Unlted States Patent [19]
`Bell et a1.
`
`[11] Patent Number:
`[45
`Date of Patent:
`
`4 908 801
`9
`9
`Mar. 13 1990
`
`9
`
`[54] REAL-TIME SIMULATION OF THE
`FAR-FIELD SIGNATURE OF A SEISMIC
`SOUND SOURCE ARRAY
`
`[75] Inventors: Robert R. Bell; Stephen M. Whitley,
`both of Houston’ Tex_
`
`[73] Assigneez Teledyne Exploration, Houston, Tex’
`
`[21] Appl' No" 358’435
`
`[22 F1 d
`1
`1e '
`
`M 30 1989
`ay ’
`
`[51] Int. cu ............................................. .. 001v 1/38
`[52] US. Cl. .................................... .. 367/23; 181/ 111;
`181/118
`[58] Field of Search ..................... .. 367/23, 20, 21, 22;
`181/111, 118, 113, 120
`
`[56]
`
`References Cited
`
`,
`U.S. PATENT DOCUMENTS
`4,326,271 4/ 1982 Ziolkowski ......................... .. 367/ 16
`4,476,550 10/1984 Ziolkowski et a1.
`367/21
`4,500,978 2/ 1985 Ziolkowski .......... ..
`367/ 142
`4,644,507 2/ 1987 Ziolkowski
`.. 367/23
`4,648,080 3/ 1987 Hargreaves .......... ..
`.. 367/20
`4,658,384 4/1987 Dragoset et a1. ..
`.. 367/23
`4,827,456 5/1989 Brac .................................... .. 367/23
`Primary Examiner—Thomas H. Tarcza
`Assistant Examiner-Ian J. Lobo
`Attorney, Agent, or Firm—Wi1liam A. Knox
`
`ABSTRACT
`[57]
`A method for extrapolating the far ?eld signature of a
`seismic sound source array from near ?eld measure
`ments, by taking into consideration, the instantaneous
`?ring and environmental parameters that exist locally
`for each individual sound source in the array.
`
`8 Claims, 1 Drawing Sheet
`
`/8
`
`I6,
`_
`
`|
`
`|
`
`22
`tt
`28/1
`26/1 30b
`
`I
`
`28b
`26b 30a
`
`28a
`26a 30
`
`l
`
`1
`
`tr
`25
`26
`
`I
`
`30"
`
`32”
`
`32b
`
`32a
`
`32
`
`'
`
`I. ___________ __
`
`34
`1, El
`36
`
`I
`l
`1
`l
`
`g
`g,
`g
`
`42‘
`
`3;
`03
`
`TDl
`
`PGS Exhibit 1016
`PGS v. WG
`
`

`
`US. Patent
`
`Mar. 13, 1990
`
`4,908,801
`
`_
`
`_ _ _ _ _ _ _ _
`
` _| I | | | I | l l I I 1 l I I | I l l I l I I | l 11.1 |_ . .
`
`
`
`
`
`
`
`
`
`. NM. 5 1 ,
`
`
`
`“ mm“ < a w 9? mm mmm Qmm :mw
`
`_ A a I _ 1|
`
`n F] l .. n mm 0
`_ mm H
`
`_ Om ,0? WV) 1. N _
`
`
`
`_ - .
`
`(0 L}
`
`w
`
`\ NEArIQ
`1 1 mick‘;
`
`
`I _ _ 7% |J>|| ‘H
`" cm .0 Col‘ 0
`
`_ 4... m \. (NV LHAV
`
`vw .. Q
`
`_ vm _ Q
`
`m ow mom 8w tom ~ h h J. LIL TL
`
`
`
`_ ~ ~ \vd + ‘ . [H [V W _
`
`“ 81 w lilcll 21 @w on www mom 8w , 8m 6m 5m
`
`%, @m g, N GE
`
`
`
`8 “m 3v\ 51 “8| #8! l
`
`
`
`_ mv . mm mwm Qwm =0“?
