CN104090299A - Surface seismic data amplitude compensation method based on VSP primary waves - Google Patents

Abstract

The invention provides a surface seismic data amplitude compensation method based on VSP primary waves. The method includes the following steps that downlink waves in VSP data are separated, initial time of each VSP datum in the downlink waves of the VSP data is automatically picked up, amplitudes corresponding to the initial time of the VSP data are obtained, and a data sequence is formed; the data sequence is fitted by means of an exponential curve, and an exponent x of the exponential curve is solved in a non-linear inversion method; amplitudes of sampling points of the seismic trace of surface seismic data are multiplied by an X power of time values corresponding to the sampling points respectively, so that time function gain is carried out on the surface seismic data, and amplitude compensation is achieved. According to the method, deep energy of the surface seismic data can be effectively compensated, true amplitude recovery of the seismic data is achieved, and prediction precision of an oil and gas reservoir is improved.

Description

Surface seismic data amplitudes compensation method based on VSP primary wave

Technical field

The present invention relates to oil exploration technology field, more particularly, relate to a kind of amplitude compensation or restoration methods of geological data.

Background technology

Along with going deep into gradually of oil-gas exploration, strengthen gradually for the input of complex lithology hydrocarbon-bearing pool and subtle reservoir, therefore more and more higher to the requirement of accuracy of seismic exploration.Seismic event is in the time of Propagation, and the energy of ripple is energy transform into heat energy due to the interaction between particle, thereby causes earthquake wave amplitude constantly to be decayed.Medium is roughly directly proportional to the wave propagation time to the absorption of seismic event, on the record section of seismic event, reflection wave from shallow to, the dominant frequency amplitude step-down of deep layer, resolution also decreases.Therefore at surface seismic Data processing, must be by someway the amplitude of this part decay being carried out to certain compensation.

For the processing of surface seismic data amplitudes compensation, conventional method has single track amplitude equalization, spherical diffusion compensation, earth's surface-consistent amplitude compensation and residual amplitude compensation etc., but the effect of said method is all not good.

Summary of the invention

For the deficiencies in the prior art, one of object of the present invention is to solve the one or more problems that exist in above-mentioned prior art.For example, one of object of the present invention is to provide a kind of and can realizes better the method for ground geological data being carried out to compensation and the recovery of real amplitude.

To achieve these goals, the invention provides a kind of surface seismic data amplitudes compensation method based on VSP primary wave.Said method comprising the steps of: A, isolate the descending ripple in VSP data, and pick up the first break time of per pass VSP data in the descending ripple of VSP data, and obtain the amplitude corresponding to first break time of per pass VSP data, form a data sequence; Data sequence described in B, employing exponential curve fitting, and adopt nonlinear inversion method to obtain the index x of described index curve; C, the amplitude of each sampling point of surface seismic data seismic trace is multiplied by respectively to x power of the time value that described each sampling point is corresponding, ground geological data carried out to function of time gain and to realize amplitude compensation.

An embodiment of the surface seismic data amplitudes compensation method based on VSP primary wave according to claim 1, described steps A comprises:

A1, read in zero-offset VSP data; A2, the upward traveling wave that separates described zero-offset VSP data and descending ripple; A3, obtain time corresponding to per pass VSP data according to the direct wave speed of well depth corresponding to per pass VSP data and given lower shape ripple; A4, time corresponding to described per pass VSP data is moved down to the schedule time after and window while forming between time corresponding to described per pass VSP data; A5, per pass VSP data time add up peak swing in window; A6, ask for the first break time of per pass VSP data, and from extract the corresponding amplitude A of first break time of per pass VSP data through the descending ripple of described steps A 2 VSP data after treatment _iwherein, certain one corresponding amplitude of VSP data first break time, Yu Gai road VSP data first break time are adjacent and be positioned at two corresponding amplitudes of time point before this road VSP data first break time and the VSP data first break time corresponding amplitude sum of two time points adjacent and that be positioned at after this road VSP data first break time in Yu Gai road is greater than the value of described peak swing divided by correction factor gained; Wherein i is the road serial number of VSP data, and 1≤i≤N and i are positive integer, and N is the road number of zero-offset VSP data maximum.

According to an embodiment of surface seismic data amplitudes compensation method who the present invention is based on VSP primary wave, described steps A 2 comprises: to described zero-offset VSP data do two-dimensional Fourier transform with by the data transformation in Space Time territory in frequency-wavenumber domain, figure after conversion is done to filtering processing, so that upward traveling wave decay, and then the data after conversion are done to two-dimensional Fourier transform so that the data transformation of frequency-wavenumber domain is returned in Space Time territory.

