Influence of Drive Level on the Fundamental Vibrator Signal

Pełen tekst

(1)Th-P11-02 Influence of Drive Level on the Fundamental Vibrator Signal R.P. Noorlandt* (Delft University of Technology), G.G. Drijkoningen (Delft University of Technology) & C.A.M. Faber (ION Geophysical). SUMMARY In this abstract we show the influence of vibrator drive level on the signal it produces. For that purpose a field survey was carried out using an INOVA's AHV-IV vehicle with a modified 266kN (60.000 lbf) vibrator. A single linear sweep was repeated at 10 different drive levels ranging from 5 to 90% at two locations. Each drive level was repeated 10 times and each run was repeated twice per location. In total 400 sweeps were carried out. From this data set we conclude that; the vibrator signal is very repeatable for a given setting; the repeatability is less comparing different runs with the same setting; the ground-base plate interaction depends on the drive level in a non-linear way and not taking these effects into account produce arrival time and amplitude errors in seismic records.. 75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013 London, UK, 10-13 June 2013.

(2) Introduction It is know that seismic vibrators do not convert a designed signal, i.e. the pilot, to a seismic signal linearly, see for example Martin and Jack (1990), Walker (1995), Lebedev and Beresnev (2004), Lebedev et al. (2006), Meunier (2011) and Wei and Phillips (2012). The non-linearity in the total system, figure 1, contributes to the distortion of the source wavelet from the pilot, as well as from its estimates measured at the source. If the interactions at the source side would be linear and in the vertical direction only; the.

(3)

(4) .

(5)

(6) . .

(7) .

(8) .

(9)

(10) . . . .

(11)

(12) . Figure 1 Stylized model of common components in a vertical seismic vibrator survey masses would be rigid, the coupling perfect and the hold-down part can be measured or neglected, the force wavelet can be obtained using the weighted sum ground force method (Castanet et al. (1965) and Sallas (1984)). In an isotropic homogeneous earth with a linear stress-strain relation this force wavelet equals the far field displacement wavelet (Miller and Pursey (1954)). In real live these assumptions do not hold, i.e. the system is more complex and non-linear. We are particular interested in the distortions caused by the coupling and stress-strain relationship near the vibrator and their influence on the repeatability of the vibrator signal. To gain more insight we carried out a number of tests at a test site in Texas, USA. The results of these experiments are described in this abstract.. Experiments For our experiments we used an INOVA’s AHV-IV vehicle with a modified 266kN (60.000 lbf) vibrator. The actuator from this vibrator has been specially designed to produce lower harmonic distortion. Since we are mainly interested in the coupling distortion we performed all experiments in reaction mass phase lock, while the amplitude was still controlled by the weighted sum ground force signal. In this way the amplitude is fixed, but the phase is not adapted for information gained at the base plate. The pilot signal used was a standard linear up-sweep, which went from 8 to 80 Hz in 12 seconds. The vibrator was placed at a new location, without pre-compaction or pre-sweeping, and the drive level was consecutively increased from 5 to 90%. At each drive level 10 sweeps were performed. The whole sequence was done twice at each location to see how compaction influenced the vibrator signal. This experiment was then repeated for two locations, about 20 meters apart. Figure 2 shows the spectra of the signals belonging to the vibrator at the first location at its first run. From these graphs it is clear that the ten sweeps performed at each drive level have very similar spectra. The controller is able to produce a relative flat amplitude spectrum, left graph of figure 2, but interestingly the relative contribution of the reaction mass and base plate differs with drive level. The acceleration spectra contain a swell that shifts from about 50–55 Hz at 5% drive level to about 30–35Hz at 90% drive level and becomes more pronounced with drive level. This swell is unrelated to the produced harmonics, we acquired similar amplitude spectra only taking the fundamental signal into account. 75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013 London, UK, 10-13 June 2013.

