Application of Low Frequency Passive Seismic Method for Hydrocarbon Detection in S Field, South Sumatra Basin

Transcription

1 Abstract Application of Low Frequency Passive Seismic Method for Hydrocarbon Detection in S Field, South Sumatra Basin Andika Perbawa, Danar Yudhatama, and M. Aidil Arham PT. Medco E&P Indonesia, Jakarta andika.perbawa@medcoenergi.com andika.perbawa@gmail.com This paper describes a feasibility study in applying low-frequency passive seismic method in S Field, South Sumatra Basin. Four wells were used to validate the result of the spectral data. This method is also considered as a prospect ranking tool in the vicinity of the S field. Eighteen measurement points were collected and grouped into 6 clusters. Four clusters are located near S- 1, S-2, S-3, and S-4 wells. One cluster is located on prospect K and the other one on prospect G. Standard signal processing flows were conducted such as band-pass filter, FFT, and moving average. The result shows that the maximum amplitude low-frequency between 2-4 Hz of K and S-1 is less than On the other hand, S-2, S-3, S-4 and G show relatively high amplitude of more than 0.02 which indicates a greater possibility of hydrocarbon accumulation when compared with K and S-1. This result was confirmed by gas production in S-2 and oil production in S-3. S-4 has not been tested yet, but the refined well correlation it indicates that there is a limestone reservoir of about 60 feet above OWC. S-1 shows low amplitude which indicates low potential. The completion log confirmed that the well did not penetrate the reservoir target. Prospect G which has high amplitude of low-frequency anomaly is more interesting than prospect K. To conclude, low-frequency passive seismic method was successfully applied in distinguishing both of hydrocarbon and non-hydrocarbon zones. It is feasible to employ this method as a tool for hydrocarbon detection and also as a tool to help in prospect ranking. Keywords: passive seismic, low frequency, hydrocarbon detection 1. Introduction Passive seismic survey is a geophysical method that utilizes a spectral frequency from seismicity data to identify a subsurface fluid behavior. The spectral frequency is derived from transformation of seismicity record at certain time intervals. The seismicity comes from the earth s oscillation then generates low and relatively stable amplitude. When the seismicity is propagating through a reservoir which contains hydrocarbon, it will amplify the amplitude of low frequency between 1-6 Hz (Dangel et al., 2003) or 1-4 Hz (Hardage, 2008) or 2-4 Hz (Suryanto and Wahyudi, 2008). In the future we will refer to this as low frequency anomaly. This anomaly is shown in Figure 1. The amplitude between frequencies 1-4 Hz which is relatively high indicates a strong HC accumulation in the subsurface. Medco has conducted a survey in S Field, South Sumatra area to test and to confirm the methodology and feasibility of passive seismic. The measurement points are located near the wells whose status is already known. If the low frequency anomaly contradicts the data of hydrocarbon accumulation in the well, we can assume that the method has failed. However, if the result shows a positive response, we can assume the method has succeeded. A total of 18 points are grouped into 6 clusters (Figure 2), divided into: 4 clusters in S Field, 1 cluster in G structure located 20 Km to the northeast and 1 cluster in K structure located 10 Km to the southwest. The points in G and K structures are subject to identifying hydrocarbon accumulation inside those prospect structures. The low frequency anomaly in these structures will be compared to anomaly in S structure. If the anomaly is similar with the anomaly in 1

2 the HC producing well, there will be an HC accumulation in the structure. 2. Data and Method The spectral anomaly (frequency anomaly) is defined by two theories which were proposed by Spectraseis AG, a research institution in Zurich. Firstly, a resonant amplification occurring at the pore scale within the reservoir, whereby seismic energy is trapped in a multiphase fluid system and then emitted with a detectable energy level and secondly, a resonant scattering at macro scale due to complex impedance contrasts between hydrocarbon-bearing rocks and the surrounding media altering the ambient seismic wave field and making it detectable at a low frequency (Suntsov et al., 2006). The illustrations of those theories are shown in Figure 3. Figure 1: Spectral amplitude with HC signature above (top) and away (bottom) from HC reservoir or around discovery well (Alwi et al., 2011) Figure 2: Location map of passive seismic acquisition survey. The observation sites are indicated by red circles. 2

