1
`
`A-04 A UNIVERSAL SIMULTANEOUS SHOOTING
`TECHNIQUE
`
`ROB DE KOK and DIANA GILLESPIE
`WesternGeco, 10001 Richmond Avenue, Houston TX 77042, USA
`
`Summary
`A novel method for the simultaneous recording of seismic data from two or more sources is
`presented. The coding for the different sources consists of the introduction of either polarity or time
`delay changes in the field. Separation is achieved through polarity decoding followed by averaging
`and/or interpolation during processing. For purposes of averaging, not only dedicated techniques, but
`also standard processes such as stacking and migration can be relied on.
`
`The method is applicable to all possible types of sources and has no restriction as to their
`locations. For instance, standard air-gun arrays may fire in combination with high-resolution sources.
`The method can be applied to impulsive sources as well as vibrators, both in marine and land
`environments, and can be used in combination with other simultaneous techniques for additional
`discrimination. Preliminary results from an experiment in the Gulf of Mexico yielded a discrimination
`of 15 dB through 40-fold stacking. Additional dedicated processing methods have indicated the
`potential for a substantial higher source separation.
`
`Introduction
`Due to the high cost of seismic acquisition, it has become common practice to acquire data at a
`low but still acceptable density of surface locations. On land, both the source and the receiver
`locations are sparsely sampled, while in the marine environment it is the shooting density that is
`generally lower than desirable. The quality of the data, recorded in this way, mostly proves acceptable
`for the initial intended purpose, such as reconnaissance, wildcat drilling, and others. However, when
`more detailed studies such as hydrocarbon identification and reservoir characterization are needed, the
`data quality often proves insufficient.
`
`The simultaneous firing of multiple sources into the same recording system, can be an
`attractive option to increase the data density at relatively low additional field effort. Simultaneous
`firing is particularly economical when additional sources can readily be deployed at a low incremental
`cost, as is, for example, the case with seismic vibrators. Unfortunately, the separation of the
`information pertaining to the individual sources is cumbersome and/or imperfect when using existing
`methods [1,2]. Also, there are very few methods suitable for use with impulsive sources such as air-
`gun arrays [1].
`
`Two alternative solutions, applicable in both marine and land environments, are presented.
`The first uses signals that are coded through introduction of small time delays in the firing sequence.
`The second uses signals that are coded through positive and negative polarities and unlike other
`polarity methods, is not restricted to vibratory sources or stationary locations. In both methods, the
`
`EAGE 64th Conference & Exhibition — Florence, Italy, 27 - 30 May 2002
`
`PGS Exhibit 2014, pg. 1
`WesternGeco v. PGS (IPR2015-00309, 310, 311)
`
`

`
`2
`enhancement of the desired source energy is accomplished through mixing, or averaging after
`horizontal time alignment of the coded signals.
`
`Principle of time delay coding and decoding
`The first coding principle consists of the introduction of a pattern of time shifts in a sequence
`of shots. The patterns are different for the individual sources that fire simultaneously. Although the
`sources do not strictly fire at the same time, it is still considered that they emit simultaneously because
`the timing differences are a few milliseconds only. The minimum cycle in a shooting pattern is four,
`allowing for the simultaneous firing of three sources. In principle, larger patterns can be used,
`allowing for the deployment of a larger number of sources. The emission pattern of a minimal
`configuration is depicted in Figure 1a. The signals from the three sources, A, B, and C, are shown
`individually, but overlay in an actual acquired record. The early signals have been emitted on the
`reference time, while the other signals have been delayed. The decoding consists of the application of
`polarity reversals to the delayed signal traces and mixing those with the non-delayed signal traces.
`Figure 1b shows the traces after polarity decoding for source A, while Figure 1c shows the result after
`subsequent mixing. Only the signal from source A is preserved, while the two other source signals
`have been suppressed completely. Two other polarity decoding schemes pertaining to the delay
`patterns of sources B and C will preserve their corresponding signals. The success of the method is
`independent of differences in the individual source signals, of their time delays, and of their strength.
`However, the discrimination level does depend on the stability of the emitted signal of a particular
`source. Also, the choice of mixing process(es) is a critical factor in the source discrimination result.
`Several alternative processing methods have been tested.
`1
`4
`1
`4
`1
`4
`1
`4
`1
`(a)
`(b)
`
`4 1
`
`4 1
`
`4
`
`4 1
`
`4
`
`1
`(c)
`
`t [s]
`
`0.2
`
`0.4
`
`Source
`A
`
`Source
`B
`
`Source
`C
`
`Source
`A
`
`Source
`B
`
`Source
`C
`
`Source
`A
`
`Source
`B
`
`Source
`C
`
`1
`
`4
`
`1
`
`4
`
`1
`
`4
`
`1
`
`4
`
`1
`
`Shot sequence number
`
`Figure 1. Time delay coding and decoding — (a) minimal set of four traces shown for three sources
`individually, (b) after polarity decoding, and (c) after polarity decoding and mixing.
`
`Figure 1c shows that the time shift method introduces a filter response equivalent to that of a
`sea surface ghost in marine applications. When the time delay is selected equal to the surface ghost
`delay time, no extra notches are introduced. The already existing ghost response is effectively squared
`in that case, and the rather complex source wavelet of Figure 1c becomes a better-behaved and more
`symmetric wavelet.
`
`Principle of polarity coding and decoding
`The second method employs polarity reversals, rather than time delays, as a field coding
`process. Similar techniques have been used with the Vibroseis method [2]. Derivatives such as phase-
`
`PGS Exhibit 2014, pg. 2
`WesternGeco v. PGS (IPR2015-00309, 310, 311)
`
`

