Introduction
Underground pumped storage hydropower (UPSH) is an alternative energy storage
system (ESS) for flat regions (Pujades et al., 2016; Pummer and
Schüttrumpf, 2018). UPSH plants consist in two reservoirs, one is
underground while the other is located at the surface (Barnes and Levine,
2011). The excess of electricity generated during low demand energy periods
is used for pumping water from the underground to the surface reservoir, and
when the demand of energy increases, water is released into the underground
reservoir through turbines for generating electricity. Although there are not
bibliographical evidences of UPSH constructed plants, this technology has
been investigated in different parts of the world: the Netherlands (Min,
1984), Singapore (Wong, 1996), USA (Allen et al., 1984; Severson, 2011),
Germany (Beck and Schmidt, 2011; Zillman and Perau, 2015; Alvarado et al.,
2016), Belgium (Bodeux et al., 2016; Poulain et al., 2018), Spain
(Menéndez et al., 2017) and South Africa (Winde and Stoch, 2010a, b; Khan
and Davidson, 2016; Winde et al., 2017), Finland and Australia (Academy of
Science of South Africa, 2016).
Although it would be possible to drill the underground reservoir, the
alternative considered in this paper, which may be more efficient and have
positive effects for local communities after the cessation of mine
activities, would consist in re-using abandoned mines. The main concern of
UPSH using abandoned mines is the water exchanges between the underground
reservoir and the surrounding porous medium because they can affect the
environment and the efficiency of the UPSH plant. Most studies focused on
water exchanges consider flow related issues (Bodeux et al., 2017; Pujades et
al., 2017a). However, recently, Pujades et al. (2017b) have suggested the
importance of considering hydrochemical changes induced by UPSH plants. These
changes may impact on the environment and affect the efficiency of the plant.
General view of the problem (a) and the whole modelled domain
(b).
![]()
Water is aerated when it is pumped, discharged and stored in the surface
reservoir. As a result, its chemistry evolves to reach equilibrium with the
atmosphere. Similarly, when water is discharged into the underground
reservoir, its chemistry evolves to reach equilibrium with the surrounding
porous medium. These hydrochemical changes may produce pH variations,
especially in coal mine contexts where pyrite is a common mineral. Its
oxidation leads to pH lowering. Low pH values would affect the environment
(decreasing the quality of groundwater and surface water bodies) and the
efficiency of the plant (corroding UPSH facilities such as pipes, turbines,
pumps or concrete structures).
Although the general behaviour of the system has been previously stated
(Pujades et al., 2018), there is not any study in which the influence of the
aquifer hydraulic parameters on a UPSH system is assessed. To establish the
role of aquifer hydraulic parameters will be meaningful for selecting the
most suitable places where constructing future UPSH plants. Thus, the main
objective of this work is to investigate the importance of the hydraulic
parameters on the pH variations occurring when abandoned coal mines (with
presence of pyrite) are used for UPSH.
Methods
Problem statement
A 200 m thick domain with an underground reservoir in the middle is
considered (Fig. 1a). The reservoir (50×50 m and 10 m of height) is
saturated in natural conditions with the top and bottom located respectively
at 95 and 105 m depth. The water table is located at 92.5 and 97.5 m depth
in the upgradient and downgradient boundaries, respectively. Thus, the
underground reservoir is located at the top of an unconfined porous medium
whose saturated thickness ranges between 107.5 and 102.5 m. The outer
boundaries are located at 500 m from the underground reservoir. The
hydraulic gradient under natural conditions is 0.005.
Frequency of pumping and discharging phases is chosen according to day/night
cycles (i.e., 12 h pumping and 12 h discharging water). Pumping and
discharging rates are 43 000 m3 d-1. These rates allow decreasing and increasing the hydraulic head inside the
underground reservoir up to 8.6 m during each pumping-discharging cycle.
It is assumed that the modelled cavity belongs to an abandoned coal mine.
Coal deposits usually contain sulphide minerals, whose oxidation may entail
important consequences for water chemistry. Pyrite is the most common
sulphide mineral in this kind of deposits (Akcil and Koldas, 2006), and
thus, it is assumed that the porous medium contains 1 % pyrite. Reaction
rates for the other minerals (e.g. silicates) are assumed very low (White
and Brantley, 1995), and are neglected.
Numerical model
The code PHAST (Parkhurst et al., 1995; Parkhurst and Kipp, 2002) is used to
simulate the problem. This code solves multicomponent, reactive solute
transport in three-dimensional saturated groundwater flow (Parkhurst et al.,
2010). The watershed divide crossing the domain from the west to the east
boundaries (Fig. 1a) allows modeling only half of the domain without
affecting the results. The modeled “half-domain” is divided in 15 600
elements whose size ranges from 2 to 100 m (they are refined towards the
underground reservoir) (Fig. 1). Dirichlet boundary conditions (BCs) are
implemented in the west and east boundaries with head prescribed at 92.5 and
97.5 m depth, respectively. Flow-rate BCs are adopted in nodes located
inside the underground reservoir for simulating the pumpings and discharges.
