CN100501450C - Method for performing waveform stretching compensation for seismic signal - Google Patents

Abstract

The present invention provides a method for executing waveshape drafting compensation to the seismic signal, wherein the waveshape drafting amount is accurately calculated firstly. Then the original record data is executed with re-sampling with the sampling rate after waveshape drifting in the range of a main frequency wavelength with stronger reflected signal. Finally the original sampling rate is used for restoring the data thereby realizing the waveshape drafting compensation. The obtained waveshape compensated by waveshape drafting is used for judging and explaining the underground geological structure thereby increasing the exploration precision. The invention helps to increase the precision of the seismic data explainer for searching the advantageous oil-gas trap and confirming the well position.

Description

A kind of method of seismic signal being carried out waveform stretching compensation

Technical field

The present invention relates to the method that a kind of geological data is handled, be specifically related to a kind of method of seismic signal being carried out waveform stretching compensation.The present invention adopts the calculating of a series of complexity, and disposal routes such as resampling are finished the correction of the waveform stretching that normal moveout correction is produced, thereby has realized not losing resolution in the compacting noise.

Background technology

Oil-field development depends on the success ratio of probing to a great extent, and the precision and the reliability of the raising of probing success ratio and seismic prospecting achievement are closely related.Present exploration engineering mainly utilizes exploration to obtain underground structure information, fully utilize other existing data again and determine drilling position, but in the seismic prospecting of reality, data quality can be subjected to the influence of various factors, background interference such as generations such as train, automobile, factory, oil well and strong wind, these influence factors cause data quality sometimes and have a strong impact on, and especially single-point receives data.

At first pass through explosive source or ground vibroseis excitation wave in the seismic prospecting, receive with single detector or string combination, obtain surface seismic wave field information, pass through disposal route targetedly again, obtaining more tectonic structure, lithological change and hydrocarbon occurrence information, is effective geophysical exploration method of oil-gas field development.Seismic prospecting can be divided into three big links substantially: seismic data acquisition, seismic data are handled, seismic data interpretation.

The main task that seismic data is handled is the seismic data that utilizes open-air single wave detector or string combination to obtain, the communication theory of base area seismic wave, utilize equipment and corresponding seismic data disposal routes such as robot calculator, transformed wave seismic data to field acquisition carries out various processing processing, to obtain to reflect information such as " seismic cross-section " of subterranean formations and seismic amplitude that the reflection formation lithology changes, frequency, velocity of propagation, for the favourable hydrocarbon trap of explanation personnel searching, determine the well location use.

In order to guarantee sufficiently high signal to noise ratio (S/N ratio), industrially must adopt multi-fold technology in seismic data is handled, consequent normal moveout correction and level stack are the conventional processing art means that seismic data is handled.What adopt in the reflection wave seismic prospecting is that any blows out the method that multiple spot receives.Acceptance point is dispersed on the rectangular node point of an exploration targets zone scope on the ground that comprises shot point.Source wavelet is through the acceptance point on underground medium reflection arrival ground.Owing to the horizontal range different differences that produce each seismic trace due in of each acceptance point with shot point.In the time-space domain, wavelet central point spread is the hyperbolic curve (two dimension) or the hyperboloid of revolution (three-dimensional).Realize the purpose of same-phase stack compacting noise, must be transformed to horizontal line (face) to hyperbolic curve (face), be called normal moveout correction.Normal moveout correction is a kind of nonlinear transformation, and it has fairly obvious stretching effect, and the waveform of acceptance point seismic trace promptly far away more apart from shot point elongates seriously more.Just make the resolution of signal reduce significantly the result who superposes through the data that distort like this.This negative effect is industrial normally unacceptable.Geophysics educational circles is paying close attention to this problem for a long time, and the achievement in research of delivering out is also a lot.But fail to tackle the problem at its root always.

