CN102279416B - Method for matching far-and-near offset-distance stacked profile phase frequencies - Google Patents

Abstract

The invention discloses a method for matching far-and-near offset-distance stacked profile phase frequencies, which comprises the steps of: extracting zero-phase wavelets with a same frequency through a mixed-phase wavelet extraction technology, and respectively performing wavelet deconvolution operations by using the extracted zero-phase wavelets with the same frequency and far-and-near offset-distance stacked profiles, so that data phase frequencies of the far-and-near offset-distance stacked profiles are matched; and representing far-and-near offset-distance stacked data x(t) as x(t) = w(t)*(t)+n(t), selecting a zero-phase wavelet with a same frequency under the condition that noises n(t) are omitted, respectively performing wavelet deconvolution by using the wavelet and seismic channels of the far-and-near offset-distance stacked data, to eliminate the differences between frequencies and phases of the far-and-near offset-distance stacked data, so that the wavelet w(t) of a far-and-near offset-distance profile is a zero-phase wavelet w0(t) with the same frequency, and then performing a convolution operation by using the w0(t) and the r(t), so that a matched seismic record is obtained.

Description

The matching process of far-and-near offset-distance stacked profile phase frequencies

Technical field

The present invention relates to a kind of petroleum detection technical field, be specifically related to a kind of matching process of far-and-near offset-distance stacked profile phase frequencies.

background technology

In seismic data is processed, stack is a kind of more effective denoising method.But, because NMO velocity when carrying out normal moveout correction is generally the interpolation speed of discrete point, be difficult to extract accurately.Make seismic section after normal moveout correction and static correction certainly exist undercorrection or proofread and correct excessive, there is certain residue time difference, can't accomplish with superimposed, cause the decrease resolution of Seismic Stacked Section and, in the high frequency resolution data is processed, the high frequency loss of stacked section is serious.Because nmo stretching causes offset distance section lineups far away chap, far and near offset distance section frequency is not mated simultaneously.The inconsistent problem of phase place caused for solving residual normal moveout at present, method commonly used is to carry out residual NMO correction, to undercorrection or proofread and correct excessive being proofreaied and correct again.The method of the residual NMO correction occurred at present be mostly using residual normal moveout as part random when static residual quantity process and need high precision velocity to extract simultaneously.But the accuracy limitations of the speed data extracted due to prior art, be difficult to meet the demand of residual correction.For nmo stretching, mostly adopt the way of excision or matching.This is because in common seismic data is processed, and nearly offset distance often mostly has underground reflection preferably, and, because offset distance is less, the nmo stretching be subject to is also less, therefore excision is little on the impact of reflection stack result.But processing can often cause shallow layer resolution to reduce like this, and energy loss is also larger.Therefore, need to develop a kind of VELOCITY EXTRACTION that do not need and protect the method that amplitude is high simultaneously, solve far and near offset distance superposition of data phase place and the inconsistent problem of frequency.After obtaining the consistent far and near offset distance superposition of data of frequency plot, when finding reservoir, predicting oil/gas, the far and near offset distance superposition of data of application can realize the across comparison that far and near superposition of data predicts the outcome, combination between well shake simultaneously is tightr, makes the prediction of reservoir and oil gas more accurate.

summary of the invention

The matching process that the purpose of this invention is to provide a kind of far-and-near offset-distance stacked profile phase frequencies, apply this matching process and far and near offset distance superposition of data is extracted to the wavelet of same frequency and phase place, the wavelet deconvolution utilize extracted, reach the purpose of eliminating the unmatched problem of far and near offset distance partial stack data frequency phase place caused due to residual normal moveout and nmo stretching.

In order to reach above-mentioned purpose of design, the technical solution used in the present invention is as follows:

A kind of matching process of far-and-near offset-distance stacked profile phase frequencies, by the mixed phase wavelet extractive technique, extract the zero-phase wavelet of same frequency, do the wavelet deconvolution computing with far and near offset distance superposition of data respectively with the zero-phase wavelet of the same frequency extracted, the far and near offset distance superposition of data phase frequency made is complementary;

Far and near offset distance superposition of data x (t) is expressed as x (t)=w (t) * r (t)+n (t),

Ignore in the situation of noise n (t), extract the zero-phase wavelet of a same frequency, do wavelet deconvolution by this wavelet respectively at the seismic trace of far and near offset distance superposition of data, to reach the difference of the frequency plot of eliminating far and near offset distance superposition of data, the wavelet w (t) that makes far and near offset distance section is the zero-phase wavelet w of same frequency ₀, then use w (t) ₀(t) do convolution operation, the seismologic record x (t) after being mated with r (t).

