ROCK PROPERTY MAPPING FOR IMPROVED PLANNING OF GEOTHERMAL INSTALLATIONS

Transcription

1 PROCEEDINGS, Thirty-First Workshop on Geothermal Reservoir Engineering Stanford University, Stanford, California, January 30-February 1, 2006 SGP-TR-179 ROCK PROPERTY MAPPING FOR IMPROVED PLANNING OF GEOTHERMAL INSTALLATIONS Andreas Hartmann, Renate Pechnig and Christoph Clauser Applied Geophysics, RWTH Aachen University Lochnerstr Aachen, 52056, Germany ABSTRACT A major hindrance for the exploration of geothermal energy is the high risk of failure due to the unknown properties of the target rocks at depth. Usually, conservative assumptions are made about these properties, resulting in larger drilling depths and increased exploration costs. We systematically study hydraulic and thermal properties for the Molasse Basin, a region in Southern Germany, in order to improve this situation. This should allow building of better models of geothermal installations in the planning stage. INTRODUCTION In general, the ranges of thermal and hydraulic properties given in compilations of rock properties (e. g. Haenel, 1988; Clauser, 1995) are too wide to be useful to constrain properties at a specific site. To overcome the insufficient knowledge of thermal properties at depth, a project has been initiated to provide a database of thermal and hydraulic properties for the subsurface of Germany. An important aspect, in addition to obtaining new core samples and measurements, is the use of data from hydrocarbon exploration that are largely untapped for the purpose of geothermal exploration. Goals of the work are defined as follows: 1. Determination of thermophysical and hydraulic data for the stratigraphic units based on laboratory and borehole measurements. 2. Analysis of the dependence of the properties on pressure, temperature, and facies change. 3. Extrapolation of thermophysical and hydraulic properties to PT-conditions at target depth. For this purpose a combination of laboratory measurements and well log interpretation is used. Laboratory data provides the basis of the study by providing the petrophysical properties, measured on a representative collection of samples obtained from outcrops and boreholes in that area. Measured properties encompass thermal conductivity and diffusivity as well as porosity and permeability. Variations arising from temperature and pressure increase with depth are studied. Logging data add knowledge about the problem by developing relationships linking core and wireline measurements as well as delineating geographical regions of different petrophysical facies type. A major aspect is the use of data from abandoned hydrocarbon exploration wells, establishing an additional data source to constrain the problem. Figure 1. Map of the survey area (brown) in Southern Germany. Shown are locations for core sampling from outcrops (triangles) and boreholes (squares). This data base will help to reduce the risk of failure for geothermal installations as it will be possible to

2 base scenarios of reservoir development on improved data. show a strong increase in the beginning that can be attributed to pressure effects. SURVEY AREA The study is focused on South German Western Molasse region and the Jurassic and Triassic landscapes north of the Molasse (Figure 1). This area is predestined for a detailed due to its high technical potential for geothermal use (Erbas, 1999) and the general geologic situation of the area. The subsurface of the Molasse region is well known due to hydrocarbon exploration in the past. The northern part of the working area is characterized by a sequence of southward dipping Mesozoic rocks (Figure 2). This allows progressively older rocks to be sampled by moving north. Comparisons of Mesozoic samples taken from outcrops or shallow boreholes with those taken from the Molasse region enable to study PT-dependence and possible facies changes of the rocks. Figure 2. Simplified geologic cross section of the survey area (after Geyer and Gwinner, 1991). DATA AND METHODS A total of about 400 core samples provide the base of the study. About two thirds were obtained from core archives of the Geological survey of Baden- Württemberg, the University of Tübingen, and the Wintershall AG. These samples where complemented by cores taken from outcrops that complete the coverage. The geologic sequence sampled ranges from Buntsandstein to Miocene rocks. The following measurements were routinely performed in the laboratory on all samples: Figure 3. Example for the scanning measurement of petrophysical properties of core samples. Thermal conductivity, p-wave velocity, Density and porosity. The use of core scanning devices allows the rapid measurement of a large number of samples in a timely fashion (Figure 3). On subsets of the samples additional properties were measured: Specific heat capacity, Natural spectral gamma-radiation, Pressure dependent permeability, Pressure- and temperature dependent thermal conductivity, Mineralogical analysis. An example for the pressure- and temperature dependent measurement of thermal conductivity is shown in figure 4. The PT-conditions simulate a burial depth of the sample of up to 7 km depth. It is apparent that the decrease of thermal conductivity with increasing temperature is the main controlling factor for most of the samples. Two gypsum samples Figure 4. Dependence of thermal conductivity on pressure and temperature. Values are normalized to ambient conditions. For porosity and permeability the use of data from hydrocarbon exploration wells is particularly important. A large body of data exists that has been compiled in this and a previous study. The data have been used to calibrate a k-φ relationship to the sandstones