A kind of normal moveout correction stacking method of avoiding NMO stretching is disclosed at Chinese patent application CN101131435A, it utilizes CMP stack altogether and the stack of the seismic event of different time of arrival to calculate the record value in each sampling time in the self excitation and self receiving record, and it mainly is with the t on the CMP _i'=t _iThe speed that-m τ point is corresponding is asked each road self excitation and self receiving time t _i' dynamic correction value (the t that locates _iThe writing time of expression seismic trace; t _iThe self excitation and self receiving time of arrival of ' expression seismic event; τ is the step-length with variable search time of arrival wavelet component; M is the maximum step number of searching for forward).There is the defective of the explanation of the speed that is too dependent in this patented claim, and is inaccurate if speed is explained, then can obtain inaccurate dynamic correction value, and the result not only may produce stretching, also may produce moving school deficiency or excessive.

In addition, in the reflection wave seismic prospecting, adopt repeatedly to cover by level stack and improve signal to noise ratio (S/N ratio), the multiple that signal to noise ratio (S/N ratio) improves is

Wherein n is the seismic trace number of participation level stack.Carry out level stack and must do normal moveout correction earlier, just the seismic channel data of each different geophone offsets is corrected to and is equivalent to the time of reception corresponding lineups identical with the zero shot-geophone distance seismic trace.Calibration on this time shaft is with superimposed prerequisite, yet has but produced the significant stretching effect of seismic wave waveform, causes wave form distortion, makes basic frequency of signal move to low frequency end, has reduced the resolution of signal significantly.This is a problem that must solve.

Summary of the invention

In order to address the above problem, we propose a kind of effective waveform stretching compensation technology, just seismic signal are carried out the method for waveform stretching compensation.

In the waveform stretching compensation method of processing seismic data of the present invention, may further comprise the steps:

(i) can utilize seismic exploration equipment, acquiring seismic data:

In the reflection wave seismic prospecting, adopt and a bit blow out the conventional method that multiple spot receives, in pre-survey area a plurality of acceptance points are set, described a plurality of acceptance point is dispersed on the rectangular node point of an exploration targets zone scope on the ground that comprises shot point, uses earthquake detection equipment to obtain the waveform that blast involves seismic event;

(ii), accurately calculate the waveform stretching amount according to the Wave data that in step (i), obtains;

Sampling rate after (iii) in a stronger dominant frequency wavelength coverage of reflected signal original recorded data being utilized waveform stretching resamples, and utilizes the crude sampling rate to come restore data again, and then the waveform that is stretched is realized waveform stretching compensation;

(iv) use waveform that the waveform stretching compensation that (iii) obtains through step crosses and judge and explain and improved surveying accuracy by subsurface geological structure, improved the favourable hydrocarbon trap of seismic data interpretation personnel searching, determined the precision of well location.

Preferably, accurately calculating waveform stretching amount according to the Wave data that obtains in step described in (ii) in step (i) may further comprise the steps:

(21) suppose that earthquake lineups wavelet center is at x ₀Time on=0 the seismic trace is t ₀, be on the seismic trace of x ≠ 0 at geophone offset, according to following normal moveout correction computing formula (1), the dynamic correction value of the central point correspondence of this wavelet is dt ₀:

${\mathrm{<figure-callout\; id="0"\; label="dt"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">dt</figure-callout>}}_0=\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_0^2+\frac{x^2}{v^2\left(t_0\right)}}-t_0---\left(1\right)$

V (t wherein ₀) be time t ₀Pairing NMO velocity;

(22) 1 i of consideration wavelet central point top supposes that it is at x ₀=0 place's time corresponding is t ₀(i)＜t ₀, on the t direction, the distance of this i and wavelet central point is:

δ＝t ₀-t ₀(i) (2)

(23) time corresponding is on x ≠ 0 place's seismic trace:

t ₁＝t ₀+dt ₀-δ＝t ₀(i)+dt ₀ (3)

(24) after normal moveout correction, with time point t ₁Value be corrected to time t ₀₁, that is to say that this point is at x before the normal moveout correction ₀=0 place's time corresponding is:

t ₀₁＝t ₁-dt ₁ (4)

Dt wherein ₁Be dynamic correction value, thereby the distance that obtains this point and wavelet central point is changed into:

δ ₁＝t ₀₁-t ₀

＝t ₀(i)+dt ₀-dt ₁-t ₀

＝-δ+dt ₀-dt ₁

≠δ

δ ₁+δ＝dt ₀-dt ₁ (5)

Dt wherein ₀-dt ₁Be the waveform stretching amount.