`
`
`
`_ mm mmvm Qm? awn
`
`PGS Exhibit 1016
`PGS v. WG
`
`

`
`1
`
`REAL-TIME SIMULATION OF THE FAR-FIELD
`SIGNATURE OF A SEISMIC SOUND SOURCE
`ARRAY
`
`4,908,801
`2
`the ghost and the direct wave?eld equals or exceeds
`0.95. For all practical purposes, the far ?eld is consid
`ered to exist at distances in excess of 250-300 meters
`from the source. As explained in the ‘384 patent, for
`various reasons, it is usually impractical to attempt to
`measure, experimentally, the far-?eld signature of a
`sound source array. The far-?eld signature is more con
`veniently extrapolated from near-?eld experiments.
`One such extrapolation method is taught by the "384
`patent. Another reference of interest is U.S. Pat. No.
`4,648,080, issued 03/03/87 to N. D. Hargeaves. Other
`references of interest are U.S. Pat. Nos. 4,644,507 issued
`02/17/87 to A. M. Ziolkowski, 4,476,550 issued 10/84
`to Ziolkowski et al., 4,500,978 issued 02/ 19/85 to Ziolk
`owski et al., 4,326,271 issued 04/20/ 82 to Ziolkwski. ‘
`Instead of using standard source arrays such as would
`be employed in routine operations, the references re
`quired special source and receiver arrangements to
`make the near-?eld measurements. Furthermore, in the
`references, it was assumed that environmental and ?ring
`conditions such as sea state, source power output and
`?re-time delays remained ideal and/or constant. The
`alleged far-?eld signature derived by the prior art was
`applied willy-nilly in all data-processing algorithms as
`though it were Gospel but without regard to the chang
`ing operating conditions that are inevitably present
`during a routine geophysical survey.
`As an example, consider the effect of wave height on
`an array measuring 30-50 meters on a side. In heavy
`seas with large swells, one air gun of the array might
`easily be ?ve or six meters deeper than a gun at the
`other end of the array (see FIG. 2). The ghost would
`arrive at the deeper gun 8 ms later than the ghost arriv
`ing at the other gun. A composite of the individual
`source signatures of the array taken as a whole, would
`not resemble a source signature as derived from the
`previous-art methods that assumed ideal, constant con
`
`10
`
`25
`
`35
`
`BACKGROUND OF THE INVENTION
`1. Field of the Invention
`This invention is concerned with seismic sound
`source arrays such as arrays of air guns. In particular, it
`is concerned with the simulation, in real time, of the far
`?eld signature of the array, based upon the instanta
`neous ?eld conditions that exist at the time the array is
`activated. Such art may be found in classes 367/20-23.
`2. Discussion of the Prior Art
`As is well known in the art of marine seismic survey
`ing, a sound source is towed behind a ship beneath the
`surface of a body of water. The sound may be generated
`by a small explosive charge, an electric spark or arc, a
`vibrator or, preferably, an array of several air guns. The
`air guns each contain a volume of air compressed to
`about 2000 psi or more. Upon command, the guns ab
`rupty release their volumes of compressed air to create
`a thunderous sound wave in the water. The resulting
`pressure wave?eld propagates downwardly, into the’
`earth beneath the sea ?oor, to the sub-bottom strata,
`whence the wave?eld is re?ected back up towards the
`water surface. The re?ected wave?eld is detected by a
`hydrophone array that it towed behind the ship just
`beneath the water surface. The hydrophone array may
`extend three thousand meters or more behind the ship
`and may include several thousand hydrophones. The
`detected re?ected wave?elds are recorded on time
`scale recordings or seismograms.
`When the seismic source it triggered or ?red, it pro
`duces a complex output pressure pulse. The hyrophones
`feel that pressure pulse and convert the pressure varia
`tions to an electrical wave train. Typically, the electri
`cal wave train consists of a short, initial, fast positive
`rise in amplitude, followed by several rapidly-decaying
`oscillations. The wavetrain might be 150 to 200 millisec
`onds (ms) long and it termed the “signature” of the
`sound source.
