SIMULTANEOUS VIBROSEIS RECORDING1

An experiment was undertaken at BP's Fulbeck Geophysical test site to compare the viability of various simultaneous vibroseis recording techniques, which are often recommended as a means of improving data acquisition production rates for 3D seismic surveys. Of particular interest were: (a) the ability to separate the signals from each source during processing, (b) the generation and suppression of harmonics and (c) the effects of any source interaction. Two vibrators were deployed with a baseplate separation of 10 m, about a borehole containing a vertical array of geophones. Our analysis concentrated on the groundforce signals measured at each vibrator and the far-field signatures measured using a vertical geo-phone at a depth of 204 m. By comparing single vibrator records with similar but separated records from a simultaneous recording sequence, signal separability, harmonic suppression and vibrator interaction could be fully studied. Separated far-field signatures from simultaneous vibroseis methods using combinations of up and downsweeps exhibited unsuppressed harmonics and substantial energy from the undesired source which leaked through the correlation process. The ‘up/down’ method was capable of separating the signal from each source by only 12.7 dB, and is therefore unsuitable as a field technique. The variphase simultaneous vibroseis methods studied afforded some harmonic suppression and gave signal separations of about 30.0 dB. Use of variphase simultaneous vibroseis methods will compromise the quality of the data recorded, when compared with single-source acquisition methods. None of the simultaneous vibroseis methods tested provided adequate signal separation and, therefore, cannot be recommended as data acquisition techniques. The ‘alternate sweeping’ method coupled with multispread recording will give the desired improvement in data acquisition rates, while preserving the necessary quality of our seismic data.     

Filtering vibroseis data in the precorrelation domain

Vibroseis data recorded at short source–receiver offsets can be swamped by direct waves from the source. The signal-to-noise ratio, where primary reflections are the signal and correlation side lobes are the noise, decreases with time and late reflection events are overwhelmed. This leads to low seismic resolution on the vibroseis correlogram. A new precorrelation filtering approach is proposed to suppress correlation noise. It is the ‘squeeze-filter-unsqueeze’ (SFU) process, a combination of ‘squeeze’ and ‘unsqueeze’ (S and U) transformations, together with the application of either an optimum least-squares filter or a linear recursive notch filter. SFU processing provides excellent direct wave removal if the onset time of the direct wave is known precisely, but when the correlation recognition method used to search for the first arrival fails, the SFU filtering will also fail. If the tapers of the source sweeps are badly distorted, a harmonic distortion will be introduced into the SFU-filtered trace. SFU appears to be more suitable for low-noise vibroseis data, and more effective when we know the sweep tapers exactly. SFU requires uncorrelated data, and is thus cpu intensive, but since it is automatic, it is not labour intensive. With non-linear sweeps, there are two approaches to the S,U transformations in SFU. The first requires the non-linear analytical sweep formula, and the second is to search and pick the zero nodes on the recorded pilot trace and then carry out the S,U transformations directly without requiring the algorithm or formula by which the sweep was generated. The latter method is also valid for vibroseis data with a linear sweep. SFU may be applied to the removal of any undesired signal, as long as the exact onset time of the unwanted signal in the precorrelation domain is known or determinable.     

Elimination of source-generated noise from correlated vibroseis data (the ‘ghost-sweep’ problem)

The most common noise-reduction methods employed in the vibroseis technique (e.g. spike and burst reduction, vertical stacking) are applied in the field to reduce noise at a very early stage. In addition, vibrator phase control systems prevent signal distortions produced by non-linearity of the source itself. However, the success of these automatic correction methods depends on parameter justification by the operator and the actual characteristics of the distorting noise. More specific noise-reduction methods (e.g. Combisweep (Trade mark of Geco-Prakla), elimination of harmonics) increase production costs or need uncorrelated data for the correction process. Because the field data are usually correlated and vertically stacked in the field to minimize logistical and processing costs, it is not possible to make subsequent parameter corrections to optimize the noise reduction after correlation and vertical stacking of a production record. The noise-reduction method described here uses the final recorded, correlated and stacked vibroseis field data. This method eliminates signal artifacts caused e.g. by incorrect vibroseis source signals being used in parameter estimation when a frequency–time analysis is combined with a standard convolution process. Depending on the nature of the distortions, a synthetically generated, nearly recursive noise-separation operator compresses the noise artifact in time using a trace-by-trace filter. After elimination of this compressed noise, re-application of the separation operator leads to a noise-corrected replacement of the input data. The method is applied to a synthetic data set and to a real vibroseis field record from deep seismic sounding, with good results.     

