SUPERVISED ROCK-MASS STORAGE OF THERMAL ENERGY PRODUCED BY A SOLAR THERMAL PANEL Fabien Delaleux 1-2, Xavier Py 1, Régis Olives 1, Antoine Dominguez 2, Sandra Lanini 3, Denis Nguyen 3 1 Laboratoire PROcédés Matériaux Energie Solaire, PROMES-CNRS UPR 8521, Université de Perpignan, Rambla de la Thermodynamique, Tecnosud, 66100 Perpignan, France, 04 68 68 27 05, Fabien.delaleux@promes.cnrs.fr 2 Dominguez-Energie, 18 rue des Martins pêcheurs, 66700 Argeles sur Mer, France 3 Bureau de Recherche Géologiques et Minières, BRGM, 1039 rue de Pinville, 34000 Montpellier, France Abstract Solargeotherm is a 3-years French project aiming to study experimentally and modelize the possibility of storing the energy produced by solar collectors into dry rock, without any underground aquifer so the water resource could not be exposed. The heat transfer to the rock mass is achieved using three geothermal probes that will heat the soil near them during summer. The energy stored will be recovered by the same probes during winter thanks to a dry cooler simulating experimentally a geothermal heat pump heating a conventional house. An experimental site has been implemented to assess the storage and recovery of the thermal energy produced by flat solar thermal collectors transferred to the underlying rock thanks to vertical probes. The aim of the project is to prove the viability of such a process with both an experimental and a theoretical (modelling) part. Keywords: solar energy, geothermal probe, energy storage, rock-mass 1. Introduction As acknowledged by everybody today, fossil energy resources are limited and the exponential increase of the global energy consumption will enhance the end of their availability. So, for a decade, the use of renewable energies, like solar or geothermal, has been an interesting alternative to the former sources. These renewable energies have a lot of well known advantages (clean, endless...); but they have also a principal drawback: the intermittency of their sources (over one day or one year). Considering the particular solar source, solar thermal collectors provides a very intermittent heat production that is time offset with respect to requirement. In order to solve this problem, an energy storage unit can be used. One of the largest available storage systems to be considered is the underground itself. As a matter of fact, even if thermal losses will be induced, a huge potential of free storage material without need in storage envelope can be mobilized. Some projects have been already made on this subject; the most well known is the project GEOSOL [1] applied to an individual house like the project GEOHELIOS [2], project led in altitude in south of France, where coupling solar panels and geothermal allowed to heat a house of 240m² with only one borehole of 100m depth. A great part of the projects made in this subject are explained by Sanner [3]. An experimental site has been implemented to assess the storage and recovery of the thermal energy produced by flat solar thermal collectors transferred to the underlying rock thanks to vertical probes. The aim of the project is to prove the viability of such a process with both an experimental and a theoretical (modelling) part.
2. Experimental device The site where the project is implemented is located in Montauriol, south of France, the underground of which is mainly composed of schist rocks. The facility is compound of 42m² of thermal solar panels, 3 vertical geothermal probes of 200m depth each (A, B, C), 3 measurement boreholes of 20m depth each (D, E, F) and a 6kW dry cooler simulating the thermal load of a conventional house (approximately 120m², 4 persons etc ). (a) (b) Fig.1: Experimental site: (a) Thermal solar panels, (b) Geothermal probes The 3 probes are spaced by 5 meters from one to each other, forming an equilateral triangle. This value of 5 meters was first calculated by modelling using the software COMSOL thanks to a first simple model in which the geothermal probes are simplified as holes exchanging heat with the underground a mean thermal power provided by the solar panels. The modelling allows the estimation of the thermal influence zone of the heat transferred around the boreholes. Nearer the boreholes are from one to each other, more the temperature near the borehole will increase. Nevertheless concerning this configuration, when boreholes are very close to each other, heat losses by diffusion of the heat produced through the rock-mass are enhanced. In the opposite, if the boreholes are far to each other, the obtained rock temperature level is low and so is the amount of heat stored. So, the value of 5 meters between boreholes has been estimated as the optimal solution to obtain an important rock-mass temperature raise and to avoid the risk of collision between the boreholes in depth. Fig.2: Simulation of a borehole thermal influence The experimental facility is composed of three main parts: the solar field, the geothermal circuit and the rock thermal load. During spring and summer, solar thermal energy is stored into the rock-mass using the geothermal heat exchange probes and during autumn and winter this stored energy is released into the dry cooler simulating
