Proceedings World Geothermal Congress 2010 Bali, Indonesia, 25-29 April 2010

1

Elaboration of Multi-Scale Fluid Flow Modeling System in Fractured Rocks for Exploitation of Geothermal Energy

(Revised, extended version)

János Viszkok, Tamás Földes, Tivadar M. Tóth, Éva Kun, István Gyenese

Mária út 10, H-5000 Szolnok, HUNGARY

jviszkok@centralgeo.hu, t.foldes@t-online.hu, mtoth@geo.u-szeged.hu, kuneva@freemail.hu, gyenese.istvan@upcmail.hu

Keywords: EGS, multi-scale fracture flow, fracture modeling, computer tomography

ABSTRACT

The Pannonian Basin in Hungary represents one of the areas of highest geothermal energy potential in Europe. The basin is a deep Neogene depression, surrounded by the Carpathian Mountains, filled at places by more than 6000 m of Miocene – Pliocene sediments. These are underlain by pre-Neogene metamorphic crystalline basement rocks.

In order to efficiently and sustainably exploit this geothermal resource, both water and heat balances need to be fully understood. One component of this understanding is knowledge of the geothermal fluid (i.e. hot groundwater) flow regime in the fracture systems of the underlying basement. This, in turn, requires information on the tectonic framework that has shaped these fracture systems and controls their permeability distribution. Unfortunately, direct study of the crystalline basement is impractical and glimpses of the basement can only be caught in deep exploration boreholes.

Heat transport and flow modeling process was tested and found useful as a standard workflow to evaluate, to compare and to rank the geological/ hydrogeological potential of different geothermal prospects.

1. INTRODUCTION

A recently published studies financed by EU FP6 ENGINE (Dövényi and Rybach, 2006) confirmed that the Pannonian basin has a great potential for establish EGS systems. There are several fractured fluid reservoirs in the metamorphic basement of the Pannonian Basin, but their research is rather problematic, because the crystalline bodies are inhomogeneous both lithologically and structurally. The main aim of the present pilot study using a dataset of an abandoned HC reservoir is to combine diverse approaches for better understanding a typical metamorphic aquifer: apply and test relatively new tools; establish a standard evaluation workflow to assess, to compare and to rank different EGS prospects; determine a standard workflow for the prospect evaluation; help geothermal engineers and decision makers or investors.

The methods involved taking into account the multi-scale concept are as follows (some overlappings in scale are enabled): core scale: - classical petrology and structural geology for rock

classification; - Computer Tomography to understand internal structure

of core samples as well as measure fracture network geometric parameters;

- rock mechanics to measure deformation parameters of diverse rock types;

well scale: - electric logs (CBIL measurement) to follow fracture

intensity along selected wells; - fracture network simulation to upscale geometric

information; - well-tests and interference tests; field scale: - seismic section and attribute evaluation to identify large

scale tectonic structure and combine small vs. large scale data;

- flow and heat transport modeling.

The information acquisition and interpretation was focused on the input data claim of the heat transport and flow modeling: geometry/structure; attribute model and boundary conditions.

2. GEOLOGICAL SETTING The study area (Fig. 1) is the western flank of a large metamorphic dome uplifted during the Neogene. The dome itself consists of different high grade metamorphic rocks, essentially of diverse gneiss and amphibolite varieties. At the deepest structural position an orthogneiss block can be found, which at places contains large enclaves of mafic composition. Because of rheological reasons, the enclaves are more intensely fractured than the host gneiss. Above the orthogneiss zone a paragneiss zone follows along a wide shear zone. Previous studies suggest that this gneiss type is hardly fractured. Along a next shear zone a medium grade amphibole gneiss follows, which is the most intensely fractured among all typical lithologies.

Tisz

a

Dan

ube

BÜKK

VIENNABASIN

TRANSYLVANIANBASIN

WESTERN CARPATHIANS

HUNGARY

MECSEK

SOUT

HERN

CARPATHIANS

Lake Balaton

EASTERNCARPATHIANS

Papuk

Dinarides

Eastern A

lps

Bohemian Massif200 km

PANN

ONIAN BASIN

APUSENI

Figure 1: The study areas

However, in other parts of the world, similarly fractured bedrock aquifers, shaped by analogous extensional tectonic regimes, can be observed in the shallow subsurface. In particular, study of the rocks of the Oslo palaeograben (Norway) offer insight into the fracture systems likely to be

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

2

found in the sub-Pannonian basement. Fortunately, the granitoid basement rocks of south-east Norway are found at the surface, exposed by glacial activity and post-Tertiary erosion, and may be examined in detail and used an analogue for understanding of deep rocks.