Elimination of Noise Caused by Spikes and Bursts in Vibroseis Data

—Seismic recording systems without a telemetry system have often been affected by electromagnetic induced spikes or bursts, which lead to strong data distortions combined with the correlation process of the vibroseis method. Partial or total loss of the desired seismic information is possible if no automatic spike and burst reduction is available in the field prior to vertical stacking and correlation of the field record.¶Currently, combined with the use of modern telemetry recording systems, the most common noise reduction methods in vibroseis techniques (e.g., spike and burst reduction, diversity stack) are already applied in the field to reduce noise in a very early state. The success of these automatic correction methods depends on the fundamental principles of the recording situation, the actual characteristic of the distorting noise and the parameter justification by the operator. Since field data are usually correlated and already vertical stacked in the field to minimize logistical and processing costs, no subsequent parameter corrections are possible to optimize the noise reduction after correlation and vertical stacking of a production record.¶The noise reduction method described in this paper uses final recorded and stacked vibroseis field data at the correlated or uncorrelated stage of processing. The method eliminates signal artifacts caused by spikes or bursts combined with a standard convolution process. A modified correlation operator compresses the noise artifact in time using a single trace convolution process. After elimination of this compressed noise, re-application of the convolution process leads to a noise-corrected replacement of the input data. The efficiency of the method is shown with a synthetic data set and a real vibroseis field record. Furthermore, several thousand records from a 2-D deep seismic reflection project could be corrected with good results using this method.     

Distance separated simultaneous sweeping, for fast, clean, vibroseis acquisition

Distance separated simultaneous sweeping DS³ is a new vibroseis technique that produces independent records, uncontaminated by simultaneous source interference, for a range of offsets and depths that span all target zones of interest. Use of DS³ on a recent seismic survey in Oman, resulted in a peak acquisition rate of 1024 records per hour. This survey employed 15 vibrators, with a distance separation of 12 km between simultaneous active sources, recorded by 8000 active channels across 22 live lines in an 18.5 km × 11 km receiver patch. Broad distribution of simultaneous sources, across an adequately sized recording patch, effectively partitions the sensors so that each trace records only one of the simultaneous sources. With proper source separation, on a scale similar to twice the maximum usable source receiver offset, wavefield overlap occurs below the zone of interest. This yields records that are indistinguishable from non-simultaneous source data, within temporal and spatial limits. This DS³ technique may be implemented using a wide variety of acquisition geometries, optimally with spatially large recording patches that enable appropriate source separation distances. DS³ improves acquisition efficiency without data quality degradation, eliminating the requirement for special data processing or noise attenuation.     

Blind deconvolution methodology for site-response evaluation exclusively from ground-surface seismic recordings

A novel blind deconvolution methodology for identification of the local site characteristics based on two seismograms recorded on the free surface of a sediment site is presented. The approach does not require recordings at depth nor at a nearby rock outcrop, and eliminates the need for any prior parameterization of source and site characteristics. It considers that the surface recordings are the result of the convolution of the ‘input motion at depth' with transfer functions (channels) representing the characteristics of the transmission path of the waves from the input location to each recording station. The input motion at depth is considered to be the common component in the seismograms (same input in a statistical sense). The channel characteristics are considered to be the part in the seismograms that is non-common, since the travel path of the waves from the input motion location at depth to each recording station is different, due to spatially variable site effects. By means of blind deconvolution, the algorithm eliminates what is common in the seismograms, namely the input motion at depth, and retains what is different, namely the transfer functions of the site from the input location to each recording station. It estimates the site response in both frequency and time domains, and identifies the duration of the site's transfer functions. The methodology is applied herein to synthetic data at realistic sites for performance validation. The blindly estimated results are in almost perfect agreement with the actual site characteristics, indicating that the approach is a promising new tool for seismic site-response identification from recorded data.     

