54 ates in fractured chalk. site investigation techniques to model

54 ates in fractured chalk. site investigation techniques to model

ATES IN FRACTURED CHALK: SITE INVESTIGATION TECHNIQUES TO
MODEL SYSTEM SUSTAINABILITY
R. Law
Geothermal Engineering Ltd
82 Lupus St, London, SW1V 3EL
[email protected]
Duncan Nicholson
Arup Geotechnics
13 Fitzroy St, London, W1T 1BQ
[email protected]
ABSTRACT
Aquifer thermal energy storage (ATES) systems are being installed in increasing numbers in
the United Kingdom. This is particularly true in central London, where planning regulations
have steered developers towards more environmentally friendly ways of providing heating
and cooling to buildings. This paper examines some sustainability issues that may occur due
to a combination of geology, small building footprints and regulatory frameworks. It also
outlines an approach to site investigation and numerical modelling to aid system design.
This approach is then applied to a project located in central London. The results of the
testing and subsequent modelling imply that the injection and extraction boreholes are linked
by a small number of fractures. Given these conditions, models were used to predict the
short and long term sustainability of the planned ATES system.
1. BACKGROUND
Changes to the planning regulations in London were introduced in 2004 (Mayor of London,
2004). The regulations required developers meet at least 10% of the total energy demands of
a proposed building from renewable sources, including ATES. Consequently there has been
a steady growth in installation. ATES systems use groundwater as a heat ‘source’ during the
heating (Winter) cycle and a heat ‘sink’ during the cooling (Summer) cycle.
A common strategy in London has been to install ATES systems that use groundwater from
the Chalk aquifer. Recent regulations from the Environment Agency in England state that
licenses for extraction from the Chalk are unlikely to be granted for extractions greater than
0.2 Mega litres/day (525 gallons per minute) (Environment Agency, 2005a). For an ATES
system of any size therefore, the majority of the water abstracted from the central London
aquifer will have to be re-injected. This water will be at a different temperature to the
aquifer.
2. GEOLOGY AND HYDROGEOLOGY OF LONDON
The central London area falls within the London Basin, an asymmetric syncline in Cretaceous
and Tertiary deposits. The Basin is faulted in a number of places. The faulting affects
aquifer continuity and the flow of water. A typical sequence of strata at a site can be
summarised as, Made Ground, Alluvium & Gravels, London Clay / Lambeth Group, Thanet
Sand Formation and the Chalk. The Chalk is located at approximately 90m below ground
level and forms the main aquifer. The intact chalk is generally considered to be impermeable
(Bloomfield 1995, Bose 1985, Macdonald 2001) but fissures and fractures can make the
overall rock mass highly permeable. Water velocities can be rapid where significant
fracturing occurs.
The degree of fissuring in the Chalk can be highly variable. The mass permeability of the
Chalk typically decreases with depth and it is generally considered, for engineering purposes,
that almost all the water is yielded by the upper 60–100 m of the chalk. The Chalk is widely
exploited for water extraction in Central London with typical well yields of 15 litres/second
(240 gallons/minute). The background hydraulic gradient in the Chalk of the London basin is
typically 1/1000 with groundwater flow in the direction of the central London groundwater
depression (Environment Agency, 2005).
3. SUSTAINABILITY OF ATES SYSTEMS
The operation of an ATES system has the capacity, during the lifetime of the system, to
change the temperature of the aquifer beneath a site. The greater the net energy withdrawn
or rejected into the aquifer, the greater the long term potential for temperature change.
Generally speaking, long term changes in the temperature of the aquifer have a detrimental
effect upon the efficiency of an ATES system.
An ATES system comprising pairs of extraction and recharge wells (doublets) is designed to
operate either on the hot well – cold well principle, or on the basis of using the flow of
groundwater between the two wells to allow energy transfer between the groundwater and the
aquifer matrix. The hot well – cold well principle involves pumping in one direction when
the system is in cooling mode and reversing the flow direction when the system is in heating
mode. Over time this process changes the localized temperature in the aquifer around each
borehole, improving the overall efficiency of the system. Provided that the heat rejected to
the aquifer is approximately balanced by the ‘coolth’ withdrawn from the aquifer this type of
system should prove sustainable in the long term.
