2012 SEG SEG Las Vegas 2012 Annual Meeting Page 1

Transcription

1 Full-wavefield, towed-marine seismic acquisition and applications David Halliday, Schlumberger Cambridge Research, Johan O. A. Robertsson, ETH Zürich, Ivan Vasconcelos, Schlumberger Cambridge Research, Dirk-Jan van Manen, WesternGeco, Robert Laws, Schlumberger Cambridge Research, Kemal Özdemir and Halvor Grønaas, WesternGeco Summary A four-component (4C) streamer recording pressure, as well as the three-component (3C) particle velocity vector, addresses long-standing geophysical problems such as receiver-side sampling and deghosting. In this paper, we introduce multicomponent marine seismic sources generating monopole and dipole responses in the water. We describe a few different alternatives for generating such a source using existing technology. Three different application areas are described in some detail: source-side deghosting, source-side wavefield reconstruction and, finally, a vector-acoustic reverse-time imaging approach that requires monopole and dipole data on both the source and receiver side of the acquisition. Introduction Robertsson et al. (2008a) introduced the concept of a fourcomponent (4C) streamer. Such a streamer includes recording elements of monopole character (the hydrophones) as well as elements of dipole character (3C particle velocity/acceleration sensors). Robertsson et al. (2008a) demonstrated that benefits of a 4C streamer include reconstruction of the received wavefield in the crossline direction and its decomposition into up- and downgoing wavefields (deghosting). Robertsson et al. (2008b, 2012) discuss how the equivalent source-side elements would provide similar benefits for the generated wavefields. Fullwavefield towed-marine seismic acquisition may, therefore, include as many as 16C source/receiver data. Marine seismic sources are almost exclusively composed of sub-elements that best can be characterized as emitting a monopole wavefield (including air-gun or explosive sources). A dipole source, on the other hand, can be thought of as two monopole sources located an infinitesimal distance apart radiating the same wavefield but with opposite polarity. Whereas a monopole source radiates an isotropic wavefield, the directionality of the dipole response results in three possible, independent, orthogonal sources. Because the sea surface has a reflection coefficient of -1 (in terms of pressure) the farfield response from a monopole source under the sea surface is similar to that of a vertically oriented dipole. However, in practice, this fact does not help us as most applications will require both monopole and dipole sources at the same source locations. In this paper, we expand on the work of Robertsson et al. (2012) who investigated different approaches that would allow us to generate both monopole and dipole responses at the same source locations. We also outline three application areas for such data: source-side deghosting, source-side reconstruction, and full-wavefield vector acoustic imaging. Figure 1: (a, b) Two sequences of 18 signatures, for air-guns of varying size. These sequences are designed using a simulated annealing approach that minimises the maximum value of the deconvolution of one sequence by the other. (c) The blue line shows the deconvolution of sequence (a) by itself, the black line shows the deconvolution of sequence (b) by sequence (a). The energy of the second sequence is distributely evenly in time when deconvolved by the energy of the first sequence. SEG Las Vegas 2012 Annual Meeting Page 1

2 designed using constructions developed in number theory that produce sets of sequences with optimal properties (Fan and Darnell 1996). Alternatively, optimization techniques such as simulated annealing can be used to design sequences with optimal properties. Figure 2: An illustration of the deghosting property of the vertical source gradient. The farfield radiation patterns of a single monopole source (red line) and a single weighted vertical dipole source (blue line) deployed at the same depth (10 m in this case) have complimentary ghost functions. When combined as described in equation 1, the result is a ghost free amplitude spectrum (black line). Methods for acquiring multicomponent marine seismic source data The most commonly used marine seismic source is the airgun array. A typical air-gun array contains several subarrays each with a string of air-gun elements in it. To retrieve the dipole response, we suggest generating the separate responses from two closely spaced monopole sources. This can be done in several ways. First, we can tow two or more air-gun arrays at different depths to generate the response from a vertical dipole, or towed with laterally-offset positions to generate horizontal dipole responses. By using simultaneous source separation techniques (Moore et al., 2008), by using orthogonal optical codes (OOC; see below), or a combination of both, we can separate the responses from individual air-gun arrays. The difference in response between two closely spaced (and separated) shot points will give us the dipole response from a point located between the two shot points (Robertsson et al., 2008). We can also treat individual subarrays in an air-gun array as separate sources and fire them independently. By configuring air-gun subarrays so that they are closely separated horizontally and/or vertically, dipole responses in different directions can thus be generated. Source signatures can be encoded using digital sequences (OOCs) where each air-gun pop represents a spike in the sequence, and the sequence from one array will be orthogonal to the other sequences. By utilizing the orthogonality property of the OOCs, the signal from each array can be decoded from the recorded data. OOCs can be In conventional seismic surveys, air-gun arrays are tuned to have properties that are desirable for seismic surveying. The arrays firing orthogonal sequences are not tuned in this way. This may present a problem if the source signature of this detuned array cannot be deconvolved adequately. Ziolkowski (1987) showed that it is possible to deconvolve the source signature from a detuned array (the so-called machine-gun array) and produce an acceptable wavelet as a result. This observation is of great importance here, as the deconvolution of the combined OOCs from the two closely spaced sources yields the monopole response at a point between the two source arrays. In Figure 1a and b, we show a pair of such sequences, each composed of 18 individual air-gun signatures. These are designed using a simulated annealing approach, and the design criteria minimizes the peak amplitude of the deconvolution of one sequence by the other (because the designature operation is in practice a deconvolution rather than a crosscorrelation). The deconvolved sequences are shown in Figure 1c. Finally, we note that because source signatures in marine vibrators can be controlled with great precision, arrays of marine vibrators lend themselves favourably to generate monopole and dipole responses as described above by means of orthogonal sweeps. Applications: Source-side deghosting Consider a single recording location in a marine seismic experiment with two types of sources in the same horizontal plane: monopole sources and vertical dipole sources. The configuration of sources that would be required to emit a downgoing wavefield on the source side is:,,,,,,,,, (1) where is the density, is angular frequency, and are the horizontal components of the wavenumber vector, and is the unsigned vertical component of the wavenumber vector, related to the two horizontal components and the velocity in water,, through the dispersion relation,,. SEG Las Vegas 2012 Annual Meeting Page 2

