1
INTEGRATED ANALYSIS OF SEISMIC RESPONSE OF GAS RESERVOIRS IN ORDOS BASINXIANG-YANG LI1, XIU-CHENG WEI1,2, ENRU LIU1,
XI-ZHE LI3, SONG-QUN SHI4 AND XIAN-BIN ZENG3
1British Geological Survey, West Mains Road, EH9 3LA
2University of petroleum, Beijing, 102200, China
3Lang Fang Research Institute, Petro-China, Hebei, China
4Chang Qing Research Institute, Petro-China, XiAn, China
Introduction
Recently, the use of converted waves to image reservoirs with low P-wave impedance but high S-wave impedance has attracted considerable interest in offshore exploration. Here we extend this technology to onshore gas reservoirs in the Ordos basin in Northwest China. The Ordos basin is the largest gas province in China. The main gas reservoirs (about 3300m in depth) are upper Paleozoic sandstones which have a low P-wave impedance and are immediately above a coal seam. This makes it very difficult to image the gas reservoirs using conventional P-wave data. This paper performs an integrated analysis of core, log and seismic data to examine the seismic response of the gas reservoirs, focusing on the feasibility of using converted PS waves to image the reservoirs.
The work involves the analysis of a large number of sandstone core samples with different gas saturation in order to establish the relationship of P- and S-wave velocities with fluid saturation. This is followed by a detailed analysis of the log characteristics of the gas reservoirs. Synthetic modeling is then performed to model the seismic response of the gas reservoirs. Based on these analyses, a large number of offset VSPs are processed and interpreted to verify the P-wave and converted-wave response of the reservoirs.
Analysis of core samples
There were 45 cores with four types of gas saturations: 100%, 65-75%, 35% and 0% (100% water saturation). Both P- and S-wave velocities were measured at confining pressure between 35and 42 MPa, and temperature between 78 and 100 oC. Experimental frequencies are 800 khz and 600 khz for P- and S-waves, respectively.
The variation of measured P- and S-wave velocities with porosity is shown in Figures 1 and 2 for all four different saturations. The P-wave velocity clearly decreases as the porosity increases, and also decreases as the gas saturation increases (Figure 1). However, there are no such clear trends of variation in the S-wave velocity, and it shows some small random variations as both the porosity and saturation change (Figure 2).
The following empirical equations are used to describe the variation of the P- and S-wave velocities with porosity and saturation,
where is the porosity and Sg is the gas saturation. From the data in Figures 1 and 2, the coefficients are obtained as:
The fitted P-wave velocity curves as a function of porosity and saturation are shown in Figures 3 and 4 for information.
Analysis of wireline logs
Wireline logs from eleven boreholes are examined to identify the characteristics of the gas reservoirs. The major diagnostic features of the sandstones saturated with gas are low gamma-ray values, low density, low wave impedance, but high electric impedance, and high compensated neutron logging values, as well as a huge anomaly in the spontaneous potential logs.
We have also examined the relationship of the P- and S-wave velocity with porosity and saturation using log data. Unfortunately, there is only one well (Well A) that has both P- and S-wave sonic logs. Figure 5 shows the cross plot of the P- and S-wave sonic velocities against the logging porosity from well A. Again it shows a clear decrease of the P-wave velocity as porosity increases. In contrast, there is no clear variation in the S-wave velocity. This confirms the findings of core data analysis.
Figure 6 shows the cross plot of P-wave sonic velocity against the logging porosity from the eleven wells across the basin. One can see a clear non-linear variation of the P-wave velocity with porosity, and this justifies the use of equation (1) to describe the variation of the P-wave velocity with porosity. Figure 7 shows the two curves fitted to Figure 6 and a clear separation of the P-wave velocity for different gas saturations is seen.
Modelling studies
Three types of sandstone models are studied. For each model, five layers of sandstones with equal thickness are embedded in a mudstone. The P- and S-wave velocities of the mudstone are 4500m/s and 2400 m/s, respectively. The sandstone layers are 10 metres thick and separated by 100 metres. The first model consists of tight sandstones with P- and S-wave velocities of 5000m/s and 2500 m/s, respectively.
The second model consists of the same mudstone as in the first one, but the five layers of tight sandstones are replaced by porous sandstones saturated with water. The porosity in each sandstone layer is, respectively, 2%, 5%, 8%, 11% and 14% from top to bottom. The corresponding P-wave velocities in each sandstone layer are calculated as 4954m/s, 4884m/s, 4814m/s, 4743m/s and 4671m/s, respectively using the empirical relationship in equation (1), whilst the S-wave velocities are unchanged.