The Pannonian basin is characterized by high heat flow (90-120 mW/m2) causing high geothermal gradient (50°C/km) in contrast to the surrounding areas, where the heat flow is normal. The high heat flow caused by lithospheric stretching can be attributed to the tectonic processes related to the formation of the basin (Horváth, F., et al, 2004).

On this area, 130°C – 160°C was measured between 2300 and 3200 m depth.

Figure 2: Lithospheric thickness beneath the Pannonian basin

3. METHODS For establishing a method for evaluation and ranking of potential site for enhanced geothermal system we elaborated two Hungarian depleted hydrocarbon fields. The multi-scale aspect was approached from high-scale core measurements (rocks mechanics, using the most up-to-date computer tomography technique and the rupture tests) through well-scale pressure development and production test to low-scale “field-sized” seismic interpretation and interference test. To simulate the fluid flow in fractured rocks first a sophisticated fracture modeling was used, then this fracture system was mounted in the flow and heat transport model. Using these fracture flow models the aptitude of the site for EGS could be tested and evaluated.

3.1 Rock mechanics

To simulate the fracturing, twenty undeformed samples were picked up for rock mechanical measurements. These cores represent all diverse rock types of the reservoir. The cylindrical samples were tested under various uniaxial loading conditions. In order to increase the damage, accumulation and avoid violent rupture an irregular cyclic loading was applied were the peak compressive pressure was about 70-80% of the uniaxial compressive strength. The axial (and lateral) deformations were measured. The characteristic rock parameters are the following (Table 1).

Table 1

Uniaxial compressive strength [MPa]

Young’s modulus E [GPa]

Poisson coefficient

[-] 19,18 18,17 0,16

The pressure-axial deformation diagram of a selected sample (Fig 2) exhibits large hysteresis loops with small softening effects. The accumulated permanent strain at zero pressure indicates considerable accumulated damage. The sample rested relatively intact after the test (M. Tóth et al., 2008).

0.05 0.10 0.15 0.20 0.25

Εa @%D

5

10

15

P @MPaD

50 100 150 200 250

t

5

10

15

P

Figure 3: Compressive pressure as a function of axial deformation of the cylindrical sample.

(The inset shows the time dependence of the applied pressure. The sampling rate is 10Hz.)

3.2 Computer Tomography

The goal of Computer Tomography (CT) measurements and interpretations was the characterization of core samples concerning petrophysical assessments, determination of microtectonic pattern and structural parameters of fracture networks in 3D. According to saturation profiles, two main types of fracture rock samples can be distinguished. Some samples behave as a breccia, while most of them are fractured (micro-fractured, open fractured, combined type; Fig. 4) (Földes et al., 2004).

Results of the uniaxial mechanical crush tests were combined with CT evaluations. Based on these results it was possible to determine the ability of fracturing in case of the different lithologies. On the CT images the running and ending of fracture network in each zone could be observed and mapped. The results were compared with the mechanical measurement evaluations and the petrophysical parameter alteration was compared to those detected prior to crushing (Fig. 5).

Type CT image + porosity map Eff volume, Por (based on total volume) , Por (based on effective volume)

50.55 % 4.39 % 8.61 %Breccia

Fractured I

Micro fractured

Apart surface

Fractured II

Mixed

43.2 % 4.59% 11.18 %

32.68 % 1.74 % 5.24 %

28.6 % 1.57% 5.24 %

47.12 % 3.48% 7.4 %

40.9 % 5.17% 12.24 %

Figure 4: Damaged rock classification based on CT

image evaluation

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

3

CT image Generated FracturesCompressive pressure as

a function of axial deformation

Lithology 1

Lithology 2

Figure 5: Comparison of CT images and stress-strain

diagrams for different rock types

3.3 Well-log

The wells drilled in Hungary for oil exploration are logged with different probes: POL, PLR, PLH, ML, MLL, SP, TG, NG, ATL, CEBL, CETEL, Diplog, CBIL, BHTV, etc. These data were used for porosity, permeability estimation and layer/ fracture identification along the borehole.