According to an embodiment of surface seismic data amplitudes compensation method who the present invention is based on VSP primary wave, described step B comprises:

B1, utilize random function random m value X of generation in the span of index x _j, 1≤j≤m, and j is positive integer, obtains exponent data group { X ₁, X ₂..., X _m, wherein, the span of described index x is [1,3];

B2, carry out iteration for the first time and, to ask for optimality index x, comprise the steps:

B21, the exponent data group { X producing from described step B1 ₁, X ₂..., X _mchoose arbitrarily a value X _jand calculate calculated amplitude value A corresponding to every one VSP data first break time according to formula (1) below _i':

(f _i) ^x＝A _i' (1)

B22, set up objective function, and the A calculating according to described step B21 _i' and amplitude A corresponding to per pass VSP data first break time that pick up from VSP data _icalculate E _j:

$E_j=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\left(A_i{-A_i}^{\′}\right)}^2---\left(2\right)$

In formula (1) and (2), f _ibe the first break time of i road VSP data, i is the road serial number of VSP data, and 1≤i≤N and i are positive integer, and N is the road number of zero-offset VSP data maximum;

B23, according to described step B21 and B22, calculate m the X that step B1 produces _jm the E accordingly of institute _j, and find out m E _jin minimum value E _mincorresponding index X _bestwith m E _jin maximal value E _maxcorresponding index X _worst;

B24, as a described m E _jin minimum value E _minwhile being less than predictive error precision ε, by index X _bestas optimality index x, otherwise, iteration next time continued;

B3, carry out iteration the τ time, τ >=2 and be positive integer, to ask for optimality index x, comprise the following steps:

B31, m the value X that described step B1 is produced at random _jcalculate new exponent data X' according to formula (3) below:

$X^{\′}=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{a}_j{X}_j---\left(3\right)$

In formula (3), -0.5≤a _j≤ 1.5;

B32, by formula (1) and (2) described in X' substitution, if the E value calculating is less than a described m E _jin maximal value, will with X' substitute described exponent data group { X ₁, X ₂..., X _min X _worst, form new exponent data group;

B33, taking described new exponent data group as benchmark, recalculate and obtain minimum value E according to described step B21～B23 _min', maximal value E _max', index X _best' and index X _worst';

If B34 is E _min' < ε, by index X _best' as optimality index x, otherwise, iteration next time continued.

Compared with prior art, the present invention is by VSP first arrival automatic Picking and adopt the method for non-linear inversion to ask for index x, and according to index x, ground geological data is carried out to function of time gain compensation, deep energy that can effective compensation geological data, realize compensation and the recovery of ground geological data being carried out to real amplitude, contributed to improve the precision of prediction of oil-bearing reservoir.

Embodiment

Hereinafter, describe in detail according to the surface seismic data amplitudes compensation method based on VSP primary wave of the present invention in connection with exemplary embodiment.

VSP (Vertical Seismic Profiling, being vertical seismic profile (VSP)) seismologic record (or is called VSP geological data, VSP seismic data, VSP data) excite the observed pattern of downhole receiving for ground, be that shot point is arranged on ground and geophone station is arranged in well, its descending direct wave is affected by noise less, can reflect more accurately that compared with surface seismic record underground medium is to amplitude fading effect, therefore inventor proposes to utilize the through descending ripple of VSP to extract earth-attenuation index, and utilize the surface seismic data that this index collects the same area to carry out function of time gain compensation.Here, descending ripple and upward traveling wave are the directionizations point that propagates into acceptance point according to ripple, wherein, and from the descending ripple that cries of acceptance point top, from the upward traveling wave that cries below acceptance point.

Surface seismic data (or being called surface seismic record, surface-seismic data) refer to the observed pattern that ground excites ground to receive, i.e. shot point setting on the ground and geophone station be also located at ground.

Particularly, inventor has proposed a kind of surface seismic data amplitudes compensation method based on VSP primary wave, first the method isolates the descending ripple in VSP data, the first arrival (being first break time) of automatic Picking per pass VSP data in the descending ripple of VSP data again, and amplitude corresponding to first break time that obtains per pass VSP data takes out a DS of formation, adopt these data of exponential curve fitting, thereby obtain definite index x, again according to this index x, the surface seismic data that collect are carried out to function of time gain compensation, thereby realize ground geological data is carried out to compensation and the recovery of real amplitude.Wherein, primary wave and first break time refer to: when first seismic event wavefront arrives certain observation station, the particle of this medium starts to occur the moment of vibration, be called the first break time of ripple, and the ripple of recording, be called primary wave, that is to say, the ripple of first arrival of recording on seismologic record is called primary wave.

