A New Ocean Bottom Seismic Node System Peter Maxwell, Sergio Grion, Shuki Ronen, Tor Haugland / CGGVeritas, Town Park Dr., Houston, USA,TX 77072

Transcription

1 OTC Paper Number OTC PP A New Ocean Bottom Seismic Node System Peter Maxwell, Sergio Grion, Shuki Ronen, Tor Haugland / CGGVeritas, Town Park Dr., Houston, USA,TX Copyright 2007, Offshore Technology Conference This paper was prepared for presentation at the 2007 Offshore Technology Conference held in Houston, Texas, U.S.A., 30 April 3 May This paper was selected for presentation by an OTC Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Offshore Technology Conference, its officers, or members. Papers presented at OTC are subject to publication review by Sponsor Society Committees of the Offshore Technology Conference. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Offshore Technology Conference is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, OTC, P.O. Box , Richardson, TX , U.S.A., fax Abstract New seismic technology aims at imaging the blind spots of conventional methods. One such new technology is Ocean Bottom Station (OBS) nodes. Such nodes record four component (4C) data; the water-borne pressure and the three components of seabed motion. Oceanographic Research groups have long been using OBS. After testing one of the leading brands of academic type OBS technology in 2002, and getting excellent data, we developed a commercial system which is deployed using Remotely Operated Vehicles (ROV). In 2006 we tested the commercial OBS in Louisiana. Because there are no cables joining the nodes together, they are well suited for working in extremely congested waters, even right up to or perhaps under production structures. Seismic monitoring surveys (4D) are practical because the ROV can re-position the units very closely to their previous sites to ensure a high level of repeatability. The ability to operate near obstacles is for surveys in congested areas typical of development 3D and seismic monitoring (4D) surveys. The survey geometry enables improved imaging of complex structures under complex overburden such as salt. Wide azimuth geometry and well sampled common receiver gathers are natural for OBS. These provide better illumination and facility for Wave Equation Migration (WEM). The P waves are recorded on both the hydrophone and the geophones and enable separation of the waves to down- and up-going waves, which in turn enables demultiple and imaging of waves which are conventionally considered useless multiples. Shear waves are recorded by the geophones on the seabed. Such shear waves are generated by P to S conversion so they do not require special shear sources. The converted PS reflections provide additional information to what is in the PP reflections. The additional information provided by the shear waves is useful for improved characterization of lithology and fractures. We present the engineering, design and field data from the 2002 and the 2006 tests. We used these data to test and demonstrate demultiple, and imaging of multiples and converted waves. Introduction The world demand for energy is accelerating, while its reserves of energy are diminishing. Producers are compelled to explore and produce in more challenging places and to maximize recovery in existing reservoirs. The application of technology has always been a key to success in such situations. Advances in seismic technology aim at imaging the blind spots of conventional methods and offering new options for reservoir monitoring (4D). One such new technology is from recent developments in ocean bottom seismometers (OBS) or nodes. OBS nodes offer better data than conventional acquisition systems, how do they do this? One reason is that they provide wide azimuth geometry. We now know that wide azimuth coverage is important for imaging structures under complex overburden such as in sub-salt imaging. Salt bodies act like huge lenses distorting seismic waves propagating through them. To image sub-salt targets we must have the data processing capability to image through complex overburdens, but even the best imaging technology alone is not enough, we also have to illuminate the targets. What we do not illuminate we cannot image. Conventional towed-streamer surveys are operated with a single seismic vessel and fixed source-receiver geometry resulting in a narrow azimuthal coverage. If either the source or the receiver is located above an overburden anomaly, the illumination of the target is likely to be poor. Wide azimuth geometry is much more likely to be successful because we record a dataset well-populated in azimuth and offset. So, if we don t illuminate a target with one azimuth, we still have a good chance to illuminate it with other azimuths. Two techniques that provide wide azimuth geometry offshore are multi-vessel towed-streamer and OBS. Wide azimuth towed-streamer (WATS) surveys are more economical for covering large areas while OBS are more economical for small areas, up to about 400 km 2. OBS nodes are deployed by ROV, and can be placed very accurately. This makes them more practical to deploy in the presence of obstacles such as