3.4 Well & Interference tests

Pressure Build-Up Measurement

The overwhelming majority of reservoir parameters can be determined from pressure build-up measurements. In case of pressure build-up measurements in ideal reservoirs (fractured-cavernous carbonates) we can expect the best results if we use electronic memory manometers with a resolving power of 70 Pa, because of the high over-all permeability. The processing of the enormous quantity of data recorded by electronic manometers demands modern softwares. The Pan System program package used for the interpretation contains a selection of the evaluation procedures deemed most necessary and useful from the published supply throughout the world. The joint application of memory manometer and software gives the flow parameters of both reservoir and the near-well zone.

From the pressure build-up test of a high-temperature well one can obtain information about heterogeneities of the reservoir – fractures, linear flow and reservoir boundary – at distances from well ranging up to several hundreds of meters. It is reasonable to have pressure build-up testing done in each well of the producing/injecting system. This is the basis also for planning and executing pulse tests. In Fig.6, the log-log (diagnostic evaluation results of a pressure build-up data (PBU) series, measured in thermal water well are shown (evaluation model: radially homogeneous medium of infinite extent).

Figure 6: Log-log curve of a pressure build-up data series evaluated with Pan System software

Capacity (Productibility) Tests

After successful perforations, the execution of capacity test at least 4 cycles is advisable. On their basis the production rate equation of the given well can be defined; under given limiting conditions this can give the ’produced fluid rate – bottom pressure under production’ relationship. Capacity tests start usually with shut-in well state, and are performed in steps of gradually increasing rate of production. Recorded temperature changes during properly chosen production rates and rate changes can give information about heat conductivity at given depths of the well, and wellhead temperature data make possible calculations of heat losses along the whole well equipment.

Pressures and pressure changes during re-injection may – even in case of fractured reservoirs – differ from those during production. Such differences in injection wells can be investigated by means of, and their reasons concluded from, capacity and pressure drawdown tests made during injection. Temperature against temperature differences measured while injecting constitutes the basis of planning comprehensive measures for heat-household managing.

Pulsation Interference Measurements

On the basis of pressure build-up tests in production and injection wells the distance of flow heterogeneities from well axes can be calculated, their impacts estimated, but their setting (orientation) is uncertain.

As for hydrodynamic connection of production and injection wells and its extent, the pulsation interference tests are capable of providing that information. The measurements with high-resolution pressure gauges have shown that the measured pressure contain noise components, the order of magnitude of which is about the same as that of the interference signals to be expected. We measure the pressure resultant of overall pressure changes within the reservoir, of atmospheric pressure and temperature, and of the tidal effect of Earth’s crust. The generation of pulsed signals, meaning equal production and shut-in time periods, filters in itself the overall pressure changes; however, only the evaluation method with noise filtering – which can detach stochastic changes – has significantly improved evaluability: by experience in Hungary: from 17 to 78 per cent. The evaluation results of the test are: transmissibility (T = kּh/µ) and storage (S = Φּctּh). In Fig. 7, the evaluation results of a thermal water well-pair with PulsEx software (noise-filtered) are presented. Twofold transformation of the measured pressure data series was executed. The first transformation

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

4

filters out higher-frequency noises, the second will drop out those having larger time constants. By the least squares method we have found the theoretical curve having the best fit to the filtered pressure changes (automatic fitting); from its parameters we calculated transmissibility (T = 4.20ּ105 µm2ּPsּs) and storage (S = 2.48ּ10-2 m/MPa). Distance between wells, a = 1508 m. (Pulsation-related change of the production rate, ∆q = 2100 m3/d; amplitude of the pulsed pressure wave, ∆p = 2.80 kPa).

By comparing for well-pairs (a) the transmissibility as calculated from the pressure build-up curves, and storage as calculated from ’working’ (= effective) bed thickness and porosity from geophysical (electrical etc.) well logs with (b) the same parameters as evaluated from the pulse tests, one can draw conclusions about the heterogeneity of the reservoir. From the point of view of geothermal energy production it is very important to know the existing flow pattern, because if outstandingly high transmissibility was found between the two wells, the cooling effect of the cooled-water swallowing injection well can quickly appear also at the production well (Árpási et al., 1999).

