CN103630935A - Method for removing alternating current interference signal in seismic data - Google Patents

Abstract

The invention provides a method for removing an alternating current interference signal in seismic data. The method comprises the steps of acquiring at least one path of seismic data, and performing Fourier transform on each path of seismic data to obtain at least one path of frequency domain seismic data; in a set frequency range, selecting one piece of frequency domain seismic data with a maximum amplitude value from each path of frequency domain seismic data, comparing the amplitude value of the selected frequency domain seismic data with a preset amplitude value, forming new frequency domain seismic data with the frequency domain seismic data of which the amplitude value is greater than or equal to the preset amplitude value, and performing Fourier inversion on the new frequency domain seismic data to obtain an observation vector of a time domain; determining a best disjunct matrix, and using a decomposition model to calculate the alternating current interference signal; subtracting the alternating current interference signal from the observation vector of the time domain to obtain an effective signal. By adopting the method, the alternating current interference signal can be effectively removed, the effective signal is effectively separated from the alternating current interference signal, the processing efficiency is high, and the operation speed is fast.

Description

Remove the method for alternating current disturbance signal in geological data

Technical field

The invention belongs to petroleum seismic Data Mining field, more particularly, relate to the method for alternating current disturbance signal in a kind of removal geological data of efficient, high-fidelity.

Background technology

The resolution and the signal to noise ratio (S/N ratio) that improve geological data are very important in each step of geological data processing, in earthquake data acquisition, if there is high voltage transmission line to pass through above survey line, the strong alternating current disturbance ripple that just there will be 50Hz in geological data, this interference wave remains unchanged to deep layer frequency, phase place and amplitude substantially from shallow-layer, can badly influence accuracy of seismic exploration.

The method of removing at present alternating current disturbance is mainly to adopt filtering method.Method is a notch filter method, and this filtering method is realized in frequency field, designs notch filter group many absorption points, that have certain resistance band, in order to eliminate first-harmonic and the harmonic components of alternating current disturbance signal.If the frequency spectrum of the frequency spectrum of useful signal and alternating current disturbance signal has aliasing, notch filter also can cause the loss of useful signal in filtering alternating current disturbance signal.Another kind method is auto adapted filtering method, and the prerequisite of the method is that useful signal is uncorrelated with alternating current disturbance signal.On this basis, utilize adaptive filter algorithm (as, lowest mean square or least square method), automatically adjust the coefficient of sef-adapting filter, to follow the tracks of the variation of input process, and realize the adaptive cancellation of alternating current disturbance signal.This filtering method is to carry out in time domain, and its essence is consistent with notch filter, can cause the loss of useful signal when removing alternating current disturbance signal.

Summary of the invention

Object of the present invention is for setting a frequency range, geological data within the scope of setpoint frequency is removed to the processing of alternating current disturbance signal, and needn't all process per pass geological data, can save a large amount of computing times, more meet the needs of actual production.

Another object of the present invention, for by determining optimal separation matrix, calculates alternating current disturbance signal, deducts alternating current disturbance signal from the measurement vector of time domain, realizes the effectively separated of useful signal and alternating current disturbance signal.

An aspect of of the present present invention provides a kind of method of removing alternating current disturbance signal in geological data, and described method comprises: (1) obtains at least one road geological data, and per pass geological data is carried out to Fourier transform, obtains at least one frequency field geological data; (2) in the frequency range of setting, from per pass frequency field geological data, choose the frequency field geological data of a peak swing value, the amplitude of the frequency field geological data of choosing and default amplitude are compared, the frequency field geological data that amplitude is more than or equal to default amplitude forms new frequency field geological data, new frequency field geological data is carried out to Fourier inversion, obtain the measurement vector of time domain; (3) determine optimal separation matrix, utilize decomposition model to calculate alternating current disturbance signal; (4) from the measurement vector of time domain, deduct alternating current disturbance signal, obtain useful signal.

Preferably, the frequency range of described setting is 48～52Hz.

Preferably, determine that the step of optimal separation matrix is,

(a) measurement vector of time domain is carried out to pre-service;

(b) utilize pretreated measurement vector, obtain separation matrix.

(c) separation matrix is normalized, obtains optimal separation matrix.

Preferably, measurement vector is carried out changing and albefaction centered by pretreated method.