2 2 OTC PP production facilities. For the purpose of seismic monitoring with repeat surveys (4D), OBS have better positioning repeatability than streamers. Last but not least, OBS nodes provide multi-component data. We can use such data for separating up- and down-going waves at the seabed for multiple attenuation and for imaging using the multiples, while shear waves can be used to provide information about lithology and fractures. Shear waves can sometimes allow us to image targets which have low reflectivity or are under gas clouds. In 2002 Veritas decided to review technology that would best suit deep-water exploration. Universities and Oceanographic Research groups have been using OBS nodes for many years. The nodes are autonomous recording units that are released and allowed to fall through the water to the seafloor where they sit and record seismic activity over long periods of time (months). The nodes can be triggered to release their anchors and rise to the sea surface for collection. Typically only a small number of units are deployed by these research organizations. Such nodes are termed self-landing and ascending (SLA) OBS. Initial tests Figure 1: Academic type SLA- OBS unit, tested in 2002 over the Britannia field, North Sea. Veritas tested one of the leading brands of SLA- OBS on two projects in the North Sea during 2002 (Figure 1). The SLA-OBS has a three component external geophone and a hydrophone mounted on a deployment mechanism. These are connected to an autonomous four channel recorder with integral clock and batteries. An anchor weight is used to sink the node to the seafloor and an acoustic release is provided to disconnect the anchor when the node is to be recovered. A hardened foam floatation unit is designed to withstand the pressure without crushing and to provide buoyancy to bring the recorder to the surface when the anchor is released. The SLA-OBS tested has a complicated mechanism to deploy the external geophone when the unit landed on the seabed. This type of SLA-OBS was chosen for the test because we believed the external geophone would provide better data quality than other SLA-OBS types with internal geophones. The external geophone used had been designed and evolved over many years to provide good coupling with the seabed. The SLA-OBS were dropped from a chase boat and they descended to the seabed slowly at a rate of about one meter per second, drifting laterally in the currents. Upon landing, the seabed pushed up on the leg of the geophone deployment mechanism, which then released an arm that deployed the external geophone. This arm and a leg apparatus worked most of the time. However, for about one deployment in ten the soft seabed would not push the leg all the way up and the external geophone would not deploy at all. This was alone was considered an unacceptable rate of failure, but moreover even when the arm did deploy, the geophone would swing and sometimes land at large tilt angles. We used omni-directional geophone elements which work at any tilt, but the package itself, in the shape and size of a Frisbee, was designed to couple properly while being approximately flat. The units were left in place to record streamer source activity. After a few weeks of continuous recording each node was pinged with its unique acoustic release code. The anchor was released and the OBS would start ascending leaving the anchor behind on the sea floor. The OBS ascended at a rate of about one meter per second to the sea surface where we would be watching to retrieve them, getting plenty of man-overboard (MOB) maneuver practice in the process! Sometimes, probably where the seabed was very soft, the leg would be deeply embedded in the mud and it would take the OBS a long time to break free and start ascending, while we would be anxiously waiting above. One condition of operation imposed by the DTI was that an attempt be made to retrieve the anchors. A float and rope technique was tried but proved essentially unviable. A Veritas crew operated the SLA-OBS units successfully throughout this project without either injury or loss with 45 OBS-node- deployments and retrievals Overall, we acquired good data from about 80% of the geophone deployments. This, although disappointing, was an improvement from the reported academic rate of success of about 60%, because we used the omni-directional geophones. The academic units used lower natural frequency geophones designed to record to lower frequencies, but which tolerate just a few degrees tilt. Of course, we did not know if the external geophone deployed at all until we retrieved them, and even then we would not know if they deployed correctly until the data was processed. The hydrophone success rate was 100% as they only require contact witht the water. At the end of the test we concluded that the SLA-OBS were not a practical system for us, but that we did indeed get excellent data. In particular, the external geophones provided better data than any ocean bottom cable (OBC) or other OBS that we had seen. In OBC, the