Method of interpreting seismic data

• US 4,780,859 A
• Filed: 03/09/1987
• Issued: 10/25/1988
• Est. Priority Date: 03/09/1987
• Status: Expired due to Term

• Current Assignee: Exxonmobil Oil Corporation (Exxon Mobil Corporation)
• Original Assignee: Mobil Oil Corporation (Exxon Mobil Corporation)
• Inventors: Hadidi, Mohamed T., Linville, Andrew F. Jr., Jurkevics, Andrejs
• Primary Examiner(s): Tarcza, Thomas H.
• Assistant Examiner(s): Swann, Tod R.
• Application Number:
  US07/023,313
• Time in Patent Office: 596 Days
• Field of Search: 367/38, 367/43, 367/42, 367/47, 367/63, 364/421, 73/602
• US Class Current: 367/43
• CPC Class Codes: G01V 1/366 by correlation of seismic signals
• CPC Class Codes: G01V 2210/3246 Coherent noise, e.g. spatially coherent or predictable

Abstract

Method to process seismic data to yield useful information regarding the subsurface area explored. Seismic data which is acquired from the seismic exploration is autocorrelated and the amplitude spectrums estimated, stacked and averaged to reduce the effects of nonwhite reflectivity upon the data. The amplitude spectrum is cepstral windowed a first time to reduce the effects of reflectivity leakage on the amplitude spectrum, amortized at both low and high frequencies so that the amplitude spectrum will approximate the analog wavelet and cepstral windowed a second time to eliminate any remaining rapidly varying components of the amplitude spectrum. The resultant amplitude spectrum is then used to estimate the phase spectrum for the wavelet and a two-sided least squares filter is then designed to shape the estimated seismic wavelet into a zero-phase, low sidelobe output wavelet.

Claims

1. In seismic exploration of a subsurface formation wherein seismic source are activited at spaced locations along a line of exploration and the amplitude of reflections are recorded as a function of time along said lines to produce a series of seismic traces, a method of compensating the amplitude spectrum of a selected one of said seismic traces for the effects of non-white reflectivity of said subsurface formation comprising:

whitening effective reflectivity of said amplitude spectrum of said selected seismic trace;

reducing the effects of reflectivity leakage on said amplitude spectrum of said selected seismic trace, said step of reducing the effects of reflectivity leakage further including the steps of transforming said amplitude spectrum from the frequency domain into the frequency domain, selecting a first cepstral window length, applying first cepstral window to said frequency domain amplitude spectrum, and transforming said amplitude spectrum from said frequency domain to said frequency domain;

compensating said amplitude spectrum of said selected seismic trace to approximate that of an analog seismic wavelet; and

smoothing said compensated amplitude spectrum to eliminate rapidly varying components, said step of smoothing said compensated amplitude spectrum further including the steps of transforming said amplitude from the frequency domain into the frequency domain, selecting a second cepstral window length, applying said second cepstral window to said frequency domain amplitude spectrum, and transforming said amplitude spectrum from said frequency domain to said time domain.

2. The method of claim 1 wherein the step of compensating said amplitude spectrum of said selected seismic trace to approximate that of an analog seismic wavelet further comprises the steps of:

amortizing low frequencies of said amplitude spectrum to remove the effect of low fequency noise such as ground roll from the estimate of the wavelet amplitude spectrum; and

amortizing high frequencies of said amplitude spectrum to remove the effect of high frequency noise from the estimate of the wavelet amplitude spectrum.

3. The method of claim 2 wherein the step of amortizing said low frequencies of said amplitude spectrum to eliminate low frequency ground roll noise further comprises the steps of:

selecting a low frequency value at which to commence low end amortization; and

applying the slope of the low cut filter to the portion of said amplitude spectrum having a frequency less than said low frequency value.

4. The method of claim 2 wherein the step of amortizing the high frequencies of said amplitude spectrum to prevent aliasing of said seismic data further comprises the steps of:

selecting a first and second high frequency value to commence high end amortization;

applying said characteristics of the earth filter to the portion of the amplitude spectrum between said first and said second high frequency values; and

applying the slope of the high cut filter to the portion of said amplitude spectrum between said second high frequency value and an upper limit frequency.