Preferably, sampling rate step is utilized waveform stretching to original recorded data described in (iii) in a stronger dominant frequency wavelength coverage of reflected signal after resamples, utilize the crude sampling rate to come restore data again, and then the waveform that is stretched realized that waveform stretching compensation may further comprise the steps:

(31) make dt ₁≈ dt ₀(i) (6)

$t_{01}\&ap;t_0\left(i\right)+{\mathrm{dt}}_0-{\mathrm{dt}}_0\left(i\right)$

$=2t_0\left(i\right)+d{t}_0-\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_0^2\left(i\right)+\frac{x^2}{v^2\left(i\right)}}---\left(7\right)$

Wherein v (i) is a NMO velocity of i correspondence;

Draw time t after the wavelet point normal moveout correction according to formula (7) ₀₁

(32) stretching effect that the normal moveout correction time is produced according to formula (7) compensates:

The amplitude of the geological data sampled point k of hypothetical record is g (k), if sampling rate is a Δ, then can draw the amplitude of g (k) at time t according to sampling thheorem from the data of record:

$g\left(t\right)=\underset{k=k1}{\overset{k2}{\mathrm{\&Sigma;}}}g\left(k\right)\frac{\sin ({\mathrm{\&omega;}}_{N}t-\mathrm{k\&pi;})}{{\mathrm{\&omega;}}_{N}t-\mathrm{k\&pi;}}---\left(8\right)$

Wherein

${\mathrm{\&omega;}}_N=\frac{\mathrm{\&pi;}}{\mathrm{\&Delta;}};$

If get t=t ₀₁, can draw after normal moveout correction the amplitude of each point in the wavelet, the seismologic record of the reconstruct on these aspects is exactly the amplitude of each point in the seismic wavelet;

Point after if these are stretched is interpreted as the real point of original wavelet, promptly t ₀₁Be reduced to t ₀(i), then at time t ₀₁Last through the seismologic record g (t after the normal moveout correction ₀₁) be exactly the seismologic record of eliminating after stretching;

Drawing the seismologic record of eliminating after stretching is:

$f\left(t_0\left(i\right)\right)=\underset{k=k1}{\overset{k2}{\mathrm{\&Sigma;}}}g\left(k\right)\frac{\sin ({\mathrm{\&omega;}}_N{t}_{01}-\mathrm{k\&pi;})}{{\mathrm{\&omega;}}_N{t}_{01}-\mathrm{k\&pi;}}---\left(9\right)$

Wherein,

$t_{01}=2t_0\left(i\right)+{\mathrm{dt}}_0-\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_0^2\left(i\right)+\frac{x^2}{v^2\left(i\right)}}---\left(10\right)$

The summation interval [k here ₁, k ₂] determine by the length N of seismic wavelet;

At geophone offset is on the seismic trace of x, determines k respectively by the dynamic correction value of wavelet terminal correspondence ₁, k ₂Value:

$k_1=-{\mathrm{dt}}_0-2{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}+\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}^2+\frac{x^2}{v^2\left({\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}\right)}}-\frac{N-1}{2}---\left(11\right)$

$k_2=-{\mathrm{dt}}_0-2{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{0,2}+\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{0,2}^2+\frac{x^2}{v^2\left({\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{0,2}\right)}}+\frac{N-1}{2}+1---\left(12\right)$

Wherein,

${\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}=t_0-\frac{N-1}{2}\mathrm{\&Delta;}---\left(13\right)$

t _0，2＝t _0，1+(N-1)Δ (14)

(33) waveform of a lineups correspondence is proofreaied and correct.

Use the waveform stretching compensation method of processing seismic data of the present invention, after utilizing more accurate seismic velocity that seismic signal is carried out normal moveout correction, can the stretching that produce behind the normal moving school be compensated, to remedy the influence that offset distance signal far away causes owing to factors such as speed and anisotropy.This method can accurately calculate the waveform stretching amount, and adopting resamples waits some row measures, finally obtains relatively accurate, suitable waveform stretching compensation, and especially at high frequency region, the power spectrum lifting of compensation back is bigger, and the resolution of data is significantly improved.

Accompanying drawing is described

Fig. 1 (a) and Fig. 1 (b) are that waveform changes synoptic diagram with offset distance.

Fig. 2 (a) and Fig. 2 (b) are the spectrum analysis synoptic diagram.