3 of the Western South German Molasse Basin (Figure 5). The relationship was originally derived for the North German Basin (Pape et al., 1999). The dataset for the Molasse Basin incorporates over 1000 porosity and permeability measurements performed on Tertiary and Mesozoic sandstones in the course of hydrocarbon exploration. Figure 6. Crossplot of thermal conductivity on a logarithmic scale versus slowness for laboratory measurements. Color coding represents measured porosity. Figure 5. k-φ relationship for Molasse basin sandstones compared to equations for the North German Basin (Pape et al., 1999) and Fontainebleau sandstones (Bourbie and Zinszner, 1985). T* denotes tertiary formations, J* Jurassic formations, and K* is Keuper. Analysis of wireline data is an additional important source of information. It serves two purposes. First, the variability of the petrophysical properties downhole can be better assessed based on wireline data compared to core data that might be subject to preferential sampling. Second, the large number of boreholes allows a better spatial characterization of changes in facies and according changes of petrophysical properties. Readings of wireline logs respond to the composition of the probed rock, its structure and environmental conditions. For the analysis of borehole geophysical data in terms of the quantitative description of the rock composition the assumption is made that a log reading responds mainly to the composition of the rock, given some appropriate mixing law. Using a standard inversion procedure (Doveton, 1979; Hoppie 1996) the lithological composition can be computed. In turn this data can be used to compute a thermal conductivity profile (Hartmann et al., 2005). It is first necessary to analyze the relationships between petrophysical properties in the laboratory. An example is given in figure 6, where thermal conductivity is plotted versus acoustic slowness for limestones of the Upper Jurassic. The triangle defines three end members, and any rock should fall in between the connecting lines. In this particular case, the rock model assumes the travel time average for the acoustic slowness (1) and the geometric mixing law for the thermal conductivity: (2) The model can be transferred to the borehole and used to infer thermal conductivity from geophysical logs, using two different logs, for instance Gamma- Ray (GR) and acoustic slowness (DT) (Figure 10). If temperature logs are available, these can be included in the analysis (Figure 7). This is particularly useful because temperature is of course the most sensitive measurement regarding variations in thermal conductivity. Because standard inversion procedures for wireline data cannot be used for the analysis of temperature data, a modified version was developed that is able to handle petrophysical logs as well as temperature logs (Hartmann, 2005a, 2006). Figure 7 shows an example of such an inversion. The sequence consists of Tertiary sandstones and marls down to a depth of about 1050 m. Jurassic limestones are found below that depth. The sand fraction at around 1140 to 1200 m depth is an artifact due to poor data quality. Here, density decreases unrealistically because of fracturing of the limestone. The anomaly in the temperature gradient indicates advective heat transport in these fractures.

4 Malm delta. These show only minor variations in thickness and appearance. However, units Malm zeta and epsilon show considerable variability, due to changes between reef and massive facies. Also, in the northern boreholes the topmost subunits will be eroded. The statistic variations of thermal conductivity for three of the different subunits are shown in figure 9. In comparison with figure 8 it is apparent that the varying shale content of the limestones is the main control on the thermal conductivity of the rocks. Porosity (see also figure 10) plays only a minor role, as it is small and has only minor variations. On average Malm epsilon has the highest values of thermal conductivity. Values above 3 W (m K) -1 are rarely attained. This is well below the range given for instance in Clauser (1995) for the thermal conductivity of calcite, highlighting the importance of work to characterize these properties on a regional basis. Figure 7. Example for the joint interpretation of temperature logs together with other petrophysical logs. Shown are Slowness t (µs ft -1 ); Bulk density ρ b (kg m -3 ); Thermal gradient T (K m -1 ); Gamma-ray GR (API). Color coding of lithological column: Shale grey; Sand yellow; Limestone blue; Porosity blank. GEOLOGIC INTERPRETATION Following the petrophysical work, the distribution and facies of the relevant rock units are analyzed. As an example this will be discussed for the Upper Jurassic. The Upper Jurassic consists primarily of limestones with interspersed beds of marlstone. The reference profile of the sequence is shown in Figure 11 together with Gamma-ray logs from boreholes in the survey area. Shown are both, boreholes in the central part of the Molasse basin as well as boreholes in the Swabian Alp where the Upper Jurassic outcrops. It is apparent that the general sequence is similar for these two regions. A correlation of subunits is possible using the Gamma-Ray log for the lower units, Malm alpha to Figure 8. Statistic variation of the thermal conductivity for three subunits of the Upper Jurassic. Shown are quartiles (blue boxes), median (red lines), range (maximum 1.5 x interquartile range, black whiskers), and outliers (red crosses). Using the data from a number of boreholes, the results can be analyzed with respect to their dependence on locality and depth. This is shown in figure 9 a) and b) for the Upper Jurassic. In these plots statistical information on thermal conductivity and porosity is plotted versus the mean depth of the Upper Jurassic. For thermal conductivity (Figure 9a) only a minor variation with depth can be observed, mainly for shallow boreholes below 1000 m. For the porosity, a modest trend to lower values for increasing depths can be observed (Figure 9b). However, this trend mainly reflects a decrease in variability rather than in mean value of porosity.