Preferably, separation matrix w when utilization formula acquisition J (y) is below maximal value,

J(y)∝[E{g(y)}-E{g(y _gauss)}] ²

Wherein, y is random quantity,

w is separation matrix, and T represents transposition,

for pretreated measurement vector, J (y) represents the negentropy of random quantity y, y _gaussfor average is zero, and there is the random quantity of mutually homoscedastic Gaussian distribution, g (y) and g (y with random quantity y _gauss) be nonlinear function, $g\left(y\right)=-\frac{1}{a}\exp (-a{y}^2/2),$ $g\left(y_{\mathrm{gauss}}\right)=-\frac{1}{a}\exp (-a{y_{\mathrm{gauss}}}^2/2),$ A is constant, E{g (y) } be the average computing of g (y), E{g (y) } be g (y _gauss) average computing.

Preferably, (I) utilizes the new value w+ of formula acquisition separation matrix w below,

$w_{+}=E\left\{\tilde{X}g\left(y\right)\right\}-E\left\{g^{\′}\left(y\right)\right\}w$

Wherein, w ₊for the new value of separation matrix w, g (y) is nonlinear function,

g ' is (y) derivative of g (y), for

average computing, E{g ' is (y) } be g ' average computing (y);

(II) new value w to separation matrix ₊be normalized, obtain optimal separation matrix w ₀, the computing formula of normalized is,

${\mathrm{<figure-callout\; id="0"\; label="w"\; filenames="HDA0000420161200000023.png,HDA0000420161200000024.png"\; state="\{\{state\}\}">w</figure-callout>}}_0=\frac{w_{+}}{\left|\right|w_{+}\left|\right|},$

Wherein, ‖ w ₊‖ is the new value w of separation matrix ₊mould.

Preferably, described decomposition model is,

s ₂=w ₀X，

Wherein, X is measurement vector, w ₀for optimal separation matrix, s ₂for alternating current disturbance signal.

According to the method for alternating current disturbance signal in removal geological data of the present invention, within the scope of setpoint frequency, obtain the measurement vector of time domain, and by determining optimal separation matrix, calculate alternating current disturbance signal, thereby the alternating current disturbance signal in the measurement vector of time domain is separated, reached the removal effect of alternating current disturbance signal in geological data.

Accompanying drawing explanation

By the detailed description of carrying out below in conjunction with accompanying drawing, above and other objects of the present invention, feature and advantage will become apparent, wherein:

Fig. 1 is the process flow diagram illustrating according to the method for alternating current disturbance signal in the removal geological data of the embodiment of the present invention;

Fig. 2 a and Fig. 2 b illustrate respectively and adopt method of the present invention to process single shot record figure before and afterwards;

Fig. 3 a and Fig. 3 b illustrate respectively and adopt method of the present invention to process single big gun spectrogram before and afterwards.

Embodiment

Provide with reference to the following description of accompanying drawing to help the complete understanding to the embodiments of the invention that limited by claim and equivalent thereof.Comprise that various specific detail are help to understand, but that these details are only considered to be is exemplary.Therefore, those of ordinary skill in the art will recognize without departing from the scope and spirit of the present invention, can make various changes and modifications embodiment described herein.In addition, for clarity and brevity, omit the description to known function and structure.

In removal geological data of the present invention, the method for alternating current disturbance signal comprises: (1) obtains at least one road geological data, and per pass geological data is carried out to Fourier transform, obtains at least one frequency field geological data; (2) in the frequency range of setting, from per pass frequency field geological data, choose the frequency field geological data of a peak swing value, the amplitude of the frequency field geological data of choosing and default amplitude are compared, the frequency field geological data that amplitude is more than or equal to default amplitude forms new frequency field geological data, new frequency field geological data is carried out to Fourier inversion, obtain the measurement vector of time domain; (3) determine optimal separation matrix, utilize decomposition model to calculate alternating current disturbance signal; (4) from the measurement vector of time domain, deduct alternating current disturbance signal, obtain useful signal.

Fig. 1 is the process flow diagram illustrating according to the method for alternating current disturbance signal in the removal geological data of the embodiment of the present invention.

In step 101, obtain at least one road geological data, per pass geological data is carried out to Fourier transform, obtain at least one frequency field geological data.

Obtain at least one road geological data, per pass geological data is designated as A (x _i, t), x _ibe the input signal of i road geological data, i represents the road number of geological data, i=1, and 2 ... m, and m is greater than 0 integer, carries out Fourier transform to per pass geological data as follows,

Y(x _i,ω)=fft(A(x _i,t)) （1）

In formula (1), A (x _i, t) be the geological data of time domain, Y (x _i, ω) be A (x _i, the t) geological data of corresponding frequency field, fft represents fast fourier transform, only time orientation is done to fast fourier transform, ω is angular frequency, ω=2 π f, f is frequency.