Fig. 3 is the CDP road set information of a reality.

Fig. 4 did data after the normal moveout correction to actual road collection shown in Figure 3.

Fig. 5 is to doing the source map after Fig. 4 CDP road set information after the normal moveout correction is done stretch compensation again.

Fig. 6 is the real data spectrum analysis synoptic diagram of doing behind normal moveout correction and the stretch compensation.

Embodiment

In the open air by mechanical vibroseis artificial excitation's seismic event on the big gun line of ground design, adopt and a bit to blow out the conventional method that multiple spot receives, in pre-survey area a plurality of acceptance points are set, described a plurality of acceptance point is dispersed on the rectangular node point of an exploration targets zone scope on the ground that comprises shot point, uses earthquake detection equipment to obtain the waveform that blast involves seismic event.

Above-mentioned waveform is implemented waveform stretching compensation method of the present invention.At first accurately calculate the waveform stretching amount, sampling rate after in a stronger dominant frequency wavelength coverage of reflected signal original recorded data being utilized waveform stretching then resamples, utilize the crude sampling rate to come restore data at last again, thereby realize waveform stretching compensation.

Particularly, suppose that earthquake lineups wavelet center is at x ₀Time on=0 the seismic trace is t ₀, like this, be on the seismic trace of x ≠ 0 at geophone offset, according to normal moveout correction computing formula (1), the dynamic correction value of the central point correspondence of this wavelet is dt ₀:

${\mathrm{<figure-callout\; id="0"\; label="dt"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">dt</figure-callout>}}_0=\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_0^2+\frac{x^2}{v^2\left(t_0\right)}}-t_0---\left(1\right)$

If do not produce distortion, the dynamic correction value of every bit should be the same on the wavelet so.But it is in fact really not so.

Consider 1 i of wavelet central point top, suppose that it is at x ₀=0 place's time corresponding is t ₀(i)＜t ₀, on the t direction, the distance of this i and wavelet central point is:

δ＝t ₀-t ₀(i) (2)

Correspondingly, time corresponding then is on x ≠ 0 place's seismic trace:

t ₁＝t ₀+dt ₀-δ＝t ₀(i)+dt ₀(3)

On the other hand, this time point t ₁Value, after normal moveout correction, be corrected to time t ₀₁, that is to say that this point is at x before the normal moveout correction ₀=0 place's time corresponding is:

t ₀₁＝t ₁-dt ₁ (4)

Wherein, dt ₁Be dynamic correction value.Like this, the distance of this point and wavelet central point is changed into:

δ ₁＝t ₀₁-t ₀

＝t ₀(i)+dt ₀-dt ₁-t ₀

＝-δ+dt ₀-dt ₁

≠δ

δ ₁+δ＝dt ₀-dt ₁ (5)

Can clearly be seen that from following formula normal moveout correction makes seismic wavelet that the variation that stretches take place.The high-order that changes as stretching can think that the dynamic correction value difference of stretching front and back position correspondence is little, promptly in a small amount

dt ₁≈dt ₀(i) (6)

Like this, from (3), (4), (6) formula can draw

$t_{01}\&ap;t_0\left(i\right)+{\mathrm{dt}}_0-{\mathrm{dt}}_0\left(i\right)$

$=2t_0\left(i\right)+d{t}_0-\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_0^2\left(i\right)+\frac{x^2}{v^2\left(i\right)}}---\left(7\right)$

Wherein, v (i) is a NMO velocity of i correspondence.Thereby, draw the time t after the wavelet point normal moveout correction ₀₁

The stretching effect how according to (7) formula the normal moveout correction time to be produced is discussed below to compensate.