the thermal load of a conventional house. A system of solenoid valves allows to use the whole solar field or only a part of it and to connect the three geothermal probes in parallel or in series with the possibility to switch off one or more of them. The experimental facility is entirely instrumented so values of temperatures into the boreholes can be measured thank to fibre optical cable, as well as solar irradiance and wind speed recorded. Moreover, heat transfer fluid flow-meter, CO2 and hygrometric sensors have been implemented too. Fig.3: Synoptic of the system Fig.4: Borehole geometry, (A,B,C) heat exchange boreholes, (D, E, F) measurement boreholes. 3. Modelling An important part of Solargeotherm project is the modelling of the rock mass thermal response. In the present study, the modelling is realized by the software COMSOL. The first part of the modelling is concerned by the storage of solar energy in the rock mass during 6 month and the corresponding temperature increase of the rock. The target is to be able to predict the soil temperature at the beginning of the energy release cycle during winter. The release step is modelized thanks to a link between COMSOL and MATLAB software; COMSOL is used for the thermal response of the rock mass while MATLAB is in charge of a typical house load of heating during winter (classic house of 120m² with four persons and their occupation scenarios). In the present work, the energy storage provided by the thermal solar panels into the rock mass thanks to the three geothermal probes of 200m depth is considered. Major simplification procedures were made: a theoretical temperature at the outlet of the solar field was fixed. The three dimensions modelling of the rock volume was simplified thanks to different horizontal sections in two dimensions. The advantage of this method is that we can consider the real relative position of the three boreholes; the measurements of their real trajectories have been made during holing. These measurements show us that the boreholes are not perfectly verticals, they have been deflected for many meters (nearly 50 meters for the more deflected). This important deflection of the boreholes will induce a significant effect on the prediction of the optimal energy storage obtain and the simulation of the borehole radial influence.
Fig.5: Simulation of the energy storage obtained after 6 months Figure 5 is a view of the COMSOL graphical interface. This illustrates the representation of the rock mass temperature distribution after 6 months of energy storage for one specific depth (30 m) with a temperature at the outlet of the panels of 60 C (Fig 5). A simulation is made every ten meters to obtain the temperature all along the 200 meters depth. The initial temperature (before any thermal exchange) and the local thermal conductivity are known every ten meters thanks to preliminary experiments of thermal shock and thermal response test described later in this article. The results of the simulations made every ten meters lead to the estimation of the temperature at the centre of the triangle formed by the three boreholes as a function of time. The simulation is made for a whole duration of 6 months (180 days). Fig.6: Temperature evolution during 6 months of storage at the centre of the three boreholes
In the Fig 6 three different families of curves are gathered. First, the temperature curves showing almost constant values all along the 6 months. Those correspond to the case of largest depths where initial temperature is already higher thanks to the natural geothermal gradient. In this case, the temperature remains constant because the fluid into the borehole is colder than in the upper part. The fluid heats up the ground and losses a part of its energy all along the borehole. The two others families of curves are in an upper part of the borehole with an initial temperature colder than in the previous case. The temperature increase is more important but we have two different behaviours. A part of the curves seems to be very linear, because of the local conductivity is less important and leads to a linear temperature increase. The other part of the curves follows the conventional variations of the solar panels outlet temperature between days and nights. It corresponds also to a higher conductivity of the ground, so the diffusion of heat into the environment and the reactivity of the temperature range are more important. 4. Experimental results 4.1. Thermal response test The aim of the thermal test response is to know, before to start a geothermal system design and implementation, the value of the initial ground temperature and the mean value of the effective thermal conductivity. These parameters are of particular importance as they will be used to size the necessary depth of the boreholes to reach the desired heat transfer. The principle of this test is rather simple; first of all, in order to determine the initial temperature of the soil, the system is started without heating the heat transfer fluid circulating along the probes. When steady state is reached, we can measure directly the temperature of the fluid at the outlet of the probes, which is equal to the initial temperature of the underground (Fig 7). Then, a known thermal power is injected into a geothermal probe, measurement of inlet and outlet temperatures and fluid flow are done. Thanks to these values and to a corresponding mathematical model, the mean effective thermal conductivity of the rock mass is identified. Fig.7: Fluid and air temperatures and fluid flow value during the test Different available mathematical models can used to determine the effective thermal conductivity using a thermal response test. The method is described in details by Eklöf & Gehlin [4]. We have chosen to use here the simple finite line source model and the approximation made by Eskilson in 1987 [5] and Hellstrom in 1991 [6]. This model leads to the following equation 1:
= + u q e q 4αt T ( r, t) T0 du T0 + ln γ 2 4πλ 4πλ 2 u r r 4αt (1) Where: - T 0 is the initial ground temperature ( C) - λ is ground thermal conductivity (W.m -1.K -1 ) - q is heat flux injected by unit of length (W.m -1 ) - α is the ground thermal diffusivity (m².s -1 ) - r is the radius of the geothermal probe (m) - γ is the Euler constant (0.5772) Then the mean temperature of the fluid into the geothermal probe can be plotted as a function of time using a logarithmic scale. Fig.8: Mean fluid temperature in function of time in logarithmic scale From this approach, we can consider that the mean fluid temperature is proportional to the logarithm of time as written in Eq.2: T ( t) = k ln( t) + m (2) So, Eq(1) and (2) give us q 4αt T 0 + ln γ = k t + m r ln( ) 2 4πλ In which k and m are constants. Therefore we can write the following equality: q q k = λ = (4) 4πλ 4πk (3) In the particular case of the thermal response test of Solargeotherm process, this leads to: k = 1.26 q = 51.7 W.m -1 which gives a value of mean thermal conductivity of λ = 3.26 W.m -1.K -1
As explained above, the thermal response test leads to an estimation of the initial temperature T 0 of the rock mass. In the particular case of the present study, this temperature is 16.5 C. All these parameters estimated by use of the thermal test are directly used in the model of the process. Another kind of thermal test was done for this project in order to confirm the results of the thermal response test, namely the thermal shock. 4.2. Thermal shock The concept of the thermal shock is to generate a sharp pulse of heat flux to the rock mass like a Dirac, to observe its thermal answer and the heat transferred by diffusion into the environment. As explained above, the geothermal probes are equipped with optical fibres to follow the evolution of boreholes temperature during the test. To realize the thermal shock, 1m 3 of water at 70 C is injected during one hour into one of the geothermal probes (probe B). The temperature of the borehole is measured during 6 days after the shock to see the relaxation of the rock mass and the diffusion of heat (Fig 9). Fig.9: Temperature at 60 m depth during the thermal shock The thermal shock, permit us to calculate thermal conductivity of rock mass every meter and not a mean value contrary to the thermal response test, using the line source model defined by Carlsaw & Jaeger [7]. This leads to a value of conductivity λ(z), function of the depth z. These values are illustrated in Fig.10 and compared to the mean value given by the thermal response test. Fig.10: Local value of conductivity determined by thermal shock
In Fig.10, the local thermal conductivity obtained by thermal shock ranges from 3 to 3.26 while the mean value estimated by the thermal test is 3.26. This illustrates the very good agreement of the two approaches. The local variation of the thermal conductivity can be attributed to local variation in rock composition, possible local water or air filled cracks. 5. Conclusion and perspectives The hybrid system coupling 42m² of thermal solar panels and three 200 m depth geothermal probes equipped with optical fibre was achieved on March 2010. The first measurements are efficient, but it will be very interesting to analyze these results after the summer and some months of storage. After the current summer, we will be able to release the stored energy thank to the dry cooler which simulate the thermal load of a conventional house. The Solargeotherm facility is also an efficient tool to compare the two different tests (thermal shock and thermal response test). The comparison between the two methods allows highlighting their differences; thermal response test is less precise because it gives a mean value of thermal conductivity, which can be sufficient for a lot of applications. Further, the different available corresponding models will be compared using the same data. The modelling part of the study will be also developed, coupling the rock mass response and the thermal load of the house models, in order to simulate the entire process and to predict the energy storage and release obtained for a long period. The first modelling part allows predicting the behaviour of the temperature increase during storage cycle and the differences obtained in function of the depth. The initial temperature given by the natural geothermal gradient and the local value of thermal conductivity are responsible of these differences. Acknowledgement The authors want to acknowledge the financial support of the French government through the ANR research program StockE. References [1] V.Trillat-Berdal, B.Souyri, G.Achard, (2006). Coupling of geothermal heat pumps with thermal solar collectors (GEOSOL). Applied thermal energy 27, 1750-1755 [2] Project GEOHELIOS, (2007). Dominguez-Energie and Laboratory ELIAUS, University of Perpignan [3] B.Sanner, C.Karystas, D.Mendrinos, L.Rybach, (2003). Current status of ground source heat pumps and underground thermal energy storage in Europe. Geothermics 32, 579-588. [4] C.Eklöf, S.Gehlin, (1996). TED-A mobile equipment for thermal response test. Master s Thesis, 198E, Lulea University of technology. [5] P. Eskilson, (1987). Thermal analysis of heat extraction boreholes. Doctoral thesis, Lund University, Sweden. [6] G. Hellstrom,(1991) Ground Heat Storage Thermal Analysis of Duct Storage Systems. Part I. Theory, University of Lund, Department of Mathematical Physics, Lund, Sweden. [7] H.S. Carslaw and J.C. Jaeger, (1959). Conduction of heat in solids. Oxford University Press, Oxford.