In step 102, in the frequency range of setting, from per pass frequency field geological data, choose the frequency field geological data of a peak swing value, the amplitude of the frequency field geological data of choosing and default amplitude are compared, the frequency field geological data that amplitude is more than or equal to default amplitude forms new frequency field geological data, new frequency field geological data is carried out to Fourier inversion, obtain the measurement vector of time domain.Preferably, the frequency range of setting is set to 48～52Hz, default amplitude be within the scope of setpoint frequency, there is peak swing value frequency field geological data around the mean value of the amplitude of the frequency field geological data of other frequency M doubly, M>1, M=2 preferably, this is only exemplary, and those skilled in the art can carry out appropriate change to the value of M according to the actual conditions of geological data.

Specifically, in frequency, be that the general alternating current disturbance signal of 48～52Hz(exists near 50Hz) scope in, to frequency field geological data Y (x _i, ω) add up, from per pass frequency field geological data, choose the frequency field geological data of a peak swing value, then the amplitude of the frequency field geological data of choosing and default amplitude are compared, preferably, the M of the mean value of the amplitude of the frequency field geological data of other frequency around the amplitude Yu Gai road frequency field geological data of the frequency field geological data of choosing is doubly compared, if the amplitude of the frequency field geological data of choosing be more than or equal to this road frequency field geological data around the mean value of the amplitude of the frequency field geological data of other frequency M doubly, think in this road frequency field geological data and have alternating current disturbance signal, otherwise, if the amplitude of the frequency field geological data of choosing be less than this road frequency field geological data around the mean value of the amplitude of the frequency field geological data of other frequency M doubly, think in the frequency field geological data of choosing and do not have alternating current disturbance signal.

In the scope that is 48～52Hz in frequency, all frequency field geological datas of alternating current disturbance signal that exist are formed to new frequency field geological data, by Fourier inversion, transform to time domain, obtain the measurement vector of time domain.According to the embodiment of the present invention, the measurement vector of described time domain be useful signal and alternating current disturbance signal and, concrete computing formula is,

X=s ₁+s ₂ （2）

In formula (2), X is measurement vector, X=[x ₁, x ₂... x _m] ^t, x _ibe the input signal of i road geological data, s ₁for useful signal, s ₁=[s ₁₁, s ₁₂... s _1m] ^t, s _1ibe the useful signal of i road geological data, s ₂for alternating current disturbance signal, s ₂=[s ₂₁, s ₂₂... s _2m] ^t, s _2ibe the alternating current disturbance signal of i road geological data, the input signal x of per pass geological data _iuseful signal s _1iwith alternating current disturbance signal s _2iand.Here, useful signal and alternating current disturbance signal are independent signal.

In step 103, determine optimal separation matrix, utilize decomposition model to calculate alternating current disturbance signal.

Specifically, the step of determining optimal separation matrix is:

(a) measurement vector of time domain is carried out to pre-service.

According to the embodiment of the present invention, preferably measurement vector is carried out changing and albefaction centered by pretreated method.The step of centralization is: from measurement vector X, deduct its average, obtain the measurement vector of zero-mean, make E (x _i)=0, E (x _i) be input signal x _iaverage computing.The step of albefaction is: the measurement vector X after centralization is carried out to linear transformation.Here, the common practise that centralization and albefaction are this area, the present invention has omitted the description to this part.

(b) utilize pretreated measurement vector, obtain separation matrix.

Specifically, separation matrix w when utilization formula acquisition J (y) is below maximal value,

J(y)∝[E{g(y)}-E{g(y _gauss)}] ² （3）

In formula (3), y is random quantity, w is separation matrix, and T represents transposition,

for pretreated measurement vector, J (y) represents the negentropy of random quantity y, y _gaussfor average is zero, and there is the random quantity of mutually homoscedastic Gaussian distribution, g (y) and g (y with random quantity y _gauss) be nonlinear function, $g\left(y\right)=-\frac{1}{a}\exp (-a{y}^2/2),$ $g\left(y_{\mathrm{gauss}}\right)=-\frac{1}{a}\exp (-a{y_{\mathrm{gauss}}}^2/2),$ A is constant, E{g (y) } be the average computing of g (y), E{g (y) } be g (y _gauss) average computing.

In the present invention, the source signal of geological data is Super-Gaussian Signals, therefore nonlinear function g (y) can be chosen for

the derivative g ' of g (y) (y) be g ' (y)=y exp (ay ²/ 2),, in formula, a is constant.

Specifically, when random quantity y has Gaussian distribution, negentropy J (y) is zero; The non-Gauss of random quantity y is stronger, and the value of negentropy J (y) is larger.When the non-Gauss of random quantity y reaches maximum, when namely the value of negentropy J (y) is maximum, show that useful signal is completely separated with alternating current disturbance signal.

