|Abdul Haris||Geology Study Program, FMIPA, Universitas Indonesia, Kampus UI Depok, Depok 16424, Indonesia|
|Aditya Dwi Prasetio||Reservoir Geophysics Graduate Program, Physics Department, FMIPA, Universitas Indonesia, Jl. Salemba Raya No. 4, Jakarta 10430, Indonesia|
|Agus Riyanto||Geology Study Program, FMIPA, Universitas Indonesia, Kampus UI Depok, Depok 16424, Indonesia|
|Sri Mardiyati||Department of Mathematics, FMIPA, Universitas Indonesia, Kampus UI Depok, Depok 16424, Indonesia|
Geostatistical seismic inversion has been successfully carried out to characterize a thin reservoir of the Air Benakat Formation in Indonesia’s Jambi sub-basin. The objective of this paper is to characterize detailed P-impedance of the thin reservoir in the Jambi sub-basin using geostatistical seismic inversion rather than deterministic seismic inversion. Geostatistical seismic inversion is believed to be able to enhance vertical resolution and accurately map sub-seismic features. This algorithm uses a geostatistical model, which is constrained by probability density function and a variogram as the input models. The method was applied to eight wells and three-dimensional seismic data that consist of 198 inline and 261 crossline. Prior to performing geostatistical seismic inversion, sensitivity analysis was carried out by cross-plotting petrophysical data to identify the petrophysical properties of the reservoir target. The geostatistical seismic inversion considered 50 realization models that were used as inputs in estimating the probability of the existing subsurface layer and the calculated P-impedance models to obtain the most probable P-impedance model that is useful for characterizing the detailed thin reservoir of the Air Benakat Formation in the Jambi sub-basin. The geostatistical seismic inversion results show a higher resolution of P-impedance compared to the deterministic seismic inversion and are able to resolve thin reservoirs below tuning thickness. In addition, this method is able to optimize better correlation between seismic and petrophysical properties of the thin reservoir with an average thickness below five metres, which is well modelled with reference to both seismic and well data.
Air Benakat Formation; Geostatistical seismic inversion; Jambi sub-basin; Thin reservoir
The resolving power of seismic data is highly dependent on the seismic bandwidth, which is generally lacking low frequencies and high frequencies. Seismic inversion is one of the most powerful techniques to broaden the seismic bandwidth by adding the low- and high-frequency content. In addition, seismic inversion is able to integrate the seismic and well log data to generate a quantitative geological model of the reservoir, including the P-impedance (Mukerji et al., 2001; Lang & Grana, 2017). This geological model is directly related to the layer properties rather than the interface properties, which is associated with geological data such as lithology, porosity, and net pay. Therefore, the quantitative geological model has been recognized as an industrial tool for reservoir characterization.
The seismic data contain reflection information that is associated with a P-impedance change in the subsurface (Avseth et al., 2005). The reflection data is transformed into geological information that is laterally and vertically distributed (Bacon et al., 2003; Haris et al., 2017). The development of seismic inversion techniques has been very fast, and there have been varying options for algorithms coming from the conventional up to the superior algorithm (Haris et al., 2018). A conventional algorithm such as sparse-spike of deterministic seismic inversion is only generating a single P-impedance model, which is useful for figuring out general features of the potential reservoir (Shrestha & Boeckmann, 2002; Francis, 2005). Further, sparse-spike of deterministic seismic inversion tends to produce smooth a P-impedance model that minimizes its variability due to the frequency limitation of the real seismic data. The new approach of geostatistical seismic inversion is applied to improve the limitation of deterministic inversion.
Geostatistical seismic inversion is the superior algorithm that considers the stochastic method by using Sequential Gaussian Simulation to solve non-uniqueness problems via statistical analysis to produce equally probable models (Sancevero et al., 2008; Bosch et al., 2010). This algorithm produces a high vertical resolution for imaging thin reservoirs (Torres et al., 1999). Therefore, selecting the geostatistical seismic inversion algorithm is crucial for characterizing the detailed thin reservoir of the Air Benakat Formation in the Jambi sub-basin.
In this work, we applied geostatistical seismic inversion to delineate a thin reservoir of the Air Benakat Formation in the Jambi sub-basin part of the South Sumatra Basin. The study area has a hydrocarbon potential to be explored (Bishop, 2001; Ginger & Fielding, 2005). The purpose of this paper is to increase vertical resolution and to accurately map sub-seismic features of the thin layer reservoir using a geostatistical model where probability density function (PDF) and a variogram as one of the input models. We demonstrate the comparison between the result of the deterministic seismic inversion and the result of the geostatistical seismic inversion.