The process of the zero-phase wavelet of described extraction same frequency is: due to the difference of wavelet and reflection coefficient sequence " smoothness ", the intermediary heat spectrum of wavelet is generally near initial point, and the intermediary heat spectrum of reflection sequence is away from initial point, therefore, just can in the intermediary heat spectrum, realize separating of wavelet and reflection coefficient at intermediary heat spectrum design low-pass filter, then the wavelet that the wavelet of intermediary heat spectral domain converts time domain to be reached to the purpose of wavelet w (t) extraction.

In the situation that the known amplitude spectrum is extracted zero-phase wavelet in the intermediary heat spectral domain:

Suppose x (t), w (t), r (t) meets the Di Liheli condition, and noise is white noise; Ignore noise n (t), formula x (t)=w (t) * r (t)+n (t) substitution Fourier transform formula

$X_{\mathrm{\ω}}={\∫}_{-\∞}^{+\∞}{x}_t{e}^{-\mathrm{j\ωt}}{d}_t$

Obtain formula under frequency field:

X(ω)=W(ω)R(ω)

X in formula (t), W (t), R (t) is respectively x (t), w (t), r (t) is in the expression formula of frequency field;

Formula X (ω)=W (ω) R (ω) is taken the logarithm, is converted into linear system:

lnX(ω)=lnW(ω)+lnR(ω)

Formula lnX (ω)=lnW (ω)+lnR (ω) is made to Fourier inversion, be about to it

Substitution inversefouriertransform formula:

$x_t={\∫}_{-\∞}^{+\∞}{X}_{\mathrm{\ω}}{e}^{2\mathrm{\πj\ωt}}{d}_{\mathrm{\ω}}$

Obtain formula: $\hat{x}\left(t\right)=\hat{w}\left(t\right)+\hat{r}\left(t\right)$

be called x (t), w (t), the intermediary heat spectral sequence of r (t);

In known seismic wavelet spectral amplitude situation, suppose that maximum, the minimum phase component of wavelet w (t) is respectively v (t), u (t), substitution convolution formula:

$u\left(t\right)*v\left(t\right)={\∫}_{-\∞}^{+\∞}u\left(\mathrm{\τ}\right)v(t-\mathrm{\τ})d_{\mathrm{\τ}}$

Obtain w (t)=u (t) * v (t)

Formula w (t)=u (t) * v (t) is expressed as at Fourier-domain:

Wherein, w (ω), u (ω), v (ω) is respectively w (t), u (t), the expression formula of v (t) Fourier-domain;

be respectively w (t), u (t), the phase spectrum of v (t);

Formula

logarithmic spectrum be expressed as:

By formula

being converted to linear formula is:

to formula

the substitution Fourier inversion is about to its substitution inversefouriertransform formula:

$x_t={\∫}_{-\∞}^{+\∞}{X}_{\mathrm{\ω}}{e}^{2\mathrm{\πj\ωt}}{d}_{\mathrm{\ω}}$

Obtain formula: $\hat{w}\left(t\right)=\hat{u}\left(t\right)+\hat{v}\left(t\right)$

Wherein

be the intermediary heat spectrum of wavelet amplitude, symmetry appears on the positive and negative axle of intermediary heat spectrum,

the intermediary heat spectrum of the minimum component spectral amplitude of wavelet,

the intermediary heat spectrum of wavelet largest component spectral amplitude,

appear at respectively on the positive axis and negative axle of intermediary heat spectrum;

According to formula $\hat{w}\left(t\right)=\hat{u}\left(t\right)+\hat{v}\left(t\right)$ Obtain: ${\hat{w}}_0\left(t\right)=\hat{u}\left(t\right)+\hat{v}\left(t\right)$

Wherein, it is the intermediary heat spectrum of zero phase spectral amplitude.