The present invention is achieved through the following technical solutions:

A, first isolate the descending ripple in VSP data, then in the descending ripple of VSP data the first break time of automatic Picking per pass VSP data, and amplitude corresponding to first break time that obtains per pass VSP data forms a data sequence;

Data sequence described in B, an exponential curve fitting of employing, and adopt nonlinear inversion method to obtain the index x of described index curve;

C, the various kinds point that surface seismic is recorded to seismic trace are multiplied by x power of the time value that each sampling point is corresponding, so that ground geological data is carried out to function of time gain compensation.

In one exemplary embodiment, the geological data amplitude compensation method based on VSP primary wave according to the present invention comprises the following steps:

(1) read in zero-offset VSP data V (i, t).

Wherein, be the road serial number of zero-offset VSP data, i={1,2......, N}, N is the road number of zero-offset VSP data maximum; T is the time, t={1, and 2......, T}, T is the zero-offset VSP data data dominant record times.Hereinafter, all identical about the implication of i, repeat no more.

(2) upward traveling wave and the descending ripple of separation zero-offset VSP data.

This step can adopt the upward traveling wave of separation zero-offset VSP data conventional in prior art and the method for descending ripple, for example, and vertical stack, multiple tracks pie slice, F-K filtering, medium filtering, best of breed filtering, least squares filtering etc.

In the present embodiment, proportion-wavenumber domain filtering.That is, by zero-offset VSP data S (i, t) are done two-dimensional Fourier transform by the data transformation in Space Time territory in frequency-wavenumber domain, at this moment descending ripple is in positive half-plane, upward traveling wave is in negative half-plane; Figure after conversion is done to filtering processing, and the data in negative half-plane are multiplied by decimal (for example 0.00001) and make upward traveling wave decay, and the descending ripple of positive half-plane is unaffected; Data after conversion are being made to two-dimentional Fourier transform and are getting back to the result S in Space Time territory _dw(i, t), upward traveling wave is decayed, and descending ripple strengthens.

(3) from VSP data trace header, read the well depth H that every track data is corresponding _i, the simultaneously direct wave speed V of given descending ripple, and obtain time corresponding to per pass VSP data according to the direct wave speed of well depth corresponding to per pass VSP data and given lower shape ripple,

(4) the time t obtaining according to the every track data of step (3) _i, move down time T _l, window (t while obtaining _i, t _i+ T _l).Herein, T _lvalue according to being to make the time window forming will comprise the first arrival in all roads as far as possible, but be difficult for excessively, generally get T _l=100ms.

(5) at the time window (t of per pass VSP data _i, t _i+ T _l) interior statistics peak swing, obtain MaxAmp _i.

(6) ask for the first break time of per pass VSP data, particularly:

Window (t when order is got i road _i, t _i+ T _l) in a time point t _p, by this time point t _pamplitude and with its before and after be adjacent to two time point t _p-2, t _p-1, t _p+1, t _p+2corresponding amplitude is added summation and obtains amplitude Sub, and when this amplitude, Sub is greater than MaxAmp _idivided by the value of correction factor SF gained, just think that this time point is the first break time of corresponding VSP track data.

That is: Sub=S _dw(i, t _p-2)+S _dw(i, t _p-1)+S _dw(i, t _p)+S _dw(i, t _p+1)+S _dw(i, t _p+2), if Sub> is (MaxAmp _i/ SF), t _pit is exactly the first break time of this road VSP data, wherein, SF refers to a corrected parameter, can judge that whether this correction factor SF value is reasonable by setting a SF value and observing the difference that automatic Picking obtains the first arrival of Ge road and the actual first arrival in manual observation Dao Ge road, if the unreasonable value of adjusting again SF, for example, first get SF=1, obtain in the manner described above the first break time of per pass VSP data, and the artificial judgment first break time of picking up and the difference of actual first break time, if difference is larger, again choose SF value and again pick up, until the first arrival of picking up conforms to substantially with actual first arrival, that is to say, for data with different, can adjust the precision that the adjustment of SF value is picked up, SF=10 in this example.Set in the present embodiment the method for all roads of automatic Picking VSP data first arrival, but the invention is not restricted to this, can also pick up by the method for manual observation the first break time of per pass VSP data, but according to the automatic Picking mode of the present embodiment, reduce the workload of manually picking up, the time of saving.