5. The method of claim 4 wherein said upper limit frequency is two times the Nyquist frequency.

6. The method of claim 1 wherein the step of whitening effective reflectivity of said amplitude spectrum of said selected seismic trace further comprises the steps of:

determining an autocorrelation function for each of said series of seismic traces;

determining an amplitude spectrum for each of said series of seismic traces from said autocorrelation functions;

stacking said amplitude spectrums for said series of seismic traces; and

determining an average amplitude spectrum for said selected seismic trace from said stacked amplitude spectrums.

7. The method of claim 1 wherein the stpe of smoothing said compensated amplitude spectrum to eliminate rapidly varying components further comprises the steps of:

doubling all values greater than zero in a first half period of said frequency domain amplitude spectrum; and

removing all values in a second half period of said frequency domain amplitude spectrum.

8. A method of processing geophysical data to make it more useful in interpreting subsurface characterstics of the earth comprising:

acquiring at least one seismic trace containing information related to the subsurface characteristics of the earth;

determining an amplitude spectrum for each of said at least one seismic trace;

whitening effective reflectivity of each of said at least one seismic trace;

compensating said amplitude spectrum of said at least one seismic trace to approximate an analog seismic wavelet;

estimating a phase spectrum for said seismic wavelet from said compensated seismic wavelet;

determining a two sided, least squares filter from said estimated phase spectrum; and

determining an output seismic wavelet from said compensated amplitude spectrum and said estimated phase spectrum by applying said two sided, least squares filter to said compensated amplitude spectrum in the time domain;

wherein said seismic wavelet is shaped into a zero-phase, low side-lobe output wavelet.

9. The method of claim 8 wherein the step of compensating said amplitude spectrum of said at least one seismic trace to approximate an analog seismic wavelet further comprises the step of reducing the effects of reflectivity leakage on said amplitude spectrum for said at least one seismic trace.

10. The method of claim 9 further comprising the step of smoothing said amplitude spectrum for said at least one seismic trace to eliminate rapidly varying components of said amplitude spectrum.

11. The method of claim 10 wherein the step of reducing the effects of reflectivity leakage on said amplitude spectrum for said at least one seismic trace further comprises the steps of:

transforming said at least one amplitude spectrum from the frequency domain into frequency domain;

selecting a first cepstral window length;

applying said first cepstral window to said frequency domain amplitude spectrum; and

transforming said amplitude spectrum from said frequency domain to said time domain.

12. The method of claim 11 wherein the step of smoothing said amplitude spectrum of said at least one seismic trace to eliminate rapidly varying components of said amplitude spectrum further comprises the steps of:

transforming said at least one amplitude from the frequency domain into the frequency domain;

selecting a second cepstral window length;

applying said second cepstral window to said frequency domain amplitude spectrum; and

transforming said amplitude spectrum from said frequency domain to said time domain.

13. The method of claim 12 wherein the step of smoothing said amplitude spectrum to eliminate rapidly varying components of said amplitude spectrum further comprises the steps of:

doubling all values greater than zero in a first half period of said frequency domain amplitude spectrum; and

removing all values in a second half period of said frequency domain amplitude spectrum.

14. The method of claim 8 further comprising the steps of:

determining an autocorrelation function for each of said at least one seismic trace;

determining an amplitude spectrum for each of said at least one seismic trace from said autocorrelation function;

stacking said amplitude spectrums for said at least one seismic trace; and

determining an average amplitude spectrum for said at least one seismic trace from said stacked amplitude spectrums.

15. The method of claim 8 wherein the step of compensating said amplitude spectrum of said at least one seismic trace to approximate an analog seismic wavelet further comprises the steps of:

amortizing low frequencies of said amplitude spectrum to remove the effect of low frequency noise such as ground roll from the estimate of the wavelet amplitude spectrum; and

amortizing high frequencies of said amplitude spectrum to remove the effect of high frequency noise from the estimate of the wavelet amplitude spectrum.

16. The method of claim 15 wherein the step of amortizing low frequencies of said amplitude spectrum of remove the effect of low frequency noise such as ground roll from the estimate of the wavelet amplitude spectrum further comprises the step of:

selecting a low frequency value at which to commence low end amortization; and

applying the slope of the low cut filter to the portion of said amplitude spectrum having a frequency less than said low frequency value.

