CN101266299A - Method for forecasting oil gas utilizing earthquake data object constructional features - Google Patents

Abstract

The invention provides a method for forecasting the oilgas using the earthquake data body structure feature, comprising: extracting the earthquake amplitude data sequence; establishing gray sequence forecasting model; computing the amplitude data sequence predicted value of the earthquake data; calculating the error between the predicted value and the measured value; determining the earthquake amplitude data gray abnormal value segment based on the error; the original earthquake amplitude data and each earthquake amplitude data gray abnormal value segment being subjected to dimensionless treatment to respectively obtain the parent sequence and each subsequence; computing the correlation coefficient between the parent sequence and each subsequence and then computing the association degrees betweent the parent sequence and each subsequence based on the correlation coefficient; obtaining the association sequence by sorting the association degrees by size and determining the oilgas layer based on the association sequence. The method does not have too many condition limits to the data and is suitable for each stage of the oilgas field from exploring to developing with wide application field and reliable precasted result.

Description

Utilize the method for earthquake data object constructional features predicting oil

Technical field

The present invention relates to a kind of petroleum-gas prediction method, relate in particular to a kind of method of utilizing earthquake data object constructional features to carry out predicting oil.

Background technology

The earthquake information amount includes oil gas information, the gross information content that is recorded on the seismic line is very big, at least comprising this two aspects content of oil gas information and non-oil gas information, not only variation has taken place in seismologic parameter when seismic event passes hydrocarbon zone, and different earthquake data object constructionals appearred, oily sandstone reservoir rerum natura is different with the country rock rerum natura, and the difference of fluid properties, not only can be so that the variation of primary seismic wave (P ripple) seismologic parameter when passing this hydrocarbon zone and different seismic facies (traditional Forecasting Methodology occurs, and different earthquake data object constructionals can occur the numerical value method of difference).Utilization variance information principle demonstration numerical difference between (being the earthquake Transformation Principle) predicting oil is just showed the part in this gross information content, utilizing the earthquake data object constructional features predicting oil then is another part, because earthquake data object constructional is comparatively stable, so it can be used for predicting oil preferably theoretically.

Utilize seismic data to bore preceding petroleum-gas prediction, since the forties in last century, use seismic data and bore in the research of preceding prediction of oil-gas reserve, once occur many technology of attempting directly to show oil gas such as " bright spot ", " dim spot ", AVO, pattern-recognition and neural network in succession.Go through the half a century practical proof, it is different that people recognize that gradually the geologic trap condition of depositing is composed in hydrocarbon-bearing pool, causes different classes of hydrocarbon-bearing pool (field) to present the difference of himself characteristic; Therefore, the effect difference of existing forming technique is very huge.Thereby impel scientific and technological circle to be that understanding is existing and carry out further investigation, in the hope of drilling success ratio so as to improving effectively with limitation and influence factor in earthquake reflected wave data demonstration oil gas method and the technology.About earthquake data object constructional features, be meant that each seismic trace discrete data point is arranged shown waveform character in chronological order in the seismic data volume; Utilize the method for earthquake data object constructional features predicting oil, promptly by extracting amplitude value or other property parameters of each seismic trace, study the relation of arrangement, assemblage characteristic and the oil-gas possibility of its data point, reach at last on the section and plane exactly, quantification dopes the purpose of hydrocarbon zone.Generally, traditional petroleum-gas prediction method is divided into vertical prediction and lateral prediction method, but because there is inconsistent phenomenon in two kinds of Forecasting Methodologies, causes for a long time vertically going up with petroleum-gas prediction transversely being discord, and makes predictablity rate not high.

Summary of the invention

The present invention is directed to the above-mentioned technical matters that prior art exists, a kind of method of utilizing the earthquake data object constructional features predicting oil is provided.

