Search

Summary “This paper describes a test conducted to identify the characteristics of the major sources of noise for urban seismic surveys. The strongest source of noise was mechanical equipment, although it was intermittent and restricted to a relatively small bandwidth centred on 75 Hz. Traffic-related noise was also high, with a bandwidth (10-25 Hz) within that typically of interest for seismic surveys, although it did decrease during the late evening/early morning. Aircraft resulted in short (∼1 minute) wide bandwidth (30-200 Hz) bursts of noise. We found no relationship with wind, but this is probably a result of the strength of the other sources and the relatively low wind speeds observed. Electrical noise was observed with the amplitude and frequency (∼50 Hz) varying during the study period. Based on these results we recommend that: acquisition should be carried out during the late night/early morning; receiver locations should be chosen to avoid sources of mechanical and electrical noise; cables should be as short as possible to avoid electrical noise. Although nodal systems for urban surveys are logistically advantageous some degree of real-time quality control must be in place to ensure that records are not acquired when high-amplitude, short-duration noise (e.g. from aircraft).”

Summary Distributed Acoustic Sensing (DAS) is a novel technology to acquire acoustic data. DAS is noticeably promising for Vertical Seismic Profiling (VSP) applications as it offers highly sampled data, acquired simultaneously at all levels, at relatively low cost. In this study, DAS data is acquired using two different cable deployments: cemented behind the casing, and deployed suspended in the well. Both datasets were compared to conventional 3-component geophone VSP. DAS acquired with cemented cable presented similar quality and signal to noise ratio as geophone data, although it suffers from loss of high frequencies. The suspended cable presented predominantly tube noise, however reflections can still be recognized. We hope the results from this study contribute to a broader use of DAS for VSP acquisition.

Summary Distributed Acoustic Sensing (DAS) is a novel technology that uses fibre-optic cables to detect acoustic signals. DAS provides a cost-effective solution for acquiring vertical seismic profile (VSP) surveys, but it suffers from a comparatively poor signal-to-noise ratio and uncertainties in depth. In this work, we compare the quality of a VSP survey acquired using a permanently installed fibre-optic DAS system with that acquired using a conventional 3-component geophone system. The results show that DAS is capable of producing a time-depth curve with differences of less than 0.8 ms from the geophone results but that the raw interval velocities calculated from the DAS data fluctuate wildly due to their small depth interval. The datasets themselves were consistent although the S/N of the DAS data was inferior and small differences in phase were evident. The DAS amplitudes were also found to decay with cosine-squared of the incidence angle compared to cosine for the geophone data. Corridor stacks generated from both datasets were broadly consistent but residual noise in the DAS data contaminated the stack. The results of this, and previously published surveys, should encourage the more widespread use of DAS for zero offset VSPs.

The desirable characteristics of vibrators such as a wide, controllable bandwidth and high repeatability has led to them becoming the most popular land seismic sources. The hydraulic vibrators typically used for land seismic acquisition, however, are large, costly to operate and expensive to purchase. This inhibits their use for small-scale surveys, such as those of the near surface and vertical seismic profiles, as well as Vibroseis research. In this paper we describe the construction of a simple portable vibrator from commercially available components, principally a number of low-frequency actuators, which can be assembled in days at a cost of less than $5,000. The peak-force of the vibrator is greater than 70 kg (686 N) and could easily be increased by adding additional actuators. Initial field results show that the vibrator is superior to a sledgehammer and is capable of transmitting sweeps with optimised spectral content. We hope that the ease with which such units can be built and operated will inspire others to use portable vibrators for near-surface work and foster further Vibroseis research.

Interest in the use of fibre-optic-based distibuted vibration sensing systems is rapidly expanding with users primarily drawn by its ability to make a large number of high-density strain measurements without the need for individual sensors. Unlike conventional measurements, strain is not measured at individual points but is instead measured across a section of the fibre (the gauge length). The gauge length (either an acquisition or processing parameter) is one of, if not the most, important parameters for DVS surveys having a direct influence on the resulting signal-to-noise-ratio and resolution. We have also derived a simple method for calculating its optimum value.

When a seismic hydraulic vibrator sweeps, it emits the signal we desire as well as a series of harmonics due to the non-linear behaviour of the hydraulic system and the baseplate-earth contact. The phase of the fundamental component of the transmitted signal also varies significantly from that of the pilot, which results in distortion of the Klauder wavelet. If we could record a more accurate estimate of the transmitted signal we could improve the signal itself (through the feedback loop of the vibrator’s controller) and/or reduce its effects during processing. In this paper we show that the pressure distribution beneath the baseplate is more complex than previously thought but that thin-film pressure pads are a promising method for measuring the signature of a vibrator.

