Underground » Exploration
In a recently successfully-completed ACARP project C22016 on “Enhancing fault detection by seismic diffraction imaging” (Zhou et al, 2015), a moving average error filtering (MAEF) algorithm was developed to extract diffractions from conventional reflection seismic data. It was demonstrated that extracted diffractions by this algorithm can help in the detection of potential coal seam structures such as faults and dykes of a few meters. However, this MAEF filter requires that the coal seam strata are relatively flat or subparallel. Although this is true in most coal mining situations, coal seams may dip at different angles. To accommodate these situations, ACARP funded this project C25067 to develop new filtering techniques to extract the seismic diffraction signals from reflection seismic data with multiple coal seam and strata, dipping at different angles. The new technique will improve our ability to detect the anomalies and discontinuities from seismic data in a complex geological environment.
Before committing additional funding for new algorithm development, this project was directed to further evaluation of the algorithms developed in C22016 for the detection of small faults. A small case study was carried out to assess the benefits the diffraction process brings to the interpretation of small faults.
Numerical examples have been used to illustrate the detectability of small faults by seismic data. It is demonstrated that detection of small faults by reflection seismic is a challenging task, even in the ideal situation. It is strongly dependent on our ability to visually separate reflections laterally and vertically and whether diffractions smear the reflection events. The presence of background structure complicates the situation and the fault detection limits increase, particularly if the throw of the fault opposes the trend of the background structure. However, diffraction signatures extracted from reflection seismic data could significantly improve our ability to detect small faults.
During this study, two small 3D seismic data sets from Area 1 and Area 2 of Oaky Creek Coal Mine were assessed for faults with both conventional reflection data and newly extracted diffraction signals using the MAEF-based algorithm, which then were validated against the available underground mine mapping. To enhance the confidence of structural interpretation from seismic data, the diffraction interpretations were integrated with the reflection interpretations by using the following criteria:
- Adjusting the location and extent of high confidence faults to match the phase shift location;
- Upgrading the confidence of faults that correlate across both datasets;
- Upgrading seismic anomalies that correlate across both datasets to low confidence faults;
- Deleting low confidence faults that have no diffraction anomalies;
- Delete low confidence seismic anomalies without diffraction anomalies; and
- Add the remaining diffraction anomalies that do not correlate with reflection as low confidence seismic anomalies.
Based on the above procedure, the results from the two data sets are as following.
Area 1. The diffraction interpretation in Area 1 correctly upgraded six seismic reflection anomalies to faults, and dropped three low confidence faults without diffraction signals. Importantly these low confidence faults had been highlighted as false positives during the first round of validation. The additional lineaments identified only in the diffraction images identified an additional 4 faults that had been missed previously, but also introduced 6 false positives.
Area 2. Area 2 includes three well-mapped fault systems with up to 3m throw. The faults have strong diffraction anomalies that could more precisely locate the fault and better define their extent. The diffraction interpretation correctly upgraded seismic reflection anomalies to faults, and eliminated 2 false positive fault interpretations. The additional lineaments identified only in the diffraction images identified one additional fault, but also introduced two false positives.
This case study clearly demonstrated that the benefits the diffraction process brings to the interpretation of faults especially those small ones:
- Improves the accuracy of location and the extent of larger structures;
- Adds confidence to the interpretation of faults by eliminating many false positive interpretations;
- Upgrades some seismic anomalies to faults; and
- Is more sensitive to the identification of very small faults than reflection data.
In addition the project investigated the feasibility of computing the reflection dips from seismic data. Estimation of the local dip of the seismic event on the seismic section is the first step toward extending the MAEF-based diffraction imaging algorithm for use in complex geological environments. To do this, we assume that the observed seismic waves can be considered as plane waves at observation points. Based on this assumption, a non-iterative dip estimation method is proposed by using the gradients or the derivatives of wavefields. Preliminary test results show that the proposed dip estimation method could be used to estimate the local dips of seismic reflections and provide a foundation for us to implement the MAEF-based diffraction extraction method to enhance fault detection for complex geological environments in the future projects.