The method of earthquake data object constructional features predicting oil of utilizing of the present invention may further comprise the steps: extract the seismic amplitude data sequence; Set up grey ordered series of numbers forecast model; Find the solution the amplitude data sequence prediction value of geological data; Calculate the error between above-mentioned predicted value and the measured value; Determine earthquake amplitude data grey exceptional value according to above-mentioned error; Original seismic amplitude data and each earthquake amplitude data grey exceptional value are carried out immeasurable firmization processing obtain auxiliary sequence and each subsequence respectively; Find the solution correlation coefficient between above-mentioned auxiliary sequence and each subsequence according to above-mentioned correlation coefficient, find the solution the degree of association between above-mentioned auxiliary sequence and each subsequence; The above-mentioned degree of association sorted by size obtain related preface; And determine hydrocarbon zone according to this association preface.

The described grey ordered series of numbers forecast model of setting up comprises, select arbitrary subnumber row from the amplitude data row of original earthquake data, these subnumber row are made one-accumulate generate, the function of time that the subsequence that utilizes one-accumulate to generate is represented according to following formula is set up gray model GM (1,1)

$\frac{{\mathrm{dX}}^{\left(1\right)}}{\mathrm{dt}}+a{X}^{\left(1\right)}=u$

Wherein, a is a parameter to be identified, and u is an endogenous variable to be identified.

Described amplitude data sequence prediction value of trying to achieve geological data comprises: set grey parameter

For $\hat{a}=\left[\begin{array}{c}a\\ u\end{array}\right],$ Obtain a and u with least square method according to following formula, $\hat{a}={\left(B^{T}B\right)}^{-1}{B}^T{Y}_N,$ Wherein B is the matrix that adds up, Y _NBe constant vector; With grey parameter a, the described function of time of u substitution, and differentiate reduction obtains the model value sequence, is shown below:

${\hat{X}}_{(t+1)}^{\left(0\right)}=-a(X_{\left(1\right)}^{\left(0\right)}-\frac{u}{a})e^{-\mathrm{at}};$

Utilize above-mentioned model value sequence through the tired predicted value sequence that obtains original ordered series of numbers that subtracts.

Described matrix B and the constant vector Y of adding up _NBe respectively:

$B=\left[\begin{array}{cc}-\frac{1}{2}\{X_{\left(1\right)}^{\left(1\right)}+X_{\left(2\right)}^{\left(1\right)})& 1\\ -\frac{1}{2}\{X_{\left(2\right)}^{\left(1\right)}+X_{\left(3\right)}^{\left(1\right)})& 1\\ .......& \\ -\frac{1}{2}\{X_{(N-1)}^{\left(1\right)}+X_{\left(N\right)}^{\left(1\right)})& 1\end{array}\right];$

$Y_N=\{X_{\left(2\right)}^{\left(0\right)},X_{\left(3\right)}^{\left(0\right)},...,X_{\left(N\right)}^{\left(0\right)}\}.$

Error between described predicted value and the measured value comprises absolute error and relative error, according to $e_{\left(t\right)}^{\left(0\right)}=X_{\left(t\right)}^{\left(0\right)}-{\hat{X}}_{\left(t\right)}^{\left(0\right)},$ $q^{\left(t\right)}=\frac{e_{\left(t\right)}^{\left(0\right)}}{X_{\left(t\right)}^{\left(0\right)}}$ Calculate, wherein

Be predicted value and X _(t) ⁽⁰⁾Be measured value.

Described definite earthquake amplitude data grey exceptional value comprises the one piece of data sequence that has absolute error and relative error specified data structure and the inconsistent corresponding earthquake the preceding paragraph of other interval stratum by predicted value and measured value.

The method of described immeasurable firmization processing is that first value is handled with equalization or interval relative value.

All data that described value just is treated in the data sequence are all removed the ordered series of numbers that obtains by the 1st data.

Described equalization is treated to the ordered series of numbers that all data in the data sequence obtain divided by the mean value of this data sequence.

