B006
`Acquisition Using Simultaneous Sources
`J. Stefani* (Chevron Energy Technology Company), G. Hampson (Chevron
`Energy Technology Company) & E.F. Herkenhoff (Chevron Energy
`Technology Company)
`
`SUMMARY
`We present a technique in which 2 or more shots are acquired during the time it normally takes to acquire
`1 shot. The 2 or more shots are fired in a near simultaneous manner with small random time delays
`between the component sources. A variety of processing techniques are applied to produce the same
`seismic images which would have resulted from firing the simultaneous shots separately. These processing
`techniques rely on coherency of the wavefield in the common-shot domain and unpredictability in
`common-receiver, -offset and -midpoint domains. We present results of its application in 2D in the Green
`Canyon area and in 3D on the Petronius field, both located in the Gulf of Mexico. These results
`demonstrate that in deep water with modest water-bottom reflectivity no special processing is required,
`whereas in shallower water with stronger water-bottom reflectivity the use of shot separation techniques is
`necessary. We conclude that this technique can be used robustly to improve source sampling and, for
`example, to acquire data from a range of azimuths simultaneously. This could potentially change the
`fundamental economics of wide azimuth acquisition.
`
`
`
`
`
`EAGE 69th Conference & Exhibition — London, UK, 11 - 14 June 2007
`
`
`
`
`
`PGS Exhibit 2007, pg. 1
`WesternGeco v. PGS (IPR2015-00309, 310, 311)
`
`

`
`
`
`Introduction
`There are a number of potential advantages in being able to acquire 2 or more shots in the
`time that it normally takes to acquire 1 conventional shot. An obvious benefit is an increase in
`time and/or cost efficiency. If there is freedom to locate the additional source(s) at different
`locations relative to the main seismic vessel, then data may be simultaneously acquired which
`supplements the typical seismic configuration. Thus such a technique can be used generally to
`increase data sampling via increased fold, increased range of azimuths, or extended offsets, or
`it can be used to increase the time value of data by reducing the need for infill or allowing an
`acquisition program to complete in less time. Although a number of authors have discussed
`simultaneous shooting using vibratory sources (e.g., Ward et al., 1990), few (e.g. Beasley et
`al. 1998, de Kok and Gillespie 2002) have discussed its application to impulsive sources, such
`as airguns. If a secondary source has asynchronous timing then it behaves like nearby seismic
`interference which Lynn et al. (1987) observe may be effectively attenuated by stacking or
`even more so by diversity stacking. Krey (1987) described the ability of migration and
`stacking to attenuate noise very effectively, particularly in 3D. Since it is current practice to
`use 3D migration algorithms we might expect asynchronous secondary source energy to be
`effectively attenuated simply by our conventional migration and stacking procedures. As a
`result we consider the possibility of using one or more secondary sources located at positions
`of our design with random timing delays all fired within the same conventional shotpoint
`timing cycle. We will briefly describe the technique and present results of its application in
`2D on the Sigsbee synthetic model and in the Green Canyon area and in 3D on the Petronius
`field, both fields located in the Gulf of Mexico. We will conclude that this technique can be
`used to improve source sampling and, for example, to acquire data from a range of azimuths
`simultaneously, thereby potentially changing the fundamental economics of wide azimuth.
`
`
`Method
`We will assume for the sake of discussion that 2 sources are present at separate locations. The
`“primary” source is fired at time-zero while the “secondary” source is fired with some random
`delay,τ, relative to the primary source. The random process is parameterised such that
`thi
`0
`
`τ
` ττ≤≤
`<
`i ττ ≠−1
` and for the
` shot
`. In such circumstances, a common-shot
`i
`max
`min
`gather will contain a superposition of the 2 coherent wavefields contributed by the 2 sources.
`In contrast, a common-receiver, -offset or -midpoint gather, in which successive traces come
`from different shots, will contain a superposition of the coherent wavefield from the primary
`sources and the incoherent noise from the randomly delayed secondary sources. Since we
`
`primary source aligned
`
`secondary source aligned
`
`Σ/n
`
`Σ/n
`
`0.4
`
`0.8
`
`-0.8
`
`-0.4
`
`0.4
`
`0.8
`
`0
`t
`
`-0.8
`
`-0.4
`
`0
`t
`
`
`Figure 1 - The primary (left) or secondary (right) sources may be aligned. Simple procedures
`such as stacking attenuates the unaligned energy.