Figure 7: Evaluation of a pulse test executed on a pair of thermal water wells with PulsEx software (noise-filtered evaluation)

3.5 Seismic interpretation

Seismic attributes give an excellent possibility to build reservoir models. Calculation of an attribute applied in our case uses Semblance coefficient (around a given channel) and Manhattan distance (between target channels). Values of the coherency depend on the number of the neighbor channels and the range of the time window. This choice of coherency shows significant anomalies around the productive perforations. According to our experiences, brecciated/fractured zones indicate coherency anomalies in the metamorphic basement (Fig. 8). Productive intervals are generally found near (or inside) these zones (M. Tóth et al., 2006).

Figure 8: Model porosity in the metamorphic reservoir based on seismic attributes

3.6 Fracture network simulation

Fracture intensity using CBIL data were evaluated in two different ways. On the inter fracture distance curves the most deformed intervals become visible. These zones usually show the position of shear zones between neighboring metamorphic blocks. Fractal dimension of fracture intersection series was also calculated and used to build DFN (discrete fracture network) models around the wells. Results of simulation confirm that communicating fracture network exists along the shear zones as well as the amphibolite rich segments of the basement (M. Tóth et al., 2004). Both the orthogneiss and the paragneiss is essentially impermeable (Fig. 9).

-2700 m

-2800 m

-2900 m

-3000 m

-3100 m

Figure 9: Fracture simulation

3.7 Fluid flow and heat transport modeling

General introduction to flow modeling: A model may be defined as a selected simplified version of a real system and phenomena that take place within it, which approximately simulates the system’s “excitation-response” relations that are of interest (Bear, 1995). With this definition, all that a model can do is to predict the future behavior of an investigated system, in response to the implementation of prescribed activities. Thus, the model provides the excitation-response relationship of the considered system. In this sense, the model is just a tool to find answer for a

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

5

given problem, to simulate a process, to examine or to predict an effect of a parameter in the system.

The model is the simplified and symbolic imitation of the selected features of the reality. A fundamental condition for the formulation of the model is that the depth of simplification of the real system is known and formally described (established). A reliable model is complicated enough to be realistic and simple enough to be solved and interpreted.

Although a model is only the poor simplified copy of the reality, it has some unsurpassable advantages in the theoretical and practical exploration. The knowledge inserted into the model restricts the professional intuition and prevents endless theoretical discussions. Modeling provides an inexpensive and fast tool to reproduce situations of un-limited variant numbers. Thus, a unique opportunity is given to test hypothesis, to examine effects of each selected parameter on the simulated process, to distinguish important and less important parameters and to predict possible events and their influence in the future (Király, 1996).

The modeling process can be divided into four phases (Fig.10):

The first phase includes the identification and formulation of the problem and the information required to realize the model. This process is an abstraction of knowledge based on the reality; to collect, to select, to check the available and usable data.

The second phase is the development of the conceptual model (it can be called abstract scheme or schematical model/representation, too). In this step, the physical, thermo-dynamical and chemical laws, which describe the flow system in time and in space are compiled.

The real system and its behavior may be very complicated, depending on the amount of details we wish to include. We need to simplify the description of the considered system and its behavior somewhat that will be useful for the purposes of solution of the given problem.

These simplifications are introduced in the form of a set of assumptions that expresses our understanding of the nature of the system and its behavior, as far as the considered problem is concerned.

Because the model is a simplified version of the real system, no unique model exists to describe it. Different sets of simplifying assumptions will result in different schematic models, each model is approximating the considered domain and it behaves in a different way. So, each abstract scheme is good or not only in respect of the answer of the given problem.

The selection of the abstract scheme for a particular case depends on two main factors: 1. The objectives of the investigations, i.e., what kind of

information is the model required to provide for the best solution.

2. The available resources (including input parameters, expertise, hardware and software) to construct and solve the numerical model.

The conceptual model for the description of fluid flow:

[S ∂h/∂t + div (-K grad h) = Q] requires the following dataset:

- Geographical topology of all the objects of the model, - Physical field of permeability (K field), - Physical field of storativity (S field), - Boundary condition of h(t) or Q(t), - Initial condition of h(t=t0). Building up an appropriate conceptual model is the most important and most time-consuming step in the modeling process.

