CN101266299B - 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

Method for predicting oil and gas by using seismic data volume structural features

Technical Field

The invention relates to an oil-gas prediction method, in particular to a method for predicting oil gas by using the structural characteristics of a seismic data volume.

Background

The seismic information quantity comprises oil and gas information, the total information quantity recorded on a seismic survey line is very large, and at least comprises two contents of oil and gas information and non-oil and gas information, when seismic waves pass through an oil-gas layer, not only are the seismic parameters changed, but also different seismic data body structures appear, the physical properties of an oil-gas sandstone reservoir layer and surrounding rocks are different, and the properties of fluids are different, so that not only can different seismic phases appear due to the change of the seismic parameters when primary seismic waves (P waves) pass through the oil-gas layer (namely a traditional prediction method and a numerical difference method), but also different seismic data body structures can appear. The method for predicting the oil gas by utilizing the difference information principle to display the numerical difference (namely the earthquake phase change principle) only displays one part of the total information quantity, and the method for predicting the oil gas by utilizing the structural characteristics of the earthquake data volume displays the other part of the total information quantity.

Since the last 40 th century, in the research of using seismic data to predict pre-drill oil and gas reservoirs, many technologies attempting to directly display oil and gas, such as 'bright spots', 'dark spots', AVOs, pattern recognition and neural networks, have been developed in succession. Through practice verification of half a century, people gradually realize that the geological trapping conditions existing in oil and gas reservoirs are different, so that the oil and gas reservoirs (fields) in different categories have different characteristics; therefore, the application effect of the existing molding technology is very different. Therefore, the scientific and technological field is promoted to carry out deep research for recognizing the limitations and influencing factors in the existing method and technology for displaying oil and gas by using earthquake reflection wave data so as to effectively improve the drilling success rate. The structural characteristics of the seismic data volume refer to waveform characteristics displayed by arranging discrete data points of each seismic channel in the seismic data volume according to a time sequence; the method for predicting oil and gas by using the structural characteristics of the seismic data volume is characterized in that the relation between the arrangement and combination characteristics of data points and the oil and gas containing property of each seismic channel is researched by extracting the amplitude value or other attribute parameters of each seismic channel, and finally the purpose of accurately and quantitatively predicting oil and gas reservoirs on a section and a plane is achieved. Generally, the traditional oil and gas prediction method is divided into a longitudinal prediction method and a transverse prediction method, but the phenomenon of inconsistency exists in the two prediction methods, so that the longitudinal oil and gas prediction and the transverse oil and gas prediction are not harmonious for a long time, and the prediction accuracy is low.

Disclosure of Invention

Aiming at the technical problems in the prior art, the invention provides a method for predicting oil and gas by using the structural characteristics of a seismic data volume.

The method for predicting oil and gas by using the structural characteristics of the seismic data volume comprises the following steps: extracting a seismic amplitude data sequence; establishing a grey number series prediction model; solving the amplitude data sequence prediction value of the seismic data; calculating the error between the predicted value and the measured value; determining the abnormal gray value of the seismic amplitude data according to the error; carrying out non-quantitative rigidization treatment on the original seismic amplitude data and the gray abnormal values of the seismic amplitude data to respectively obtain a mother sequence and sub sequences; solving the correlation coefficient between the mother sequence and each subsequence, and solving the correlation degree between the mother sequence and each subsequence according to the correlation coefficient; sorting the association degrees according to the magnitude to obtain an association order; and determining the hydrocarbon reservoir based on the correlation order.

The establishing of the grey number series prediction model comprises the steps of selecting any sub number series from amplitude data series of original seismic data, performing primary accumulation generation on the sub number series, establishing a grey model GM (1, 1) according to a time function represented by the following formula by utilizing the sub series generated by the primary accumulation,

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

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

The step of obtaining the amplitude data sequence prediction value of the seismic data comprises the following steps: setting ash parameters

Is composed of $\hat{a}=\left[\begin{array}{c}a\\ u\end{array}\right],$ A and u are found by the least square method according to the following formula, $\hat{a}={\left(B^{T}B\right)}^{-1}{B}^T{Y}_N,$ where B is the accumulation matrix, Y_NIs a constant vector; substituting the gray parameters a and u into the time function, and deriving and reducing to obtain a model value sequence as shown in the following formula:

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

and (4) obtaining a predicted value sequence of the original sequence of numbers by utilizing the model value sequence through accumulation and subtraction.