5 Clauser, C., and Huenges, E. (1995), Thermal Conductivity of Rocks and Minerals, in Ahrens, T. J. (ed.), Rock Physics and Phase Relations - a Handbook of Physical Constants, AGU Reference Shelf, American Geophysical Union, 3, Doveton, J. H. and Cable, H. W. (1979), Fast matrix methods for the lithological interpretation of geophysical logs, Computers & Geology, 3, Erbas, K., Seibt, A., Hoth, P. and Huenges, E. (1999), Evaluierung geowissenschaftlicher und wirtschaftlicher Bedingungen für die Nutzung hydrogeothermaler Ressourcen, Scientific Technical Report, GeoForschungsZentrum Potsdam. Geyer, O. and Gwinner, M., (1991), Geologie von Baden-Württemberg, E. Schweizerbart sche Verlagsbuchhandlung, Stuttgart. Haenel, R., Rybach, L., and Stegena, L. (ed.) (1988), Handbook of terrestrial Heat-Flow Density Determination, Kluwer Academic Publishers. Figure 9. Variation of petrophysical properties with depth for the Upper Jurassic. A) Thermal conductivity, B) Porosity. CONCLUSION The results of the combined analysis of laboratory measurements and wireline data provide base petrophysical properties for rocks from the Buntsandstein to the Miocene epoch in South Germany. In addition to the statistics our work also provides information on the main controlling factors for the variation of these properties. This information is particularly important. It will allow a more rational choice of rock properties in the design phase of a geothermal installation because information on the site-specific geologic situation can be translated into constraints that narrow the possible bounds for petrophysical properties of the rock. REFERENCES Bourbie, T., and Zinszner, B. (1985), Hydraulic and acoustic properties as a function of porosity in Fontainebleau sandstone, Journal of Geophysical Research, 90, Hartmann, A., Rath, V., and Clauser, C. (2005), Thermal conductivity from core and well log data, International Journal of Rock Mechanics & Mining Sciences, 42, Hartmann, A., Rath, V. (2005a), Simultaneous inversion of temperature and other wireline logs, presented at the 2 nd General Assembly of the EGU, Vienna, Austria, April. Hartmann, A. (2006), Inversion of geothermal parameters using borehole and core data, Ph.D. thesis, University of Aachen. Hoppie, Bryce W. (1996), High-resolution lithologic characterization of sequences on the New Jersey margin slope through inversion of leg 150 logging data for lithologies, in Mountain, G. S., Miller, K. G., Blum, P., Poag, C. W., and Twitchell, D. C. (ed.), Proceedings of the Ocean Drilling Program, Scientific Results, Texas A&M University, 150, Pape, H., Clauser, C., and Iffland, J. (1999), Permeability prediction for reservoir sandstones based on fractal pore space geometry, Geophysics, 64(5),

6 Figure 10. Example for the computation of a continuous thermal conductivity profile from wireline data, based on the model of figure 6. Input data (GR and DT) are shown in track 2. Track 4 shows the computed lithology and track 3 the computed thermal conductivity (TC) and sonic porosity (PHIDT), together with a flag for clean limestone (V SH < 0.2). Figure 11. Reference profile of the Upper Jurasic (Malm) in Southern Germany compared with Gamma-Ray logs from hydrocarbon exploration wells. Units alpha to delta can be well correlated, whereas units epsilon and zeta show considerable vaiability.