`The wave?eld generated by the sound source radi-I
`ates by spherical spreading in all directions. There is a
`downwardly-travelling component as well as an up
`going component. The air-water interface is an excel
`lent re?ecting surface with a re?ection coef?cient that
`may approach —l. The up-going component of the
`wave?eld is re?ected from the water surface, is re
`versed in polarity and becomes another downwardly
`travelling wave?eld, popularly known as the ghost.
`The seismic sound source array is usually towed'
`about 6 to 10 meters beneath the sea surface. Assuming
`a water velocity of 1500 meters per second (mps), the
`vertical two-way time lag between the direct primary
`wave?eld and the ghost will be about 8 to 14 ms. The
`ghost interferes, in opposite polarity, with the direct
`wave?eld to create a complex source signature. That
`circumstance is termed the ghost effect. The ghost is an
`integral part of the source signature.
`The signi?cance of the ghost effect is explained in
`considerable detail in U.S. Pat. No. 4,658,384, issued
`04/14/87 to W. H. Dragoset et al., which is incorpo
`rated herein by reference. In that patent, it is shown that
`in the near ?eld, that is, within a few tens of meters, the
`ghost distorts the source signature differently than in
`the far ?eld. The far ?eld is de?ned as that distance
`from the source at which the amplitude ratio between
`
`'
`ditions.
`During the course of a seismic project, many differ
`ent lines of survey are occupied. Each line may offer
`different operating problems that require differentarray
`con?gurations. Obviously, the far-?eld signature de
`rived for one array con?guration will not be the same as
`one that is derived for some other con?guration. The
`far-?eld signatures must be individually determined for
`each line of survey to ?t changing environments.
`It is a purpose of the invention to provide a method
`for deriving the far ?eld signature of an array of sound
`sources in real (or near real) time, without use of spe
`cialized equipment. The method takes into consider
`ation the instantaneous ?ring and environmental condi
`tions that exist for each of the individual units of the
`array at the time the array as a whole is triggered.
`
`SUMMARY OF THE INVENTION
`In accordance with this invention, prior to the begin
`ning of an assigned line of survey, we provide an array
`of seismic sound sources. Each of the respective sound
`sources is individually ?red and the resulting near ?eld
`signatures are measured. The measured signatures are
`scaled to compensate for differences in source power
`outputs and are corrected for erratic ?re-time delays.
`The signatures are then deghosted and stored in a ma
`trix, as a function of source location in the array, to
`provide a library of elemental source signatures for that
`line.
`
`45
`
`50
`
`55
`
`65
`
`PGS Exhibit 1016
`PGS v. WG
`
`

`
`BRIEF DESCRIPTION OF THE DRAWINGS
`The advantages and bene?ts of our invention will be
`better understood by reference to the attached detailed
`description and the drawings wherein:
`.
`FIG. 1 shows a ship towing geophysical devices
`along a line of survey;
`FIG. 2 is a schematic illustration of a sound source
`array and associated data processing devices.
`
`20
`
`25
`
`4,908,801
`4
`3
`Depth information for the respective guns, as mea
`Upon moving onto the line of survey, we sequentially
`sured by the depth sensors such as 30, is transmitted
`occupy a plurality of seismic stations along the line. The
`through takeouts 28 and cable 22 to depth-parameter
`source array is ?red at each station to produce a seismo
`register 40 over line 38 to inputs D1, D2, D3, D". Only
`gram. As each station is occupied, at the instant that the
`one line is shown to avoid confusing the drawing, but as
`array is ?red, we measure the instantaneous local ?ring
`many lines as there are guns may be provided. Alterna
`and environmental parameters for each source of the
`tively, a single, time-sharing, digital telemetric link
`array. We then extract the elemental signatures, corre
`could be used which might be a ?ber-optic link or a
`sponding to each of the source locations of those
`wireline. Depth-parameter register 40, which may be
`sources that actually ?red, from the library. Each signa
`analog or digital, receives and stores the depth of each
`ture is then re-ghosted, re-scaled and corrected for the
`element of the array at the instant the array is triggered;
`known ?re-time delays in accordance with the mea
`that is, register 40 records the instantaneous measured
`sured instantaneous local ?ring and environmental pa
`local gun depth which is the essential parameter needed
`rameters. The so-processed signatures are composited
`for deghosting. Although depth information is always
`to create a simulated far-?eld signature appropriate for
`available at the inputs to register 40, the register is pref
`that seismic station. The simulated far-?eld signature is
`erably enabled, over line 41, to receive fresh depth data
`recorded, for archival retrieval, on the seismogram.