The accumulation matrix B and the constant vector Y_NRespectively as follows:

<math><mrow><mi>B</mi><mo>=</mo><mfenced open='[' close=']'><mtable><mtr><mtd><mo>-</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><mrow><mo>{</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>+</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>)</mo></mrow></mtd><mtd><mn>1</mn></mtd></mtr><mtr><mtd><mo>-</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><mrow><mo>{</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>+</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>3</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>)</mo></mrow></mtd><mtd><mn>1</mn></mtd></mtr><mtr><mtd><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo></mtd><mtd></mtd></mtr><mtr><mtd><mo>-</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><mrow><mo>{</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mi>N</mi><mo>-</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>+</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mi>N</mi><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>)</mo></mrow></mtd><mtd><mn>1</mn></mtd></mtr></mtable></mfenced><mo>;</mo></mrow></math>

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

the error between the predicted value and the measured value includes an absolute error and a relative error, according to $e_{\left(t\right)}^{\left(0\right)}=X_{\left(t\right)}^{\left(0\right)}-{\hat{X}}_{\left(t\right)}^{\left(0\right)},$ $q^{\left(t\right)}=\frac{e_{\left(t\right)}^{\left(0\right)}}{X_{\left(t\right)}^{\left(0\right)}}$ Is calculated, wherein

Is predicted value and X_(t) ⁽⁰⁾Is an actual measurement value.

The step of determining the seismic amplitude data gray abnormal value comprises the step of determining a data sequence of a section of stratum corresponding to the earthquake, wherein the data sequence is inconsistent with other sections through the absolute error and the relative error of the predicted value and the measured value.

The method of the non-quantitative rigidification processing is initialization and equalization or interval relative value processing.

The initialization processing is a sequence obtained by dividing all data in the data sequence by the 1 st data.

The averaging process is a sequence of all data in the data sequence divided by the average value of this data sequence.

The interval relative value is processed by subtracting the minimum value of the whole data sequence from each value of the data sequence and then dividing the minimum value by the difference between the maximum value and the minimum value of the data sequence, and the specific expression 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 the kth data in the original data sequence, y (k) is the kth data of the data sequence after the non-dimensionalization processing, x (max), and x (min) respectively represent the maximum value and the minimum value of the data in the original data sequence.

Solving the correlation coefficient between the parent sequence and each subsequence according to the following formula,

<math><mrow><msub><mi>&gamma;</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>=</mo><mfrac><mrow><munder><mi>min</mi><mi>j</mi></munder><munder><mi>min</mi><mi>k</mi></munder><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo><mo>+</mo><mi>&xi;</mi><munder><mi>max</mi><mi>j</mi></munder><mrow><munder><mi>max</mi><mi>k</mi></munder><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub></mrow><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo></mrow><mrow><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo><mo>+</mo><mi>&xi;</mi><munder><mi>max</mi><mi>j</mi></munder><munder><mi>max</mi><mi>k</mi></munder><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo></mrow></mfrac></mrow></math>

wherein the parent sequence is represented as x₀＝{x₀(1)，x₀(2)，…，x₀(n) }; subsequence of x_j＝{x_j(1)，x_j(2)，…，x_j(n)}，j＝1，2，…，m；x_j(k) Is a subsequence x_jThe kth element of (1); x is the number of₀(k) Is a mother sequence x₀And x_j(k) The corresponding elements; xi is a resolution coefficient; $\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ is the second-order minimum difference and, $\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-X_j\left(k\right)|$ is the two-step maximum difference, | x₀(k)-X_j(k)|＝Δ_j(k) Is the absolute difference between the amplitude of the k-th parent sequence and the amplitude of the subsequence.