When a system is operational (whether in heating or cooling mode) it can be advantageous to
rely on the background hydraulic gradient to carry recharged groundwater offsite: in this way
a proportion of the energy deficit (heat or coolth) is exported beyond the boundaries of the
site. In practice however, in central London, the natural hydraulic gradient is seldom large
enough and flow beneath the site is dominated by the artificial gradient between the injection
and extraction wells. In the Chalk, typical drawdowns associated with an open system in
which wells are operated at 5 – 10l/s (80 to 160 gallons per minute) are within the range of 3
to 5 m (McDonald, 2001). These are matched by equivalent injection heads, so that the head
difference between pairs of wells in a doublet may be 6 – 10m. A building footprint in central
London will seldom exceed 100m at its maximum width which means that the local gradient
will be between 0.06 and 0.1.
The hydraulic gradient generated by the extraction and injection boreholes is therefore up to
100 times greater than the background gradient. Only a very small proportion of the reinjected groundwater, carrying the temperature deficit, will be carried off in the regional
groundwater flow system and the majority of the heat rejected from the building will remain
beneath the site (Figure 1) and travel towards the extraction borehole.
Thermal breakthrough in ATES systems occurs when the water from the injection borehole
interferes with the extracted water (either causing a temperature increase or decrease). If
thermal breakthrough does occur, the system will begin to encounter problems, most notably
a lowering of efficiency. If the temperature in the extraction well continues to increase /
decrease the system will eventually be forced to shut down, either due to a licence violation
or to temperature limits of the building system.
a) Sustainable system (ideal) – heat is
removed from the site
b) Unsustainable system (reality) – heat is
transferred back to the extraction well
Groundwater flow
Groundwater flow
Figure 1 Thermal breakthrough and the principle of gradients
Short term sustainability problems are caused by short circuiting between the injection and
extraction wells. Short circuiting is primarily caused by unexpectedly rapid water movement
through fractures. Fractures dominate the flow in the Chalk and are normally located in the
horizontal plane. It is conceivable that a single fracture could carry the majority of the flow
between an injection and extraction borehole causing rapid thermal interference (potentially
in a time frame under 48 hours).
4. DESIGN ISSUES AND TESTING
The issues of long term and short term sustainability should be addressed before the ATES
system is installed. Long term sustainability problems can be resolved by ensuring that the
energy demands placed upon the aquifer are approximately balanced over the year (this is not
necessarily the same as having balanced building loads). This can only be achieved through
dialogue between the building designers, M&E engineers and the ATES designer. Short term
sustainability issues are more difficult to resolve as they require some knowledge of the flow
within the aquifer. To better understand the structure of the aquifer beneath a site and thus
short term sustainability issues, a combination of Site Investigation techniques were used at a
planned site:
CCTV survey, Flow logging, Pumping tests, Tracer testing
All of the above tests require at least one and in some cases two test boreholes to be drilled at
the site. Although, if the trials are successful, the boreholes can be used for the actual ATES
system the process represents something of a Catch 22. A test borehole or boreholes are
expensive to install. The client is therefore to some degree committed to the system even
through no tests have been undertaken to prove that it will function as expected.
A CCTV survey was conducted over the entire length of the boreholes drilled at the site. A
CCTV survey serves two purposes in this context: a method of checking the way in which the
borehole has been drilled and cased and as a method of locating likely fracture zones. A
sample image of the CCTV survey is shown in Figure 2. It shows a possible fracture zone
located within the upper layers of the Chalk. Although the fracture can be identified it may
not necessarily be flowing.
Chalk matrix
Fracture
2mm
Figure 2 CCTV survey showing the potential fracture in the Chalk
Flow logging was undertaken in both boreholes under static and pumped conditions. A flow
log measures the flow at all points along the depth of the borehole. Flow rates increase where
water flows into the borehole from the aquifer. Identifying zones of flow can help to estimate
the numbers of fractures zones that carry flow into a borehole. If the flow is fairly regular
over the entire depth of the borehole then it can be assumed that many fractures are carrying
the flow and the aquifer can be treated as a homogenous medium. Conversely, if the flow
rate jumps at certain points along the length of the well then only a few fractures are likely to
be carrying the flow. Flow logs in both boreholes are shown in Figure 3. It appears as
though flow ingress occurs in both boreholes at the same depth. This is particularly apparent
in borehole 2. This implies that a significant proportion of the flow is occurring through a
single narrow horizon.
Flow ingress – fractures?