Due to their ease of use, relative low cost, and excellent spatial sampling, distributed fibre-optic- based seismic acquisition systems are gaining increased attention, particularly in vertical seismic profile (VSP) surveys. Data recorded using such systems clearly exhibit a radiation pattern effect whose characteristics depend on the effective length of the section of fibre being used for sampling. The use of long gauge lengths results in the attenuation of high frequencies and high wavenumber whereas the use of short gauge lengths results in the attenuation of low frequencies and low wavenumbers. One solution to this problem would be to process data with a variety of different gauge lengths and merge the data at a later time. Unfortunately many of the systems currently available cannot offer this option because the length is fixed either by the hardware or within the initial processing.

Hydraulic seismic vibrators are the preferred source for land vertical seismic profile (VSP) surveys. The introduction of new, more powerful vibrators wogether with the use of Maximum Displacement (MD) sweeps has resulted in the extension of the bandwidth from less than three octaves to nearly seven. One of the advantages of the design procedure for MD sweeps is that the spectrum of the resulting sweep is an input to, rather than a result of, the design process. This enables the design of sweeps where the spectrum contains sections of the bandwidth that have been enhanced. The data presented in this paper shows that the bandwidth of the data received downhole can be successfully adjusted by varying the power spectrum of the transmitted sweep. Even in highly attenuated terrains this enables us to transmit high frequency energy to significant depth.

The moving coil geophone is still the most commonly used sensor for land seismic surveys despite the introduction of other sensors. Modern geophone development has reached the stage where the signal recorded is of very high quality, but it can still be affected by the geophone being placed on an angle relative to the vertical (‘tilt’). This paper describes the acquisition and analysis of field measurements using tilted geophones. It is shown that the critical angle of the 10 Hz geophones used for this test is 55°. For data recorded using vertically placed geophones separated by only 10 cm the perturbation level (the difference between the data recorded by adjacent geophones) averaged 8% and increased to more than 50% at tilt angles of 40o. The level of perturbation is heavily dependent on the orientation of the tilt angle relative to the source-detector axis, for example, for a geophone tilted at an angle of 30°, the perturbation varied between 14% and 48%. The obvious solution to these issues is to record data using sensors that have been planted extremely carefully or to use other sensors, such as digital accelerometers or 3C sensors that can have the effect of tilt removed during processing.

Hydraulic vibrators are the most popular source for land seismic surveys despite historically having a limited bandwidth, usually less than 3.5 octaves. At low frequencies ( 80 Hz), the predominant limitation is the phase difference between the reaction-mass and the baseplate which is a result of the ground conditions. These limitations can be overcome to a large extent by reducing the drive level where appropriate while decreasing the instantaneous sweep rate to ensure that the energy is not reduced. This process has been shown to enable transmission of sweeps with bandwidths of more than seven octaves. It must always be kept in mind, however, that the vibrator is only half of the equation; if not adequately coupled to the surface, or if the surface is not compliant, then the energy will not be transmitted.

Historically, arrays have been used to attenuate ambient noise under the assumption that the level of attenuation is directly proportional to the square root of the number of sensors in the array. Given the availability of high channel-count point-receiver systems and the cost associated with laying out large arrays this assumption of ‘spatial randomness’ requires further analysis. Using measurements of ambient noise made at various sites in Perth, Australia with closely spaced geophones we show that ambient noise is strongly correlated over distances of up to 10 m. This correlation reduces the signal-to-ambient-noise performance of an array considerably. The correlation coefficient can be modelled using an exponential function and the correlation-distance used to determine the efficient geophone spacing. The optimum geophone spacing on days with a low wind speed ( 80 km/h) the optimum spacing is 2.5 m, although this wind speed is very uncommon, occurring on average less than once each year. For more than 90% of days the wind speed is such that the optimum geophone spacing required for ambient noise suppression is 7.5 m.

The targets for VSP surveys are becoming more challenging, being deeper and overlain by more complex sub-surface formations. Coupled with this is a move to more complex acquisition methodologies, 3D VSPs in particular, and a desire for wider bandwidth data, in particular enhanced low-frequency content. As the preferred source for land VSP surveys is the seismic vibrator there is a requirement for vibrators to be able to emit lower frequencies than has previously been possible. Emitting such frequencies requires a sweep design methodology that takes into account the mechanical, hydraulic and control electronics limitations of the vibrator. Through measurements made both at the vibrator and downhole we show that Maximum Displacement Sweeps can be successfully employed to considerably increase bandwidth.