Thermal utilization of shallow groundwater development and application of groundwater models in the water-rights concession process

Transcription

1 Thermal utilization of shallow groundwater development and application of groundwater models in the water-rights concession process Dr. Joachim Poppei Carl Philipp Enssle Liane Schlickenrieder AF-Colenco Ltd., Switzerland ABSTRACT: The use of groundwater for cooling and heating purposes is of interest from both the economical and the ecological perspective. For example, cooling facilities relying on groundwater help to minimize the power-intensive refrigeration capacity of other cooling systems commonly used in industrial refrigeration plants while groundwater-based heat pump facilities may cut the consumption of fossil fuels. The utilization of groundwater for cooling or heating purposes via heat transfer requires a water-rights concession which calls for a site-specific investigation to assess the potential impact on the environment. The Swiss Guideline on Water Protection recommends that the re-injection of thermally used groundwater shall not increase or decrease the ambient temperature of the groundwater body by more than 3 K beyond the immediate vicinity of the facility. The site-specific thermal environmental impact assessments normally are based on numerical groundwater models. Key aspects typically studied include the available hydraulic capacity of the aquifer (yield), the optimum number, location and operation mode of pumping and re-injection wells, and the likely resulting temperature distribution in the groundwater downstream of the site with any potential impact on wells already in place. For small and medium-sized thermal projects the effort necessary to develop a local, site-specific model is often inappropriate. However, regional-scale models have been established for many basin or river aquifers. This paper illustrates how existing regional models may be employed as an alternative to developing new local models for the purpose of optimizing facility design and assessing the local impact of the thermal groundwater utilization. The discussion concentrates on the conceptual model and the potential simplification of processes involved. The Reuss River Valley model serves as an example of a regionalscale model of an aquifer with naturally high flow velocities which has been used as a base for site-specific thermal environmental impact assessments for several small heating and cooling projects. INTRODUCTION The shallow groundwater is not new on the scene of renewable energy sources. Groundwater has been fostered as an agent for heating and cooling for more than 30 years. Heating systems for family homes and large construction projects alike have integrated heat pumps and many an industrial refrigeration plant includes a cooling component relying on groundwater. The utilization of a shallow groundwater reservoir for cooling or heating purposes becomes increasingly attractive both from an economical and an ecological point of view. Naturally, the hydrogeological setting in some areas is more suitable for thermal utilization than in others. In order to optimize the efficiency of the thermal facility while keeping the impact on the natural environment at a minimum requires a good understanding of the local conditions. Site-specific investigations are indispensable and, often, numerical groundwater modeling is the tool of choice. An issue specific to some Swiss valley aquifers is the competition for the thermal utilization of shallow groundwater. Not only is there growing demand for the commodity; in many places the aquifers high hydraulic conductivities and correspondingly high flow velocities require detailed prognoses of potentially adverse interferences between neighboring groundwater-based cooling or heating facilities. Again, numerical models serve well. Another aspect concerns licensing. The transfer of heat from or to the groundwater is subject to a water-rights concession. The concession is granted only if no public interests are violated and the rights of third parties are not affected. The Swiss Guideline for Water Protection (BUWAL, 2004), recommends that the re-injection of thermally used groundwater shall not increase or decrease the ambient temperature of the groundwater body by more than 3K. Only in the immediate vicinity of the

2 re-injection point, i.e. within a distance of 100 m is a temperature change of more than 3K tolerated. Most concession applications call for some kind of groundwater model. Various software packages are available today to generate the much coveted numerical groundwater models. However, despite user friendly pre- and post-processing tools, the overall effort expended for setting up and calibrating a model and for running the simulations continues to be relatively large. Oftentimes the effort would outgrow the budget allocated to planning a small- or medium-sized facility. The technical designers and modelers are then forced to design an adequate model which accounts for the budgetary constraints and sufficiently represents the most important processes and local conditions with the necessary resolution and level of detail. SPECIFIC TASK DESCRIPTION AND MODEL CONCEPT Facilities for the thermal utilization of groundwater for heating and/or cooling purposes are always designed specifically for one site. The design needs to account for the local hydrogeological conditions and is characterized by the specific technical requirements posed on the aquifer in terms of the required groundwater quantity, the required energy content and the intended purpose (i.e. for cooling or heating). In terms of hydrogeology, the thermal utilization of a valley aquifer requires consideration of the following aspects: Valley aquifers, i.e. Quaternary gravels or river sediments, typically have high hydraulic conductivities (10-3 >10-2 m/s), in general with little spatial variability, a relatively steep gradient of the groundwater table (several to a few %), tend to be bound to the valley in their lateral extent, and are usually influenced by hydraulic communication with rivers or creeks. The high flow velocities often make it necessary to extend the model over relatively large areas of several km 2 that encompass a correspondingly large array of recharge and discharge features for which data have to be acquired and integrated into the model. Locally induced thermal changes of the groundwater disperse quickly, often in narrow plumes downstream of the injection point, and extending beyond the actual area of investigation and/or subject to a concession. As a result, competing claims and licensing aspects may carry significant weight and even become liable for trial in a court of justice (which doesn t make things easier for a modeler seeking to provide an adequate model). According to conceptual considerations the model may be developed with the following key processes: The dominating transport process is the advective transport associated with flow. It is retarded by the heat exchange between water and matrix. Secondary processes are the heat exchange with low-conductivity lenses in the aquifer, underlying units or unsaturated sedimentary layers above, heat conduction within the aquifer and dispersion. Flow and heat transport may be de-coupled; the dependence of hydraulic properties on temperature is negligible for the temperature range considered. Of interest are only changes in the middle-ranged, unaffected conditions, e.g. annual averages of groundwater table and temperature. Flow quickly becomes steady-state under constant conditions but the ongoing injection of thermally altered water has long-term consequences and if there are no adequate sinks needs to be considered as transient. This requires that a suitable representation of the consumption-dependent utilization schemes be defined. Flow in highly conductive aquifers can usually be represented in a simplified manner in 2 dimensions, however, in general, with due consideration of the correct depth of the aquifer s base. This implies a non-conservative disregard of transport along preferential flow paths while taking advantage of thorough mixing throughout the full thickness of the aquifer. Due to this non-conservative model assumption the consideration of heat exchange with the underlying units, sedimentary cover or the ground surface as an additional retarding factor should be carefully assessed. Where the local hydrostratigraphy justifies the assumption, one can do this even in a 2D model with a simplified implementation of the time-dependent heattransfer concept (cf. Diersch/Kolditz/Jesse, 1989).