Described interval relative value is treated to minimum value that each numerical value with data sequence deducts whole data sequence poor divided by data sequence maximal value and minimum value then, and the formula of embodying is $y\left(k\right)=\frac{x\left(k\right)-x\left(\min \right)}{x\left(\max \right)-x\left(\min \right)},$ Wherein, x (k) is k data in the original data sequence, and y (k) is k data of the data sequence of nondimensionalization after handling, x (max), and x (min) represents the maximal value and the minimum value of data in the original data sequence respectively.

Find the solution correlation coefficient between auxiliary sequence and each subsequence according to following formula,

${\mathrm{\γ}}_j\left(k\right)=\frac{\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-x_j\left(k\right)|+\mathrm{\ξ}\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-x_j\left(k\right)|}{|x_0\left(k\right)-x_j\left(k\right)|+\mathrm{\ξ}\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-x_j\left(k\right)|}$

Wherein, auxiliary sequence is designated as x ₀={ x ₀(1), x ₀(2) ..., x ₀(n) }; Subsequence is x _j=x _j(1), x _j(2) ..., x _j(n) }, j=1,2 ..., m; x _j(k) be subsequence x _jK element; x ₀(k) be auxiliary sequence x ₀And x _j(k) elements corresponding; ξ is a resolution ratio; $\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ Be second level lowest difference, $\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-X_j\left(k\right)|$ Be two-stage maximum difference, | x ₀(k)-x _j(k) |=Δ _j(k) be the absolute difference of k point auxiliary sequence amplitude and subsequence amplitude.

Find the solution the degree of association of separating the correlation coefficient between auxiliary sequence and each subsequence according to following formula, $\stackrel{\‾}{r_j}=\frac{1}{n}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{r}_j\left(k\right)$ Wherein, r _jBe the degree of association of subsequence to auxiliary sequence.

Determine hydrocarbon zone according to this association preface, relative and absolute error according to prediction all is maximum exceptional value section, and with the degree of association of the target zone of drilling well be sample, determine hydrocarbon zone according to this association preface, the interval of coupling is the oily interval of prediction with it.

Beneficial effect of the present invention is:

(1) the present invention is applied in single track and multiple tracks earthquake data object constructional features in the production practices, realized that vertically and laterally can quantize demarcation and chart simultaneously contrasts mutually, the new technology of quantitative interpretation hydrocarbon zone has remedied the not high deficiency of some petroleum-gas prediction method predictablity rates of tradition.Wherein, the earthquake data object constructional features of single track geological data row is meant that each seismic trace discrete data point arranges shown single track waveform character in chronological order.The earthquake data object constructional features of multiple tracks seismic data volume be meant and the road between discrete data point arrange the architectural feature at shown numerous waveform consecutive numbers strong point in chronological order.

(2) petroleum-gas prediction of the present invention is mainly on the architectural feature based on the different permutation and combination of data point, on section, the position that the architectural feature of the numerous seismic channel data points of foundation can be predicted oil-gas possibility comparatively exactly, in the plane, can iris out the hydrocarbon zone distribution range comparatively exactly by vector association analysis between road and the road, thereby realized indulging and the horizontal novel petroleum-gas prediction technological means that can link and contrast, overcome the inconsistent not high difficulties of predictablity rate of bringing such as vertical prediction and horizontal supposition technical method.

(3) the present invention is aspect the expression of petroleum-gas prediction achievement, adopt figure and table contrast and dimensionless black and white to demarcate, make to predict the outcome more simply, clear, help doping exactly the distribution range and the feature of HYDROCARBON-BEARING REGION, reduce artificial factor, reduce the uncertainty of oil and gas detection.

(4) the required data of Forecasting Methodology of the present invention only is conventional geological data and a small amount of log data, all can use at oil field prospecting and each stage of exploitation, even also can carry out petroleum-gas prediction in the newly developed area of few well.Because forecast model is a continuous type, rather than discrete type, so can more long-range, dynamic prediction continuously.

