Author: Ana Kamenski, MSc.
Proper assessment of the distribution of lithological composition in the subsurface is one of the key elements when evaluating the hydrocarbon potential of an area, as well as geothermal potential and possibility for the CO2 geological storage. Spatial definition of the distribution of lithological composition is only one of the steps in the characterization of the subsurface.
The data obtained from the exploration of surface outcrops (hard data) and the subsurface characterization (very little hard data is available, e.g. core material) are conditionally compatible. The lithological composition in the inter-well area is conventionally evaluated on the basis of data obtained from the surrounding wells (cuttings, cores, logs) using either the conventional lithofacies mapping approach [1] where interpretation depends solely on the experience of the interpreter, or by making use of mathematical algorithms [2]. Such procedures have high dose of uncertainty in regional surveys where the wells are very distant from each other and irregularly distributed, and comparatively smaller uncertainty in areas with hydrocarbon reservoirs where there is a large number of relatively closely spaced wells. Following the trend of technological development, it is needed to turn to mathematical and statistical tools to eliminate subjectivity when interpreting lithology, although general understanding of the geology is always invaluable [3]. In every subsurface exploration, one of the most important assignments are determining key factor—age, structural settings and lithology [4]. These have a very large influence on scientific results, as well as economic implications if the results are applied to any type of resource estimates.
The purpose of this paper [5] was to analyse the data using both geostatistics and geological knowledge as objectively and realistically as possible. For this purpose, the area of the depleted oil field (Figure 1) was selected for the process, which is located within the Drava Depression, and belongs to the Croatian part of the Pannonian Basin (northern Croatia).
Figure 1 Pannonian Basin System and surrounding tectonic and geographic units with outline of the North Croatian Basin and study area with well locations [6, 7, 8]
This object was chosen due to the available data for lithology interpretation in the wells and 3D seismic coverage needed to define the lithological composition throughout the seismic volume. Clastic Pannonian interval (CPI) was selected for the analysis (Figure 2, 3) as the lithology of this unit can be generalized to three classes—sandstones and marls that occur through the whole interval and coals that are most often found in the top of the interval.
Figure 2 Schematic representation of stratigraphy, lithology and major tectonic events in the surroundings of the exploration area [9]
Figure 3 Mapping model boundaries on classic seismic profiles [5]
The lithological composition of the subsurface (Figure 4) is simplified in accordance with the general geological composition of the Pannonian age sediments in the research area.
Figure 4 Interpreted and upscaled lithology classes within four wells [5]
For the purpose of lithology modeling, selected seismic volume was analysed by using artificial neural networks. Two approaches to artificial neural networks (ANN) were used to observe the influence on result prediction of changing the type of the approach. First approach (DAANN) used a large number of different architecture networks, regarding different number of neurons in the hidden layer and different activation functions. Second approach (SAANN) employed the same architecture network but with different distribution of cases within the training, test and selection datasets, and with a different starting point (case) for the analysis. Out of a 1000 total cases, 100 realizations of each approach were singled out upon which the data points with probability of 50%, 75% and 90% of occurrence of certain lithology category were upscaled in the model. Six models were generated by indicator kriging (Figure 5).
Figure 5 Lithology models performed by indicator kriging on upscaled lithology cells [5]
Although in theory, the higher accuracy data should provide a more accurate result, the geologically most sound results were obtained by 50% accuracy data. In higher accuracy results, sandstone lithology was unrealistically over emphasized as a result of the upscaling process, variography and statistical analysis. Considering that majority of hydrocarbon reservoirs discovered so far are in clastic sediments, the methodology presented in this paper represents one of the possible ways of determining subsurface lithology, that can lead to new discoveries not only in the study area, but also in other sedimentary basins. Presented research can be used in all geoenergy-related subsurface explorations, including hydrocarbon and geothermal explorations, and subsurface characterization for CO2 storage potential and underground energy storage potential as well.
Reference:
[1]. Forgotson, J.M.: Review and classification of quantitative mapping techniques. Am. Assoc. Pet. Geol. Bull. 44, 83–100 (1960).
[2]. Feng, R., Luthi, S.M., Gisolf, D., Angerer, E.: Reservoir lithology classification based on seismic inversion results by Hidden Markov Models: applying prior geological information. Mar. Pet. Geol. 93, 218–229 (2018). https ://doi.org/10.1016/j.marpe tgeo.2018.03.004.
[3]. Hohn, M.E.: Geostatistics and Petroleum Geology. Springer, Dordrecht (1999).
[4]. Selley, R.C., Sonnenberg, S.A.: Methods of Exploration. Elements of Petroleum Geology, pp. 41–152. Elsevier, New York (2015).
[5]. Kamenski, A.; Cvetković, M.; Kolenković Močilac, I.; Saftić, B. (2020): Lithology prediction in the subsurface by artifcial neural networks on well and 3D seismic data in clastic sediments: a stochastic approach to a deterministic method // GEM - International journal on geomathematics, 11, 8; 1-24 doi:10.1007/s13137-020-0145-3.
[6]. Cvetković, M., Matoš, B., Rukavina, D., Kolenković Močilac, I., Saftić, B., Baketarić, T., Baketarić, M., Vuić, I., Stopar, A., Jarić, A., Paškov, T.: Geoenergy potential of the Croatian part of Pannonian Basin: insights from the reconstruction of the pre-Neogene basement unconformity. J. Maps. 15, 651–661 (2019). doi:10.1080/17445647.2019.1645052
[7]. Dolton, G.L.: Pannonian Basin Province, Central Europe (Province 4808)—Petroleum Geology, Total Petroleum Systems, and Petroleum Resource Assessment. (2006)
[8]. Schmid, S.M., Bernoulli, D., Fügenschuh, B., Matenco, L., Schefer, S., Schuster, R., Tischler, M., Ustaszewski, K.: The Alpine-Carpathian-Dinaridic orogenic system: Correlation and evolution of tectonic units. Swiss J. Geosci. 101, 139–183 (2008). doi:10.1007/s00015-008-1247-3
[9]. Malvić, T., Cvetković, M.: Lithostratigraphic units in the Drava Depression (Croatian and Hungarian parts) – a correlation. Nafta. 63, 27–33 (2013)
Ana Kamenski, MSc., PhD student/scholarship holder of the Faculty of Mining, Geology and Petroleum Engineering, University of Zagreb, and expert associate at the Department of Geology of the Croatian Geological Survey. In December 2018, she enrolled in the doctoral study of Applied Geosciences, Mining and Petroleum Engineering with the topic of her doctoral thesis: Improvement of the deep-geological characterization in the eastern area of the Drava Depression – spatial prediction of lithological properties based on seismic and well data.
E-portfolio Link
ResearchGate Link
Google Scholar Link
CROSBI Link
Stranica projekta Link