The amplitude of the geological data sampled point k of hypothetical record is g (k), if sampling rate is a Δ, then can draw the amplitude of g (k) at time t according to sampling thheorem from the data of record:

$g\left(t\right)=\underset{k=k1}{\overset{k2}{\mathrm{\&Sigma;}}}g\left(k\right)\frac{\sin ({\mathrm{\&omega;}}_{N}t-\mathrm{k\&pi;})}{{\mathrm{\&omega;}}_{N}t-\mathrm{k\&pi;}}---\left(8\right)$

Wherein,

${\mathrm{\&omega;}}_N=\frac{\mathrm{\&pi;}}{\mathrm{\&Delta;}}.$

If get t=t ₀₁, promptly can draw after normal moveout correction the amplitude of each point in the wavelet.The seismologic record of the reconstruct on these aspects is exactly the amplitude of each point in the seismic wavelet.If the point after we are stretched these is interpreted as the real point of original wavelet, promptly t ₀₁Be reduced to t ₀(i), so, at time t ₀₁Last through the seismologic record g (t after the normal moveout correction ₀₁) be exactly the seismologic record f (t that eliminates after stretching ₀(i)).Like this, the seismologic record after we can draw and eliminate stretch is:

$f\left(t_0\left(i\right)\right)=\underset{k=k1}{\overset{k2}{\mathrm{\&Sigma;}}}g\left(k\right)\frac{\sin ({\mathrm{\&omega;}}_N{t}_{01}-\mathrm{k\&pi;})}{{\mathrm{\&omega;}}_N{t}_{01}-\mathrm{k\&pi;}}---\left(9\right)$

Wherein,

$t_{01}=2t_0\left(i\right)+{\mathrm{dt}}_0-\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_0^2\left(i\right)+\frac{x^2}{v^2\left(i\right)}}---\left(10\right)$

The summation interval [k here ₁, k ₂] determine by the length N of seismic wavelet.At geophone offset is on the seismic trace of x, determines k respectively by the dynamic correction value of wavelet terminal correspondence ₁, k ₂Value.

$k_1=-{\mathrm{dt}}_0-2{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}+\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}^2+\frac{x^2}{v^2\left({\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}\right)}}-\frac{N-1}{2}---\left(11\right)$

$k_2=-{\mathrm{dt}}_0-2{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{0,2}+\sqrt{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{0,2}^2+\frac{x^2}{v^2\left({\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{0,2}\right)}}+\frac{N-1}{2}+1---\left(12\right)$

Here,

${\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="C200810098060E00171.png,C200810098060E00192.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{0,1}}=t_0-\frac{N-1}{2}\mathrm{\&Delta;}---\left(13\right)$

t _0，2＝t _0，1+(N-1)Δ (14)

So just can realize the waveform of a lineups correspondence is proofreaied and correct.In fact, the rally of the road after normal moveout correction has many lineups all to need to eliminate the correction of stretching.For this reason, we determine by investigating the energy in first road specifically which lineups need be proofreaied and correct, and avoids that same lineups are decomposed into several lineups and proofreaies and correct.Concrete way is, on the energy trace in this road collection first road, finds several peak values, when these peak values during greater than given threshold value, during such as 0.2 times of this road average energy, is interpreted as that then new lineups need eliminate the correction of stretching.

The experiment of model data synthetic information

According to actual seismic data, we have set up a horizontal layer rate pattern that has 16 layers, and the concrete model parameter sees Table 1.

Fig. 1 (a) and Fig. 1 (b) are that waveform changes synoptic diagram with offset distance, and x is an offset distance, and y is a waveform.In the certain deviation distance, waveform changes (in the a-quadrant of Fig. 1 (a)) in certain scope, and after through the high-order normal moveout correction, offset distance seismic event far away produces deformation owing to being stretched, and makes waveform ANOMALOUS VARIATIONS (in the B zone of Fig. 1 (a)) occur.After doing stretch compensation, reduce as can be seen by the B region area of Fig. 1 (b), use this technology to make the result behind the stretch compensation, the wave form distortion that the stretching effect of normal moveout correction causes is eliminated, and is better with the waveform consistance of shortcut a long way.

Fig. 2 (a) and Fig. 2 (b) are the spectrum analysis synoptic diagram, and p represents energy, and f represents frequency.The low frequency range frequency spectrum has certain stability, and after finishing nmo stretching, near the frequency spectrum the 100HZ suddenly descends (Fig. 2 (a)), use this technology to compensate after, near the frequency spectrum the 100HZ has obvious lifting (Fig. 2 (b)).This technology also has certain destruction near other lineups the reflection horizon, if the selection appropriate measures of taking during use, so ruined should be coherent noise, and can not reduce the quality of data.