(5) Forecasting Methodology of the present invention utilizes the seismic channel data sequence to judge oil gas, reduces the restriction of seismic interpretation achievement, need not the situation of worry about tomography and layer position.Only need to extract the amplitude data sequence of original earthquake data body target zone, and, can carry out the prediction of oil gas in conjunction with the situation of well-log information.This will reduce the workload of seismic interpretation in conventional oil field prospecting, the performance history greatly, also can reduce simultaneously in the seismic interpretation process explanation personnel to the interpretation results artificial factor, so achievement is more objective.

(6) the earthquake data object constructional features predicting oil theory that proposes of the present invention has changed in the past in the petroleum-gas prediction method in order to the earthquake value data being the situation of leading, the waveform of geological data and numerical value combined carry out petroleum-gas prediction, make that seismic data is omnibearing to have obtained application.

Description of drawings

Accompanying drawing described herein is used to provide further understanding of the present invention, constitutes a part of the present invention, does not constitute limitation of the invention.In the accompanying drawings:

Fig. 1 is the process flow diagram that utilizes the method for earthquake data object constructional features predicting oil of the present invention.

Embodiment

For the purpose, technical scheme and the advantage that make the embodiment of the invention is clearer,, the embodiment of the invention is described in further details below in conjunction with embodiment and accompanying drawing.At this, illustrative examples of the present invention and explanation thereof are used to explain the present invention, but not as a limitation of the invention.

The invention provides a kind of method of utilizing the earthquake data object constructional features predicting oil.The present invention is described in detail below in conjunction with Fig. 1.

(1) extracts the seismic amplitude data sequence, set up grey ordered series of numbers forecast model and definite grey exceptional value

At first,, set up grey ordered series of numbers forecast model (GM model) according to the geological data sequence of extracting, gray model be expressed as GM (1, n), 1 expression single order, n represents the dimension of variable, adopts GM (1,1) model usually, comprising:

The first step from the amplitude data of original earthquake data, selects arbitrary seismic amplitude data rows to be listed as subnumber.For example, the amplitude data row of the original earthquake data of formula (1) expression original seismic data:

$X\left(0\right)=\{X_{\left(1\right)}^{\left(0\right)},X_{\left(2\right)}^{\left(0\right)},X_{\left(3\right)}^{\left(0\right)},...{,X}_{\left(N\right)}^{\left(0\right)}\}---\left(1\right)$

Wherein, subscript is represented the variable number, and subscript is represented number of operations.

From the amplitude data row of the original earthquake data of formula (1) expression, select arbitrary subnumber row, as the formula (2):

$X^{\left(0\right)}=\{X_{\left(2\right)}^{\left(0\right)},X_{\left(3\right)}^{\left(0\right)},X_{\left(4\right)}^{\left(0\right)},...,X_{\left(N\right)}^{\left(0\right)}\}---\left(2\right)$

Second step, these subnumber row to be made one-accumulate generate, the concrete operation operation that one-accumulate generates is $X_{\left(k\right)}^{\left(0\right)}+X_{(k-1)}^{\left(1\right)}=X_{\left(k\right)}^{\left(1\right)},$ Obtain

$X^{\left(1\right)}=\{X_{\left(2\right)}^{\left(1\right)},X_{\left(3\right)}^{\left(1\right)},X_{\left(4\right)}^{\left(1\right)},...,X_{\left(N\right)}^{\left(1\right)}\}---\left(3\right)$

In the 3rd step, the sequence of utilizing one-accumulate to generate is set up gray model GM (1,1), as the formula (4):

$\frac{{\mathrm{dX}}^{\left(1\right)}}{\mathrm{dt}}+{\mathrm{aX}}^{\left(1\right)}=u---\left(4\right)$

Wherein: a is development coefficient (parameter to be identified), reflects original ordered series of numbers and the developing state of the ordered series of numbers that adds up; U is grey action (endogenous variable to be identified), and generally speaking, the systemic effect amount should be external or predetermination, and GM (1,1) is single-row modeling, only uses the behavior sequence of system, and not outer effect sequence.Grey action among the GM (1,1) is the data of excavating out from background value, the relation of its reflection data variation, and its definite intension is grey.The grey action is the imbody of intension extensionization, and its existence is the watershed divide of difference grey modeling and general modeling, also is the important symbol of distinguishing gray system viewpoint and ash bin viewpoint.

