Oil and gas identification method and system

Abstract

The present invention provides a method and system for oil and gas identification. The oil and gas identification method includes: Step 1, obtaining seismic attribute data of a research area and processing the seismic attribute data; Step 2, initializing parameters; Step 3, solving the non-oil and gas representation vector and the oil and gas target representation vector; Step 4, predicting the oil and gas information of unknown points. Starting from the perspective of pixel reconstruction of seismic attribute images, the oil and gas identification method and system reconstruct the information of unknown points using a dictionary structure, which has strong interpretability. At the same time, the sparse expression of the non-oil and gas sample dictionary and the dense expression of the oil and gas sample dictionary are introduced, which can improve the expression ability and discrimination ability of features simultaneously, thereby improving the identification accuracy.

Description

Technical Field

[0001] The present invention relates to the technical fields of structural geology and petroleum geology, and particularly relates to a method and system for oil and gas identification. Background Art

[0002] Oil and gas prediction based on seismic data has always been an important issue in the field of oil and gas exploration. In recent years, the application of artificial intelligence technology has injected new impetus into oil and gas prediction.

[0003] In the Chinese patent application with the application number 200910236634.6, a method for oil and gas prediction is involved. The method includes: 1) performing wavelet transform on the longitudinal wave signals at multiple locations in a region to obtain the wavelet transform coefficients of each longitudinal wave signal; 2) for each longitudinal wave signal, comparing the wavelet transform coefficients of the longitudinal wave signal under different scale parameters σ, and determining the scale parameter σbest corresponding to the maximum wavelet transform coefficient; 3) for each longitudinal wave signal, calculating the main frequency and / or quality factor of the longitudinal wave signal under the determined scale parameter σbest of the longitudinal wave signal, where the quality factor is a parameter reflecting the degree of vibration energy loss of the longitudinal wave signal; and 4) predicting the favorable regions of oil and gas distribution based on the calculated main frequencies and / or quality factors of the multiple longitudinal wave signals. By utilizing the good local variation characteristics of the wavelet function in the time domain and frequency domain, the invention can calculate the main frequency and / or quality factor of the longitudinal wave signal under the optimal scale parameter, and the main frequency and / or quality factor can accurately reflect the attenuation characteristics of the longitudinal wave signal, so as to accurately predict the favorable regions of oil and gas distribution.

[0004] In the Chinese patent application with the application number 201410608508.X, a method and device for predicting the oil and gas production of horizontal wells in tight oil and gas reservoirs are involved. The method includes: establishing a reservoir structure index model and a productivity heterogeneity index model of the drilled vertical well according to the drilling data of the drilled vertical well; establishing a maximum monthly production prediction model of the drilled vertical well by using the reservoir structure index model and the productivity heterogeneity index model; establishing a maximum monthly production prediction model of the horizontal well to be drilled at the corresponding position according to the maximum monthly production prediction model of the drilled vertical well; and establishing a production prediction model of the horizontal well to be drilled varying with production time according to the maximum monthly production prediction model of the horizontal well to be drilled, so as to predict the oil and gas production of the horizontal well to be drilled varying with production time. The invention can obtain a predicted production that is more consistent with the actual oil and gas production of horizontal wells in tight oil and gas reservoirs.

[0005] In the Chinese patent application with the application number 201710744444.X, a method and device for predicting the distribution of dolomite oil and gas reservoirs are involved, including: obtaining seismic data and logging data of the reservoir to be measured; performing anisotropy elimination processing on the seismic data to obtain CRP and CMP gather data volumes; performing environmental correction on the logging data to obtain density logging data; interpreting the porosity and acoustic travel time of the reservoir to be measured; establishing a rock physics model for the dolomite section of the reservoir to be measured; performing cross-analysis of rock physics elastic parameters for the target dolomite formation series to determine the types and threshold ranges of rock physics elastic parameters of the reservoir to be measured; performing joint inversion based on the seismic sub-stack data volume with different incident angles and logging data; and interpreting the inversion result data volume of the rock physics elastic parameters of the reservoir to be measured to obtain the cumulative thickness corresponding to the reservoir section and the cumulative thickness corresponding to the effective reservoir section of the reservoir to be measured. The method and device for predicting the distribution of dolomite oil and gas reservoirs provided by the invention improve the accuracy of predicting the distribution of dolomite oil and gas reservoirs.