Wherein, S _dw(i, t _p-2) represent the t of i road VSP data _p-2the amplitude that moment is corresponding, S _dw(i, t _p+1), S _dw(i, t _p), S _dw(i, t _p+2) implication in like manner.

(7) obtain according to above (1)～(6) step the first break time f that per pass is corresponding _i, will be from step (2) VSP data after treatment S _dwcorresponding amplitude A corresponding to per pass first break time that take out in (i, t) _i.

(8) set up following relational expression (1), ask for index x:

(f _i) ^x＝A _i' (1)

Wherein, f _ibe the first break time of i road VSP data, for example, f ₁represent the first break time t of the 1st road VSP data _p, f ₂represent the first break time t of the 2nd road VSP data _p.

(9), in the span [1,3] of given index x, x, in 1～3 scope, produces m value X at random with random function _j, 1≤j≤m, and j is positive integer, obtains exponent data group { X ₁, X ₂..., X _m, and carry out iteration for the first time, τ=1 according to step (10)～(15).

(10) the exponent data group { X that index x produces from step (9) ₁, X ₂..., X _mchoose arbitrarily a value X _jand calculate x according to formula (1) above and be taken as X _jtime calculated amplitude value A corresponding to every one VSP data first break time _i';

(11) set up objective function, and the A calculating according to described step (10) _i' and amplitude A corresponding to every track data first break time that pick up from VSP data of step (7) _icalculate x and be taken as X _jtime E _j:

$E_j=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left(A_i{-A_i}^{\&prime;}\right)}^2---\left(2\right)$

(12), according to described step (10) and (11), calculate m the X that (9) produce _jm the E accordingly of institute _j, and find out m E _jin minimum value E _minwith and corresponding index X _best, m E _jin maximal value E _maxwith and corresponding index X _worst

(13) given error precision ε=0.001.

(14) work as E _min> ε, repeats step (15)～(19) below; Otherwise, by minimum value E _mincorresponding index X _bestas optimum index x, and carry out step (20).

(15) τ=τ+1, carries out next iteration.

(16) according to formula (3) below to { X ₁, X ₂..., X _mrecombinating generates new data X':

$X^{\&prime;}=\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{a}_j{X}_j---\left(3\right)$

In formula (3), -0.5≤a _i≤ 1.5.

(17) by this new data X' substitution formula (1) and (2), if the E value that fruit calculates is less than X _worstcorresponding E value (is E _max), will substitute described exponent data group { X with X' ₁, X ₂..., X _min X _worst, form new exponent data group.

(18) taking the new exponent data group of gained as benchmark, according to step (8), (10), (11) and (12), again according to formula (1) and (2), recalculate the minimum value E that obtains E _min', the maximal value E of E _max' and index X corresponding to its difference _best', index X _worst';

(19) if E _min' < ε, by index X _best' as optimality index x, and carry out step (20); Otherwise, continue iteration next time, repeat above step (15)～(18).

(20) according to index x, ground geological data is carried out to the recovery of real amplitude, the amplitude of each sampling point of seismic trace is multiplied by x power of its time value, that is:

A ₀＝A _tt ^x (4)

Wherein, A ₀the amplitude after compensation, A _tthe amplitude that refers to a sampling point of seismic trace, t refers to A _tcorresponding time value (moment in other words).That is to say, above (1) to (19) step be all in order to obtain index x, per pass surface seismic data are all with identical index x, formula (4) is that ground geological data seismic trace is carried out to amplitude compensation, and other seismic trace also compensates according to this formula (4).

According to the present invention, by VSP first arrival automatic Picking, and quantitatively obtain index x, need not rule of thumb or test given index x, avoid losing time and the bad problem of surface seismic compensation data effect, realized ground geological data is carried out to compensation and the recovery of real amplitude.During the amplitude compensation that applies the present invention to surface-seismic data is processed, deep energy that can effective compensation surface seismic data, realizes the true amplitude recovery of surface seismic data, contributes to like this to improve the precision of prediction of oil-bearing reservoir; Simultaneously this technology handled easily, travelling speed are fast, can meet the demand of actual production.Adopt the present invention in seismic data treatment process, to bring into play further positive role, application prospect is good.

Although above by having described the present invention in conjunction with exemplary embodiment, it will be apparent to those skilled in the art that in the case of not departing from the spirit and scope that claim limits, can carry out various amendments and change to exemplary embodiment of the present invention.