In the 4th step, set grey parameter For $\hat{a}=\left[\begin{array}{c}a\\ u\end{array}\right],$ Can obtain a and u with least square method, then:

$\hat{a}={\left(B^{T}B\right)}^{-1}{B}^T{Y}_N---\left(5\right)$

Wherein B is the matrix that adds up, Y _NBeing constant vector, is the derivation result that obtains by the differential equation of separating front albefaction model, and respectively suc as formula shown in (6) and (7),

$B=\left[\begin{array}{cc}-\frac{1}{2}(X_{\left(1\right)}^{\left(1\right)}+X_{\left(1\right)}^{\left(2\right)})& 1\\ -\frac{1}{2}(X_{\left(1\right)}^{\left(1\right)}+X_{\left(1\right)}^{\left(2\right)})& 1\\ .......& \\ -\frac{1}{2}(X_{\left(1\right)}^{\left(1\right)}+X_{\left(1\right)}^{\left(2\right)})& 1\end{array}\right]---\left(6\right)$

$Y_N=\{X_{\left(0\right)}^{\left(1\right)},X_{\left(0\right)}^{\left(2\right)},...,X_{\left(0\right)}^{\left(n\right)}\}---\left(7\right)$

Find the solution the predicted value of the amplitude data sequence of geological data according to the grey ordered series of numbers forecast model of setting up above, comprising:

The first step by the function of time with the expression of grey parameter a, u substitution formula (4), and obtains the model value sequence to its differentiate reduction, as the formula (8):

${\hat{X}}_{(t+1)}^{\left(0\right)}=-a(X_{\left(1\right)}^{\left(0\right)}-\frac{u}{a})e^{-\mathrm{at}}---\left(8\right)$

Second step, the model value sequence of utilizing above-mentioned model to predict to obtain, and subtract the predicted value sequence that obtains original ordered series of numbers through tired, as the formula (9):

${\hat{X}}^{\left(0\right)}=\{{\hat{X}}_{\left(2\right)}^{\left(0\right)},{\hat{X}}_{\left(3\right)}^{\left(0\right)},...,{\hat{X}}_{\left(N\right)}^{\left(0\right)}\}---\left(9\right)$

Then, utilize predicted value

With measured value X _(t) ⁽⁰⁾Between absolute error and existing to error, calculate seismic amplitude data gray exceptional value, promptly have the one piece of data sequence on absolute error and relative error specified data structure and the inconsistent corresponding earthquake the preceding paragraph of other interval stratum by predicted value and measured value.

Wherein, X _(t) ⁽⁰⁾With

Absolute error and relative error calculate by formula (10):

$e_{\left(t\right)}^{\left(0\right)}=X_{\left(t\right)}^{\left(0\right)}-{\hat{X}}_{\left(t\right)}^{\left(0\right)},$ $q^{\left(t\right)}=\frac{e_{\left(t\right)}^{\left(0\right)}}{X_{\left(t\right)}^{\left(0\right)}};---\left(10\right)$

The predicted value of ash rate response gained can reflect the distortion that seismic wave waveform is produced owing to the contained fluid of reservoir is different with the absolute error between the measured value, and the relative error between them then embodies the degree that deviation takes place.Therefore absolute error and relative error are analyzed, can be made the prediction that whether contains oil gas in the stratum.Actual prediction can be made according to seeking the little layer position section of relative error.