The beneficial effect of the matching process of far-and-near offset-distance stacked profile phase frequencies of the present invention is: the unmatched problem of phase place, frequency that will process the superposition of data of different offset distances is thought because seismic wavelet causes, process original earthquake data by the zero phase seismic wavelet of introducing same frequency, in the inconsistent while of phase frequency of eliminating different offset distance superposition of data, can guarantee the resolution of seismic section.Can realize the contrast of earthquake and well, between the well shake, in conjunction with tightr, make the prediction of reservoir and oil gas more accurate simultaneously.There is good practical value.

embodiment

The available convolution model of the relation of described seismologic record and seismic wavelet and reflection coefficient sequence is expressed as:

x(t)=w(t)*r(t)+n(t) （1）

[0040]x in formula (t), w (t), r (t), n (t) is respectively seismologic record, seismic wavelet, reflection coefficient and noise, in the situation that reflection coefficient r (t) and noise n (t) are definite, the far and near offset distance superposition of data phase place caused due to residual normal moveout and nmo stretching and frequency inconsistent is due to the phase place of seismic wavelet w (t) and frequency is inconsistent causes; Simultaneously, due to mixed phase wavelet, by this wavelet, carry out deal with data x (t), can cause the decrease resolution of x (t), can not meet the demand of real work; Adopt zero phase phase place wavelet, carry out deal with data x (t) by this wavelet, can guarantee that the resolution of x (t) meets the demand of real work, therefore extract the zero-phase wavelet of a same frequency, do wavelet deconvolution by this wavelet respectively at the seismic trace of far and near offset distance superposition of data, to reach the difference of the frequency plot of eliminating far and near offset distance superposition of data.

In the situation that the known amplitude spectrum is extracted zero-phase wavelet in the intermediary heat spectral domain:

Suppose (1) x (t), w (t), r (t) meets the Di Liheli condition, and (2) noise is white noise; Ignore noise n (t), formula (1) substitution Fourier transform formula

$X_{\mathrm{\ω}}={\∫}_{-\∞}^{+\∞}{x}_t{e}^{-\mathrm{j\ωt}}{d}_t$

Obtain formula under frequency field:

X(ω)＝W(ω)R(ω) (2)

X in formula (t), W (t), R (t) is respectively x (t), w (t), r (t) is in the expression formula of frequency field;

Formula (2) is taken the logarithm, and formula (2) is converted into linear system:

ln X(ω)＝ln W(ω)+ln R(ω) (3)

Formula (3) is made to Fourier inversion, is about to formula (3) substitution inversefouriertransform formula:

$x_t={\∫}_{-\∞}^{+\∞}{X}_{\mathrm{\ω}}{e}^{2\mathrm{\πj\ωt}}{d}_{\mathrm{\ω}}$

Obtain formula: $\hat{x}\left(t\right)=\hat{w}\left(t\right)+\hat{r}\left(t\right)---\left(4\right)$

be called x (t), w (t), the intermediary heat spectral sequence of r (t);

In known seismic wavelet spectral amplitude situation, suppose that maximum, the minimum phase component of wavelet w (t) is respectively v (t), u (t), substitution convolution formula:

$u\left(t\right)*v\left(t\right)={\∫}_{-\∞}^{+\∞}u\left(\mathrm{\τ}\right)v(t-\mathrm{\τ})d_{\mathrm{\τ}}$

Obtain w (t)=u (t) * v (t) (5)

Formula (5) is expressed as at Fourier-domain:

Wherein, w (ω), u (ω), v (ω) is respectively w (t), u (t), the expression formula of v (t) Fourier-domain;

be respectively w (t), u (t), the phase spectrum of v (t);

The logarithmic spectrum of formula (6) is expressed as:

Formula (7) is converted to linear formula is:

To formula (8) substitution Fourier inversion, be about to formula (8) substitution inversefouriertransform formula:

$x_t={\∫}_{-\∞}^{+\∞}{X}_{\mathrm{\ω}}{e}^{2\mathrm{\πj\ωt}}{d}_{\mathrm{\ω}}$

Obtain formula: $\hat{w}\left(t\right)=\hat{u}\left(t\right)+\hat{v}\left(t\right)---\left(9\right)$

Wherein

be the intermediary heat spectrum of wavelet amplitude, symmetry appears on the positive and negative axle of intermediary heat spectrum,

the intermediary heat spectrum of the minimum component spectral amplitude of wavelet, the intermediary heat spectrum of wavelet largest component spectral amplitude,

appear at respectively on the positive axis and negative axle of intermediary heat spectrum;

According to formula (10), obtain:

Wherein,

it is the intermediary heat spectrum of zero phase spectral amplitude.