The detailed characterization of the thin reservoir of the Air Benakat Formation in the Jambi sub-basin was carried out by applying geostatistical seismic inversion. To get a better understanding of the advantage of geostatistical seismic inversion, we compared its result to the deterministic seismic inversion result. These two seismic inversion algorithms were applied to three-dimensional post-stack seismic data and data from eight wells. Each well contains gamma ray, density, neutron porosity, sonic, and resistivity data. This work was performed using CGG Jason software. A well–seismic tie was applied to eight wells, and we obtained the average of the correlation coefficient of 0.75. This means that the seismic and well data have a good match in terms of geological data and seismic features.
The deterministic seismic inversion is based on the Constrained Sparse Spike Inversion for calculating P-impedance over the entire survey area (trace gate), which is a type of trace-based inversion. This inversion was based on the convolutional model between the reflection coefficient and seismic wavelet. To reduce the uncertainty, a low-frequency model was used as a guide. This low-frequency model was built from the low-pass filter of P-impedance log interpolation and used the geological horizon as a guide.
Geostatistical inversion inverts the reflection seismic data with the most complex geostatistical algorithm (Havelia et al., 2017). The essential step in this inversion is the geostatistical modelling that generates the PDF and experimental variogram from well data in every target reservoir (Haas & Dubrule, 1994; Sulistiono et al., 2015). The step is continued by simulating the number of initial models based on the PDF and variogram (Robinson, 2001). The initial models were analysed and we decided the best parameter to be chosen. We had to pay much attention to the decisive step of the geostatistical seismic inversion by setting the noise level and sampling rate.
The key differences between the deterministic and geostatistical seismic inversion were in the realization model, as the deterministic seismic inversion only resulted in one P-impedance model, whereas geostatistical inversion resulted in multiple P-impedance models (Sams & Saussus, 2010; Nunes et al., 2017). The multiple models of realization provided the quantity of the non-uniqueness and uncertainty of the inversion result (Francis, 2006; McCrank et al., 2009). The geostatistical inversion was generating a realization model that was bounded by the probability density function from seismic and well data. The uncertainty model was determined based on the multiple realizations model.
Geostatistical seismic inversion has been successfully applied to the thin reservoir of the Air Benakat formation in the Jambi sub-basin by producing high vertical resolution and accurately mapping sub-seismic features of the thin reservoir rather than by using the deterministic seismic inversion. The reservoir target was identified as glauconitic sandstone with a relatively high P-impedance, which was based on the cross-plot analysis. The geostatistical seismic inversion model was not limited to the tuning thickness of seismic data, as it was based on simulations by considering the PDF and variogram model. In addition, the geostatistical seismic inversion was constrained by the probability model and the discrete properties model, which was useful in delineating between the target zone and non-target zone. The detailed analysis of the inverted P-impedance showed that the geostatistical seismic inversion illustrates continous target zone rather than the deterministic seismic inversion. The thin layer distribution was indicated by the high P-impedance with the certainty of 60%, which was based on the simulation model run 50 times.
The authors would like to thank Basic Research Excellent Grant for University from Kemenristekdikti under the contract number: 384/UN2.R3.1/HKP05.00/2018 for supporting this fund’s research. Moreover, we would like to express our appreciation to CGG Jason for providing the software to process the geostatistical inversion.