The treatment effect of real data

Fig. 3 is the CDP road set information of a reality; Fig. 4 did data after the normal moveout correction to actual road collection shown in Figure 3; Fig. 5 is to doing the data after Fig. 4 CDP road set information after the normal moveout correction is done stretch compensation again; Fig. 6 is the real data spectrum analysis synoptic diagram of doing behind normal moveout correction and the stretch compensation, and A is that low frequency is than the stabilized zone among the figure; A represents the stretched compensation of HFS, the curve after the frequency spectrum lifting; B represents the spectrum curve after HFS is done the high-order normal moveout correction, and curve suddenly descends.This figure shows that power spectrum has obvious lifting after using this technology near the 100Hz, that is to say that the resolution of data is improved.

Especially be apparent that, use the waveform stretching compensation method of processing seismic data of the present invention, after utilizing more accurate seismic velocity that seismic signal is carried out normal moveout correction, can the stretching that produce behind the normal moving school be compensated, to remedy the influence that offset distance signal far away causes owing to factors such as speed and anisotropy.This method can accurately calculate the waveform stretching amount, and adopting resamples waits some row measures, finally obtains relatively accurate, suitable waveform stretching compensation, and especially at high frequency region, the power spectrum lifting of compensation back is bigger, and the resolution of data is significantly improved.Like this, use waveform that the waveform stretching compensation obtain crosses and judge and explain and improved surveying accuracy by subsurface geological structure, improved the favourable hydrocarbon trap of seismic data interpretation personnel searching, determined the precision of well location.

As above-mentioned, the clear method that has described the waveform stretching compensation of the present invention's proposition in detail.Although the present invention is described and explained to the preferred embodiments of the present invention in detail, but this area those skilled in the art be appreciated that, under the situation of the spirit and scope of the present invention that do not deviate from the claims definition, can in form and details, make multiple modification.

Claims (3)

1, a kind of waveform stretching compensation method of processing seismic data, this waveform stretching compensation method may further comprise the steps:

(i) can utilize seismic exploration equipment, acquiring seismic data:

In the reflection wave seismic prospecting, adopt and a bit blow out the conventional method that multiple spot receives, in pre-survey area a plurality of acceptance points are set, described a plurality of acceptance point is dispersed on the rectangular node point of an exploration targets zone scope on the ground that comprises shot point, uses earthquake detection equipment to obtain the waveform that blast involves seismic event;

(ii), accurately calculate the waveform stretching amount according to the Wave data that in step (i), obtains;

Sampling rate after (iii) in a stronger dominant frequency wavelength coverage of reflected signal original recorded data being utilized waveform stretching resamples, and utilizes the crude sampling rate to come restore data again, and then the waveform that is stretched is realized waveform stretching compensation;

(iv) use waveform that the waveform stretching compensation that (iii) obtains through step crosses and judge and explain and improved surveying accuracy by subsurface geological structure, improved the favourable hydrocarbon trap of seismic data interpretation personnel searching, determined the precision of well location.

2, the method in the claim 1, wherein step accurately calculates the waveform stretching amount according to the Wave data that obtains described in (ii) and may further comprise the steps in step (i):

(21) suppose that earthquake lineups wavelet center is at x ₀Time on=0 the seismic trace is t ₀, be on the seismic trace of x ≠ 0 at geophone offset, according to following normal moveout correction computing formula (1), the dynamic correction value of the central point correspondence of this wavelet is dt ₀:

${\mathrm{dt}}_0=\sqrt{t_0^2+\frac{x^2}{v^2\left(t_0\right)}}-t_0---\left(1\right)$

V (t wherein ₀) be time t ₀Pairing NMO velocity;

(22) 1 i of consideration wavelet central point top supposes that it is at x ₀=0 place's time corresponding is t ₀(i)＜t ₀, on the t direction, the distance of this i and wavelet central point is:

δ＝t ₀-t ₀(i) (2)

(23) time corresponding is on x ≠ 0 place's seismic trace:

t ₁＝t ₀+dt ₀-δ＝t ₀(i)+dt ₀ (3)

(24) after normal moveout correction, with time point t ₁Value be corrected to time t ₀₁, that is to say that this point is at x before the normal moveout correction ₀=0 place's time corresponding is:

t ₀₁＝t ₁-dt ₁ (4)

Dt wherein ₁Be dynamic correction value, thereby the distance that obtains this point and wavelet central point is changed into:

δ ₁＝t ₀₁-t ₀

＝t ₀(i)+dt ₀-dt ₁-t ₀

＝-δ+dt ₀-dt ₁

≠δ

δ ₁+δ＝dt ₀-dt ₁ (5)

Dt wherein ₀-dt ₁Be the waveform stretching amount.

