Machine Learning as a Silent Observer of Advanced Geoscience Interpretation

Machine learning is proving its successes in all fields of life including medical, automotive, planning, engineering, etc. In the world of geoscience, ML showed impressive results in seismic fault interpretation, advance seismic attributes analysis, facies classification, and geobodies extraction such as channels, carbonates, and salt, etc. One of the challenges faced in geoscience is the availability of label data which is one of the most time-consuming requirements in supervised deep learning. In this paper, an advanced learning approach is proposed for geoscience where the machine observes the seismic interpretation activities and learns simultaneously as the interpretation progresses. Initial testing showed that through the proposed method along with transfer learning, machine learning performance is highly effective, and the machine accurately predicts features requiring minor post prediction filtering to be accepted as the optimal interpretation.

Assessing the Accuracy of Fault Interpretation using Machine Learning Techniques when Risking Faults for CO2 Storage Site Assessment

Generating an accurate model of the subsurface for the purpose of assessing the feasibility of a CO2 storage site is crucial. In particular, how faults are interpreted is likely to influence the predicted capacity and integrity of the reservoir; whether this is through identifying high risk areas along the fault, where fluid is likely to flow across the fault, or by assessing the reactivation potential of the fault with increased pressure, causing fluid to flow up the fault. New technologies allow users to interpret faults effortlessly, and in much quicker time, utilizing methods such as Deep Learning. These Deep Learning techniques use knowledge from Neural Networks to allow end-users to compute areas where faults are likely to occur. Although these new technologies may be attractive due to reduced interpretation time, it is important to understand the inherent uncertainties in their ability to predict accurate fault geometries. Here, we compare Deep Learning fault interpretation versus manual fault interpretation, and can see distinct differences to those faults where significant ambiguity exists due to poor seismic resolution at the fault; we observe an increased irregularity when Deep Learning methods are used over conventional manual interpretation. This can result in significant differences between the resulting analyses, such as fault reactivation potential. Conversely, we observe that well-imaged faults show a close similarity between the resulting fault surfaces when both Deep Learning and manual fault interpretation methods are employed, and hence we also observe a close similarity between any attributes and fault analyses made.

Automated Fault Detection and Extraction under Gas Chimneys Using Hybrid Discontinuity Attributes

3D-seismic data have increasingly shifted seismic interpretation work from a horizons-based to a volume-based focus over the past decade. The size of the identification and mapping work has therefore become difficult and requires faster and better tools. Faults, for instance, are one of the most significant features of subsurface geology interpreted from seismic data. Detailed fault interpretation is very important in reservoir characterization and modeling. The conventional manual fault picking is a time-consuming and inefficient process. It becomes more challenging and error-prone when dealing with poor quality seismic data under gas chimneys. Several seismic attributes are available for faults and discontinuity detection and are applied with varying degrees of success. We present a hybrid workflow that combines a semblance-based fault likelihood attribute with a conventional ant-tracking attribute. This innovative workflow generates optimized discontinuity volumes for fault detection and automatic extraction. The data optimization and conditioning processes are applied to suppress random and coherent noise first, and then a combination of seismic attributes is generated and co-rendered to enhance the discontinuities. The result is the volume with razor sharp discontinuities which are tracked and extracted automatically. Contrary to several available fault tracking techniques that use local seismic continuity like coherency attributes, our hybrid method is based on directed semblance, which incorporates aspects of Dave Hale’s superior fault-oriented semblance algorithm. The methodology is applied on a complex faulted reservoir interval under gas chimneys in a Malaysian basin, yet the results were promising. Despite the poor data quality, the methodology led to detailed discontinuity information with several major and minor faults extracted automatically. This hybrid approach not only improved the fault tracking accuracy but also significantly reduced the fault interpretation time and associated uncertainty. It is equally helpful in detecting any seismic objects like fracture, chimneys, and stratigraphic features.