Vertical seismic profiling using a daisy‐chained deployment of fibre‐optic cables in four wells simultaneously – Case study at the Ketzin carbon dioxide storage site

The geological storage of carbon dioxide is considered as one of the measures to reduce greenhouse gas emissions and to mitigate global warming. Operators of storage sites are required to demonstrate safe containment and stable behaviour of the storage complex that is achieved by geophysical and geochemical monitoring, combined with reservoir simulations. For site characterization, as well as for imaging the carbon dioxide plume in the reservoir complex and detecting potential leakage, surface and surface‐borehole time‐lapse seismic monitoring surveys are the most widespread and established tools. At the Ketzin pilot site for carbon dioxide storage, permanently installed fibre‐optic cables, initially deployed for distributed temperature sensing, were used as seismic receiver arrays, demonstrating their ability to provide high‐resolution images of the storage formation. A vertical seismic profiling experiment was acquired using 23 source point locations and the daisy‐chained deployment of a fibre‐optic cable in four wells as a receiver array. The data were used to generate a 3D vertical seismic profiling cube, complementing the large‐scale 3D surface seismic measurements by a high resolution image of the reservoir close to the injection well. Stacking long vibro‐sweeps at each source location resulted in vertical seismic profiling shot gathers characterized by a signal‐to‐noise ratio similar to gathers acquired using geophones. A detailed data analysis shows strong dependency of data quality on borehole conditions with significantly better signal‐to‐noise ratio in regions with good coupling conditions.     

Vibroseis productivity: shake and go

We use both model and field data to compare three methods for increasing vibroseis productivity and decreasing acquisition costs. The first method, HFVS (high-fidelity vibratory seismic), allows us to separate the responses from individual vibrators when multiple vibrators are operating simultaneously. The data quality of the separated records is superior to that of conventional correlated data because they are processed with measured ground-force signals, but the number of sweeps must be greater than or equal to the number of vibrators. The second method, cascaded sweep, eliminates the listening time between multiple sweeps and partially mitigates harmonic noise observed at later times on near-offset traces. Finally, a combined method, continuous-HFVS (C-HFVS), allows source separation with a single, long, segmented sweep. Separation is as good as with HFVS and interference noise is limited to times near the end of a sweep-segment length. All three methods produce acceptable seismic images for post-stack and prestack amplitude interpretation.
The choice of which option to use depends upon the area being investigated. HFVS has numerous benefits, especially when fine sampling is required to mitigate static problems and elevation changes. Due to the ability to separate individual responses, fine sampling can be achieved without sacrificing productivity. For deeper targets, cascaded sweep can be more efficient but data quality suffers from harmonic noise. C-HFVS, which combines features of HFVS and cascaded sweep, has the potential to result in the highest productivity, without sacrificing either fine sampling or data quality.     

Source measurement effect on high fidelity vibratory seismic separation

High Fidelity Vibratory Seismic (HFVS) acquisition and separation can play an important role in today's land acquisition schemes. The method – in which multiple vibrators are swept simultaneously using sweeps with known phase encoding and then the data are inverted and separated into individual records – can improve productivity in the field and at the same time improve signal characteristics in the data. It relies on the measured weighted sum of accelerations (base plate and reaction mass) to invert the acquired data and separate the individual vibrator responses. Separation can be sub-optimal if the measured motions vary from the 'true source' input into the ground. Differences in true source and measured source can arise due to poor coupling between vibrators and ground, soil compaction or other factors. Using both a synthetic model and real data, we show that if the true source changes between sweeps but is not measured, vibrator responses can leak into adjacent vibrator responses upon separation. In a recent survey with HFVS acquisition, we observed a 25–30 dB separation between adjacent vibrators, which could be improved with greater reliability of the source measurement. The vibrator leakage can reduce the data quality considerably. We discuss the results of this survey and show that separation is affected by source measurement error. Further, we conclude that it is necessary either 1) to use source measurements that can capture the variability of the true source between sweeps or 2) to compensate for the source measurement variations in processing or in acquisition.     

伪随机编码震源信号的地震响应

常规的可控震源Chirp扫描信号的地震响应存在着相关信号旁瓣大、分辨率低的缺点.本文利用数值模拟技术研究可控震源伪随机编码信号地震响应的规律和特点.伪随机编码信号的地震响应剖面的旁瓣电平与常规的Chirp扫描信号情形相比有明显地降低,但是伪随机编码信号地震响应剖面的相关噪声仍然没有得到有效地压制.利用伪随机编码信号的周期性特点把一个长地震记录以信号周期为间隔进行分时叠加可以提高地震记录的信噪比和地震数据的利用率.     