The correlation degree of the correlation coefficient between the solved mother sequence and each subsequence is solved according to the following formula, <math><mrow><mrow><mover><msub><mi>r</mi><mi>j</mi></msub><mo>&OverBar;</mo></mover><mo>=</mo><mfrac><mn>1</mn><mi>n</mi></mfrac><munderover><mi>&Sigma;</mi><mrow><mi>k</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>r</mi><mi>j</mi></msub></mrow><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow></mrow></math> wherein,

the association degree of the subsequence to the parent sequence is shown.

And determining the hydrocarbon reservoir according to the correlation sequence, determining the hydrocarbon reservoir according to the abnormal value section with the maximum predicted relative and absolute errors and the correlation degree with the drilled target layer as a sample, and determining the hydrocarbon reservoir according to the correlation sequence, wherein the most matched hydrocarbon reservoir section is the predicted hydrocarbon reservoir section.

The invention has the beneficial effects that:

(1) the invention applies the structural characteristics of single-channel and multi-channel seismic data volumes to production practice, realizes the simultaneous quantitative calibration and chart comparison of the longitudinal direction and the transverse direction, quantitatively explains the new technology of oil-gas reservoirs, and makes up the defect of low prediction accuracy of the traditional oil-gas prediction methods. The seismic data volume structure characteristics of the single-channel seismic data column refer to single-channel waveform characteristics displayed by arranging discrete data points of each seismic channel according to a time sequence. The seismic data volume structural feature of the multi-channel seismic data volume is the structural feature of a plurality of waveform adjacent data points displayed by arranging discrete data points between channels according to the time sequence.

(2) The oil-gas prediction method is mainly based on the structural characteristics of different arrangement combinations of data points, on a section, the position of oil-gas containing property can be accurately predicted according to the structural characteristics of a plurality of seismic channel data points, and on a plane, the distribution range of an oil-gas layer can be accurately enclosed through the vector correlation analysis between channels, so that a novel oil-gas prediction technical means that longitudinal and transverse can be in linkage comparison is realized, and the difficulty that the prediction accuracy is low due to the inconsistency of longitudinal prediction and transverse prediction technical methods and the like is overcome.

(3) In the aspect of oil and gas prediction result expression, the invention adopts graph-to-table comparison and dimensionless black-and-white calibration, so that the prediction result is simpler and clearer, the distribution range and the characteristics of an oil and gas area can be predicted accurately, the influence of human factors is reduced, and the uncertainty of oil and gas detection is reduced.

(4) The data required by the prediction method of the invention are only conventional seismic data and a small amount of logging data, and the prediction method can be applied to various stages of exploration and development of oil fields, and can also be used for oil and gas prediction even in a new area with few wells. Because the prediction model is continuous, rather than discrete, it can be predicted for long periods of time, continuously and dynamically.

(5) The prediction method of the invention judges oil gas by using the seismic channel data sequence, reduces the limitation of seismic interpretation results, and does not need to consider the conditions of faults and horizons too much. The oil and gas can be predicted only by extracting the amplitude data sequence at the destination layer of the original seismic data body and combining the conditions of logging data. The method can greatly reduce the workload of seismic interpretation in the conventional oil field exploration and development process, and can also reduce the influence of interpreters on human factors for interpreting results in the seismic interpretation process, so the results are more objective.

(6) The seismic data volume structure characteristic prediction oil-gas theory provided by the invention changes the situation that the traditional oil-gas prediction method mainly utilizes the numerical value of the seismic data, combines the waveform and the numerical value of the seismic data together to perform oil-gas prediction, and enables the seismic data to be applied in all directions.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention. In the drawings:

FIG. 1 is a flow chart of a method of the present invention for predicting hydrocarbons using structural features of a seismic data volume.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention are described in further detail below with reference to the embodiments and the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.

The invention provides a method for predicting oil and gas by using structural characteristics of a seismic data volume. The present invention will be described in detail with reference to fig. 1.