17. The method of claim 15 wherein the step of amortizing high frequencies of said amplitude spectrum to remove the effect of high frequency noise from the estimate of the wavelet amplitude spectrum further comprises the steps of:

selecting a first and second high frequency value to commence high end amortization;

applying said characteristics of the earth filter to the portion of the amplitude spectrum between said first and said second high frequency values; and

applying the slope of the high cut filter to the portion of said amplitude spectrum between said second high frequency value and an upper limit frequency.

18. The method of claim 17 wherein said upper limit frequency is two times the Nyquist frequency.

19. In seismic exploration wherein seismic sources are activated at spaced locations along a line of exploration and where the amplitude of reflections are recorded as a function of time along said line to produce a series of seismic traces, a method of interpreting said seismic traces to provide information related to subsurface characteristics of the earth comprising the steps of:

determining an amplitude spectrum for each of said series of seismic traces;

whitening effective reflectivity of said series of seismic traces;

reducing the effects of relectivity leakage on said amplitude spectrums of said series of seismic traces, said step of reducing the effects of reflectivity leakage further including the stpes of transforming said amplitude spectrums from the frequency domain into the frequency domain, selecting a first cepstral window length, applying said first cepstral window to said frequency domain amplitude spectrums, and transforming said amplitude spectrums from said frequency domain to said frequency domain;

compensating said amplitude spectrums of said series of seismic traces to approximate an analog seismic wavelet, said step of compensating said amplitude spectrum of said series of seismic traces to approximate an analog seismic wavelet further including the steps of amortizing low frequencies of said amplitude spectrums to remove the effect of low frequency noise such as ground roll from the estimate of the wavelet amplitude spectrum and amortizing high frequencies of said amplitude spectrum to remove the effect of high frequency noise from the estimate of the wavelet amplitude spectrum;

smoothing said compensated amplitude spectrums to eliminate rapidly varying components, said step of smoothing said compensated amplitude spectrums further including the steps of transforming said amplitude spectrums from the frequency domain into the frequency domain, selecting a second cepstral window length, applying said second cepstral window to said frequency domain amplitude spectrums, and transforming said amplitude spectrums from said frequency domain to said time domain;

estimating a phase spectrum for said seismic wavelet from said smoothed compensated seismic wavelet;

determining a two sided, least squares filter from said estimated phase spectrum; and

determining an output seismic wavelet from said smoothed compensatedd amplitude spectrum said estimated phase spectrum by applying said two sided, least squares filter to said smoothed compensated amplitude spectrum in the time domain

wherein said seismic wavelet is shaped into a zero-phase, low side-lobe wavelet.

20. The method of claim 19 further comprising the steps of:

determining an autocorrelation function for said series of seismic traces;

determining an amplitude spectrum for said series of seismic traces from said autocorrelation functions;

stacking said amplitude spectrums for said series of seismic traces; and

determining an average amplitude spectrum for said series of seismic traces from said stacked amplitude spectrums.

21. The method according to claim 19 wherein the step of amortizing low frequencies of said amplitude spectrum to remove the effect of low frequency noise such as ground roll from the estimate of the wavelet amplitude spectrum further comprises the steps of:

selecting a low frequency value at which to commence low end amortization; and

applying the slope of the low cut filter to the portion of said amplitude spectrum having a frequency less than said low frequency value.

22. The method to claim 19 wherein the step of amortizing high frequencies of said amplitude spectrum to remove the effect of high frequency noise from the estimate of the wavelet amplitude spectrum compriese the steps of:

selecting a first and second high frequency value to commence high end amortization;

applying said characteristics of the earth filter to the portion of the amplitude spectrum between said first and said second high frequency values; and

applying the slope of the high cut filter to the portion of said amplitude spectrum between said second high frequency value and an upper limit frequency.

23. The method of claim 19 wherein the step smoothing said amplitude spectrums to eliminate rapidly varying components of said amplitude spectrums further comprises the steps of:

doubling all values greater than zero in a first half period of said frequency domain amplitude spectrums; and

removing all values in a second half period of said frequency domain amplitude spectrums.