编码源地电阻率观测试验

地电阻率的高精度观测是实现地震预测预报的前提之一.面对日益严重的电磁干扰,提出了基于编码源循环互相关辨识技术的地电阻率观测方法,其实质是将待测地质体视为待辨识系统,利用编码源信号激励供电电极A和B产生的电流信号作为系统输入信号,测量电极M与N之间的电压信号作为系统响应输出,将输入和输出信号严格同步采样为时间序列,分别与参考信号进行循环互相关运算并转换至频率域计算获得待探测地电阻率谱(幅度和相位).由于系统环境的干扰和随机噪声与编码源信号不相关,通过循环互相关运算可以达到抑制环境随机噪声和干扰的目的.这种地电阻率观测体系在环境干扰较大的甘肃省兰州观象台和陇南汉王地震台站利用现有的观测场地和线路进行了观测试验,测量结果显示,数据的一致性好、均方差小,说明该方法在强干扰环境下具有较好的抗电磁干扰能力,观测频带较以往直流电法测量有较大的扩展.该方法为现有地面地电阻率台站持续发展提供了技术保障,可为地震预报与科学研究提供高质量的地电阻率观测数据.      

脉冲编码可控震源信号设计

通过脉冲编码技术使震源向地下激发出可人为控制的变步长的脉冲序列,采用解码技术将这种连续多次小能量激发情形的原始地震记录转换为等价的单次大能量激发的地震记录.文中通过数值模拟技术对脉冲编码震源信号地震响应的规律和特点进行了研究.模拟结果表明,脉冲编码信号冲击间隔的增量是压制相关噪声的关键因素,调整冲击间隔的增量对相关噪声的影响要比增加冲击次数对相关噪声的影响灵敏得多,而重复的冲击间隔对压制相关噪声不起任何作用.     

消除滑动扫描地震数据中谐波干扰的反相关方法(英文)

滑动扫描技术是高效、高保真、环保的可控震源勘探技术之一,是下一组震源不必等待上一组震源震动结束即可开始震动的高效采集方法。该技术由于缩短了相邻两炮的等待时间,使得生产效率得到显著提高。但是后一炮的谐波畸变与前一炮的基波信号混叠在一起,不易分离,在相关后的地震记录上形成了严重的谐波干扰,降低了地震资料的质量。本文提出一种反相关方法来压制滑动扫描地震数据中的谐波干扰。该方法首先把地面力信号分解为基波和各阶谐波分量;然后将后一炮的相关前数据分别与各分量相关,只选取正时间轴中对应分量的自相关部分,利用各分量的反相关算子提取各阶谐波信息;最后从前一炮数据中减去提取出的高阶谐波,得到压制谐波后的地震记录。该方法对有效信号影响小,可同时处理相关前和相关后数据,而且算法简单稳定,计算效率高。本文分别对理论模型和实际数据进行处理,验证了该方法消除谐波干扰的有效性。     

Vibroseis deconvolution: A comparison of pre and post correlation vibroseis deconvolution data in real noisy data

Vibroseis is a source used commonly for inland seismic exploration. This non-destructive source is often used in urban areas with strong environmental noise. The main goal of seismic data processing is to increase the signal/noise ratio where a determinant step is deconvolution. Vibroseis seismic data do not meet the basic minimum-phase assumption for the application of spiking and predictive deconvolution, therefore various techniques, such as phase shift, are applied to the data, to be able to successfully perform deconvolution of vibroseis data.This work analyzes the application of deconvolution techniques before and after cross-correlation on a real data set acquired for high resolution prospection of deep aquifers. In particular, we compare pre-correlation spiking and predictive deconvolution with Wiener filtering and with post-correlation time variant spectral whitening deconvolution. The main result is that at small offsets, post cross-correlation spectral whitening deconvolution and pre-correlation spiking deconvolution yield comparable results, while for large offsets the best result is obtained by applying a pre-cross-correlation predictive deconvolution.     