`know the delays we may randomise either the primary or the secondary source wavefield in
`these gathers.
`
`
`EAGE 69th Conference & Exhibition — London, UK, 11 - 14 June 2007
`
`PGS Exhibit 2007, pg. 2
`WesternGeco v. PGS (IPR2015-00309, 310, 311)
`
`

`
`
`Figure 1 (above) schematically shows the traces recorded from 9 shots. The primary sources
`are aligned on the left, whereas the secondary sources have been aligned on the right. The
`bottom trace is the stack which illustrates the idea that summation processes may be used in
`conjunction with wavefield alignment to separate the primary and secondary source energy.
`Modern processing techniques typically involve one or more such summation processes
`(migration and stacking) in which the traces from different shotpoints are summed. Therefore,
`simply aligning the source of choice and applying conventional processing is an effective
`source separation procedure. We may consider the effectiveness of this separation by noting
`that, if we assume that the unaligned wavefield behaves as random noise then we should
`n/1
`anticipate its attenuation to be as
`. The situation is enhanced greatly by 2D or, more
`particularly, 3D migration in which very large numbers of traces are summed (Krey, 1987). In
`any case, the gain in superposing diffractor-consistent signal must exceed the noise from
`inconsistent interfering events. Although we find that exploiting incoherency is effective, we
`also have the opportunity to work with the superposition of the 2 coherent source wavefields
`on the common-shot gathers. In order for any form of separation to be effective here, there
`needs to be a degree of orthogonality between the 2 wavefields, at least locally, if not
`globally. This typically requires that the location of the 2 sources in time and space be such
`that the resulting wavefields locally occupy substantially separate areas of dip space. We have
`found that exploiting a combination of the incoherent/coherent features of the pre-stack
`wavefield produces good results. Although these simple ideas seem to be reasonably
`effective, in the future we plan to examine several potentially better techniques.
`
`Field experiments
`As a proof of concept we first applied this technique on synthetic acoustic data from the well
`known Sigsbee model. Although not shown in this expanded abstract, the PSDM stacked
`images from the conventional acquisition and from the simultaneous source acquisition (with
`
`
`
`Conventional CMP gatherConventional CMP gather
`
`
`
`Simultaneous CMP gatherSimultaneous CMP gather
`
`
`
`asynchronous secondary source energyasynchronous secondary source energy
`
`
`
`coco
`
`herent primary source energy
`herent primary source energy
`
`
`
`9km9km
`
`
`
`11
`
`432
`432
`
`time
`time
`
`5 6
`5 6
`
`Figure 2 - Conventional CMP gather (left) and simultaneous CMP gather (right).
`no specially applied noise removal) were practically indistinguishable.
`
`Fortified with these synthetic results, we then tested this idea in 2D on real data from the
`Green Canyon area in the Gulf of Mexico. A seismic line was acquired several times with
`different configurations with reasonable effort to match the cable feathering for each pass.
`The area lies in 790m water and the near water bottom reflectivity is relatively low. Figure 2
`(above) compares a conventional CMP gather with one from the simultaneous source
`
`EAGE 69th Conference & Exhibition — London, UK, 11 - 14 June 2007
`
`PGS Exhibit 2007, pg. 3
`WesternGeco v. PGS (IPR2015-00309, 310, 311)
`
`

`
`
`experiment exhibiting incoherent noise from all the secondary sources contributing to this
`CMP.
`
`Figure 3 (below) shows the migrated image of the conventional data compared with the
`simultaneous data without any special noise processing. It is clear that the sources were easily
`separated simply by migration and stacking providing data quality as good as the control
`dataset. In our analysis of these data we also investigated the use of diversity-based
`techniques and dip discrimination. We found that these made almost negligible contributions.
`Although not shown, post migration narrow angle substacks, important for AVA analysis,
`
`
`
`ConventionalConventional
`
`
`
`SimultaneousSimultaneous
`
`
`
`22
`
`
`
`33
`
`
`
`44
`
`time
`time
`
`
`
`55
`
`66
`
`
`
`17.2 km17.2 km
`Figure 3 – PSTM of conventional data (left) and simultaneous source data (right) from the
`Green Canyon area. No special separation processing has been applied.
`were also insignificantly affected by the simultaneous source technique.