geophones are usually in a cylindrical package, which is subject to rolling. Alternatively, if the cable is buried in the seabed while being dragged into position, the coupling is affected by the tension in the cable. Any of these effects cause a difference between the coupling of the horizontal in-line and cross-line components. The Frisbee shaped external geophones on the other hand did not show any differences in in-line or cross-line coupling It was clear to us that the SLA-OBS that we used were not suitable for commercial OBS operations. Perhaps we could improve the geophone deployment and anchor retrieval mechanisms, but the SLA-OBS were practical only in small numbers. For full scale 3D OBS surveys, or even for undershooting platforms, we needed many hundreds of OBS and we knew that the MOB maneuver retrieval was impractical with such large numbers. However, the 4C data

3 [OTC PP] 3 was enticingly good. We decided to put our efforts into developing a system designed to be capable of full-scale 3D operations. As we got busy with the development, SLA-OBS were left aside. Development and testing Ambitious project goals were set for the development of the nodes. The prime objectives were safety and operational efficiency. To operate OBS in large numbers it would be necessary to use a different deployment/retrieval method. Remotely Operated Vehicle (ROV) deployment was chosen as the best method for deep water and so contacts were initiated contacts with ROV operators. One critical technical goal was to reduce the power consumption of the data recorder and clock, in order to increase the operating time-endurance to 3 months. Low power requirements also reduce the size and weight of the unit and might allow redeployment of the nodes without charging or changing batteries. In parallel to developing a new recorder, we wanted to redesign everything else in the OBS. With ROV operations, the floatation, the anchor and the acoustic release were redundant. For HSE and operational reasons, an integrated package implying an internal geophone, was preferred. To develop the necessary technology two consultant companies were engaged: Carrack for the node chassis and SEND for the recorder electronics. Carrack has years of experience designing OBS units and had already designed the very successful remotely-deployed geophones used in the 2002 tests. SEND has a strong reputation and extensive experience in supplying OBS recorders. The design of the chassis incorporates symmetry and a low centre of gravity, since these optimize coupling to the seabed. Essentially our chassis is an up-scaled disc shaped case with all the key features of the well accepted remote phones used in the academic OBS systems. However, it is not possible to simply up-scale the SLA geophone design, since this would result in excessive weight and compromise the required pressure rating for operation at the desired water depth (4000m). Carrack s solution was to make a rigid circular baseplate on which 3 separate pressure cylinders are mounted symmetrically (Figure 2). Figure 2: OBS node, right with cowling and top-plate removed to see internal structure. The three cylinders contain the recorder and two battery packs. In this design, the pressure cylinders are kept as small as possible resulting in the smallest wall thickness to support the external pressure; this in turn results in a considerable weight saving over larger cylinder designs. A fourth pressure housing contains a three component Gal Perin geophone; the latter is arrangement to be consistent with the triangular symmetry of the unit. A hydrophone is mounted externally. The housing penetrates the circular baseplate coaxially and the upper surface of the node is streamlined by a cowling that avoids snagging of the internal cables. The centre of the cowling is closed by a flat top-plate to allow an ROV suction handling system to be used for deployment and retrieval. The underside of the base-plate (Figure 3) has a symmetric arrangement of ribs and grooves that is similar to that used on the SLA-OBs external geophone. These ribs and grooves help in channelling liquid mud out from under the unit as it is deployed, thereby allowing the unit to sink a few centimetres into the sediment and to contact more Figure 3: Underside of baseplate showing grooves. consolidated material. The base and cowl have holes for water ingress to allow free flooding of the unit. Only the pressure cylinders are water-tight. The triangular symmetry of the design and operating environment led to the adoption of the name Trilobit for these new nodes. A full time experienced team was formed to manage the development project, test the prototypes, and to start commercializing the Trilobits. A number of other companies have recently developed OBS nodes, notably Fairfield in co-operations with BP on their Atlantis project (Beaudoin and Michell, 2006; Mitchell and Grisham, 2006). While there are obvious superficial similarities among the models, (e.g. circular shape, yellow colour for visibility, handling method), each has been developed independently and the internal design details are quite different. The technology development has been on a rapid timescale with a simple field test of two prototype units conducted in the UK during August By the end of 2005, three prototype units were tested with explosive sources in a pond in Kansas. These valuable tests confirmed excellent data quality and highlighted the need for improvements that have been incorporated into a production design. An alpha test was completed with 20 nodes piggy-backed on a transition zone project in Louisiana in Thirty more units are being built and a number of test projects are being discussed with clients. Test Data The first field tests of production units took place in May July Twenty Trilobits were deployed within the area of