(13) Weighted sum ground force. Reaction mass acceleration. 005% 010% 020% 030% 040% 050% 060% 070% 080% 090%. −6 −9 −12 −15 −18 −21 −24 −27 −30 −33. BP Loc1 RM−Control Run1. 0. 005% 010% 020% 030% 040% 050% 060% 070% 080% 090%. −3 −6 −9 dB: 10 log10( data / max ). 0 −3. dB: 10 log10( data / max ). Base plate acceleration. RM Loc1 RM−Control Run1. −12 −15 −18 −21 −24 −27 −30 −33. 0 −6 −9 −12 −15 −18 −21 −24 −27 −30 −33. −36. −36. −36. −39. −39. −39. −42. −42. −45. −45 10. 20. 30. 40 50 Frequency (Hz). 60. 70. 80. 005% 010% 020% 030% 040% 050% 060% 070% 080% 090%. −3. dB: 10 log10( data / max ). GF Loc1 RM−Control Run1. −42 −45 10. 20. 30. 40 50 Frequency (Hz). 60. 70. 80. 10. 20. 30. 40 50 Frequency (Hz). 60. 70. 80. Figure 2 dB Spectra of different source measurements. Colour indicates drive level percentage. The experiment was repeated at the same spot to see the impact of first 100 sweeps on the coupling and the repeatability of the experiment. Figure 3 shows the spectra of the signals belonging to the vibrator at the first location for the first and second run. Again it is clear that, except for the 5% drive level, the controller was able to fix the output amplitude. In the second run, especially for the higher frequencies, the influence is shifted from base plate to reaction mass for all drive levels. At the second location, not shown here, the opposite effect was found, i.e. the base plate gained more importance over the reaction mass signal. The difference between the runs becomes smaller for larger drive levels. Weighted sum ground force. Reaction mass acceleration. 005% Run1 005% Run2 010% Run1 010% Run2 020% Run1 020% Run2 030% Run1 030% Run2 040% Run1 040% Run2 050% Run1 050% Run2 060% Run1 060% Run2 070% Run1 070% Run2 080% Run1 080% Run2 090% Run1 090% Run2. −6. dB: 10 log10( data / max ). −9 −12 −15 −18 −21 −24 −27 −30 −33 −36 −39 −42 −45. BP Loc1 RM−Control. 0. 005% Run1 005% Run2 010% Run1 010% Run2 020% Run1 020% Run2 030% Run1 030% Run2 040% Run1 040% Run2 050% Run1 050% Run2 060% Run1 060% Run2 070% Run1 070% Run2 080% Run1 080% Run2 090% Run1 090% Run2. −3 −6 −9 dB: 10 log10( data / max ). 0 −3. −12 −15 −18 −21 −24 −27 −30 −33 −36 −39 −42 −45. 10. 20. 30. 40 50 60 Frequency (Hz). Base plate acceleration. RM Loc1 RM−Control. 70. 80. 0. 005% Run1 005% Run2 010% Run1 010% Run2 020% Run1 020% Run2 030% Run1 030% Run2 040% Run1 040% Run2 050% Run1 050% Run2 060% Run1 060% Run2 070% Run1 070% Run2 080% Run1 080% Run2 090% Run1 090% Run2. −3 −6 −9 dB: 10 log10( data / max ). GF Loc1 RM−Control. −12 −15 −18 −21 −24 −27 −30 −33 −36 −39 −42 −45. 10. 20. 30. 40 50 60 Frequency (Hz). 70. 80. 10. 20. 30. 40 50 60 Frequency (Hz). 70. 80. Figure 3 dB Spectra of different source measurements. First run results are plotted with solid lines, second run results are plotted with dashed lines. Colour indicates drive level percentage.. Ground filtering of the signal The spectra shown in figure 2 and 3 show that this vibrator experiment does not scale in a linear way, but it is hard to discriminate between effects produced by the mechanics and/or hydraulics of the vibrator and those produced by the base plate ground interaction. If we consider the interactions shown in figure 1 as one-dimensional forces, we see that the base plate experiences five different forces from; gravity, the hold down system, the driving engine, the reaction mass support system and the ground force. If we now ignore gravity and the hold down system, since both act below the frequency of interest, we see that the weighted sum ground force is just the filtered version of the force acting on the reaction mass. Figure 4 shows the spectra of the weighted sum ground force divided by the reaction mass acceleration to be able to see the ground force spectra as if the mechanics would have generated a flat force spectrum. From these graphs it is clear that if we correct for the variability of the mechanical force on the base plate, i.e. correct for the controller, the response varies considerable with drive level.. 75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013 London, UK, 10-13 June 2013.