Figure 3 Flow logs for both boreholes under static and pumped conditions
In fractured aquifers, yields can vary within a short distance. Pumping tests help to confirm
the predicted yields. Three different types of pump tests were undertaken at the site:
1. Step drawdown tests
2. Constant rate pumping tests
3. Extraction and recharge trial
The step drawdown tests and the constant rate pumping tests enable the transmissivity of the
aquifer to be calculated. The calculated transmissivity is an important value as it helps to
determine whether the aquifer contains significant flowing fractures. The calculated
transmissivity from the tests is compared with the expected transmissivity for the site
(literature values and other borehole records). If the transmissivity is higher than expected
then this is a good indication that the Chalk is fractured. At the site the calculated
transmissivity was between 700 and 1000m²/day. This is significantly above that which
would be expected for the area (250m²/day) (Monkhouse, 1995) implying a significant
fracture flow.
Tracer tests are a method of measuring the velocity of water travelling between an injection
and extraction well. This test requires two boreholes to be drilled; one injection and one
extraction. At the injection borehole a known quantity of tracer (often fluorescein) is injected
in the borehole. A detector is placed at the extraction borehole and the amount of time that
the fluorescein takes to travel from the injection borehole to the extraction borehole is
measured. This travel time can be used to estimate the velocity of the water which is then
used to back calculate the transmissivity of the aquifer and a range of possible fracture
numbers and apertures.
Figure 4 Tracer test results
The results from the tracer test are shown in Figure 4. The spacing between the doublet
boreholes is 122 m. The results show three interesting features. First a clear rapid early
breakthrough curve at the extraction borehole. The first breakthrough occurs at 65 minutes
after start of injection, which is approximately 60 minutes after the tracer is first injected to
the aquifer (as a result of the delay in transmission down the borehole in the injection pipe).
The second is the apparent secondary breakthroughs at later times. Closer inspection of the
secondary breakthroughs show a periodic pattern with decaying amplitude. The secondary
breakthroughs imply the recycling of the tracer around the injection/extraction loop. The
third, and somewhat surprising result, is the smoothness of the breakthrough curves. A
fracture network would be expected to give a ‘noisier’ response and this suggests flow in
continuous and relatively uniform planes or channels.
A transmissivity of about 1000 m2/d has been calculated based on the interpretation of the
head differences between the boreholes during pumping.
To satisfy the combined
transmissivity and aperture constraints presented by the pumping and tracer tests implies that
about 4 equal fractures of approximately 1.5mm would yield the appropriate breakthrough.
This calculation assumes that the hydraulic aperture and the mechanical aperture for the
fractures are the same. This calculation also assumes that the fractures are planar rather than
channelled and it assumes that all the flow is through the upper horizon. To test these
assumptions it is necessary to do more than simply investigate the first breakthrough time, it
is necessary to model the full breakthrough curve. This was carried out using a particle
tracking model to model conservative tracer breakthrough to the discharge borehole in
steady-state flow conditions (analytically modelled) around a doublet borehole in an infinite
aquifer. The results of this modelling indicate that about 70% of the injected water is
travelling through the fractures in the upper section of the Chalk. The model also suggests a
total aperture of 7 mm and, therefore, assuming that 70% of the transmissivity (ie 666 m2/d)
is provided by the upper section and uniform fracture properties, then roughly 7 fractures of
aperture 1.1mm are required to meet both the transmissivity and fracture porosity constraints.
This is different to the first, simpler interpretation.
5. MODELLING
The interpretations of the tests were used as inputs to a number of numerical models. The
program used to construct the numerical model was SUTRA 3D (Saturated-Unsaturated
Transport). SUTRA 3D is a modelling code that can be used to simulate density-dependent,
saturated or unsaturated, water movement and the transport of either energy or dissolved
substances in a subsurface environment. SUTRA 3D was released by the United Stated
Geological Survey (USGS) in September 2003.
The models were designed to focus upon the thermal interaction between the boreholes over
time and the thermal breakthrough time for different supposed aquifer structures. The models
provided limited information on the environmental impacts associated with the general
thermal migration from the site. As the principal result of interest is the change of
temperature at the extraction wells, the predicted temperature at the abstraction boreholes
were plotted against time. The models operated under the assumption that if the temperature
of the extraction borehole was raised by a degree then the corresponding injection
temperature was also raised by a degree.