Experience from the Ecors program in regions of complex geology

The French Ecors program was launched in 1983 by a cooperation agreement between universities and petroleum companies. Crustal surveys have tried to find explanations for the formation of geological features, such as rifts, mountains ranges or subsidence in sedimentary basins. Several seismic surveys were carried out, some across areas with complex geological structures. The seismic techniques and equipment used were those developed by petroleum geophysicists, adapted to the depth aimed at (30–50 km) and to various physical constraints encountered in the field. In France, Ecors has recorded 850 km of deep seismic lines onshore across plains and mountains, on various kinds of geological formations. Different variations of the seismic method (reflection, refraction, long-offset seismic) were used, often simultaneously. Multiple coverage profiling constitutes the essential part of this data acquisition. Vibrators and dynamite shots were employed with a spread generally 15 km long, but sometimes 100 km long.Some typical seismic examples show that obtaining crustal reflections essentialy depends on two factors: (1) the type and structure of shallow formations, and (2) the sources used. Thus, when seismic energy is strongly absorbed across the first kilometers in shallow formations, or when these formations are highly structured, standard multiple-coverage profiling is not able to provide results beyond a few seconds. In this case, it is recommended to simultaneously carry out long-offset seismic in low multiple coverage.Other more methodological examples show: how the impact on the crust of a surface fault may be evaluated according to the seismic method implemented (vibroseis 96-fold coverage or single dynamite shot); that vibrators make it possible to implement wide-angle seismic surveying with an offset 80 km long; how to implement the seismic reflection method on complex formations in high mountains.All data were processed using industrial seismic software, which was not always appropriate for records at least 20 s long. Therefore, a specific procedure adapted to deep seismic surveys was developed for several processing steps. The long duration of the vibroseis sweeps often makes it impossible to perform correlation and stack in the recording truck in the field. Such field records were first preprocessed, in order to be later correlated and stacked in the processing center. Because of the long duration of the recordings and the great length of the spread, several types of final sections were replayed, such as: (1) detailed surface sections (0–5 s), (2) entire sections (0–20 s) after data compression, (3) near-trace sections and far-trace sections, which often yield complementary information.Standard methods of reflection migration gave unsatisfactory results. Velocities in depth are inaccurate, the many diffractions do not all come from the vertical plane of the line, and the migration software is poorly adapted to deep crustal reflections. Therefore, migration is often performed graphically from arrivals picked in the time section. Some line-drawings of various onshore lines, especially those across the Alps and the Pyrenees, enable to judge the results obtained by Ecors.     

WHERE IS ZERO TIME?*

Mis-ties are all-too-common results of seismic surveys made at the same place but at different times with different equipment or by different organizations. Even after removal of positioning or polarity errors, reflection times often appear to differ by several tens of milliseconds. Zero time appears to fluctuate. How can zero time differ on surveys with only minor differences in acquisition or processing? What can be done to identify the true zero time for each survey? The first step toward establishing zero time is to record the source pulse. It is well-known that the different sources currently used in reflection seismic prospecting (propane-oxygen explosions, compressed-air discharges, explosives, steam bubbles, mechanical implosions, vibrations, etc.) yield different pressure wavefronts as the input to the seismic reflection system. By recording this wavefront we capture the basic pulse shape and we establish the initial time delay. The second step is to process the recorded source pulse as if it were reflection data to establish the additional time and shape changes introduced by data processing. Then, display the recorded and processed source pulse as an auxiliary variable at the ends of the seismic section. From this display the interpreter can systematically establish the time shifts appropriate to each picked event. He can determine also whether the pick should be a peak or a trough. He can see why surveys which appear to tie for shallow reflections appear to mis-tie for deep reflections. The display of the processed source pulse constitutes a major interpretation aid which, in a readily useable form, increases the information content of the basic seismic section.     

Identification of High Frequency Pulses from Earthquake Asperities Along Chilean Subduction Zone Using Strong Motion