(1) Extracting seismic amplitude data sequence, establishing grey number sequence prediction model and determining grey abnormal value

First, from the extracted seismic data sequence, a gray series prediction model (GM model) is built, the gray model being expressed as GM (1, n), 1 representing the first order and n representing the dimension of the variable, typically using the GM (1, 1) model, including:

in the first step, any seismic amplitude data column is selected as a sub-number column from the amplitude data of the original seismic data. For example, equation (1) represents the amplitude data column for the raw seismic data of the raw seismic record:

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

wherein, the subscript represents the number of variables and the superscript represents the number of operations.

Selecting any one of the amplitude data columns of the original seismic data represented by the formula (1), as shown in the formula (2):

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

secondly, the sub-sequence is accumulated once, and the specific operation of the accumulation is $X_{\left(k\right)}^{\left(0\right)}+X_{(k-1)}^{\left(1\right)}=X_{\left(k\right)}^{\left(1\right)},$ To obtain

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

Thirdly, establishing a gray model GM (1, 1) by using a sequence generated by the primary accumulation, as shown in formula (4):

$\frac{d{X}^{\left(1\right)}}{\mathrm{dt}}+a{X}^{\left(1\right)}=u---\left(4\right)$

wherein: a is a development coefficient (parameter to be identified) reflecting the development situation of the original sequence and the accumulated sequence; u is the gray effect (endogenous variable to be identified), and in general, the systemic effect should be exogenous or predetermined, while GM (1, 1) is a single-column modeling, using only the behavioral sequence of the system, and no exogenous effect sequence. The amount of gray contribution in GM (1, 1) is data mined from background values, which reflects the relationship of data changes, the exact connotation of which is gray. The gray effect quantity is embodied by the content extension, and the existence of the gray effect quantity is a watershed for distinguishing gray modeling from general modeling and is an important mark for distinguishing a gray system view from a gray box view.

The fourth step, setting the ash parameter

Is composed of $\hat{a}=\left[\begin{array}{c}a\\ u\end{array}\right],$ A and u can be found by the least squares method, then:

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

where B is the accumulation matrix, Y_NIs a constant vector, is a derivation result obtained by solving the differential equation of the previous whitening model, and is respectively shown in the formulas (6) and (7),

<math><mrow><mrow><mi>B</mi><mo>=</mo><mfenced open='[' close=']'><mtable><mtr><mtd><mo>-</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><mrow><mo>(</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>+</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></msubsup><mo>)</mo></mrow></mtd><mtd><mn>1</mn></mtd></mtr><mtr><mtd><mo>-</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><mrow><mo>(</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>+</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></msubsup><mo>)</mo></mrow></mtd><mtd><mn>1</mn></mtd></mtr><mtr><mtd><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo><mo>&CenterDot;</mo></mtd><mtd></mtd></mtr><mtr><mtd><mo>-</mo><mfrac><mn>1</mn><mn>2</mn></mfrac><mrow><mo>(</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow></msubsup><mo>+</mo><msubsup><mi>X</mi><mrow><mo>(</mo><mn>1</mn><mo>)</mo></mrow><mrow><mo>(</mo><mn>2</mn><mo>)</mo></mrow></msubsup><mo>)</mo></mrow></mtd><mtd><mn>1</mn></mtd></mtr></mtable></mfenced><mo>-</mo><mo>-</mo><mo>-</mo></mrow><mrow><mo>(</mo><mn>6</mn><mo>)</mo></mrow></mrow></math>

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

solving the prediction value of the amplitude data sequence of the seismic data according to the gray number sequence prediction model established above, wherein the prediction value comprises the following steps:

firstly, substituting the ash parameters a and u into a time function represented by an equation (4), and deriving and reducing the time function to obtain a model value sequence, as shown in equation (8):

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

secondly, a model value sequence obtained by predicting by using the model is accumulated and subtracted to obtain a predicted value sequence of an original sequence, as shown in 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, the predicted value is used

And measured value X_(t) ⁽⁰⁾And calculating the gray abnormal value of the seismic amplitude data according to the absolute error and the current pair error between the absolute error and the current pair error, namely determining a section of data sequence of a section of stratum on the earthquake, which corresponds to the inconsistency of the data structure with other sections, according to the absolute error and the relative error between the predicted value and the measured value.