`only when ?re controller F emits a trigger signal to the
`guns. Old depth data is shifted out.
`Because of mechanical de?ciencies, the sources do
`not necessarily all ?re at the same instant; relative varia
`tions in ?re time of one or two milliseconds or less
`between guns are not uncommon. As before mentioned,
`the signature hydrophones such as 32 measure both the
`signature of the corresponding gun and the instant that
`the gun actually ?red. Usually, the relative ?re-time
`delays between guns is consistent. The outputs from the
`respective signature hydrophones 32, 32a, 32b, 32n are
`sent through line 42 to gun controller 44 at inputs TD1,
`TDZ, TD3 TD”. As before, only a single line 42 is shown
`to avoid complicating the drawing. Gun controller 44,
`which may be an LRS-lOO unit, made by Litton Re
`sources Systems of Houston, TX, measures the relative
`?re~time delays by monitoring the ?rst arrival times of
`the gun signatures. It then applies, over line 43, a cor
`rective time offset to the trigger pulses that are emitted
`by ?re-control module F to the respective guns, so that
`all of the guns will effectively ?re simultaneously.
`A data processor 46 is provided which is preferably a
`digital computer or the like. Data processor 46 receives
`as its inputs, the air pressure existing in air supply mani
`fold P, the ?re time instant from ?re control module F,
`The respective instantaneous depth readings from depth
`parameter register 40, and the relative gun ?re-time
`delay data from gun controller 44. During the pre-line
`calibration operation, switch 48 is closed so that the
`signatures from the individual guns may be recorded
`and stored for further processing. Otherwise, during
`routine operations, switch 48 is open.
`Considering now, the pre-line calibration operation,
`each gun is ?red individually. Switch 48 is closed. The
`measured air pressure is entered into data processor 46
`as is also the appropriate gun depth from depth parame
`ter register 40. The signature as measured by a signature
`hydrophone 32 is received by processor 46. If a recent
`delay-time history is available for the gun under consid
`eration, that too may be entered from gun controller 44.
`In processor 46 over line 49, the amplitude of the
`measured signature is scaled to compensate for varia
`tions in the power output level of the gun with respect
`to a reference. We consider that the applied air pressure
`is a measure of the power output level of an air gun. In
`that matter, We differ from Ziolkowski, who considers
`that the air gun volume constitutes such a measure. We
`choose a reference pressure such as 2000 psi. The scal
`ing coef?cient is the difference, raised to the third
`power, between the actual pressure and the reference
`pressure. The scaled measured signature may now be
`corrected for ?re-time delay as required.
`
`DETAILED DESCRIPTION OF THE
`PREFERRED EMBODIMENT
`With reference to FIG. 1, there is shown a ship 10
`towing a hydrophone streamer cable 12 and a sound
`source array 14 through a body of water 16 near the
`surface 18 thereof. Streamer cable 12 includes a plural
`ity of hydrophones 20, 20a, 20b, 20n of which four are
`shown but as many as 1000 such instruments may be
`provided.
`Sound source array 14 is shown to include four ele
`ments or sources, which may be air guns, but 32 to 64
`such sources are not uncommon. The spacing between
`elements in the array is greater than one wavelength of
`the seismic wave?elds of interest to inhibit wave?eld
`interaction between elements. Air supply hoses and
`electrical control and data lines are contained within a
`cable assembly 22, which is connected to gun-control
`and data-handling interface 24 mounted aboard ship 10.