3 From equation 1 and the dispersion relation, we see that reconstructing the downgoing wavefield from a monopole source requires the monopole response from that point combined with spatially filtered vertical dipole source data in the same horizontal plane. A vertical incidence approximation to the dispersion relation results in and removes the need to filter spatially the dipole source data. Such a simple approximation will clearly have some limitations and may not always be adequate. In Figure 2 we show the farfield amplitude spectrum of an impulsive monopole source in red and the equivalent spectrum of the (weighted) vertical dipole source in blue. The farfield spectrum of the (deghosted) downgoing wavefield is shown in black. This spectrum is notch free as desired. Applications: Source-side reconstruction Figure 3: Common receiver gathers showing (a) monopole and (b) inline dipole responses generated by acquiring data using closely spaced sources, with timing delays allowing for simultaneous source separation. The shot sampling is 50 m. The monopole is the result of summing the two separated sources and the dipole is the result of differencing the two separated sources. Note, despite the level of separation noise being low, the differencing involved in computing the gradient boosts the level of this noise with respect to the signal. Figure 4: (a) Directly modelled common reciever gather for a monopole source sampled at 12.5 m. (b) Result of combining the monopole and dipole data in Figure 3 to interpolate from 50m to 12.5 m. Multichannel interpolation (i.e., using monopole and dipole responses) is used where the impact of separation noise is low, and single-channel interpolation (using only the monopole data) is used where the impact of separation noise is high. Just as the recorded particle velocities of the received wavefield are proportional to spatial derivatives or gradients of the pressure wavefield, we can think of dipole sources as representing source-side derivatives (or gradients). Identical methods to those described by Robertsson et al. (2008a) and Vassallo et al. (2010) are directly applicable to source-side interpolation. Sourceside reconstruction would be highly desirable as it would enable more efficient acquisition of marine seismic data and address the present imbalance between receiver sampling and source sampling densities in typical marine seismic surveys. In particular, it would address longstanding challenges in imaging and multiple attenuation, as well as aiding acquisition in areas with surface obstructions. We consider a synthetic example of inline source reconstruction. For simplicity, we assume that the receiver spread is fixed. There are 201 sources, sampled at 50 m, and 281 receivers sampled at 25 m. Sources and receivers are located at 20 m and 5 m below the air/water interface, respectively. The model has a realistic 2.5D geology based on a tilted fault block (invariant in the cross-line direction). A monopole source is used and the signature is a Ricker wavelet with a central frequency of 25 Hz. To measure inline dipole responses, two sources are fired; the first at a regular interval of 50 m, and the second source is fired 5 m behind the first. A small timing dither is applied to the second source, resulting in a spatial variation of plus or minus 1 m, enabling simultaneous source separation (Moore et al., 2008). These closely spaced sources can then be used to find the finite-difference approximation to the inline spatial derivative of the source, i.e., a dipole pointing in the inline direction. In Figure 3a, we show a single common-receiver gather for the monopole responses extracted by summing our dual-source recordings after simultaneous source separation, and the equivalent dipole SEG Las Vegas 2012 Annual Meeting Page 3