The third phase is the numerical modeling or mathematical simulation: the studied rock volume is subdivided in smaller units, and the hydrogeological equations with the chosen initial and boundary conditions in the nodes of elements are solved. In this phase, we get the consequence of our hypothesis. The mathematical simulation includes the software, the used database (measured and hypothetical values, as well), the presentation of the results.

The fourth phase is the analysis of the results of the numerical simulation. Here, the consequences of the hypothesis, the conceptual model and the application of the mathematical simulation to dataset are verified. The result will not be identical to the conceptual model, but within a desired accuracy the hypothesis can be confirmed or refused. In the case of the rejection, the abstract scheme on basis of the comparison between the reality and the results derived from the mathematical simulation have to be modified (Viszkok, 2000).

Figure 10: The main phases of modeling process

If the result is refused, the following questions will have to been answered: - Is the constructed conceptual model appropriate to

exhibit the features of the processes that take place within the real problem domain that are of interest to us to an acceptable degree of approximation?

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

6

- Are the values of the various parameters correct, reliable and sufficient?

- Are the selected boundaries appropriate? - Are the conditions assumed to prevail within domain

boundaries correct? Building up our pilot model: The data referring different rock volume (Section 3.1 – 3.6) can be used in different part of the second phase of the modeling process. As presented above a special emphasis was put on the quality and authenticity of the input data of the model because the result will depend on them. Generally, the core scale data (fracture pattern, permeability tensor, conductivity) were used to determine the “attribute model” (K values) while field scale data specified the model geometry and boundary conditions.

Figure 10: Model geometry

For fluid flow and heat transport simulation, the well-known and widely accepted FEFLOW v5.4, – a monophasic, finite element code – was used.

In the model 68’184 3D triangle prisms some 2D elements and 40’239 mesh nodes can be found. The model area is 1600 km2.

The model boundary was determined on the basis of a main structural line of the basement to the North, two less regional faults to West and to South. An artificial line was chosen to the East boundary of the model. The first three boundaries were inferred as no-flow boundary conditions, while on the east side some extra inflow was added. The main interesting area is quite far from boundaries to minimize the effect of the uncertain boundary conditions. The mesh geometry follows the inner geological structure, the expected hydraulic gradient and the well locations (Fig. 10). The applied triangulated prism mesh is well suitable to make a replica of even more complex geological structure. At the mesh design, the fault system revealed by seismic was taken into account, too. Later, during the simulation scenario, vertical 2D elements were inserted between two 3D prisms along the fault line in the basement layers representing the faults and tested its effect on the flow pattern (Fig. 10 and Fig. 11). 9 horizontal surfaces (slices) and 8 layers were implemented to the model, which correspond to the topography Quaternary and Pliocene (Upper-Pannonian) coarse grained continental clastic sediments – 2 layers, Upper-Miocene (Lower-Pannonian) and Middle-Miocene fine grained sediments – 2 layers and 4 layers of metamorphic rock types, from top to bottom respectively. The model bottom is a plane at -6000 m bsl.

The basement is build with low grade amphibole gneiss (AG), sillimanite biotite gneiss, paragneiss (SG) and orthogneiss (OG) with mafic xenoliths and granite aplites (Fig. 11) (M.Tóth et al., 2008).

Figure 11: Mosaic-like structure of the basement

The preliminary studies are confirmed that the lithological units can be characterized different fracture pattern (M. Tóth and Zachar, 2006) and the different fracture patterns (Földes et al. 2004) represent well separated hydrogeologic units having different hydrodynamic features as permeability, connectivity, storativity, porosity, etc.

The producing well situated on a better fractured amphibole-gneiss zone – a more permeable “pocket” in a less permeable matrix – straddled by lesser fractured sillimanite gneiss as presented in Fig. 11. This pocket can prevent huge volume of water loss, direct the injected water toward the producing well and assure the heat exchange between rocks and water.

Simulation: After calibrating the model a scenario was set up to investigate the effect of re-injected water in different distance from the producing well.