`Referring now to FIG. 2, source array tow cable
`assembly 22 is coupled to the respective sound sources
`26, 26a, 26b, 26n by takeouts 28, 28a, 28b, 28n. To each
`sound source, there is coupled a depth sensor 30, 30a,
`30b, 30n, such as the TXC-Dl as made by the assignee
`of this invention. Additionally coupled to each sound
`source, there is a signature hydrophone such as 32, 32a,
`32b, 32n, that may also serve as a ?re-time detector.
`Usually, the signature hydrophones are mounted 10 or
`15 centimeters from the gun exhaust ports.
`The air supply to the guns is provided through an air
`line 34, that is part of tow cable assembly 22, from air
`supply manifold P. Air supply P also includes a suitable
`pressure gage (not shown) that will monitor the air
`pressure provided to the guns. It is assumed, for all
`practical purposes, that the air pressure is equalized
`between the manifold and the individual guns. At each
`gun, some of the air is tapped off and fed to the gun
`through the takeouts such as 28. A trigger or ?ring
`‘ signal is transmitted to the guns over line 36, from ?re
`control module F, through tow cable 22 and takeouts
`28. Although only one ?ring line is shown, it should be
`understood that the guns may be ?red individually or in
`unison as desired.
`
`35
`
`45
`
`65
`
`PGS Exhibit 1016
`PGS v. WG
`
`

`
`15
`
`20
`
`25
`
`30
`
`5
`The measured, scaled, time-corrected signature must
`now be deghosted. As mentioned earlier, the signature
`hydrophone is quite close to the air gun exhaust ports
`relative to the depth of the gun beneath the water sur
`face. In the near ?eld, the primary pulse will be consid
`erably stronger than the later-arriving ghost. Since the
`gun depth and the water velocity are known, it is a
`simple matter for the processor to compute the ghost
`lag time. It looks for a negative-going pulse centered
`around the computed lag time and then simply zeros out
`that negative-going pulse. The deghosted signature is
`stored in a matrix 50 as a function of the gun location as
`part of a library of elemental signatures such as E1, E2,
`E3, E", in the form of_ an essentially positive-going pulse
`such as 52. The steps supra are repeated for all of the
`guns in the array. Since instrumental artifacts may dis
`tort the signature, we prefer to deconvolve the elemen
`tal signatures prior to storage in matrix 50.
`In the above procedure, it has been assumed that the
`signature hydrophones are properly calibrated and in
`good condition. In an alternative method, it is possible
`to provide separate, calibrated hydrophones for the
`pre-line calibration operation.
`We next sequentially occupy the seismic stations
`along the assigned line of survey in a manner well
`known. For routine operations, switch 48 is open so that
`the inputs to processor 46 will be pressure, ?re time,
`depth readouts and relative ?re-time delays between
`sources. The readouts will be the instantaneous values
`of the ?ring and environmental parameters existing
`locally at at each gun at the time of the shot.
`At each seismic station, the sound source array taken
`as a whole, is ?red. The hydrophones in streamer cable
`12 receive and record the re?ected wave?elds as a
`seismogram 58 by use of a recording system 60 that is
`part of control and data handling interface 24. For each
`source location in the array. We extract from matrix 50,
`the elemental signature corresponding to that location
`and place it in processor 46. There, the elemental signa
`ture is re-scaled in proportion to the measured power
`output of that source as described above. The re-scaled
`elemental signature is time-shifted in accordance with
`the known ?re-time delay stored in gun controller 44.
`Finally, the elemental signature is reghosted in propor
`tion to the measured instantaneous depth of the corre
`sponding source. That step is done using the method
`described in US. Pat. No. 4,658,384 that was earlier
`cited. In that method, the elemental signature is in
`verted; the inverted pulse is then time shifted in propor
`tion to twice the measured gun depth divided by the
`velocity of sound in water. The process is repeated for
`all of the guns of the array and the resulting processed
`elemental signatures are composited by simple summa
`tion. The processing steps recited supra may be per
`formed in any convenient order.