Claims (4)

1. the surface seismic data amplitudes compensation method based on VSP primary wave, is characterized in that, said method comprising the steps of:

A, isolate the descending ripple in VSP data, and pick up the first break time of per pass VSP data in the descending ripple of VSP data, and obtain the amplitude corresponding to first break time of per pass VSP data, form a data sequence;

Data sequence described in B, employing exponential curve fitting, and adopt nonlinear inversion method to obtain the index x of described index curve;

C, the amplitude of each sampling point of surface seismic data seismic trace is multiplied by respectively to x power of the time value that described each sampling point is corresponding, ground geological data carried out to function of time gain and to realize amplitude compensation.

2. the surface seismic data amplitudes compensation method based on VSP primary wave according to claim 1, is characterized in that, described steps A comprises:

A1, read in zero-offset VSP data;

A2, the upward traveling wave that separates described zero-offset VSP data and descending ripple;

A3, obtain time corresponding to per pass VSP data according to the direct wave speed of well depth corresponding to per pass VSP data and given lower shape ripple;

A4, time corresponding to described per pass VSP data is moved down to the schedule time after and window while forming between time corresponding to described per pass VSP data;

A5, per pass VSP data time add up peak swing in window;

A6, ask for the first break time of per pass VSP data, and from extract the corresponding amplitude A of first break time of per pass VSP data through the descending ripple of described steps A 2 VSP data after treatment _iwherein, certain one corresponding amplitude of VSP data first break time, Yu Gai road VSP data first break time are adjacent and be positioned at two corresponding amplitudes of time point before this road VSP data first break time and the VSP data first break time corresponding amplitude sum of two time points adjacent and that be positioned at after this road VSP data first break time in Yu Gai road is greater than the value of described peak swing divided by correction factor gained;

Wherein i is the road serial number of VSP data, and 1≤i≤N and i are positive integer, and N is the road number of zero-offset VSP data maximum.

3. the surface seismic data amplitudes compensation method based on VSP primary wave according to claim 2, is characterized in that, described steps A 2 comprises:

To described zero-offset VSP data do two-dimensional Fourier transform with by the data transformation in Space Time territory in frequency-wavenumber domain, figure after conversion is done to filtering processing, so that upward traveling wave decay, and then the data after conversion are done to two-dimensional Fourier transform so that the data transformation of frequency-wavenumber domain is returned in Space Time territory.

4. the surface seismic data amplitudes compensation method based on VSP primary wave according to claim 1, is characterized in that, described step B comprises:

B1, utilize random function random m value X of generation in the span of index x _j, 1≤j≤m, and j is positive integer, obtains exponent data group { X ₁, X ₂..., X _m, wherein, the span of described index x is [1,3];

B2, carry out iteration for the first time and, to ask for optimality index x, comprise the steps:

B21, the exponent data group { X producing from described step B1 ₁, X ₂..., X _mchoose arbitrarily a value X _jand calculate calculated amplitude value A corresponding to every one VSP data first break time according to formula (1) below _i':

(f _i) ^x＝A _i' (1)

B22, set up objective function, and the A calculating according to described step B21 _i' and amplitude A corresponding to per pass VSP data first break time that pick up from VSP data _icalculate E _j:

$E_j=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left(A_i{-A_i}^{\&prime;}\right)}^2---\left(2\right)$

In formula (1) and (2), f _ibe the first break time of i road VSP data, i is the road serial number of VSP data, and 1≤i≤N and i are positive integer, and N is the road number of zero-offset VSP data maximum;

B23, according to described step B21 and B22, calculate m the X that step B1 produces _jm the E accordingly of institute _j, and find out m E _jin minimum value E _mincorresponding index X _bestwith m E _jin maximal value E _maxcorresponding index X _worst;

B24, as a described m E _jin minimum value E _minwhile being less than predictive error precision ε, by index X _bestas optimality index x, otherwise, iteration next time continued;

B3, carry out iteration the τ time, τ >=2 and be positive integer, to ask for optimality index x, comprise the following steps:

B31, m the value X that described step B1 is produced at random _jcalculate new exponent data X' according to formula (3) below:

$X^{\&prime;}=\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{a}_j{X}_j---\left(3\right)$

In formula (3), -0.5≤a _j≤ 1.5;

B32, by formula (1) and (2) described in X' substitution, if the E value calculating is less than a described m E _jin maximal value, will with X' substitute described exponent data group { X ₁, X ₂..., X _min X _worst, form new exponent data group;

B33, taking described new exponent data group as benchmark, recalculate and obtain minimum value E according to described step B21～B23 _min', maximal value E _max', index X _best' and index X _worst';

If B34 is E _min' < ε, by index X _best' as optimality index x, otherwise, iteration next time continued.