Enhanced ant tracking: Using a multispectral seismic attribute workflow to improve 3D fault detection

Fault interpretation is a complex task that requires time and effort on behalf of the interpreter. Moreover, it plays a key role during subsurface structural characterization either for hydrocarbon exploration and development or well planning and placement. Seismic attributes are tools that help interpreters identify subsurface characteristics that cannot be observed clearly. Unfortunately, indiscriminate and random seismic attribute use affects the fault interpretation process. We have developed a multispectral seismic attribute workflow composed of dip-azimuth extraction, structural filtering, frequency filtering, detection of amplitude discontinuities, enhancement of amplitude discontinuities, and automatic fault extraction. The result is an enhanced ant-tracking volume in which faults are improved compared to common fault-enhanced workflows that incorporate the ant-tracking algorithm. To prove the effectiveness of the enhanced ant-tracking volume, we have applied this methodology in three seismic volumes with different random noise content and seismic characteristics. The detected and extracted faults are continuous, clean, and accurate. The proposed fault identification workflow reduces the effort and time spent in fault interpretation as a result of the integration and appropriate use of various types of seismic attributes, spectral decomposition, and swarm intelligence.

Fault interpretation uncertainties using seismic data, and the effects on fault seal analysis: a case study from the Horda Platform, with implications for CO₂ storage

Significant uncertainties occur through varying methodologies when interpreting faults using seismic data. These uncertainties are carried through to the interpretation of how faults may act as baffles or barriers, or increase fluid flow. How fault segments are picked when interpreting structures, i.e. which seismic line orientation, bin spacing and line spacing are specified, as well as what surface generation algorithm is used, will dictate how rugose the surface is and hence will impact any further interpretation such as fault seal or fault growth models. We can observe that an optimum spacing for fault interpretation for this case study is set at approximately 100 m, both for accuracy of analysis but also for considering time invested. It appears that any additional detail through interpretation with a line spacing of ≤ 50 m adds complexity associated with sensitivities by the individual interpreter. Further, the locations of all seismic-scale fault segmentation identified on throw–distance plots using the finest line spacing are also observed when 100 m line spacing is used. Hence, interpreting at a finer scale may not necessarily improve the subsurface model and any related analysis but in fact lead to the production of very rough surfaces, which impacts any further fault analysis. Interpreting on spacing greater than 100 m often leads to overly smoothed fault surfaces that miss details that could be crucial, both for fault seal as well as for fault growth models. Uncertainty in seismic interpretation methodology will follow through to fault seal analysis, specifically for analysis of whether in situ stresses combined with increased pressure through CO2 injection will act to reactivate the faults, leading to up-fault fluid flow. We have shown that changing picking strategies alter the interpreted stability of the fault, where picking with an increased line spacing has shown to increase the overall fault stability. Picking strategy has shown to have a minor, although potentially crucial, impact on the predicted shale gouge ratio.

Fault Interpretation Uncertainties using Seismic Data, and the Effects on Fault Seal Analysis: A Case Study from the Horda Platform, with Implications for CO₂ storage

Significant uncertainties occur through varying methodologies when interpreting faults using seismic data. These uncertainties are carried through to the interpretation of how faults may act as baffles/barriers or increase fluid flow. How fault segments are picked when interpreting structures, i.e. what seismic line spacing is specified, as well as what surface generation algorithm is used, will dictate how detailed the surface is, and hence will impact any further interpretation such as fault seal or fault growth models. We can observe that an optimum spacing for fault interpretation for this case study is set at approximately 100 m. It appears that any additional detail through interpretation with a line spacing of ≤ 50 m adds complexity associated with sensitivities by the individual interpreter. Further, the location of all fault segmentation identified on Throw-Distance plots using the finest line spacing are also observed when 100 m line spacing is used. Hence, interpreting at a finer scale may not necessarily improve the subsurface model and any related analysis, but in fact lead to the production of very rough surfaces, which impacts any further fault analysis. Interpreting on spacing greater than 100 m often leads to overly smoothed fault surfaces that miss details that could be crucial, both for fault seal as well as for fault growth models. Uncertainty in seismic interpretation methodology will follow through to fault seal analysis, specifically for analysis of whether in situ stresses combined with increased pressure through CO2 injection will act to reactivate the faults, leading to up-fault fluid flow/seep. We have shown that changing picking strategies alter the interpreted stability of the fault, where picking with an increased line spacing has shown to increase the overall fault stability. Picking strategy has shown to have minor, although potentially crucial, impact on the predicted Shale Gouge Ratio.