(2) association analysis

The grey exceptional value of earthquake data object constructional features is carried out the association analysis of five steps, and oil-gas-water layer is determined in ordering then.At first, the first step is carried out the nondimensionalization processing to original seismic amplitude data.The method of nondimensionalization, commonly used have first value and equalization, an interval value relatively.Value just is meant that all data all remove with the 1st data, obtains a new ordered series of numbers then, and the value that this new ordered series of numbers promptly is the variant moment is with respect to the number percent of the value in the 1st moment.Equalization is meant the mean value of all data divided by this data sequence.Interval relativization is meant the relative value of all data intervals: the minimum value that each numerical value of data rows deducts whole data rows is poor divided by data rows maximal value and minimum value then, and the formula of embodying is $y\left(k\right)=\frac{x\left(k\right)-x\left(\min \right)}{x\left(\max \right)-x\left(\min \right)}.$ Wherein, x (k) is k data in the original data sequence, and y (k) is k data of the data sequence of nondimensionalization after handling, x (max), and x (min) represents the maximal value and the minimum value of data in the original data sequence respectively.

In second step, ask two differential in the correlation coefficient.A difference in the compute associations coefficient process is that following compute associations coefficient is prepared, and comprising:

| x ₀(k)-x _j(k) |, wherein footmark is represented two ordered series of numbers each vector that neutralizes, the footmark value is got not income value difference simultaneously.Specifically, | x ₀(k)-x _j(k) |=Δ _j(k) be called K point x ₀With x _jAbsolute difference, $\underset{j}{\min }\underset{j}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ Be the two-stage lowest difference, $\underset{k}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ Be first order lowest difference, it is illustrated on the j bar curve and looks for each point and X ₀Lowest difference, $\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ Be second level lowest difference, be illustrated on the basis of the lowest difference of finding out in each bar curve, find out the lowest difference in all curves again. $\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-X_j\left(k\right)|$ Maximum poor for the second level, meaning is identical with second level lowest difference, is illustrated on the basis of the maximum difference of finding out in each bar curve, and the maximum of finding out again in all curves is poor.

In the 3rd step, ask correlation coefficient.Seismic amplitude data after the nondimensionalization processing are made as auxiliary sequence and are designated as x ₀, i.e. x ₀={ x ₀(1), x ₀(2) ..., x ₀(n) }; Subsequence is designated as x _jBe x _j={ x _j(2), (2) ..., x _j(n) }, j=1,2 ..., m.Contact between each ordered series of numbers is called the ash relation.The tightness degree of ash relation can embody with grey correlation coefficient, and its expression formula is:

${\mathrm{\γ}}_j\left(k\right)=\frac{\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-x_j\left(k\right)|+\mathrm{\ξ}\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-x_j\left(k\right)|}{|x_0\left(k\right)-x_j\left(k\right)|+\mathrm{\ξ}\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-x_j\left(k\right)|}---\left(11\right)$

Show subsequence x _jK element x _j(k) with auxiliary sequence x ₀Respective element x ₀(k) relative difference, ξ are resolution ratio.

In the 4th step, ask the degree of association.Because the number of correlation coefficient is more, information is not concentrated, and is not easy to comparison.For this reason, the correlation coefficient under each element is averaged be r _jWith r _jBe defined as the degree of association of subsequence, that is: to auxiliary sequence

$\stackrel{\‾}{r_j}=\frac{1}{n}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{r}_j\left(k\right)---\left(12\right)$

Article two, the shape of curve is similar more each other, and the degree of association is just big more, otherwise then the degree of association is more little.Key wherein is that grey incidence matrix is analyzed, and finds out wherein active factor.

In the 5th step, discharge related preface.When reference sequence more than one, when being compared factor, just can carry out benefit analysis more than one.Claim that below reference sequence is female ordered series of numbers (or female factor), comparand is classified subnumber row (sub-factor) as, and it is old to constitute related square by female ordered series of numbers (or female factor) and subnumber row (sub-factor).