|R3-CE-2088-20180923064051.pdf||Scribendi proofread certificate|
Avseth, P., Mukerji, T., Mavko, G., 2005. Quantitative Seismic Interpretation: Applying Rock Physics Tools to Reduce Interpretation Risk. Cambridge: Cambridge University Press
Bacon, M., Simm, R., Redshaw, T., 2003. 3-D Seismic Interpretation. Cambridge: Cambridge University Press
Bishop, M.G., 2001. South Sumatera Basin Province, Indonesia: The Lahat/Talang Akar-Cenozoic Total Petroleum System. Denver: United States Geological Survey
Bosch, M., Mukerji, T., Ezequiel, F.G., 2010. Seismic Inversion for Reservoir Properties Combining Statistical Rock Physics and Geostatistics: A Review. Geophysics, Volume 75(5), pp. 75A165–75A176
Francis, A.M., 2005. Limitation of Deterministic and Advantages of Stochastic Seismic Inversion. Recorder, Volume 30(2), pp. 5–11
Francis, A.M., 2006. Understanding Stochastic Inversion: Part 1. First Break, Volume 24(11), pp. 69–77
Ginger, D., Fielding, K., 2005. The Petroleum Systems and Future Potential of the South Sumatra Basin. In: The 30th Annual Indonesian Petroleum Association Conference Proceedings
Haas, A., Dubrule, O., 1994. Geostatistical Inversion: A Sequential Method of Stochastic Reservoir Model Constrained by Seismic Data. First Break, Volume 12(11), pp. 561–569
Haris, A., Murdianto, B., Susattyo, R., Riyanto, A., 2018. Transforming Seismic Data into Lateral Sonic Properties using Artificial Neural Network: A Case Study of Real Data Set. International Journal of Technology. Volume 9(3), pp. 472–478
Haris, A., Novriyani, M., Suparno, S., Hidayat, R., Riyanto, A., 2017. Integrated Seismic Stochastic Inversion and Multi-attributes to Delineate Reservoir Distribution: Case Study MZ Fields, Central Sumatra Basin. AIP Conference Proceedings. Volume 1862(1), pp. 030180-1–030180-4
Havelia, K., Aly, O., Mukherjee, A., Zeng, R., Nyein, G., Figuera, L.G., Aamir, M., Al Hosani, K.M., Alkatheeri, F., Al Raeesi, M., Al Hamedi, A., 2017. Characterization of the Lower Cretaceous Thamama Group Reservoirs through Stochastic Inversion and Rock Physics Modeling. In: Fourth EAGE Workshop on Rock Physics 2017 Nov 11
Hermana, M., Ghosh, D.P., Sum, C.W., 2017. Discriminating Lithology and Pore Fill in Hydrocarbon Prediction from Seismic Elastic Inversion using Absorption Attributes. The Leading Edge, Volume 36(11), pp. 902–909
Lang, X., Grana, D., 2017. Geostatistical Inversion of Prestack Seismic Data for the Joint Estimation of Facies and Impedances using Stochastic Sampling from Gaussian Mixture Posterior Distributions. Geophysics, Volume 82(4), pp. M55–M65
McCrank, J., Lawton, D., Mangat, C., 2009. Geostatistical Inversion of Seismic Data from Thinly Bedded Ardley Coals. CSPG CSEG CWLS Convention. Calgary, Alberta, Canada
Mukerji, T., Jørstad, A., Avseth, P., Mavko, G., Granli, J.R., 2001. Mapping Lithofacies and Pore-fluid Probabilities in a North Sea Reservoir: Seismic Inversions and Statistical Rock Physics. Geophysics, Volume 66(4), pp. 988–1001
Nunes, R., Soares, A., Azevedo, L., Pereira, P., 2017. Geostatistical Seismic Inversion with Direct Sequential Simulation and Co-simulation with Multi-local Distribution Functions. Mathematical Geosciences, Volume 49(5), pp. 583–601
Robinson, G., 2001. Stochastic Seismic Inversion Applied to Reservoir Characterization. Canadian Society of Exploration Geophysicists Recorder. Volume 26(1), pp. 36–40
Sams, M., Saussus, D., 2010. Comparison of Lithology and Net Pay Uncertainty between Deterministic and Geostatistical Inversion Workflows. First break, Volume 28(2), pp. 35–44
Sancevero, S.S., Remacre, A.Z., Mundim, E.C., de Souza Portugal, R., 2008. Application of Stochastic Inversion to Reservoir Characterization Process. Earth Science Frontiers, Volume 15(1), pp. 187–195
Shrestha, R.K., Boeckmann, M., 2002. Stochastic Seismic Inversion for Reservoir Modeling. In: 2002 SEG Annual Meeting. Society of Exploration Geophysicists
Sulistiono, D., Vaughan, R., Ali, M. Rasoulzadeh, S., 2015. Integrating Seismic and Well Data into Highly Detailed Reservoir Model through AVA Geostatistical Inversion. In: Abu Dhabi International Petroleum Exhibition and Conference. Society of Petroleum Engineers
Torres-Verdin, C., Victoria, M., Merletti, G., Pendrel, J., 1999. Trace-based and Geostatistical Inversion of 3-D Seismic Data for Thin-sand Delineation: An Application in San Jorge Basin, Argentina. The Leading Edge, Volume 18(9), pp. 1070–1077