Wherein, X_(t) ⁽⁰⁾And

the absolute error and the relative error of (2) are calculated by equation (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 absolute error between the predicted value and the measured value obtained by the ash differential response can reflect the distortion of the seismic waveform caused by different fluids contained in the reservoir, and the relative error between the predicted value and the measured value shows the deviation degree. Thus, by analyzing the absolute and relative errors, a prediction can be made as to whether the formation contains hydrocarbons. The actual prediction can be made from finding horizon segments with small relative errors.

(2) Association analysis

And performing five-step correlation analysis on the grey abnormal values of the structural features of the seismic data volume, and then sequencing to determine the oil, gas and water layers. First, raw seismic amplitude data is subjected to dimensionless processing in a first step. The dimensionless method is commonly used for initialization, equalization and interval relative quantization. Initialization means that all data are divided by the 1 st data to obtain a new sequence of numbers, i.e. the percentage of the values at different times relative to the 1 st timeAnd (4) the ratio. Averaging refers to dividing all data by the average of this data sequence. The interval relatization means that all data intervals are relatively valued: subtracting the minimum value of the whole data column from each value of the data column and dividing the value by the difference between the maximum value and the minimum value of the data column, wherein the specific expression 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 the kth data in the original data sequence, y (k) is the kth data of the data sequence after the non-dimensionalization processing, x (max), and x (min) respectively represent the maximum value and the minimum value of the data in the original data sequence.

And secondly, solving two-stage difference in the correlation coefficient. One difference in the process of calculating the correlation coefficient provides for calculating the correlation coefficient as follows, including:

|x₀(k)-x_j(k) and the angle mark represents the sum of each vector in the two arrays, and the values obtained when the angle mark values are different. Specifically, | x₀(k)-X_j(k)|＝Δ_j(k) Referred to as Kth point x₀And x_jThe absolute difference of the absolute values of the two, $\underset{j}{\min }\underset{j}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ is the minimum difference of the two levels, $\underset{k}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ is the first-order minimum difference, which means that each point is found on the jth curve and X₀The minimum difference of the sum of the difference of the two, $\underset{j}{\min }\underset{k}{\min }|x_0\left(k\right)-X_j\left(k\right)|$ the second-order minimum difference represents that the minimum difference in all the curves is found on the basis of the minimum difference found in each curve. $\underset{j}{\max }\underset{k}{\max }|x_0\left(k\right)-X_j\left(k\right)|$ The maximum difference is the second-level maximum difference, has the same meaning as the second-level minimum difference, and represents that the maximum difference in all the curves is found on the basis of the maximum difference found in each curve.

And thirdly, solving a correlation coefficient. The seismic amplitude data after dimensionless processing is set as a parent sequence and recorded as x₀I.e. x₀＝{x₀(1)，x₀(2)，…，x₀(n) }; denote the subsequence as x_jI.e. x_j＝{x_j(1)，_j(2)，…，x_j(n) }, j ═ 1, 2, …, m. The relationship between the arrays is called gray relationship. The closeness of the gray relationship can be represented by a gray correlation coefficient, which is expressed as:

<math><mrow><mrow><msub><mi>&gamma;</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>=</mo><mfrac><mrow><munder><mi>min</mi><mi>j</mi></munder><munder><mi>min</mi><mi>k</mi></munder><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo><mo>+</mo><mi>&xi;</mi><munder><mi>max</mi><mi>j</mi></munder><munder><mi>max</mi><mi>k</mi></munder><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo></mrow><mrow><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo><mo>+</mo><mi>&xi;</mi><munder><mi>max</mi><mi>j</mi></munder><munder><mi>max</mi><mi>k</mi></munder><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo></mrow></mfrac><mo>-</mo><mo>-</mo><mo>-</mo></mrow><mrow><mo>(</mo><mn>11</mn><mo>)</mo></mrow></mrow></math>

indicates the subsequence x_jThe kth element x of_j(k) With the mother sequence x₀Corresponding element x of₀(k) Is the resolution factor.