The third model is the same as the second model but the sandstones are saturated with gas, and the P-wave velocities in each sandstone layer are calculated as 4917m/s, 4791m/s, 4664m/s, 4536m/s and 4407m/s accordingly.
Figures 8 and 9 show the zero-offset P- and S-wave reflection sections, respectively. The dominant frequency of the zero phase wavelet is 40 hz. As shown in Figure 8, for water saturation, the P-wave reflection amplitude decreases slightly as the porosity increases. However, for gas saturation, the P-wave reflection amplitude decreases sharply as porosity increases. The reflection amplitude is nearly zero with 11% porosity, and a polarity reversal is observed with 14% porosity. In contrast, the S-wave reflection amplitude remains unchanged for all three models. Therefore, it is possible to use the S-wave to image gas reservoirs and to combine the P- and S-waves to distinguish fluid saturations.
VSP seismic response
VSP seismic data are processed to examine if the above characteristic P- and S-wave responses of the gas reservoirs can be observed in real data. We have analyzed more than 20 VSPs from seven wells across the basin, and the results from well B are presented here.
Data from well B consists of one P- and one S-wave source zero-offset VSPs, and four P-wave source offset VSPs for four different offsets and azimuths. The offset and azimuth distribution of the four offset VSPs are: offset VSP 1- (1070m, 90o), offset VSP 2 - (1400m, 180o), offset VSP 3 - (930m, 270o) and offset VSP 4 - (1350m, 350o), respectively. The azimuth is measured from North.
Figure 10 shows the corridor stacks of the P- and S-wave zero-offset VSPs. The sections are displayed in two-way P-wave time. The target horizon is identified at about 1.96s two-way time, immediately above the strong coal seam reflection at about 2.04s. Comparing the P- with the S-wave response of the gas reservoir, one can see a weak P-wave reflection but a relative strong S-wave reflection.
Figure 11 shows the VSP-CDP sections obtained from the four offset VSPs. Both PP and PS converted-waves sections are produced, and displayed in two-way P-wave time. Again, the reservoir shows a weak P-wave reflection in the section, and it is often hidden behind the strong coal reflection. This makes it very difficult to interpret the gas reservoir in the P-wave section reliably. In contrast, the gas reservoir shows a strong PS wave reflection with amplitudes comparable to the coal seam reflection, and it can be clearly identified in the converted-wave VSP CDP sections. This confirms that converted-waves can be used to image the gas reservoirs in the Ordos basin.
Conclusions
We have performed an integrated study of the seismic response of the gas reservoirs in the Ordos basin using core, log and seismic data. Analysis of the core and logging data shows that the variation in porosity and gas saturation has a strong effect on the P-wave velocity of the reservoir sandstone, whereas the effect on the S-velocity is small. The sandstone gas reservoirs are surrounded by mudstone, and the modelling reveals a weak P-wave but a strong S-wave response of the reservoir. This is confirmed by the analysis of P- and S-wave zero-offset VSPs. Utilizing this concept, the gas reservoirs are successfully mapped using the converted-waves from offset VSPs.
Acknowledgements
We thank PetroChina for permission to show the data. This work is supported by the DTI/Trade Partners (UK) international collaboration programme and PetroChina through the Edinburgh Anisotropy Project (EAP) of the British Geological Survey, and is published with the approval of all project partners and the Executive Director of British Geological Survey (NERC).
Figure 1. P-wave velocity (Vp) from core data. Figure 2. S-wave velocity (Vs) from core data.
Figure 3. Vp vs. porosity from equation (1). Figure 4. Vp vs. gas saturation from equation (1).
Figure 5. Vp and Vs from logs of well A. Figure 6. Vp from logs of 11 wells.
Figure 7. Vp curves fitted to Figure 6 using equation (1).
Figure 8. Synthetic P-wave section. Figure 9. Synthetic S-wave section. Figure 10. Corridor stacks from well B.
P-wave VSP-CDP PS-wave VSPCDP P-wave VSP-CDP PS-wave VSPCDP
VSP 1 VSP 2
P-wave VSP-CDP PS-wave VSPCDP P-wave VSP-CDP PS-wave VSPCDP
VSP 3 VSP 4
Figure 11. VSP CDP sections from the offset VSPs at well B.
EAGE 64th Conference & Technical Exhibition — Florence, Italy, 27 - 30 May 2002