(14) Source Location 1 Run 1. Source Location 1 Run 2. Loc1 RM−Control Run1. Loc1 RM−Control Run2 GF/RM 005% GF/RM 010% GF/RM 020% GF/RM 030% GF/RM 040% GF/RM 050% GF/RM 060% GF/RM 070% GF/RM 080% GF/RM 090%. 0. −1. 10. −1. 10. dB: 10 log ( data / max ). 0. dB: 10 log ( data / max ). GF/RM 005% GF/RM 010% GF/RM 020% GF/RM 030% GF/RM 040% GF/RM 050% GF/RM 060% GF/RM 070% GF/RM 080% GF/RM 090%. −2. −3. −2. −3. 10. 20. 30. 40 50 60 Frequency (Hz). 70. 80. 10. 20. 30. 40 50 60 Frequency (Hz). 70. 80. Amplitude spectra Loc1 RM−Control Run1. Loc1 RM−Control Run2 GF/RM 005% GF/RM 010% GF/RM 020% GF/RM 030% GF/RM 040% GF/RM 050% GF/RM 060% GF/RM 070% GF/RM 080% GF/RM 090%. unwrapped Phase in Degree. 0. −10. −20. −30. −40. −50. 10. GF/RM 005% GF/RM 010% GF/RM 020% GF/RM 030% GF/RM 040% GF/RM 050% GF/RM 060% GF/RM 070% GF/RM 080% GF/RM 090%. 0 unwrapped Phase in Degree. 10. −10. −20. −30. −40. 10. 20. 30. 40 50 60 Frequency (Hz). 70. 80. −50. 10. 20. 30. 40 50 60 Frequency (Hz). 70. 80. Phase spectra Figure 4 Weighted sum ground force corrected for reaction mass acceleration spectra. Colour indicates drive level percentage.. Influence on seismic records The influence of the drive level on seismic data is larger than might be expected from the flat spectra of the weighted sum ground force. We plotted common receiver gathers of the different drive levels for the first location and first run in figure 5 after being correlated with the pilot signal and scaled per trace maximum. Within each constant force panel there is little variation in the signals, but from panel to panel the signals do change. In more detail we see that the positive time ground force signal is more affected than the negative time. The near field geophone shows a shift of about 10 ms and amplitude and bandwidth variations. The far field and borehole geophone also show time, amplitude and bandwidth variations. Not all events seem to be effected equally strong, some events present at low drive level are vanish with larger drive level and vice versa, while others are persistent for all drive levels.. Conclusions We have shown from field data that • For a given setting the vibrator signal is very repeatable. (figure 2 and 5) • The vibrator signal spectra depend on the history of sweeps. (figure 3) • The ground - base plate interaction depends on drive level. (figure 4) • Not taking these effects into account results in arrival time and amplitude errors in seismic records. (figure 5) 75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013 London, UK, 10-13 June 2013.

(15) Weighted sum Groundforce. Near field geophone. Far field geophone. Borehole geophone. Figure 5 Common receiver gathers for four different signals, correlated with the pilot signal. The plots show 10 panels with drive levels at 5,10,20,30,40,50,60,70,80 and 90 % from left to right. Each panel contains 10 consecutive sweeps. The top two plots have a time axis from -125 to 125ms, while the lower two plots contain 250 to 1700ms. Colour indicates signal amplitude.. Acknowledgements We are grateful Inova Geophysical offered us the possibility to perform these experiments and allowing us to present these data. We would also like to thank Shell for funding these experiments. This work has been sponsored by the Research School of Integrated Earth Sciences (ISES). It is also supported by the Nederlandsche Aardolie Maatschappij (NAM). The technical support of the field crew during the experiments was appreciated a lot.. References Castanet, A., Rueil-Malmaison and Lavergne, M. [1965] Vibrator controlling system. U.S. patent 3 208 550. Lebedev, A. and Beresnev, I. [2004] Nonlinear distortion of signals radiated by vibroseis sources. Geophysics, 69, 968–977. Lebedev, A., Beresnev, I. and Vermeer, P. [2006] Model parameters of the nonlinear stiffness of the vibratorground contact determined by inversion of vibrator accelerometer data. Geophysics, 71, H25–H32. Martin, J. and Jack, I. [1990] The behaviour of a seismic vibrator using different phase control methods and drive levels. First Break, 8, 404–414. Meunier, J. [2011] Seismic Acquisition from Yesterday to Tomorrow: 2011 Distinguished Instructor Short Course. Distinguished instructor series, Society of Exploration Geophysicists, ISBN 9781560802815. Miller, G. and Pursey, H. [1954] The field and radiation impedance of mechanical radiators on the free surface of a semi-infinite isotropic solid. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 223, 521–541. Sallas [1984] Seismic vibrator control and the downgoing p-wave. Geophysics, 49, 732–740. Walker, D. [1995] Harmonic resonance structure and chaotic dynamics in the earth-vibrator system. Geophysical Prospecting, 43, 487–507. Wei, Z. and Phillips, T. [2012] Break through the limits of vibroseis data quality. Geophysical Prospecting, 60, 373–393.. 75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013 London, UK, 10-13 June 2013.

(16)