(c) separation matrix is normalized, obtains optimal separation matrix.

(I) utilizes the new value w of formula acquisition separation matrix w below ₊,

$w_{+}=E\left\{\tilde{X}g\left(y\right)\right\}-E\left\{g^{\&prime;}\left(y\right)\right\}w---\left(4\right)$

In formula (4), w ₊for the new value of separation matrix w, g ' is (y) derivative of g (y),

for average computing, E{g ' is (y) } be g ' average computing (y);

(II) new value w to separation matrix ₊be normalized, obtain optimal separation matrix w ₀, the computing formula of normalized is,

$w_0=\frac{w_{+}}{\left|\right|w_{+}\left|\right|}---\left(5\right)$

In formula (5), ‖ w ₊‖ is the new value w of separation matrix ₊mould.

If comprise a plurality of independent signals in geological data, need to carry out repeatedly normalized.For what guarantee to be normalized at every turn, it is all different independent signals, preferably after the new value of each acquisition separation matrix, first carry out decorrelation processing, and then be normalized, until separation matrix convergence, the separation matrix of final convergence is exactly optimal separation matrix.

In one example, the initial value of known separation matrix is w _p-1, the number that P is independent signal, p≤2, can obtain the initial value w of separation matrix according to formula (4) _p-1new value w _pfor,

$w_P=E\left\{\tilde{X}g\left(w_{P-1}^T\tilde{X}\right)\right\}-E\left\{g^{\&prime;}\left(w_{P-1}^T\tilde{X}\right)\right\}{w}_{P-1}---\left(6\right)$

Initial value w to the separation matrix in formula (6) _p-1new value w _pcarry out decorrelation processing, obtain the separation matrix w after decorrelation is processed _pfor,

$w_p=w_p-\underset{j=1}{\overset{p-1}{\mathrm{\&Sigma;}}}{w}_p^T{w}_j{w}_j---\left(7\right)$

In formula (7), w _jbe j separation matrix, j=1,2 ..., p-1, and j is integer.

Separation matrix w after decorrelation in formula (7) is processed _pbe normalized for,

$w_p=\frac{w_p}{\left|\right|w_p\left|\right|}---\left(8\right)$

Judgement separation matrix w _pwhether restrain, as separation matrix w _prestrain computing and finish, as separation matrix w _pdo not restrain, make p=p+1, the computing of recurring formula (6)～formula (8), until separation matrix w _p+1till convergence, the separation matrix w now restraining _p+1be optimal separation matrix.

More particularly, the present invention includes two independent signals (useful signal exchanges undesired signal with), namely the number p=2 of independent signal, when definite optimal separation matrix, need carry out twice-normalization processing.For example, can first given p=1, now the initial value w of separation matrix _p-1known, can utilize formula (7) to calculate the initial value w of the separation matrix after decorrelation is processed _p-1new value w _p, separation matrix w namely ₁; And then make p=p+1, now separation matrix w _pknown, w namely ₁known, can utilize formula (7) to calculate the separation matrix w after decorrelation is processed _pnew value w _p+1, separation matrix w namely ₂, by the separation matrix w calculating for twice ₁with separation matrix w ₂carry out dot product, dot product result converges on 1 or-1, thinks separation matrix w ₂convergence, now the separation matrix w of convergence ₂be optimal separation matrix.

According to the decomposition model of the embodiment of the present invention, be,

s ₂=wX （9）

Wherein, X is measurement vector, and w is optimal separation matrix, s ₂for alternating current disturbance signal.

The optimal separation matrix calculating is brought in formula (9), obtains alternating current disturbance signal s ₂.

In step 104, from the measurement vector of time domain, deduct alternating current disturbance signal, obtain useful signal.

In measurement vector, only have two kinds of signals, a kind of is alternating current disturbance signal, and a kind of is useful signal, so measurement vector is exactly useful signal s ₁with alternating current disturbance signal s ₂linear sums.By obtaining in step 103 after alternating current disturbance signal, alternating current disturbance signal is brought in the computing formula (2) of measurement vector, from measurement vector, deduct alternating current disturbance signal, just can obtain useful signal.

According to the method for alternating current disturbance signal in removal geological data of the present invention, within the scope of setpoint frequency, choose the geological data composition instantaneous mixing model that surpasses default amplitude, and by determining optimal separation matrix, calculate alternating current disturbance signal, thereby the alternating current disturbance signal in geological data is separated with useful signal, reach the effect of removing alternating current disturbance signal from geological data.

Fig. 2 a and Fig. 2 b illustrate respectively and adopt method of the present invention to process single shot record figure before and afterwards.