3 The thermo-physical properties unlike the hydraulic properties vary relatively little and are rather insensitive in the ranges considered. It is generally sufficient to parameterize these globally. The predominance of the advective heat transport bound to groundwater flow, in combination with the site-specific variability of the factors affecting this flow requires that one starts with a sufficiently large flow model. Any subsequent consideration like the local thermal utilization scenario is then relatively simple. This is exemplified in the following discussion of a regional model which has been used repeatedly in the past few years for various assignments. THE REUSSTAL AQUIFER MODEL The numerical groundwater model of the Reusstal river valley was developed in 2004/05 as a groundwater flow model of an important valley aquifer in the Swiss Canton of Lucerne, north of Lake Lucerne. The model is based on a comprehensive hydrogeological investigation and research conducted in the 1990 s. It covers an area of about 20 km 2. The model is a steady-state 2D flow model with an unconfined surface (Fig. 1). It takes into account all factors affecting the groundwater flow field such as groundwater replenishment, inflow from the valley slopes, hydraulic communication with the Reuss River and several creeks via infiltration and exfiltration and more than 20 drinking water wells. Figure 2 shows the average balances of sources and sinks in the model area. Figure 1: Overview of the sedimentary valley aquifers in Switzerland (BWG 2005) and location of the Reusstal aquifer model (in the Canton of Lucerne; Basis: Digitaler Übersichtsplan des Kantons Luzern).

4 Replenishment Inflow upstream Inflow from valley slopes Exfiltration into rivers Outflow downstream Infiltration from the Reuss Wells Figure 2: Groundwater sources (left) and sinks (right) in the Reusstal aquifer model. The model was coarsely calibrated with data collected at 55 observation points in a reference-day measurement campaign (Fig. 3). The Reusstal model was developed for the Office of the Environment of the Canton of Lucerne (Colenco 2004) with the objective to serve for various types of assignments or to at least provide the hydrogeological boundary conditions for local project tasks. In the past few years these included not only issues of thermal utilization but also flood protection and construction projects with an impact on the groundwater system. One prerequisite for the application of a standby, i.e. always ready to use, model is the continual updating and maintenance of the model, particularly when the catchment characteristics of the water wells are changed, new drilling activities reveal new insights etc. Figure 3: Groundwater flow field of the Reusstal valley aquifer after calibration (isolines), and point measurements (black numbers) in m a.s.l.

5 USING MODELS FOR THERMAL APPLICATIONS A CASE STUDY In the model area the aquifer is presently utilized by a number of small-scale heat pumps for single family homes, and by about half a dozen medium- to large-scale facilities for heating and cooling purposes. The planning and assessment of one of these facilities is illustrated in the following. An international trade chain planned to utilize groundwater for heating and cooling purposes at one of its new branches. The branch includes offices, a cold store and dry storage facility. Its location within the Reusstal aquifer model is shown in Figure 4. The construction planners quoted the heat demand for air conditioning, ventilation and heating as 2100 kw of which 500 kw were to be contributed by a groundwater heat pump. The required refrigeration capacity of about 200 kw was to be obtained via direct heat exchange. Figure 4: Model area of the Reusstal valley aquifer and location of the case study for thermal groundwater utilization. Once the site had been investigated sufficiently and the feasibility proven in principle it was time to prepare for the concession application. The preparatory modeling included various scenarios for potential well locations, utilization schemes and temperatures for the water re-injected into the ground. The Reusstal aquifer model served as a base. It was locally refined and augmented with thermal properties (Table 1). The most recent results from field investigations such as the depth of the aquifer base as determined by drilling were integrated. The model considers the local hydrogeological and water management settings with due respect. These are, specifically, the proximity of a communal collective pump connected to 5 wells, the Reuss River and its interaction with the groundwater, and the wetland which is of particular botanical value.