`
`As a result of the successful 2D test we designed a 3D test to duplicate part of a 3D survey
`over the Petronius field in the Gulf of Mexico. The Petronius field lies in 530m of water
`although the water bottom depth varies significantly across the area. The near water-bottom
`reflectivity is stronger than in the Green Canyon area and produces strong refractions,
`reflections and turning waves. We used 2 sources, one on the main 3D vessel and one
`displaced 3200m cross-line and 900m astern of the main source. Our design was partly meant
`to determine if we could simultaneously acquire a wider range of azimuths robustly in 3D and
`partly to find out whether the stronger reflectivity could be handled during the source
`separation. The field operation proved to be very robust after careful mobilisation and crew
`briefing. The data processing consisted of dip discrimination on the common-shot records, 3D
`pre-stack time migration and stack. Figure 4 (below) compares the conventional data to the
`simultaneous source data without and with dip discrimination processing. The wavefield
`separation required significantly greater attention to detail than at Green Canyon and proved
`to be less successful. This is due to the strong near water-bottom reflectivity which appears as
`strong coherent secondary energy superposed upon the primary wavefield. Although we did
`use dip discrimination to help separate the sources, they were not completely separated due to
`the secondary source location. Nonetheless, this relatively simple approach yielded
`surprisingly good results and stimulates the search for better separation techniques.
`
`Conclusions
`We have presented a technique in which 2 or more shots are acquired during the time it
`normally takes to acquire 1 shot. The data processing exploits coherency and incoherency in
`
`EAGE 69th Conference & Exhibition — London, UK, 11 - 14 June 2007
`
`PGS Exhibit 2007, pg. 4
`WesternGeco v. PGS (IPR2015-00309, 310, 311)
`
`

`
`
`
`ConventionalConventional
`
`
`
`SimultaneousSimultaneous
`
`
`
`Simultaneous + separationSimultaneous + separation
`
`
`
`
`
`11
`
`
`
`22
`
`time
`time
`
`
`
`33
`
`
`
`12.7 km12.7 km
`
`44
`
`
`Figure 4 – 3D PSTM of conventional data (left), simultaneous source data without (centre) and
`with dip discrimination processing from the Petronius field.
`different domains to produce the same seismic images which would have resulted from firing
`the simultaneous shots separately. We have tested the technique in 2D and 3D in 2 different
`areas of the Gulf of Mexico. The results demonstrate that in deep water with modest water-
`bottom reflectivity no special processing is required, whereas in shallower water with stronger
`water-bottom reflectivity the use of shot separation techniques is necessary. We conclude that
`this technique can be used to improve source sampling and, for example, to acquire data from
`a range of azimuths simultaneously. This could be of significant economic importance for
`wide azimuth acquisition. Although the simple source separation ideas are reasonably
`effective, in the future we plan to examine several potentially better techniques and to assess
`this method’s effect on pre-stack AVA analysis.
`
`Acknowledgements
`The authors wish to express their appreciation to Chevron Energy Technology Company,
`Chevron GOMBU Deepwater Business Unit and WesternGECO. We particularly wish to
`thank the crews of GECO Diamond and the Odyssey for their valuable help, and Tim Mills
`for being an extraordinary bird-dog with the patience to translate our ideas into action in the
`field. Finally we wish to thank a number of our colleagues for their significant contributions,
`counsel and tolerance.
`
`References
`Beasley, C.J., Chambers, R.E. and Jiang, Z. [1998] A new look at simultaneous sources: 68th
`Annual International Meeting, SEG, Expanded Abstracts , 133-135.
`de Kok, R.J. and Gillespie, D. [2002] A Universal Simultaneous Shooting Technique: 64th
`Meeting, EAGE, Expanded Abstracts , A004.
`Krey, T. C. [1987] Attenuation of random noise by Two-D and Three-D CDP stacking and
`Kirchhoff migration: Geophysical Prospecting, 35 , no.2, 135-147.
`Lynn, W., Doyle, M., Larner, K. and Marschall, R. [1987] Experimental investigation of
`interference from other seismic crews: Geophysics, 52 , no.11, 1501-1524
`Ward, R.M., Brune, R.H., Ross, A. and Kumamoto, L.H. [1990] Phase encoding of Vibroseis
`signals for simultaneous multisource acquisition: 60th Annual International Meeting,
`SEG, Expanded Abstracts , 938-941.
`
`
`
`EAGE 69th Conference & Exhibition — London, UK, 11 - 14 June 2007
`
`PGS Exhibit 2007, pg. 5
`WesternGeco v. PGS (IPR2015-00309, 310, 311)