As shown in Figure 2, horizontal ordinate is Taoist monastic name, and ordinate is the time.Image in Fig. 2 a picture frame is alternating current disturbance signal, by Fig. 2 a, can find out that single big gun exists obvious alternating current disturbance signal before adopting method of the present invention to process.As shown in Figure 2 b, after adopting method of the present invention to process, the alternating current disturbance signal in single shot record has obtained effective removal.

Fig. 3 a and Fig. 3 b illustrate respectively and adopt method of the present invention to process single big gun spectrogram before and afterwards.

As shown in Figure 3, horizontal ordinate represents frequency, and ordinate represents amplitude.If the image in Fig. 3 a frame is alternating current disturbance signal, by Fig. 3 a, can find out that adopting method of the present invention to process front single big gun frequency spectrum exists obvious spike pulse.As shown in Figure 3 b, after adopting method of the present invention to process, the spike pulse in figure in single big gun frequency spectrum has obtained compacting, illustrates that the alternating current disturbance signal in single shot record has obtained effective decay.

According to the method for alternating current disturbance signal in removal geological data of the present invention, in the frequency range of setting, there is alternating current disturbance signal in the geological data that has first judged which road, and then to existing the geological data of alternating current disturbance signal to remove the processing of alternating current disturbance signal, need not all remove to every track data the processing of alternating current disturbance signal, therefore a large amount of operation time can be saved, the needs of actual production can be met.

Method of the present invention is applied to, in the alternating current disturbance signal removal of seismic data, can effectively remove alternating current disturbance signal, fidelity is effective, and treatment effeciency is high simultaneously, and travelling speed is fast, can meet the demand of actual production.Adopt method of the present invention in seismic data treatment process, to bring into play further positive role, application prospect is good.

Although specifically shown with reference to its exemplary embodiment and described the present invention, but it should be appreciated by those skilled in the art, in the situation that do not depart from the spirit and scope of the present invention that claim limits, can carry out the various changes in form and details to it.

Claims (7)

1. remove a method for alternating current disturbance signal in geological data, described method comprises:

(1) obtain at least one road geological data, per pass geological data is carried out to Fourier transform, obtain at least one frequency field geological data;

(2) in the frequency range of setting, from per pass frequency field geological data, choose the frequency field geological data of a peak swing value, the amplitude of the frequency field geological data of choosing and default amplitude are compared, the frequency field geological data that amplitude is more than or equal to default amplitude forms new frequency field geological data, new frequency field geological data is carried out to Fourier inversion, obtain the measurement vector of time domain;

(3) determine optimal separation matrix, utilize decomposition model to calculate alternating current disturbance signal;

(4) from the measurement vector of time domain, deduct alternating current disturbance signal, obtain useful signal.

2. the method for claim 1, wherein the frequency range of described setting is 48～52Hz.

3. the method for claim 1, wherein the step of definite optimal separation matrix is,

(a) measurement vector of time domain is carried out to pre-service;

(b) utilize pretreated measurement vector, obtain separation matrix;

(c) separation matrix is normalized, obtains optimal separation matrix.

4. method as claimed in claim 3, wherein, carries out changing and albefaction centered by pretreated method to measurement vector.

5. method as claimed in claim 3, wherein, step (b) comprising: the separation matrix w while utilizing formula below to obtain J (y) for maximal value,

J(y)∝[E{g(y)}-E{g(y _gauss)}] ²

Wherein, y is random quantity, w is separation matrix, and T represents transposition,

for pretreated measurement vector, J (y) represents the negentropy of random quantity y, y _gaussfor average is zero, and there is the random quantity of mutually homoscedastic Gaussian distribution, g (y) and g (y with random quantity y _gauss) be nonlinear function, a is constant, E{g (y) } be the average computing of g (y), E{g (y) } be g (y _gauss) average computing.

6. method as claimed in claim 3, wherein, step (c) comprising:

(I) utilizes the new value w of formula acquisition separation matrix w below ₊,

Wherein, w ₊for the new value of separation matrix w, g (y) is nonlinear function,

g ' is (y) derivative of g (y),

for

average computing, E{g ' is (y) } be g ' average computing (y);

(II) new value w to separation matrix ₊be normalized, obtain optimal separation matrix w ₀, the computing formula of normalized is,

Wherein, ‖ w ₊‖ is the new value w of separation matrix ₊mould.

7. the method for claim 1, wherein described decomposition model is,

s ₂=w ₀X，

Wherein, X is measurement vector, w ₀for optimal separation matrix, s ₂for alternating current disturbance signal.

Images