4 4 OTC PP a conventional Transition Zone (TZ) project. The survey was in shallow water (2-3 m), over a salt dome in southern Louisiana. We recorded data from some of the TZ survey dynamite shots (5 kg at 50 m). Nineteen Trilobits were of the production design and one prototype unit had the internal geophone removed and was connected to an external SLA phone of the same type that provided good results in The objective was to compare the data from the new untested Trilobits to the data from the tried and tested external Carrack phone. We made two deployments (Figure 4). In the first the water bottom was muddy. The second deployment was in an area with pond weed and mud. Figure 4: Deploying a node in shallow water. The test was conducted alongside an exclusive proprietary survey and we thank our client for the permission to deploy the Trilobits and record data from the shots. The client prefers to remain anonymous, but we have permission to present un-stacked data without imaging and stacking. polarized. Significantly, there is no leakage of shear waves into the vertical geophone components. This has been a problem with OBC and some other node designs. The noise cone is from surface waves (Love) in the lake bed and is strongest on the two horizontal components. Figure 8 shows the hodogram of a Love wave. It is elliptically polarized in the horizontal plane. Figure 6: Hodograms of a PP reflection Results A sample of the data from the Trilobits is shown in Figure 5. This data is a common receiver gather from a sngle shot line. Figure 7: Hodograms of a PS reflection. Figure 5: Test data from a Trilobit OBS node. PP reflections are clearly seen on the vertical component (Z) and hydrophone. Converted shear waves (PS) have slower moveout than PP waves and are visible on the two horizontal components (X, Y). Hodograms of a PP event are shown in Figure 6. Hodograms are cross-plots of recorded X, Y and Z amplitudes. The data window used to compute the hodograms is highlighted by a black box. As expected, the near-offset PP event is vertically polarized while the PS event in Figure 7 is horizontally Figure 8. Hodograms generated for a Love wave.

5 [OTC PP] 5 The second deployment was in an area with water weeds (Figure 9). We speculate that the weeds prevented good coupling of the external SLA phone to the lake bed. The main symptom of the poor coupling is the resonant nature of the data without clear P or S waves separated by polarization into vertical and horizontal components (Figure 10). We show these data to demonstrate that not all data from all OBS are good. The data in Figures 5-8 are from Trlobits with internal geophone deployed in a weed-free area. Although not shown, Trilobits with internal geophones did have good coupling in the weedy area with data comparable to Figures 5-8. illumination, especially of reflectors whose depth under the seabed is less than the node intervals (Figure 11). In addition, if an OBS node fails, this problem is greatly exacerbated. Figure 9: A benchmark OBS, with a prototype Trilobit with and external geophone. Unfortunately this deployment was in and area with weeds which evidently interfered with the coupling Figure 11: Poor illumination of sparse OBS. Note the gaps in shallow reflectors coverage. This problem is exacebrated if any OBS nodes fail. Notation: shots are shown as black stars, receivers dead and alive are in yellow and black respectively. The processing of OBS data must address this sparsity in receiver sampling. Fortunately, there is a good solution for this problem. The OBS data recorded by hydrophones and geophones can be combined to separate up-going from downgoing waves (White, 1965; Barr and Sanders, 1989). Conventionally, only up-going primary reflections are imaged. However, down-going receiver ghosts bounce from the same reflectors as the primary waves. Previous authors have successfully imaged ghosts recorded in various contexts: by ocean bottom hydrophones (Godfrey et al., 1998), VSP (Jiang et al., 2005) and OBS nodes (Ronen et al., 2005). This is not yet, but we expect will soon be, standard processing procedure for OBS and OBC data. OBC cables offer dense sampling in the cable direction but sparse crossline sampling, comparable to the OBS sampling. Down-going receiver ghosts are recorded after reflection from the sea