This embodiment is the preferred embodiments of the present invention, can not be limited the present invention, and concrete every rights protection scope is defined by the claims.

Claims (3)

1. the matching process of a far-and-near offset-distance stacked profile phase frequencies, it is characterized in that: by the mixed phase wavelet extractive technique, extract the zero-phase wavelet of same frequency, do the wavelet deconvolution computing with far and near offset distance superposition of data respectively with the zero-phase wavelet of the same frequency extracted, the far and near offset distance superposition of data phase frequency made is complementary;

Far and near offset distance superposition of data x (t) is expressed as x (t)=w (t) * r (t)+n (t),

Ignore in the situation of noise n (t), extract the zero-phase wavelet of a same frequency, do wavelet deconvolution with the seismic trace of far and near offset distance superposition of data respectively by this wavelet, to reach the difference of the frequency plot of eliminating far and near offset distance superposition of data, the wavelet w (t) that makes far and near offset distance section is the zero-phase wavelet w of same frequency ₀, then use w (t) ₀(t) do convolution operation, the seismologic record after being mated with r (t).

2. according to the matching process of the far-and-near offset-distance stacked profile phase frequencies of claim 1, it is characterized in that: the process of the zero-phase wavelet of described extraction same frequency is: due to the difference of wavelet and reflection coefficient sequence " smoothness ", the intermediary heat spectrum of wavelet is generally near initial point, and the intermediary heat spectrum of reflection sequence is away from initial point, therefore, just can in the intermediary heat spectrum, realize separating of wavelet and reflection coefficient at intermediary heat spectrum design low-pass filter, then the wavelet that the wavelet of intermediary heat spectral domain converts time domain to be reached to the purpose of wavelet w (t) extraction.

3. according to the matching process of the far-and-near offset-distance stacked profile phase frequencies of claim 2, it is characterized in that: in the situation that the known amplitude spectrum is extracted zero-phase wavelet in the intermediary heat spectral domain:

Suppose x (t), w (t), r (t) meets the Di Liheli condition, and noise is white noise; Ignore noise n (t), formula x (t)=w (t) * r (t)+n (t) substitution Fourier transform formula

Obtain formula under frequency field:

X(ω)＝W(ω)R(ω)

X in formula (t), W (t), R (t) is respectively x (t), w (t), r (t) is in the expression formula of frequency field;

Formula X (ω)=W (ω) R (ω) is taken the logarithm, is converted into linear system:

ln X(ω)＝ln W(ω)+ln R(ω)

Formula ln X (ω)=ln W (ω)+ln R (ω) is made to Fourier inversion, is about to formula ln X (ω)=ln W (ω)+ln R (ω) substitution inversefouriertransform formula:

Obtain formula:

be called x (t), w (t), the intermediary heat spectral sequence of r (t);

In known seismic wavelet spectral amplitude situation, suppose that maximum, the minimum phase component of wavelet w (t) is respectively v (t), u (t), substitution convolution formula:

Obtain w (t)=u (t) * v (t)

Formula w (t)=u (t) * v (t) is expressed as at Fourier-domain:

Wherein, w (ω), u (ω), v (ω) is respectively w (t), u (t), the expression formula of v (t) Fourier-domain;

be respectively w (t), u (t), the phase spectrum of v (t);

Formula

logarithmic spectrum be expressed as:

By formula being converted to linear formula is:

To formula

The substitution Fourier inversion is about to its substitution inversefouriertransform formula:

Obtain formula:

Wherein

be the intermediary heat spectrum of wavelet amplitude, symmetry appears on the positive and negative axle of intermediary heat spectrum,

the intermediary heat spectrum of the minimum component spectral amplitude of wavelet,

the intermediary heat spectrum of wavelet largest component spectral amplitude,

appear at respectively on the positive axis and negative axle of intermediary heat spectrum;

According to formula

obtain:

Wherein, it is the intermediary heat spectrum of zero phase spectral amplitude.