3, the method in the claim 1, sampling rate after wherein step is utilized waveform stretching to original recorded data described in (iii) in a stronger dominant frequency wavelength coverage of reflected signal resamples, utilize the crude sampling rate to come restore data again, and then the waveform that is stretched is realized that waveform stretching compensation may further comprise the steps: (31) make dt ₁≈ dt ₀(i) (6)

$t_{01}\&ap;t_0\left(i\right)+{\mathrm{dt}}_0-{\mathrm{dt}}_0\left(i\right)$

$={2t}_0\left(i\right)+{\mathrm{dt}}_0-\sqrt{t_0^2\left(i\right)+\frac{x^2}{v^2\left(i\right)}}---\left(7\right)$

Wherein v (i) is a NMO velocity of i correspondence;

Draw time t after the wavelet point normal moveout correction according to formula (7) ₀₁

(32) stretching effect that the normal moveout correction time is produced according to formula (7) compensates:

The amplitude of the geological data sampled point k of hypothetical record is g (k), if sampling rate is a Δ, then can draw the amplitude of g (k) at time t according to sampling thheorem from the data of record:

$g\left(t\right)=\underset{k=k1}{\overset{k2}{\mathrm{\&Sigma;}}}g\left(k\right)\frac{\sin ({\mathrm{\&omega;}}_{N}t-\mathrm{k\&pi;})}{{\mathrm{\&omega;}}_{N}t-\mathrm{k\&pi;}}---\left(8\right)$

Wherein ${\mathrm{\&omega;}}_N=\frac{\mathrm{\&pi;}}{\mathrm{\&Delta;}};$

If get t=t ₀₁, can draw after normal moveout correction the amplitude of each point in the wavelet, the seismologic record of the reconstruct on these aspects is exactly the amplitude of each point in the seismic wavelet;

Point after if these are stretched is interpreted as the real point of original wavelet, promptly t ₀₁Be reduced to t ₀(i), then at time t ₀₁Last through the seismologic record g (t after the normal moveout correction ₀₁) be exactly the seismologic record of eliminating after stretching;

Drawing the seismologic record of eliminating after stretching is:

$f\left(t_0\left(i\right)\right)=\underset{k=k1}{\overset{k2}{\mathrm{\&Sigma;}}}g\left(k\right)\frac{\sin ({\mathrm{\&omega;}}_N{t}_{01}-\mathrm{k\&pi;})}{{\mathrm{\&omega;}}_N{t}_{01}-\mathrm{k\&pi;}}---\left(9\right)$

Wherein,

$t_{01}={2t}_0\left(i\right)+{\mathrm{dt}}_0-\sqrt{t_0^2\left(i\right)+\frac{x^2}{v^2\left(i\right)}}---\left(10\right)$

The summation interval [k here ₁, k ₂] determine by the length N of seismic wavelet;

At geophone offset is on the seismic trace of x, determines k respectively by the dynamic correction value of wavelet terminal correspondence ₁, k ₂Value:

$k_1={-\mathrm{dt}}_0-{2t}_{\mathrm{0,1}}+\sqrt{t_{\mathrm{0,1}}^2+\frac{x^2}{v^2\left(t_{\mathrm{0,1}}\right)}}-\frac{N-1}{2}---\left(11\right)$

$k_2={-\mathrm{dt}}_0-{2t}_{0,2}+\sqrt{t_{\mathrm{0,2}}^2+\frac{x^2}{v^2\left(t_{\mathrm{0,2}}\right)}}+\frac{N-1}{2}+1---\left(12\right)$

Wherein,

$t_{\mathrm{0,1}}=t_0-\frac{N-1}{2}\mathrm{\&Delta;}---\left(13\right)$

t _0，2＝t _0，1+(N-1)Δ (14)

(33) waveform of a lineups correspondence is proofreaied and correct.

Images