surface. Therefore, when imaging ghosts, the sea surface acts as a mirror reflecting the image of subsurface structure (Figure 12). Figure 10: Data from the external geophone shown in Figure 9 Mirror imaging Deploying OBS nodes on the seabed takes considerable time and ROV boats have high operating cost. Therefore, an economic geometry for OBS acquisitions is a sparse grid of nodes (400 x 400m) and a dense grid of shots (50 x 50m). However, the sparse node geometry provides poor Figure 12 The sea surface acts as a mirror for primary reflections. The down-going receiver ghost is an up-going primary reflected downward from the sea surface. With mirror imaging we image the down-going receiver ghost as if it were recorded not on the seabed but on a surface as high as the sea is deep. The source is kept at the original level

6 6 OTC PP (Figure 13). paper we used only the one shot line which was approximately above the OBS. The conventional image produced from the primaries and the image produced from the ghosts, are shown in Figure 16. Figure 13 Imaging the ghost is equivalent to imaging primaries recorded on a surface as high as the sea is deep The down-going receiver ghosts consist of up-going primaries that reverberate once in the water layer. These are in general more informative then the up-going primaries because they offer extended illumination of subsurface reflectors (Figure 14). Figure 14 Illumination of the upgoing wave (a) is narrower than that of the downgoing waves (b). In particular, the seabed cannot be imaged with the upgoing waves but it can be imaged with the downgoing waves (c). Another factor contributing to an improved image from the ghosts is velocity anomalies and scattering just under the seabed. The receiver in (b) is in effect further away from the seabed anomalies than the receiver in (a). A third factor is that the ghosts (b) are traveling in effect closer to vertical than the primaries (a). We illustrate the advantages of mirror imaging using two field data examples from the North Sea (Figure 15). These two OBS datasets have comparable receiver spacing but different water depth. The advantages of mirror imaging are evident in both cases and more dramatic for the deep water example, as expected. The first dataset was recorded in 2002 in the North Sea offshore Norway, over the Britannia field. This data was acquired to test leading brands of academic type SLA-OBS, as part of the Trilobits research and development process.figure 1 shows one of the OBS units used. We imaged data from 9 OBS deployed at an interval of about 500m and at a depth of about 150m. For the purpose of this Figure 11 Location of the Britannia and Storegga surveys. As a second example, we applied mirror imaging to OBS data recorded offshore Norway on the headwall of the Storegga Slide. This data was acquired using academic OBS nodes similar to those in Figure 1. Storegga, which means Great edge in Norwegian, is the largest known submarine slide. It is located 100km offshore Norway above the giant gas field Ormen Lange. About 8100 years ago, an area of continental shelf about the size of Ireland collapsed in a series of retrogressive events creating a large tsunami in the North Atlantic. Sediments in this region contain gas hydrates, an ice like substance composed of methane trapped within a lattice of water molecules. They are stable in cold deep water environments where conditions are favorable. At a certain depth under the seabed, primarily due to the increased geothermal temperature, the hydrates change from a solid ice phase to a liquid gas phase. A Bottom Simulating Reflector (BSR) caused by the resulting velocity contrast marks the boundary between the two phases. Methane hydrates are important for predicting seabed stability and may also become a huge hydrocarbon reserve if and when production technology is developed. To study methane hydrates and other features of Storegga, the OBS data object of this study provides wide azimuth P and S data on the steepest part of the headwall. The survey targets are shallow sediments containing gas hydrates. These were imaged using down-going waves, demonstrating how a target less than 300m under the seabed can be imaged with OBS deployed 400m apart, even in the presence of node failure (Figure 17). We find that the image from the ghosts is better than the image from the primaries (Figure 16, Figure 17). The main reason is that the illumination is usually better, especially for shallow targets. Illumination is discussed in more detail in the next paragraph. Another reason might be that the ghosts are

