Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
Results
Probabilistic Neural Network Inversion For Characterization of Coalbed Methane
Yenugu, Malleswar (ConocoPhillips School of Geology and Geophysics, The University of Oklahoma) | Fisk, Jeremy C. (ConocoPhillips School of Geology and Geophysics, The University of Oklahoma) | Marfurt, Kurt J. (ConocoPhillips School of Geology and Geophysics, The University of Oklahoma)
Summary The seismic guided estimation of reservoir properties away from the wells is a common problem that geophysicists, geologists and reservoir engineers face every day. This problem is due to low resolution of seismic data as well as the lack of proper models that link the seismic data to the borehole data. Geostatistical methods help resolve this problem, but these methods rely only on a linear fit between seismic attributes and reservoir parameters. Artificial neural networks are the best method to relate the non-linear fit between the borehole parameters to seismic volume parameters to better understand the heterogeneity of the reservoir properties. Probabilistic neural networks (PNN) are used to invert the seismic data of coalbed methane (CBM) field from northeast Australia to better understand the reservoir properties of the coals sandwiched between sands and shales. PNN has not only helped to improve the vertical resolution but also the lateral variation in the heterogeneity of the reservoir. Introduction The contribution of production from the unconventional reservoirs like shale and coal is increasing everyday to the energy needs of the World. These reservoirs have low permeability often act as both the primary source rock as well as the reservoir rocks. These are not only thin but wide spread in the basin. The characterization of these reservoirs is a challenge to the geoscientists in terms of resolution and distribution in an area. The calibration of well logs (high) resolution with 3D seismic data (low resolution) is a challenge while building the comprehensive geological models. Artificial Neural Networks (ANN) was introduced to the geosciences community in 1980?s. ANNs have the ability of recognize complex, non-linear relationships between seismic attributes and petrophysical data. These relationships are applied to seismic data to predict interwell reservoir properties. However, there is the danger that the neural network can become “over trained”. That is, the fit at the wells is excellent, but the underlying model is too complex and does not lead to physically meaningful results away from the well. This problem is addressed by using the technique of cross-validation, in which we remove wells from the training stage and then „blindly?, predict these wells in the validation stage (Herrera et al., 2006). The objective of this paper is to apply probabilistic neural networks (PNN) to invert the seismic data volume to impedance by training and validating the acoustic impedance logs at the wells. Probabilistic Neural Networks The ability of ANNs to detect and recognize data patterns and to exploit functional, complex non-linear relationships between multiple data inputs provides for a powerful exploration tool. ANNs are robust in noisy or missing data sections that solve ambiguities or differences between data inputs. ANN results are developed through repeated training. An input model may not be required to be successfully applied. Since the 1980?s many different artificial multi-layer feedforward neural networks have been successfully applied to various geophysical problems. With good quality seismic and well log data, the PNN is favored.
- North America > United States > Kentucky > Butler County (0.25)
- Oceania > Australia > Queensland (0.17)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock > Mudrock > Shale (0.76)
- Geology > Rock Type > Sedimentary Rock > Organic-Rich Rock > Coal (0.49)
- Geophysics > Seismic Surveying > Seismic Processing (0.89)
- Geophysics > Seismic Surveying > Surface Seismic Acquisition (0.55)
- South America > Ecuador > Sucumbíos > Oriente Basin > Block 11 > Bermejo Field > Tena Formation (0.99)
- South America > Ecuador > Sucumbíos > Oriente Basin > Block 11 > Bermejo Field > Hollin Formation (0.99)
- Oceania > Australia > Queensland > Central Highlands > Bowen Basin (0.99)
- (24 more...)
Summary Australia is the world''s second largest producer of coalbed methane (Faiz, 2008). Characterized as a marginal tectonic setting, the Bowen Basin of Queensland displays a thick succession of numerous thin bituminous rank coal seams. The area''s heterogeneous production is particularly perplexing; it is not unusual for production to change as much as 50-75% between neighboring wells within a few kilometers of one another. A myriad of factors can affect coalbed methane production to include coal thickness, coal cleat architecture, local maximum horizontal stress direction, and the in situ stress magnitude. We show how seismic curvature attributes illustrate lineaments that correlate to production. We use a technique to generate 3D rose diagrams from curvature attributes and show that the diagrams depict the face and butt cleat architecture. Introduction Unconventional reservoirs continue to contribute an increasing percentage of the total amount of oil and gas production in the world. Shale and coal are examples of low-permeability unconventional reservoirs that often act as both the primary source rock and the reservoir. The application of seismic attributes and multiattribute transforms are incrementally pushing the limits to seismic resolution and facilitating different data analysis perspectives to use during reservoir characterization. Geologic Background The Bowen Basin is a Permian-Triassic age major economic coal basin that extends approximately 900 km in a generally north to south direction in the eastern portion of Queensland, Australia. The Bowen Basin is one portion of the Bowen-Gunnedah-Sydney foreland system that formed as a result of the collision of the paleo-Pacific and paleo- Australian plates beginning as early as 294 Ma in the Early Permian. This orogenic belt is often referred to as the New England fold belt. Figure 1 is a map of Australia which outlines the approximate geographical limits of the Bowen Basin in Queensland. Toward the end of the Permian, the coal measures formed in environments described as fluvio-deltaic. Structurally, this reservoir lies within a large anticlinal structure that Korsch (2004) has labeled a fault-propagation fold. Figure 2 is a seismic inline view of the fault propagation fold from the 3D seismic data for this study with interpreted faults. Figure 3 shows the seismic line (A-A'') from Figure 2 and well production data for this 3D seismic survey. Seismic Data Quality The company''s primary goal for this 3D seismic survey was the improved mapping of the upper coal horizons within the basin. The survey size is 31.56 km2 of rolling farmlands. The data were recorded using 45 source lines with a 200 m interval and 23 receiver lines with a 200 m interval, resulting in a natural bin spacing of 25 m x 12.5 m. Four Mertz M26 vibrators vibrated with a sweep length of four seconds and a sweep frequency range from 6-130 Hz. Group arrays consisted of twelve sensor SM4, 10 Hz geophones in a linear array with the source in the middle of each line, resulting in a multiplicity of 36 fold. The sampling rate of the data is 2 ms.
- Phanerozoic > Paleozoic > Permian (0.88)
- Phanerozoic > Mesozoic > Triassic (0.55)
- Geology > Structural Geology > Tectonics > Compressional Tectonics > Fold and Thrust Belt (1.00)
- Geology > Rock Type > Sedimentary Rock (1.00)
- Geology > Geological Subdiscipline (1.00)
- Geophysics > Seismic Surveying > Surface Seismic Acquisition (1.00)
- Geophysics > Seismic Surveying > Seismic Interpretation (0.90)
- Oceania > Australia > Queensland > Central Highlands > Bowen Basin (0.99)
- North America > United States > New Mexico > San Juan Basin > Fruitland Formation (0.99)
- North America > United States > Colorado > San Juan Basin (0.99)
- (5 more...)