6 Parameter Value Dimension Transport effective porosity Thermal conductivity of rock 2 W/m K Thermal conductivity of water 0.65 W/m K Heat capacity of rock Ws/m 3 K Heat capacity of water Ws/m 3 K Dispersion length (longitudinal) 10 m Table 1: Thermophysical and transport parameters The conditions for the thermal utilization of the groundwater were found to be most favorable for the following scenarios: November to April: reduction of the water temperature as the groundwater is used for heating purposes; 68 m 3 /h pumping rate as a base load over 24 h/day on 7 days/week May to October: increase in the water temperature as the groundwater is used directly for cooling; 34 m 3 /h pumping rate from 7 a.m. to 6 p.m. on 6 days/week The simulation for the cooling period is based on a continual production scheme over 184 days with an average rate of about 17 m 3 /h for a constant temperature difference. The daily modifications in this scheme may be neglected for these long-term considerations. It is worth noting that all of the withdrawn water is returned to the aquifer. It is assumed for the simulation that the unaffected average groundwater temperature is 10 C. Thus, the utilized water may be assumed to be re-injected with a temperature of 6 C during the heating period and 14 C during the cooling period. The thermal influence on the groundwater conditions is assessed by looking at two points in time, namely in the 9 th and 10 th year of operation: At the end of April when the heating period has ended and there is no more re-injection of temperature-reduced water At the end of October when the cooling period has ended and there is no more re-injection of temperature-raised water Figure 5 and 6 show the isotherms at the re-injection site. It is apparent that the conditions are the same in the 10 th year as they are in the 9 th and, therefore, the influence is not expected to grow any further with time as long as the pumping scheme remains the same, the temperature reduction and temperature rise within an annual cycle are reversible, i.e. they are extinguished by the respective scheme of the following period and the thermal effects caused by the re-injection of temperature-reduced water during the heating period are stronger than those resulting in the cooling period. For the re-injection of temperature-raised water (after the cooling period), the 13 C isotherm remains in the close proximity of a few meters from the injection point (and, therefore, is not discernible in Fig. 5). When the groundwater is used for heating purposes and the water is re-injected with a reduced temperature, the 7 C isotherm drifts to a maximum distance of about 180 m from the injection point. Over the course of the alternating production scheme, the energy drawn from the groundwater during the heating period is about 5-fold of what is introduced while the groundwater is used for cooling and refrigeration. The temperature reduction makes it impossible, within an annual cycle, to stay within the recommended 3K/100 m. However, the authorities may take into account that with 4K the thermal impact is minor even at the center, the impact is neutralized over the course of an annual cycle and if the utilized water is re-injected into the ground at the recommended location, the main impact will occur within the perimeters of the land owned by the user. The authorities granting these thermal utilization concessions may impose monitoring directives.

7 Figure 5: Temperature distributions after the cooling period (isotherms: 11 C and 12 C) in the 9 th year (left) and in the 10 th year of operation. Figure 6: Temperature distributions after the heating period (isotherms: 7 C, 8 C and 9 C) in the 9 th year (left) and in the 10 th year of operation. ADVANTAGES OF THIS APPROACH The above case study of an intensively utilized valley aquifer demonstrates how different expectations and responsibilities may be distributed in a cost-effective framework. The groundwater model fulfills specific functions in the context of site planning and facility design, as well as specific functions concerning protection requirements posed by the environmental authorities and the prognosis and monitoring of potentially competing operations. The relatively large effort generally invested in setting up, calibrating and maintaining a groundwater model is shared by several stakeholders. If the labor- and cost-intensive development of a regional flow model is financed by the authorities, they own the model and can use it for various investigations with different objectives concerning the groundwater. The authorities granting the concessions for the thermal utilization of groundwater thereby retain a current overview of installations and their consequences. The possibility to resort to such a regional flow model enables potential implementers of small or medium facilities for the thermal utilization of groundwater to rely on numerical simulations for the optimization of the facility s design and to check its licensing suitability while staying within a reasonable planning budget. In Switzerland, models like the Reusstal aquifer model have been established for many river aquifers.

8 REFERENCES BUWAL 2004: Wegleitung Grundwasserschutz. Vollzug Umwelt, Bundesamt für Umwelt, Wald und Landschaft BWG 2005: Hydrogeologische Übersichtskarte der Schweiz, Bundesamt für Wasser und Geologie; reproduced by energie schweiz, Info-Geothermie, Nr. 9, July 2005 Colenco 3513/01: Qualifizierung des FEFLOW-Modells Reusstal, im Auftrage des uwe Luzern, Dezember 2004 Diersch, H.-J., Kolditz, O. and J. Jesse, 1989: Finite element analysis of geothermal circulation processes in hot dry rock fractures, Zeitsch. Angew. Math. Mech., 69, 3, , 1989