And fourthly, solving the association degree. Because the number of the correlation coefficients is large, the information is not centralized and is inconvenient to compare. For this purpose, the correlation coefficients under the individual elements are averaged to

Will be provided withDefined as the degree of association of the subsequence to the parent sequence, i.e.:

<math><mrow><mrow><mover><msub><mi>r</mi><mi>j</mi></msub><mo>&OverBar;</mo></mover><mo>=</mo><mfrac><mn>1</mn><mi>n</mi></mfrac><munderover><mi>&Sigma;</mi><mrow><mi>k</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>r</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><mo>-</mo><mo>-</mo></mrow><mrow><mo>(</mo><mn>12</mn><mo>)</mo></mrow></mrow></math>

the more similar the shapes of the two curves are to each other, the greater the degree of association, and vice versa, the smaller the degree of association. The key point is to analyze the gray correlation matrix to find out the factors playing the dominant role.

And fifthly, discharging the association sequence. If there is more than one reference number series and more than one compared factor, then an analysis of dominance can be performed. The reference number sequence is referred to as a mother number sequence (or mother factor), the comparison number sequence is a child number sequence (child factor), and the association matrix can be formed by the mother number sequence (or mother factor) and the child number sequence (child factor).

(3) Predicting oil and gas

And analyzing the correlation sequence of the seismic data volume structure characteristic grey abnormal value obtained in the first two steps, comparing the condition of the seismic data volume structure characteristic model when oil gas is contained, analyzing whether the correlation sequence of the grey abnormal value is related to the oil gas, and predicting the oil gas content of the reservoir. By relating the relationship between the elements, which factors are dominant and which factors are not dominant can be analyzed, and then the obtained association degrees of the intervals are sorted according to the size to obtain an association order. The relation between the correlation sequence and the oil gas is determined according to the specific well data of each region, the correlation degree of the drilled target layer is taken as a sample, the most matched layer section is the predicted hydrocarbon-bearing layer section, or the section of stratum with the maximum predicted relative and absolute errors is the hydrocarbon-bearing stratum.

The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (10)

1. A method for predicting hydrocarbons using structural features of a seismic data volume, the method comprising the steps of:

a) extracting a seismic amplitude data sequence and establishing a grey number sequence prediction model;

b) solving the amplitude data sequence prediction value of the seismic data;

c) calculating the error between the predicted value and the measured value;

d) determining a seismic amplitude data gray abnormal value section according to the error;

e) carrying out non-quantitative rigidization treatment on the original seismic amplitude data and the gray abnormal values of the seismic amplitude data to respectively obtain a mother sequence and sub sequences;

f) solving the correlation coefficient between the parent sequence and each subsequence;

g) according to the correlation coefficient, solving the correlation degree between the parent sequence and each subsequence;

h) sorting the association degrees according to the magnitude to obtain an association order;

i) and determining hydrocarbon reservoirs according to the correlation sequence.

2. The method of claim 1, wherein: the establishing of the grey number series prediction model comprises the steps of selecting any sub number series from amplitude data series of original seismic data, performing primary accumulation generation on the sub number series, establishing a grey model GM (1, 1) according to a time function represented by the following formula by utilizing the sub series generated by the primary accumulation,

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

wherein, a is the parameter to be identified, u is the endogenous variable to be identified, and both can be obtained by the least square method.

3. The method of claim 2, wherein said deriving amplitude data sequence predictors for seismic data comprises:

a) setting ash parameters

Is composed of

A and u are obtained by the least square method according to the following formula,

where B is the accumulation matrix, B^TIs the transposed matrix of B, Y_NIs a constant vector;

b) substituting the gray parameters a and u into the time function, and deriving and reducing to obtain a model value sequence as shown in the following formula:

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

c) and (4) obtaining a predicted value sequence of the original sequence of numbers by utilizing the model value sequence through accumulation and subtraction.