`
`3
`encoded sweeps in single or cascaded forms have been particularly successful. Because of the rather
`strong variability of land wavelets, the emission of polarity-encoded sweeps has mostly taken place at
`a single location rather than at several source locations. The method proposed here can be used in
`combination with other Vibroseis methods and can also be used in a marine environment in
`combination with end-fire sources. In the marine case, the positive signal is created by end-firing
`source elements in the downward direction, while the negative signal is generated by end-firing
`upwards, the signal becoming negative upon reflection against the sea surface. In Figure 2 the
`principle of polarity encoding using end-fire sources is illustrated. Although the signals are not
`perfectly opposite in this case, ideal discrimination can be achieved using three sources in a cycle of
`four shots. The advantage is that, not only is the sea surface ghost attenuated, but also that the ghost-
`like filter response, as seen in the time shift method, is absent.
`1
`4
`1
`4
`1
`4
`1
`4
`1
`4 1
`(a)
`(b)
`
`4
`
`1
`(c)
`
`4 1
`
`4 1
`
`4
`
`t [s]
`
`0.2
`
`0.4
`
`Source
`A
`
`Source
`B
`
`Source
`C
`
`Source
`A
`
`Source
`B
`
`Source
`C
`
`Source
`A
`
`Source
`B
`
`Source
`C
`
`Shot sequence number
`1
`4
`1
`4
`1
`4
`1
`4
`1
`Figure 2. Polarity coding and decoding — (a) minimal set of four traces shown for three sources individually,
`(b) after polarity decoding and (c) after polarity decoding and mixing.
`
`When using perfect opposite signals, the method allows for the deployment of one additional source,
`firing continuously at constant polarity. In that case, the minimum cycle has a period of two shots,
`whereby a second source emits alternately positive and negative signals.
`
`Field data example
`A field experiment was conducted in the Port Isabel area of the Gulf of Mexico. The
`WesternGeco Monarch acquired a 2D line, using two subarrays as two independent sources. The
`sources were towed with a lateral separation of 50 m and at a depth of 5 m. The time delay of 6.67 ms
`was selected to correspond with the sea surface ghost delay. Due to weather conditions, the cables had
`to be lowered to 10 m from the planned 6 m. The data were processed through a standard processing
`sequence to stack. In addition, the relevant traces were polarity reversed in three parallel flows
`according to the delay coding schemes of sources A, B, and C, as shown in Figure 1. The results,
`shown in Figure 3, were obtained without any further special processing, and leaving the required
`mixing to the CMP stack. Because one of the decoding schemes pertained to a third, not used source,
`the corresponding stacked record should show no data. In fact, there is residual energy that is
`approximately 15 dB below that of the actual sources, indicating a discriminating power of 15 dB
`through stack alone. Dedicated techniques, under development, indicate a potential prestack
`discrimination of more than 12 dB. We estimate that the compounded effects of the prestack, stack and
`migration processes allow for more than 30 dB, possibly 40 dB, in source signal separation.
`
`EAGE 64th Conference & Exhibition — Florence, Italy, 27 - 30 May 2002
`
`PGS Exhibit 2014, pg. 3
`WesternGeco v. PGS (IPR2015-00309, 310, 311)
`
`

`
`4
`
`550
`
`600
`
`650
`
`700
`
`750
`
`Missing source
`
`Port source
`
`Starboard source
`
`3.0
`
`t [s]
`3.5
`
`4.0
`
`3.0
`
`3.5
`
`4.0
`
`550
`
`600
`
`650
`
`700
`
`750
`
`550
`550
`
`600
`600
`
`650
`700
`CMP number
`
`750
`750
`
`3.0
`
`t [s]
`3.5
`
`4.0
`
`3.0
`
`3.0
`
`t [s]
`3.5
`
`3.5
`
`4.0
`
`4.0
`
`3.0
`
`3.5
`
`4.0
`
`550
`600
`750
`CMP number
`550
`600
`CMP number
`Figure 3. Data recorded from actual and missing sources — Tests involved deployment of two sources with the
`possibility for a third ‘missing’ source (top left). Shown are stacked data from the: starboard source (top right),
`port source (bottom left), and missing source (bottom right).
`
`750
`
`Conclusion and remarks
`The methods presented herein have the potential for a source discrimination of more than 30
`dB. Experiments will be carried out to test the practical limits, while various alternative uses will be
`explored in the near future. A present risk free application is found in the combination with other
`methods, such as those of Beasley et al. [1], Walker et al. [3], or with currently used simultaneous
`Vibroseis methods, for achieving additional discrimination.
`
`Acknowledgement
`We acknowledge Nick Moldoveanu and Jonathan Say for their contributions to the field test
`and the crew of the WesternGeco Monarch for conscientiously conducting the test survey in the Gulf
`of Mexico. We thank WesternGeco for permission to publish this work.
`
`References
`1. Beasley, C.J. and Chambers, R.E., 1999, Method for acquiring and processing seismic data: US
`patent 5,924,049.
`2. Silverman, D., 1979, Method of three dimensional seismic prospecting: US Patent 4,159,463.
`3. Walker, R.C., and Lindtjoern, O., 1999, Method and apparatus for marine seismic surveying: US
`Patent 5,973,995.
`
`PGS Exhibit 2014, pg. 4
`WesternGeco v. PGS (IPR2015-00309, 310, 311)