(3) predicting oil

The related preface of the earthquake data object constructional features grey exceptional value that preceding two steps are obtained is analyzed, whether the related preface of the situation analysis grey exceptional value of the earthquake data object constructional features model when contrast contains oil gas is relevant with oil gas, the oil-gas possibility of predicting reservoir.By the relation between old each element of related square, can analyze which factor is advantage, and which factor is not an advantage, and the degree of association to resulting each interval sorts by size to obtain a related preface then.The relation of related preface and oil gas is to determine according to the concrete well data of each department, with the degree of association of the target zone of drilling well is sample, the interval that mates most is the oily interval of prediction with it, or the relative and absolute error of prediction all is that one section maximum stratum is hydrocarbon-bearing formation.

Above-described specific embodiment; purpose of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the above only is specific embodiments of the invention; and be not intended to limit the scope of the invention; within the spirit and principles in the present invention all, any modification of being made, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1, a kind of method of utilizing the earthquake data object constructional features predicting oil, this method comprises the following steps:

A) extract seismic amplitude data sequence and set up grey ordered series of numbers forecast model;

B) find the solution the amplitude data sequence prediction value of geological data;

C) error between above-mentioned predicted value of calculating and the measured value;

D) determine earthquake amplitude data grey exceptional value section according to above-mentioned error;

E) original seismic amplitude data and each earthquake amplitude data grey exceptional value are carried out immeasurable firmization processing and obtain auxiliary sequence and each subsequence respectively;

F) find the solution correlation coefficient between above-mentioned auxiliary sequence and each subsequence;

G), find the solution the degree of association between above-mentioned auxiliary sequence and each subsequence according to above-mentioned correlation coefficient;

H) the above-mentioned degree of association is sorted by size obtain related preface;

I) and according to above-mentioned related preface determine hydrocarbon zone.

2, method according to claim 1, it is characterized in that: the described grey ordered series of numbers forecast model of setting up comprises, from the amplitude data row of original earthquake data, select arbitrary subnumber row, these subnumber row are made one-accumulate to be generated, the function of time that the subsequence that utilizes one-accumulate to generate is represented according to following formula is set up gray model GM (1,1)

$\frac{{\mathrm{dX}}^{\left(1\right)}}{\mathrm{dt}}+{\mathrm{aX}}^{\left(1\right)}=u$

Wherein, a is a parameter to be identified, and u is an endogenous variable to be identified, and all available least square method is tried to achieve.

3, method according to claim 2 is characterized in that, described amplitude data sequence prediction value of trying to achieve geological data comprises:

A) set grey parameter

For $\hat{a}=\left[\begin{array}{c}a\\ u\end{array}\right],$ Obtain a and u with least square method according to following formula, $\hat{a}={\left(B^{T}B\right)}^{-1}{B}^T{Y}_N,$ Wherein, B is the matrix that adds up, B ^TBe the transposed matrix of B, Y _NBe constant vector;

B) with grey parameter a, the described function of time of u substitution, and differentiate reduction obtains the model value sequence, is shown below:

${\hat{X}}_{(t+1)}^{\left(0\right)}=-a(X_{\left(1\right)}^{\left(0\right)}-\frac{u}{a})e^{-\mathrm{at}}$

C) utilize above-mentioned model value sequence through the tired predicted value sequence that obtains original ordered series of numbers that subtracts.

4, method according to claim 3 is characterized in that, described matrix B and the constant vector Y of adding up _NBe respectively:

$B=\left[\begin{array}{cc}-\frac{1}{2}\{X_{\left(1\right)}^{\left(1\right)}+X_{\left(2\right)}^{\left(1\right)})& 1\\ -\frac{1}{2}\{X_{\left(2\right)}^{\left(1\right)}+X_{\left(3\right)}^{\left(1\right)})& 1\\ .......& \\ -\frac{1}{2}\{X_{(N-1)}^{\left(1\right)}+X_{\left(N\right)}^{\left(1\right)})& 1\end{array}\right];$

$Y_N=\{X_{\left(2\right)}^{\left(0\right)},X_{\left(3\right)}^{\left(0\right)},...,X_{\left(N\right)}^{\left(0\right)}\}.$