4. The method of claim 3, wherein the accumulation matrix B and a constant vector Y_NRespectively as follows:

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

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

5. the method of claim 2, wherein: the error between the predicted value and the measured value includes an absolute error and a relative error, according to

Is calculated, whereinIs a predicted value and

is a measured valueAnd determining the seismic amplitude data gray abnormal value section according to the prediction model, and if the error of the prediction sequence obtained by the gray model established by the amplitude data of the original seismic data is too large after detection, establishing a residual error model for correcting the originally established gray model until the requirement is met.

6. The method of claim 1, wherein: the step of determining the seismic amplitude data gray abnormal value comprises the step of determining a data sequence of a section of stratum corresponding to the earthquake, wherein the data sequence is inconsistent with other sections through the absolute error and the relative error of the predicted value and the measured value.

7. The method of claim 1, wherein: the method for the non-quantity rigidization processing comprises initialization processing, equalization processing or interval relative value processing, wherein the initialization processing is a sequence obtained by dividing all data in a data sequence by the 1 st data, the equalization processing is a sequence obtained by dividing all data in the data sequence by the average value of the data sequence, the interval relative value processing is a step of subtracting the minimum value of the whole data sequence from each value of the data sequence and then dividing the minimum value by the difference between the maximum value and the minimum value of the data sequence, and the specific expression is that

Wherein, x (k) is the kth data in the original data sequence, y (k) is the kth data of the data sequence after the non-dimensionalization processing, x (max), and x (min) respectively represent the maximum value and the minimum value of the data in the original data sequence.

8. The method of claim 1, wherein if the reference number is X₀The number series to be compared, i.e. the factor number series, is X_jJ is 1, 2, …, n, curve X₀And X_jThe correlation coefficient at the k-th point is represented by:

<math><mrow><msub><mi>&gamma;</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>=</mo><mfrac><mrow><munder><mi>min</mi><mi>j</mi></munder><munder><mi>min</mi><mi>k</mi></munder><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo><mo>+</mo><msub><mi>&xi;</mi><mrow><munder><mi>max</mi><mi>j</mi></munder><munder><mi>max</mi><mi>k</mi></munder></mrow></msub><msub><mrow><mo>|</mo><mi>x</mi></mrow><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo></mrow><mrow><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo><mo>+</mo><msub><mi>&xi;</mi><mrow><munder><mi>max</mi><mi>j</mi></munder><munder><mi>max</mi><mi>k</mi></munder></mrow></msub><mo>|</mo><msub><mi>x</mi><mn>0</mn></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>-</mo><msub><mi>x</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>|</mo></mrow></mfrac></mrow></math>

wherein, the reference number sequence is marked as x₀＝{x₀(1)，x₀(2)，…，x₀(n) }; the number series being compared (factor number series) is x_j＝{x_j(1)，x_j(2)，…，x_j(n)}，j＝1，2，…，m；x_j(k) Is a subsequence x_jThe kth element of (1); x is the number of₀(k) Is a reference number sequence x₀With the compared series x_j(k) The corresponding elements; xi is a resolution coefficient, the value range is between 0 and 1, and is generally 0.5;is the second-order minimum difference and,

is the two-step maximum difference, | x₀(k)-x_j(k)|＝Δ_j(k) Is the absolute difference between the amplitude of the k-th parent sequence and the amplitude of the subsequence.

9. The method of claim 1, wherein the expression of whole X is obtained according to the following formula_jCurve and parameter curve X₀Correlation degree of correlation coefficient between sequences:

<math><mrow><mover><msub><mi>r</mi><mi>j</mi></msub><mo>&OverBar;</mo></mover><mo>=</mo><mfrac><mn>1</mn><mi>n</mi></mfrac><munderover><mi>&Sigma;</mi><mrow><mi>k</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>r</mi><mi>j</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow></mrow></math>

wherein,

the association degree of the subsequence to the parent sequence is shown.

10. The method of claim 1, wherein: and determining the hydrocarbon reservoir according to the correlation sequence according to the abnormal value section with the maximum predicted relative and absolute errors and the correlation degree with the drilled target layer as a sample, wherein the interval which is most matched with the hydrocarbon reservoir is the predicted hydrocarbon-containing interval.

Images