The values of longitudinal (αL) and transversal (αT) dispersivity are 10 and 1 m, respectively. The underground
reservoir is modelled by implementing a high value of K (106 m d-1), S of 1,
and a dispersivity of 104 m in the three directions. The validity of
the assumptions adopted for modelling the underground reservoir has been
evaluated by Pujades et al. (2017b). Three different scenarios (Sce1, Sce2
and Sce3) are simulated and compared to ascertain the influence of the
hydraulic conductivity (K) and porosity (Φ) values in the surrounding
porous medium on the system behaviour. K and Φ are 0.01 m d-1 and 0.05,
respectively, in Sce1. K is increased in Sce2 up to 0.1 m d-1, while Φ is
increased in Sce3 to 0.25. The hydraulic parameters (K and Φ) remain
constant during the simulated period. This particularity does not affect
noticeably the results because the variation of Φ is negligible
(Pujades et al., 2018).
Basic concepts
Pujades et al. (2018) stated the main trends of the system. Dissolved oxygen
increases when the water is pumped, discharged and stored in the surface
reservoir. When this water is discharged in the underground reservoir and is
exchanged with the surrounding porous medium, it oxidizes pyrite decreasing
the groundwater pH (Fig. 2). Pyrite is oxidized until all available oxygen
is consumed. Subsequently, water is pumped, discharged and stored in the
surface reservoir and the dissolved oxygen increases again. pH in the
underground reservoir decreases when it is filled with groundwater from the
surrounding porous medium (Fig. 2). As a result, when this water is pumped
to the surface reservoir, pH also decreases on it. In addition, minerals
such as ferrihydrite, goethite and schwermannite may precipitate in the
surface reservoir contributing also to the pH reduction.
pH evolution in the surrounding porous medium at 15 m from the
underground reservoir (a). pH evolution in the reservoirs (b). These
results are obtained for Sce1.
![]()
Results
Results show the differences in percentage between Sce1 and the scenarios
Sce2 and Sce3. A positive difference means that computed results (for Sce2
and Sce3) are higher than those obtained for Sce1 while differences are
negative when they are lower.
Underground and surface reservoirs
Figure 3 shows the results concerning the pH evolution in the surface (left)
and underground reservoirs (right). pH is higher for both scenarios (Sce2
and Sce3) than that computed for Sce1. pH decreases less for Sce2 and Sce3
than for Sce1 because the volume of groundwater reaching the reservoir from
the upgradient side, which is less affected by the pyrite oxidation,
increases when the values of K and Φ are incremented. In principle, pH
difference should increase constantly with time as the volume of water
reaching the underground reservoir from the upgradient side. However, after
an initial increase, the pH difference remains nearly constant, especially
between Sce3 and Sce1, and the pH difference with respect Sce1 only start to
increase constantly after 20 simulated days. This behaviour may be related
with the precipitation of schwertmannite in the surface reservoir, which
releases H+ (i.e., reduces the pH). In addition, the precipitation rate
of schwertmannite decreases with pH and stops for pH values lower than 2.8.
Therefore, given that pH in Sce2 and Sce3 are higher than in Sce1, more
schwertmannite precipitates, which contribute to decrease the pH and avoid a
constant increase of the pH difference between scenarios. However, after 20 days, the precipitation of schwertmannite stops and pH difference between
scenarios only depend on the groundwater reaching the reservoir from the
upgradient side. As a result, the pH difference starts to increase
constantly.
pH differences in the surface (a) and underground reservoirs
(b).
![]()
Surrounding medium
Figure 4 shows the results concerning the pH evolution in the surrounding
porous medium. pH is computed at a distance of 15 m from the underground
reservoir (in the downgradient side). In Sce2, pH is lower than that for
Sce1 because dissolved oxygen reaches faster the surrounding medium
(dissolving more pyrite) and groundwater with low pH flows faster until the
distance at which the pH is computed. Contrarily, pH decreases less for Sce3
than in Sce1 although the water exchanges increase and the dissolved oxygen
reaches faster the surrounding medium. In this case (Sce3), the volume of
water in the aquifer is higher than that of Sce1 and the pH reduction is
buffered (i.e., there is a dilution effect).
pH differences in the surrounding porous medium. pH is computed 15 m far away from the underground reservoir in the downstream direction.
![]()
Conclusions
This work investigates the influence of the hydraulic conductivity and the
porosity on the hydrochemical changes induced by UPSH. Results could be
helpful for defining screening strategies, which should be used for the
selection of potential abandoned coal mines to construct future UPSH plants.
Results show that pH decreases less in the reservoirs but more in the
surrounding porous medium when the value of K is raised. This behaviour in
the reservoirs is positive for mitigating corrosion problems in the plant
facilities (e.g., pipes, turbines, pumps) and the environmental impact if
some water from the surface reservoir is accidentally discharged in surface
water bodies. However, the mitigation of the corrosion and environmental
impact in surface water bodies would be very limited since pH difference
between Sce2 and Sce1 in the reservoirs are lower than 5 %. Contrary, the
pH decreases approximately the 20 % in comparison with Sce1 in the
surrounding porous medium, which would increase the environmental impact on
groundwater. Given these results, coal mines surrounded by materials with
low values of K would be preferable for mitigating the environmental impact.
A different behaviour is observed when porosity is increased. In this case,
pH decreases less in the reservoirs and also in the surrounding porous
medium. In that particular case, the adverse effects of UPSH in the presence
of pyrite would be mitigated. Thus, coal mines surrounded by materials with
high values of porosity would be preferable for constructing UPSH plants.