The input flow parameters were adjusted are as follows: - Transient flow and transient heat transport, - 6000 m3/day production/injection of water using

multilayer option at layer 5; - porosity values are between 2% and 0.2%; - K values can be isotropic or anisotropic (vertical

conductivity is higher than the horizontal one) and altering according to the scenario between 1x10-8 m/sec and 2 x10-4 m/sec (Fig.12);

- time span is 50 years, applying automatic time step control via predictor-corrector scheme;

- error tolerance 1x10-3 applied to L2 integral (RMS) norm with maximum 12 number of iterations per time step;

- initial time step length is 0.001 day;

and the initial heat parameters are as follows: - at the top of the model (topography) is 11°C as first

kind boundary condition;

injection well

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

7

- at the bottom -10’400 J/m2/d heat flux as second kind boundary condition;

- fluid viscosity is heat-dependent; - the initial temperature values by slice are 11°C, 50°C,

80°C, 100°C, 130°C, 150°C, 160°C, 180°C, 200°C, respectively, calculated from the regional geothermal gradient;

- the injected water temperature is 20°C; - other parameters offered the software are accepted due

to the lack of information.

Figure 12: Hydraulic conductivity (K value) on a section

Results: The range of the injection would not be bigger than 1 km. The change of hydraulic head on slice 5 is shown on Fig. 13. The length of the arrows is proportion to the flow velocity.

Figure 13: Hydraulic head around the production / virtual injection well pair (5. slice)

Figure 14: Hypothetic hydraulic head increment in the injection well versus time

The Fig. 14 presents a hypothetical hydraulic head increment due to cooled down water injection against time. After 500 days, the effect of injection, causing ~ 30 m of hydraulic head increase is stabilized.

4. DISCUSSION

Before the discussion, the limits of the model have to be mentioned. The results of numerical simulation are based on the actual geological, hydrogeological information of the research area. Any new data, which are in contradiction with this knowledge, require the re-structuring of the model and restarting the calibration and simulation process. The results can only be accept if the input data are free of contradiction and suit the reality on the level our actual knowledge.

The results of simulation can be represented in graphical or numerical way, but it is not an easy task to choose the most suitable cross-sections or/and slices for showing the most important and most characteristics features of flow and heat transport in the most adequate time span. Furthermore, they had to be compared to the results of other parts of the scenario.

Up to now, the scenario was concentrated on the fracture isotropy and conductivity values:

1.) Anisotropic case: Kxx= 0.044e-4 m/s, Kyy=0.004e-4 m/s, Kzz=0.0841e-4 m/s

2.) Isotropic case1: Kxx=Kyy=Kzz= 0.0841e-4 m/s

3.) Isotropic case2: Kxx=Kyy=Kzz= 0.0044e-4 m/s

Here, the temperature distribution presented along a line through the injection well in the case of scenario 1.) and 2.). The section of the anisotropic case (1.) was very similar to the isotropic case2 (3.). On the section the place of the injection well is 360 m and that of the producing well is 620 m. On the slice 5 (Fig. 17), 2 zones of the effect of injected 20 °C of water can be seen, one is the “washed zone” where the temperature is quasi equal, while the other one is the “transition zone” where fast change of temperature can be observed. The washed zone is asymmetric due to the fluid flow direction; ~160 m towards the producing well and ~60 m to the opposite direction. The transition zone is limited to ~110 m toward the producing well, while it is ~200m wide in the opposite direction. Just a few degree of temperature decrease can be seen at the producing well, which is within the numerical error.

Figure 15: Anisotropic case: Temperature (°C)

distribution along a line on slice 5

producing well

injection well

producing well

injection well

500 m

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

8

Similar features can be seen on the isotropic (case1) scenario (Fig 16. and Fig 18.), too. The size of the “washed” and the” transition” zone is a little bit different.

Figure 16: Isotropic case1: Temperature (°C) distribution along a line on slice 5

On the slice 6 (Fig. 17 and Fig. 18), the washed zone is disappeared and fast change on temperature is displayed. The temperature decrease in the producing well is detectable (150°C -> 145°C and 130°C).

Figure 17: Anisotropic case: Temperature (°C) distribution along a line on slice 6

Figure 18: Isotrop case1: Temperature (°C) distribution along a line on slice 6

The Fig. 19 and Fig. 20 show the 3D representation capabilities of the applied software: the heat distribution can be seen of the full model (Fig. 19) displayed with finite elements and on an arbitrary selected cut-off section displayed with wells.