7 [OTC PP] 7 less susceptible to velocity variations just under the seabed, which cause scattering, amplitude variations, and statics on OBS and OBC (Figure 13). The ghosts travel through the water twice (up and down) after going through the seabed anomalies. Longer travel through the water is an advantage for the same reason that deep-water streamer data have less seabed-associated statics than seabed and shallow water streamer data. Arguably, a third factor is that the ghosts travel closer to vertical than the primaries (Schuster 2005, pers. comm.). As noted above, mirror imaging usually provides improved illumination, expecially for shallow targets. However illumination depends on the velocity model and is not a property of the acquisition geometry alone. It is possible that, for a given velocity model, conventional imaging could provide better illumination than mirror imaging in some subsurface areas. The joint migration of primaries and their ghosts may provide improved results over mirror imaging alone. This migration would double the cost of mirror or conventional imaging as for each receiver at the seabottom (Figure 14-a) there would also be a mirrored receiver (Figure 14-b). It should also be noted that the wider illumination provided by mirror imaging comes at the expense of decreased illumination fold. For example, in Figure 14a and Figure 14b the number of considered rays is the same, but in the mirror imaging case (Figure 14b) they are distributed over a wider area. However, good data quality achievable with OBS acquisitions is expected to provide reliable images regardless of this fold decrease. The examples shown confirm this observation. Conventional up-going waves imaging has the advantage of being less sensitive to rough sea effects than mirror imaging. In rough sea conditions, the height of the water column is time and space variant and both source and the receiver ghosts are affected. Conventional imaging, while still influenced by the source-side ghost, is not affected by the receiver ghost. Nevertheless, wave height in normal seismic acquisitions operations is of the order of a few meters and this effect is normally considered negligible in marine seismic data processing. Techniques to attenuate the rough sea effect are discussed in Kragh and Laws (2006). The data examples shown were acquired with a self landing ascending OBS with an external geophone like the one shown in Figure 1. Based on our test in Louisiana, we expect that Trilobit data will be comparable to, or better quality than, the data from the OBS with the external geophone. signals. Due to cost constrains, OBS acquisitions necessarily have sparse receiver geometry. Poor illumination and sensitivity to velocity anomalies are problems that negatively affect the processing of such sparse data. We find that we can produce better images from the ghosts, using the mirror imaging method, than from the conventional images produced from the primary reflections. This is mainly because of improved illumination and reduced exposure to shallow inhomogeneous anomalies under the seabed. Acknowledgements The authors gratefully acknowledge the contributions of the following CGGVeritas staff to the successful development of the Trilobit OBS nodes: Mark Cartwright, Neil Catling, Stuart Denny, Helmut Jakubowicz, Allan Johnson and Frans Vellema. The Storegga data is courtesy of the European Comission funded HYDRATHECH project EVK3-CT , processing was carried out at CGGVeritas. References Barr, F. J. and Sanders, J. I. [1989] Attenuation of water-column reverberations using pressure and velocity detectors in a waterbottom cable, 59 th SEG Annual Meeting, Beaudoin, G., Michell, S. [2006] The Atlantis OBS Project: OBS Nodes Defining the Need, Selecting the technology, and Demonstrating the Solution. OTC.06. paper OTC Godfrey, R. J., Kristiansen, P., Armstrong, B., Cooper, M. and Thorogood, E. [1998] Imaging the Foinaven ghost. 68 th SEG Annual Meeting, Mitchell, S., Grisham, T. [2006] The Atlantis OBS Project: Developing and Building the OBS Node Technology. OTC.06, paper OTC White, J.E. [1965] Seismic waves: radiation, transmission and attenuation. McGraw-Hill. Ronen, S., Comeaux, L. and Miao, J.G. [2005] Imaging downgoing waves from ocean bottom stations. 75 th SEG Annual Meeting. Conclusions We have successfully developed an OBS unit that incorporates the necessary features for commercial 3D deployment. Our tests demonstrate that the unit is capable of providing highquality data containing clear primary (or pressure) P and secondary (or shear) S waves. Most importantly, the units show good coupling and do not appear to suffer from crosscoupling of energy, allowing good separation of P and S

8 8 OTC PP Figure 16 Conventional imaging of the up-going waves (left). Note the poor imaging of shallow reflectors. The OBS data were acquired with 500 meter intervals. The location of the nodes is shown in black. The image produced from the ghost (right) shows improvement especially in the shallow part. The sparse OBS receiver interval is typical of a cross line OBC interval in 3D. Similar improvements are expected when imaging 3D OBC data in the cross-line direction. We expect greater improvement in deeper water. Figure 17 Up-going (left) and ghost (right) migrated image from the Storegga slide, offshore Norway. Data acquired with seven OBS nodes 400 m apart. The location of the nodes is marked in black One node failed and is marked in yellow. The BSR reflector 300m below the sea bottom is accurately imaged using the mirror method despite the wide receiver separation.