3 Figure 3: Theory of passive seismic method based on resonant amplification (above) and resonant scattering (below) Two case studies of passive seismic survey are shown in Figure 4. Both surveys generate the same feature. The signal in low frequency shows a higher amplitude when above hydrocarbon reservoir rather than when away from hydrocarbon reservoir. The amplitude of low frequency anomaly in every area shows a different cutoff amplitude value to distinguish if it is HC or non-hc area, so validation with a known HC producing well is essential. The standard data processing conducted by Medco and Andalas Petroleum are described below and illustrated in Figure 5: 1. Record the seismicity data in the field for 30 minutes for each point measurement, 2. Perform QC of the field recording each day. Each recording consist of three component data: Vertical signal Up- Down (UD), Horizontal signal East- West (EW) and North-South (NS), 3. Digitize data from analog to digital data, 4. Merge the data into one single file because the tools recorded the seismicity data every one minute only, 5. Reformat the digital recording to ASCII file (*.txt) then put into data processing software, 6. Transform the data using FFT to frequency domain then apply noise reduction using band-pass filter from 0.5 to 10 Hz and apply low-cut filter at 0.9 Hz, 7. Perform back transform from filtered data in frequency domain to time domain in order to QC the recording data after filtering is applied, 8. Choose three interval windows in every recording time which contain minimum noise and have relatively stable amplitude, 9. Apply FFT (Fast Fourier Transform) and CWT (Continuous Wavelet Transform) to analyze the spectral low frequency amplitude and stability of frequency along the window, 10. Smooth the spectral frequency to reduce outlier, 11. Calculate the highest amplitude between 2 to 4 Hz, 12. Calculate the average the amplitude in every cluster. Figure 4: A low frequency anomaly feature in several areas (Frehner, 2007) 3. Result and Discussion Eighteen measurement points were collected to represent 6 target clusters. Each cluster contains 3 points which in fact signifies that a small area surrounding those points. By calculating the average amplitude of those points into a single number, disturbance can be reduced. Later we will obtain a single value for each cluster. The average low frequency anomaly amplitude is shown in Figure 6. Figure 6 shows average amplitude which represents a relative hydrocarbon accumulation vertically in the subsurface. Compared with the other points, G has the highest amplitude and correlates with a high probability of hydrocarbon accumulation among them. Conversely, K has the lowest probability of accumulation due to low amplitude. The second highest amplitude is S-4 followed by S-3, S-2 and S-1. The status of the production well is needed to justify the correlation between amplitude 3

4 of low frequency and hydrocarbon accumulation. S-1 is a dry hole well with no reservoir target indication. Based on low frequency amplitude, this well shows a low value of about S-2 (0.021) is produces gas while S-3 (0.31) is produces oil. Therefore the amplitude between and is a boundary of hydrocarbon presences. S-4 shows high amplitude of about In fact, the well was abandoned with no testing. Based on the well correlation (Figure 7), there are 60 feet of limestone reservoir above oil water contact and a high probability that the interval is filled with oil. Unfortunately, based on well log analysis, the interval is interpreted as a tight limestone due to high density. This hypothesis could be the reason why the interval has not yet been tested but still might produce oil. G structure is a promising prospect because it has high anomaly amplitude. Conversely, K structure is the worst prospect due to low frequency anomaly. Raw Data After Bandpass Window selection Zoom in from selected window STFT from Raw data STFT after bandpass filter Amplitude spectrum from selected window Amplitude spectrum from Raw data Amplitude Spectrum after Bandpass filter Smoothing amplitude spectrum from selected window + normalization Figure 5: Passive seismic data processing flow from raw data loading until smoothing of FFT (Red: East West, Green: North South, Blue: Up Down (vertical direction)). 4. Conclusions Figure 6: Average amplitude of frequency between 2-4 Hz in every cluster. 1. Four locations have a relatively higher amplitude frequency (>0.02) between 2-4 Hz: S-2, S-3, S-4 and G structure while S-1 and K structure shows relatively small amplitude frequency (<0.017). 2. Passive seismic confirms that low amplitude frequency in S-1 corresponds with no hydrocarbon accumulation while a high amplitude frequency in S-2 and S-3 corresponds with HC accumulation in S structure. 4

5 3. S-4 shows high amplitude of low frequency which is justified by 60 feet of limestones reservoir above OWC. It is possible that the zone is filled by oil. 4. G structure shows a high amplitude frequency similar with S-3. There is a possibility the G structure has a HC accumulation in the structure. 5. K structure shows a low amplitude frequency similar with S-1, which indicates low possibility HC accumulation in K structure. 6. Passive seismic survey is successfully applied for identifying hydrocarbon accumulation in the subsurface. Figure 7: Well correlation across S-1, S-2, S-3, S-4 and S-5. Acknowledgements We would like to thank to Medco E&P Indonesia, management, and colleagues for the dataset and support. Thanks to Andalas Petroleum as our partner who did the survey and data processing. References Alwi, M., Praptono, S.H., and Murtadho., Low Frequency Passive Seismic Study in Brantas Block, East Java, Indonesia. IPA Proceeding. Jakarta Dangel, S., M. E. Schaepman, E. P. Stoll, R. Carniel, O. Barzandji, E. D. Rode, and J. M. Singer, Phenomenology of tremor like signals observed over hydrocarbon reservoirs: Journal of Volcanology and Geothermal Research, 128, Frehner, M., Schmalholz, S., Podladchikov, Y., and Holzner, R., Low Frequency Modification of Seismic Background Noise Due to Interaction With Oscillationg fluids in Porous Rocks. EAGE. Hardrage, B., Passive Seismic Techniques, Search and Discovery Articles # Suntsov, A.E., Aroutunov, S.L., Mekhin, A.M. and Meltchouk, B.Y., Passive Infra-Frequency Microseismic Technology-Experience and Problems of Practical Use: EAGE Workshop Passive Seismic, December 10-13, Dubai, United Arab Emirates Suryanto, W., and Wahyudi., Monitoring Mikroseismik Untuk Deteksi Langsung Keberadaan Hidrokarbon, PIT HAGI Bandung. 5