`To the left side of FIG. 2, the ocean surface 18 is
`shown as a large undulating swell. Since the sound
`sources are at depth, they remain fairly stable. With the
`gun depth at 32n much greater than the depth at gun 32,
`the individual simulated far-?eld signatures 58n through
`58 show signi?cantly different delays between the pri
`mary and the ghost. Unquestionably, the composite far
`?eld signature continuously changes character during
`an operation, an effect that must be continuously moni
`tored.
`The composite signature is convolved for instrument
`response and sent to signature register 54 as far-?eld
`signature 56. Far-?eld signatuare 56 may now be re-.
`
`4,908,801
`6
`corded on seismogram 58 in real time, for the seismic
`station that was just occupied.
`Within any array, one or more guns may malfunction
`to become drop-outs. In that case, only the elemental
`signatures for the gus that actually ?red are used in the
`composite signature.
`Our invention has been described with a certain de
`gree of speci?city for illustrative purposes. Variations
`will occur to those skilled in art but which may be
`included within the scope and spirit of this invention
`which is limited only by the appended claims.
`We claim as our invention:
`1. A method for real-time simulation of the far ?eld
`signature of an array of a plurality of seismic sound
`sources, comprising:
`individually ?ring each of the respective seismic
`sound sources of the array;
`separately measuring the near ?eld signature of each
`said seismic sound source;
`scaling the respective measured signatures with re
`spect to a reference to compensate for variations in
`the power output levels of the sources;
`correcting the measured, scaled signatures for ?re
`time delays;
`, deghosting the measured, scaled, corrected signa
`tures;
`separately storing in a matrx, as a function of the
`respective source locations in the array, each of the
`measured, scaled, corrected and deghosted signa
`tures to form a library of elemental near ?eld seis
`mic source signatures;
`occupying sequentially, a plurality of seismic stations
`in an assigned area of survey;
`at each said station, ?ring the array of seismic sound
`sources, taken as a whole and recording a seismo
`gram;
`for each source location in the array, measuring the
`instantaneous local ?ring and environmental pa
`rameters at the time of ?ring;
`'
`extacting from said library the elemental signatures
`coresponding to each said source location;
`creating a simulated far ?eld signature that is charac
`teristic of the station occupied, by processing each
`of the extracted elemental signatures to compen
`sate for variations in the instantaneous local ?ring
`and environmental parameters as measured at the
`respective corresponding source locations in the
`array; and
`compositing said processed signatures.
`2. The method as de?ned by claim 1, comprising:
`recording said simulated far ?eld signature on said
`seismogram.
`'
`3. The method as de?ned by claim 1, wherein
`the insantaneous local ?ring and environmental pa
`rameters include at least the individual seismic
`source depths, the power output level of each
`source and the respective inherent relative delays
`in the ?ring times of the sources.
`4. The method as de?ned by claim 1, wherein the step
`of processing comprises-the steps of:
`reghosting each said elemental signature in propor
`tion to the measured instantaneous depth of the
`corresponding source;
`re~scaling each elemental signature in proportion to
`the measured power output of the corresponding
`source;
`
`45
`
`55
`
`60
`
`65
`
`PGS Exhibit 1016
`PGS v. WG
`
`

`
`7
`offsetting, in time, each elemental signature in pro
`portion to the known ?re-time delay associated
`with each said source.
`5. The method as de?ned by claim 1, comprising the
`step of:
`deconvolving the elemental signatures with respect
`to instrumental response prior to the step of stor
`ing.
`6. The method as de?ned by claim 1, comprising the
`step of:
`
`4,908,801
`8
`convolving the processed composited elemental sig
`natures with respect to instrumental response.
`7. the method as de?ned by claim 1 wherein said step
`of individually ?ring is performed prior to the step of
`occupying said seismic stations along said assigned line
`of survey.
`8. The method as de?ned by claim 1, wherein the step
`of ‘ extracting is accomplished only for those source
`locations corresponding to seismic sound sources that
`actually ?red.
`
`10
`
`*
`
`*
`
`*
`
`*
`
`=1!
`
`15
`
`25
`
`35
`
`40
`
`45
`
`55
`
`65
`
`PGS Exhibit 1016
`PGS v. WG