The Chilean subduction zone is one of the most active of the world with M?=?8 or larger interplate thrust earthquakes occurring every 10?years or so on the average. The identification and characterization of pulses propagated from dominant asperities that control the rupture of these earthquakes is an important problem for seismology and especially for seismic hazard assessment since it can reduce the earthquake destructiveness potential. A number of studies of large Chilean earthquakes have revealed that the source time functions of these events are composed of a number of distinct energy arrivals. In this paper, we identify and characterize the high frequency pulses of dominant asperities using near source strong motion records. Two very well recorded interplate earthquakes, the 1985 Central Chile (Ms?=?7.8) and the 2007 Tocopilla (Mw?=?7.7), are considered. In particular, the 2007 Tocopilla earthquake was recorded by a network with absolute time and continuos recording. From the study of these strong motion data it is possible to identify the arrival of large pulses coming from different dominant asperities. The recognition of the key role of dominant asperities in seismic hazard assessment can reduce overestimations due to scattering of attenuation formulas that consider epicentral distance or shortest distance to the fault rather than the asperity distance. The location and number of dominant asperities, their shape, the amplitude and arrival time of pulses can be one of the principal factors influencing Chilean seismic hazard assessment and seismic design. The high frequency pulses identified in this paper have permitted us to extend the range of frequency in which the 1985 Central Chile and 2007 Tocopilla earthquakes were studied. This should allow in the future the introduction of this seismological result in the seismic design of earthquake engineering.     

Seismic Source Characteristics of Soviet Peaceful Nuclear Explosions

—?During the period 1965 to 1988, the former Soviet Union (FSU) conducted over 120 peaceful nuclear explosions (PNE) at locations widely dispersed throughout the territories of the FSU. These explosions sample a much wider range of source conditions than do the historical explosions at the known nuclear test sites and, therefore, seismic data recorded from these PNE tests provide a unique resource for use in deriving improved quantitative bounds on the ranges of seismic signal characteristics which may require consideration in global monitoring of the Comprehensive Test-Ban Treaty (CTBT). In this paper we summarize the results of a detailed statistical analysis of broadband seismic data recorded at the Borovoye Geophysical Observatory from 21 of these PNE tests at regional distances extending from about 7 to 19 degrees, as well as the results of theoretical waveform simulation analyses of near-regional (Δ?相似文献   

Accounting for source location errors in the bayesian analysis of seismicity and seismic hazard

A method is presented for incorporating the uncertainties associated with hypocentral locations in the formulation of probabilistic models of the time and space distributions of the activity of potential seismic sources, as well as of the resulting seismic hazard functions at sites in their vicinity. For this purpose, a bayesian framework of analysis is adopted, where the probabilistic models considered are assumed to have known forms and uncertain parameters, the distribution of the latter being the result of an a priori assessment and its updating through the incorporation of the direct statistical information, including the uncertainty associated with the relations between the actual hypocentral locations and the reported data. This uncertainty is incorporated in the evaluation of the likelihood function of the parameters to be estimated for a given sample of recorded locations. For the purpose of illustration, the method proposed is applied to the modelling of the seismic sources near a site close to the southern coast of Mexico. The results of two alternate algorithms for the incorporation of location uncertainties are compared with those arising from neglecting those uncertainties. One of them makes use of Monte Carlo simulation, while the other is based on a closed-form analytical integration following the introduction of some simplifying assumptions. For the particular case studied, accounting for location uncertainties gives place to significant changes in the probabilistic models of the seismic sources. Deviations of the same order of magnitude can be ascribed to differences in the mathematical and/or numerical tools used in the uncertainty analysis. The resulting variability of the seismic hazard at the site of interest is less pronounced than that affecting the estimates of activity of individual seismic sources.     

COMBISWEEP—A CONTRIBUTION TO SWEEP TECHNIQUES*

Nonlinear sweeps have often successfully been employed in the 1960s. However, this area of sweep technology has been neglected since the introduction of digital recording techniques in the Vibroseis system. Now the advent of computerized recording instruments yields a new economical possibility of forming approximately nonlinear sweeps by combining several linear sweeps with or without time gaps to a “Combisweep”. The total duration of a Combisweep may be as long as the maximum available recording time, for example 32 s. Beside the attenuation of correlation noise, the new method has further merits, such as the weighting of predetermined frequency ranges, in order to effect a certain kind of optimum filtering on the emitter side, or in order to compensate to some degree for frequency dependent absorption. In all these applications the Combisweep is considered as one signal in the correlation process. But by correlating with the individual sweeps or a partial combination of them and by applying automatic switching at predetermined times within the gaps between the individual sweeps additional possibilities arise, such as obtaining in one run with a twenty-four channel recording unit twenty-four traces with small distances between vibrators and geophones for shallow reflections and another twenty-four traces with larger distances for deeper reflections. Various Combisweeps and their applications are presented.