4 Figure 5: Using the SEG Advanced Modelling (SEAM) model, panel (a) shows the full vector acoustic (VA) image from dual-source 4C data, while its separate contributions from pressure- only sources is shown in (b). source responses computed by differencing the dualsources after separation are shown in Figure 3b. Note, that the residual separation error is amplified relative to signal when we difference the sources (Figure 3b), rather than sum them (Figure 3a). The monopole sources are combined with the dipole sources to interpolate the shot interval from 50 m to 12.5 m, using multichannel interpolation by matching pursuit (MIMAP) (Vassallo et al., 2010). Due to the effect of separation noise in the gradient estimate, we apply the multichannel interpolation technique in regions unaffected by the separation noise, and apply the single-channel equivalent in areas with high separation noise. These areas occur later in the shot gathers, and hence the data are less aliased than at earlier times. In Figure 4a, we show the same common receiver gather as in Figure 3a, but here the monopole sources are directly modeled at a sampling interval of 12.5 m. Figure 4b shows the same gather, but computed by combining the data in Figure 3 to interpolate from 50-m to 12.5-m sampling. The MIMAP algorithm (Vassallo et al., 2010) requires only a few input points to be effective and is, therefore, particularly suitable for such applications where the receiver spread is not stationary. It also is naturally suited to operating on irregular data, such as those generated here when dithering one of the source locations to compute the inline dipole response. Applications: Full-wavefield vector acoustic imaging Assuming 4C records of seismic data generated using both pressure/monopole and gradient/dipole sources, Vasconcelos (2011) and Vasconcelos et al. (2012) propose a full-wavefield vector-acoustic (VA) reverse-time imaging method that can jointly image up- and downgoing wavefields. In this method, multicomponent data on both the source and receiver sides are required in the wavefield extrapolation as well as imaging conditions. When considered separately, VA imaging from either monopoleor dipole-only sources yields different subsurface images (Figure 5), which correctly focus all receiver-side ghost energy from the 4C data, but cannot properly focus arrivals related to source-side downgoing waves. However, when both sources are combined into a single VA image all the up- and downgoing energy is correctly migrated (Figure 5a). Conclusions We introduced methods for constructing multicomponent marine wavefields from marine sources. These methods involve recording wavefields from two closely spaced monopole sources, and using those monopoles to compute separated dipole and monopole responses. Such data can be acquired efficiently using dithered simultaneous sources, or using orthogonal optical codes that allow overlapping shot gathers to be separated. We described several applications of these multicomponent sources, including source-side deghosting and reconstruction, and full-wave vector-acoustic imaging. In a synthetic example of source-side reconstruction, we illustrated the interpolation of sources in the inline direction. This may enable increased towing speeds and improved source sampling. In summary, multicomponent receivers were shown to solve numerous problems in marine seismic data processing, and the introduction of a multicomponent marine source promises similar benefits on the source side. SEG Las Vegas 2012 Annual Meeting Page 4

5 EDITED REFERENCES Note: This reference list is a copy-edited version of the reference list submitted by the author. Reference lists for the 2012 SEG Technical Program Expanded Abstracts have been copy edited so t hat references provided with the online metadata for each paper will achieve a high degree of linking to cited sources that appear on the Web. REFERENCES Fan, P., and M. Darnell, 1996, Sequence design for communications applications: Chapter 15, optical orthogonal sequences: Research Studies Press. Moore, I., B. Dragoset, T. Ommundsen, D. Wilson, C. Ward, and D. Eke, 2008, Simultaneous source separation using dithered sources: 78th Annual International Meeting, SEG, Expanded Abstracts, Robertsson, J. O. A., D. Halliday, D-J. van Manen, I. Vasconcelos, R. Laws, K. Özdemir, and H. Grønaas, 2012, Full-wavefield, towed-marine seismic acquisition and applications: 74th Conference and Exhibition, EAGE, Extended Abstracts. Robertsson, J. O. A., I. Moore, M. Vassallo, A. K. Özdemir, D-J. van Manen, and A. Özbek, 2008a, On the use of multicomponent streamer recordings for reconstruction of pressure wavefields in the crossline direction: Geophysics, 73, no. 5, A45 A49. Robertsson, J. O. A., D-J. van Manen, D. Halliday, and R. Laws, 2008b, Seismic data acquisition and source-side derivatives generation and application: U. S. Patent 7,876,642. Vasconcelos, I., 2011, Source-receiver reverse-time imaging of vector -acoustic seismic data: 81st Annual International Meeting, SEG, Expanded Abstracts, Vasconcelos, I., J. O. A. Robertsson, M. Vassallo, and D -J van Manen, 2012, Reverse-time imaging of dual-source four-component seismic data using primaries, ghosts and multiples: 74th Conference and Exhibition, EAGE, Extended Abstracts. Vassallo, M., A. Özbek, K. Özdemir, and K. Eggenberger, 2010, Crossline wavefield reconstruction from multicomponent streamer data: Part 1 Multichannel interpolation by matching pursuit (MIMAP) using pressure and its crossline gradient: Geophysics, 75, no. 6, WB53 WB67. Ziolkowski, A. M., 1987, The determination of the far -field signature of an interacting array of marine seismic sources from near-field measurements: Results from the Delft Air Gun experiment: First Break, 5, SEG Las Vegas 2012 Annual Meeting Page 5