5, method according to claim 1 is characterized in that: the error between described predicted value and the measured value comprises absolute error and relative error, according to $e_{\left(t\right)}^{\left(1\right)}=X_{\left(t\right)}^{\left(0\right)}-{\hat{X}}_{\left(t\right)}^{\left(0\right)},$ $q^{\left(t\right)}=\frac{e_{\left(t\right)}^{\left(0\right)}}{X_{\left(t\right)}^{\left(0\right)}}$ Calculate, wherein

Be predicted value and X _(t) ⁽⁰⁾Be measured value, according to above-mentioned forecast model, determine earthquake amplitude data grey exceptional value section,, then set up master mould is made the residual error model of correction till reaching requirement if the forecasting sequence that the gray model that raw data is set up obtains is too big through verify error.

6, method according to claim 1 is characterized in that: described definite earthquake amplitude data grey exceptional value comprises the one piece of data sequence that has absolute error and relative error specified data structure and the inconsistent corresponding earthquake the preceding paragraph of other interval stratum by predicted value and measured value.

7, method according to claim 1, it is characterized in that: the method for described immeasurable firmization processing is that first value is handled with equalization or interval relative value, all data that described value just is treated in the data sequence are all removed the ordered series of numbers that obtains by the 1st data, described equalization is treated to the ordered series of numbers that all data in the data sequence obtain divided by the mean value of this data sequence, described interval relative value is treated to minimum value that each numerical value with data sequence deducts whole data sequence poor divided by data sequence maximal value and minimum value then, and the formula of embodying is $y\left(k\right)=\frac{x\left(k\right)-x\left(\min \right)}{x\left(\max \right)-x\left(\min \right)},$ Wherein, x (k) is k data in the original data sequence, and y (k) is k data of the data sequence of nondimensionalization after handling, x (max), and x (min) represents the maximal value and the minimum value of data in the original data sequence respectively.

8, method according to claim 1 is characterized in that, if reference sequence is X ₀, the ordered series of numbers that is compared (factor ordered series of numbers) is X _j, j=1,2 ..., n, then curve X ₀With X _jRepresent by following formula at the correlation coefficient that k is ordered:

${\mathrm{\γ}}_j\left(k\right)=\frac{\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-x_j\left(k\right)|+\mathrm{\ξ}\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-x_j\left(k\right)|}{|x_0\left(k\right)-x_j\left(k\right)|+\mathrm{\ξ}\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-x_j\left(k\right)|}$

Wherein, reference sequence is designated as x ₀={ x ₀(1), x ₀(2) ..., x ₀(n) }; The ordered series of numbers that is compared (factor ordered series of numbers) is x _j={ x _j(1), x _j(2) ..., x _j(n) }, j=1,2 ..., m; x _j(k) be subsequence x _jK element; x ₀(k) be reference sequence x ₀With the ordered series of numbers x that is compared _j(k) elements corresponding; ξ is a resolution ratio, and span generally gets 0.5 between 0 and 1; $\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ Be second level lowest difference, $\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-X_j\left(k\right)|$ Be two-stage maximum difference, | x ₀(k)-X _j(k) |=Δ _j(k) be the absolute difference of k point auxiliary sequence amplitude and subsequence amplitude.

9, method according to claim 1 is characterized in that, just obtains representing whole X according to following formula _jCurve and parametric line X ₀The degree of association of the correlation coefficient between the sequence:

$\stackrel{\‾}{r_j}=\frac{1}{n}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}{r}_j\left(k\right)$

Wherein, r _jBe the degree of association of subsequence to auxiliary sequence.

10, method according to claim 1, it is characterized in that: the relative and absolute error according to prediction all is maximum exceptional value section, and with the degree of association of the target zone of drilling well be sample, determine hydrocarbon zone according to this association preface, the interval of coupling is the oily interval of prediction with it.

Images