutch branche association for
geothermal energy storage, Woerden.
Schultz van Haegen, M.H., 2013. Wijzigingsbesluit
bodemenergiesystemen, in: Environment, M.o.I.a. (Ed.).
Staatscourant 23617, Den Haag.
SIKB, 2015. Userdata of ATES systems, in: NL, d.p.i. (Ed.).
recieved from 4 provinces of the Netherlands in 2015,
Gouda.
Sommer, W., 2015. Modelling and monitoring Aquifer
Thermal Energy Storage. Wageningen University,
Wageningen.
TNO, 2002a. Boringen uit Dinoloket, digitale
grondwaterkaart, Utrecht.
TNO, 2002b. Grondwaterkaart van Nederland Utrecht.
TNO, 2002c. REGIS, Utrecht.
USGS, 2000. Modflow-2000, the u.S. Geological survey
modular ground-water model—user guide to modularization
concepts and the ground-water flow process in: Harbaugh,
A.W., Banta, E.R., Hill, M.C., McDonald, M.G. (Eds.). US
Geological Survey, Virginia.
Verbong, G., van Selm, A., Knoppers, R., Raven, R.P.J.M.,
2001. Een Kwestie van Lange Adem. De Geschiedenis van
Duurzame Energie in Nederland 1970–2000 Aeneas, Boxtel.
Willemsen, N., 2016. Rapportage bodemenergiesystemen in
Nederland. RVO / IF technology, Arnhem.
Zheng, C., Wang, P.P., 1999. MT3DMS: A Modular Three-
Dimensional Multispecies Transport Model for Simulation
of Advection, Dispersion, and Chemical Reactions of
Contaminants in Groundwater Systems; Documentation and
User's Guide.
ACKNOWLEDGEMENTS
This research was supported by the joint research of
the Dutch Drinking water companies, Climate-kic E-
use (aq) and the URSES research program funded by
the Dutch organization for scientific research (NWO)
and Shell.