Figure 19: Isotropic case1: Temperature (°C) distribution on the full model body

Figure 20: Isotropic case1: Temperature (°C) distribution along a cross section cut from the model body, representing the wells

5. CONCLUSION

A former (preliminary) study (M. Tóth et al., 2008) showed that in a fractured, depleted HC reservoir hydrodynamic parameters are under lithological and structural control. Among the lithologies, the amphibolite is more porous than the gneiss types. The fracture flow parameters and patterns resulted from preliminary investigations, regarded as the input values of flow and heat transport modeling.

The size and thermo/hydrogeological characteristic of a better-fractured zone, called “pocket” has strong impact on the aptitude of the planned EGS system (water loss, heat exchange, etc).

The applied new examination and interpretation tools (CT measurements and evaluation, fracture modeling, etc)

producing well

injection well

producing well

injection well

producing well

injection well

J. Viszkok, T. Földes, T. M. Tóth, É. Kun and I. Gyenese

9

combined with the usual ones are capable to generate more reliable and more precise data.

The range of geothermal effect is limited to ~ 4 km2, so the model area could be decreased. The water injection and production produce asymmetric heat and hydraulic head change, where a “washed” and a transition zone can be distinguished.

The injection (6000 m3/day) makes the hydraulic head change is surprisingly fast (one and half year) and causes 30 m of increase.

The heat flow is extremely sensitive the change of hydraulic head. During the calibration process the error tolerance was considered 1 m at hydraulic head and 3 °C at temperature. Of course, the results can be fined more with the change of high-scale parameters, but this type of fine adjustment is just “virtual”, because it is not based on real data.

Well chosen heat transport and fracture flow modeling methods can constrain the arising hypothetical ideas by their strict formulation and identical data claim. Their results can be regarded as the consequence of the ideas and they are comparable by numerical and graphical way, as well.

A depleted fractured reservoir − using its valuable data and existing infrastructure − is capable for establishing EGS system in the Pannonian basement.

The presented modeling workflow is suitable to evaluate and to rank geothermal prospects from geological point of view and is a good basis of the consecutive feasibility study.

6. ACKNOWLEDGMENT

This paper is a synthesis of a series of preliminary studies, wide range of investigation and profound team-work. The indicated authors express their highest gratitude to all, non-mentioned team members and contributors. This paper and the poster presentation have never been appeared without the financial help of EEA and Norwegian Financial Mechanism.

REFERENCES

Árpási, M., Gyenese, I., Megyery, M. (1999) Hydro-dynamic tests in well-pairs producing geothermal energy. Kıolaj és Földgáz, May 1999. p.129.

Bear (1995) (1995) Modeling groundwater flow and contaminant transport. Web page: http://w3.technion.ac.il/ ~cvrbear

Dövényi, P. and Rybach, L. (2006) EGS prospects in Hungray. ENGINE. Launching conference – Orléans, France

Földes, T., B. Kiss, G. Árgyelán, P. Bogner, I. Repa, K. Hips (2004) Application of medical computer tomograph measurements in 3D reservoir characterization. Acta Geologica Hungarica, Vol.47/1, pp-63-73.

Horváth, F. et al. (2004) Atlas of the present-day geodynamics of the Pannonian basin. OTKA registration number: T034928

Király (1996) Introduction a l'utilisation des modeles en hydrogeologie - Bases théoriques a l'intention des hydrogéologues. Cycle postgrade en hydrologie et en hydrogéologie, Université de Neuchatel, Switzerland

M. Tóth, T., Hollós, Cs., Szőcs, É., Schubert, F. (2004) Conceptual fracture network model of the crystalline basement of the Szeghalom Dome (Pannonian Basin, SE Hungary). Acta Geol. Hung., 47/1: 19-34.

M. Tóth, T. and Zachar, J. (2006) Petrology and deformation history of the metamorphic basement in the Mezısas-Furta crystalline high (SE Hungary). Acta Geol. Hung., 49/2, 165-188.

M. Tóth T., Földes T., Kiss B., Vásárhelyi B., Viszkok J. (2008) Complex evaluation of the Mezısas-W fractured metamorphic HC reservoir (Pannonian Basin, Hungary), Poster presentation, 70th EAGE Conference & Exhibition — Rome, Italy, 9 - 12 June 2008

Viszkok (2001) Subsurface fluid flow simulation with finite element method in the East Pannonian Basin. Terre et Environnement, Vol. 31. Université de Genève.

The section break that follows the last words of the paper will cause the columns to be even.