A number of different models were run using different fracture apertures and numbers
calculated from the tracer tests and were calibrated using the injection heads and the
drawdowns measured in the pumping tests (particularly extraction and injection). Calculated
ground energy demand figures were then applied to the most appropriate model to understand
both the short and long term sustainability of the planned system (injection and abstraction
borehole separated by 108m). The models consisted of; a peak flow run (20 l/s, 836 kW)
where the system was assumed to operate for 24 hours, a model that used the weekly energy
data (Figure 5) and a long term sustainability model that used the annual heat rejection
figures to the ground (200 MWh)
40
Rejection to ground
30
20
0
-10
Absorption from ground
(including HW)
-20
-30
-40
Wk 1
Wk 2
Wk 3
Wk 4
Wk 5
Wk 6
Wk 7
Wk 8
Wk 9
Wk 10
Wk 11
Wk 12
Wk 13
Wk 14
Wk 15
Wk 16
Wk 17
Wk 18
Wk 19
Wk 20
Wk 21
Wk 22
Wk 23
Wk 24
Wk 25
Wk 26
Wk 27
Wk 28
Wk 29
Wk 30
Wk 31
Wk 32
Wk 33
Wk 34
Wk 35
Wk 36
Wk 37
Wk 38
Wk 39
Wk 40
Wk 41
Wk 42
Wk 43
Wk 44
Wk 45
Wk 46
Wk 47
Wk 48
Wk 49
Wk 50
Wk 51
Wk 52
MWh
10
Weeks
Heat into ground
Heat from ground
Figure 5 Weekly energy demands placed on the ground
.
For the continual peak flow model run, the model predicted that thermal breakthrough would
occur after approximately 5 days (thermal breakthrough in this case being defined as a 0.1°C
rise in temperature). The second model run, applying the weekly data, predicted that the
temperature in the extraction well will rise by approximately 0.75°C by the end of the
summer. The temperature was then predicted to drop back to baseline (14˚C) after the system
switches from cooling mode to heating mode. The long term sustainability model predicts
that the temperature at the extraction well will not rise by more than 1°C over 50 years of
operation.
6. CONCLUSIONS
A number of different Site Investigation techniques were undertaken at a site in central
London to understand the nature of flow within the Chalk aquifer. The results of the tests
were used as inputs for numerical models to predict the short and long term sustainability of
an ATES system. The flow within the Chalk was inferred to occur through a small number of
fractures (less than 10). Based on this assumption, thermal breakthrough was predicted to
occur at the extraction borehole in approximately 5 days if the ATES system were run
continually at peak loads. Using the predicted weekly energy profile figures in the model,
temperatures at the extraction borehole are not predicted to rise by more than 0.75°C during
the summer, the peak cooling demand time. This would suggest that the system, as planned,
will function sustainably during peak operating times. A long term sustainability model
predicted that the system, as currently proposed, should run for at least 50 years without
causing any significant temperature rise in the aquifer.
REFERENCES
ASHRAE, 2003. Applications handbook SI (2003). P 32.16
Bell F.G, Culshaw M.G and Cripps J.C, (1999). A review of selected engineering geological characteristics of
English Chalk. Engineering Geology 54 (237-269)
Bloomfield, J, (1995). Characterisation of hydrogeologically significant fractures in the Chalk; an example
from the Upper Chalk of southern England. Journal of Hydrology 184 (1996) 355-379
Bose J.E, Parker J.D and Mquiston F.C, (1985). Design/ data manual for closed loop, ground coupled heat
pump systems. ASHRAE 1985.
Environment Agency, (2005)a. Groundwater levels in the Chalk-Basal Sands aquifer of the central London
basin, May 2005.
Environment Agency, (2005)b. Chalk groundwater licensing policy for London. Feb 2005.
Law, R. (2007). Aquifer thermal energy storage in the fractured Chalk aquifer. Stanford Geothermal Congress.
Macdonald, A.M and Allen D.J, (2001). Aquifer properties of the chalk of England. Quarterly Journal of
Engineering Geology and Hydrogeology. 34, 571-384
Mayor of London, 2004. Green Light to Green Power – The Mayor’s energy strategy.
Monkhouse, R., (1995). Prediction of well yield in the confined London basin – Quarterly Journal of
Engineering Geology, 28, 171-187
Voss, C.I., and Provost, A.M, (2002). SUTRA. A model for saturated-unsaturated variable-density groundwater flow with solute or energy transport. U.S. Geological Survey Water-Resources Investigations Report 024231, 260 p