[0006] In the Chinese patent application with the application number 201210393409.5, a method for predicting carbonate rock oil and gas reservoirs using low-frequency information is involved. First, forward modeling is performed on the fluid layer using the dispersion-viscosity wave equation to obtain the low-frequency seismic reflection characteristics of the fluid layer, i.e., low-frequency strong energy anomaly. Secondly, the single-frequency time-frequency domain energy time series of the seismic signal is used as a new time signal, and a hybrid filter is applied to it, and then the instantaneous maximum time-frequency domain energy is taken. In this way, the obtained low-frequency energy is not only extremely prominent but also has a very high time resolution. All of the above are carried out in the low-frequency band, and the low-frequency band is determined as the frequency band corresponding to 15% - 35% of the cumulative energy of the instantaneous spectrum. The invention uses the dispersion-viscosity wave equation to perform forward modeling to obtain the low-frequency seismic reflection characteristics of the fluid reservoir, i.e., low-frequency strong energy anomaly. Compared with the existing low-frequency technologies, the low-frequency strong energy anomaly extracted by the invention is not only very prominent but also has a very high time resolution, improving the accuracy of predicting carbonate rock oil and gas reservoirs.

[0007] In the Chinese patent application with the application number 201310450077.4, an oil and gas prediction method based on spectral shape analysis technology is involved. The reservoir calibration is carried out by using the horizon calibration technology in the seismic data interpretation system. According to the difference in reservoir thickness, the discrete Fourier transform or wavelet transform is used to transform the seismic data of the reservoir section from the time domain to the frequency domain, obtaining the frequency domain data volume of the reservoir section. The oil and gas bearing property of a single point is determined by the spectral shape, and the planar distribution of the oil and gas bearing zone is predicted by using the frequency domain data volume of the reservoir section, obtaining the distribution range of the oil and gas bearing zone in the whole area. Compared with the frequency attenuation attribute calculated from the conventional time domain seismic data, the single-point spectral shape analysis can be clearly carried out on the frequency domain data volume and mapped on the plane, with good results, and the coincidence rate reaches more than 90%. Moreover, the calculation time is increased by 3 - 5 times. Compared with the frequency attenuation attribute calculated from the conventional time domain seismic data, the single-point spectral shape analysis can be clearly carried out on the frequency domain data volume and mapped on the plane. These methods require a lot of expert knowledge and cost a huge amount of time. At the same time, due to the differences in the experience of different experts, the inconsistency of the prediction results may occur. Therefore, the data-driven oil and gas prediction can effectively improve such problems. However, most of the existing data-driven oil and gas prediction methods adopt black box models such as neural networks. Although the accuracy is relatively high, the interpretability is poor, and it is difficult to ensure its reliability.

[0008] The above existing technologies are quite different from the present invention and fail to solve the technical problems we want to solve. Therefore, we have invented a new oil and gas identification method. Summary of the Invention

[0009] The purpose of the present invention is to provide an oil and gas identification method that can simultaneously improve the expression ability and discrimination ability of features, thereby enhancing the identification accuracy.

[0010] The purpose of the present invention can be achieved by the following technical measures: an oil and gas identification method, which includes:

[0011] Step 1: Obtain the seismic attribute data of the research area and process the seismic attribute data.

[0012] Step 2: Initialize the parameters.

[0013] Step 3: Solve the non-oil and gas representation vector and the oil and gas target representation vector.

[0014] Step 4: Predict the oil and gas information of unknown points.

[0015] The purpose of the present invention can also be achieved by the following technical measures:

[0016] In step 1, obtain the seismic attribute data of the research area. According to the geodetic coordinates, label the hydrocarbon-bearing property of the seismic attribute vector at a certain coordinate point. If the logging information, oil testing results, and manual interpretation results of this coordinate point show the presence of oil and gas, label it as having oil and gas; otherwise, label it as having no oil and gas. Obtain a certain number of seismic attribute vector sets without oil and gas. And the seismic attribute vector set with oil and gas where d is the dimension of the seismic attribute, and n0 and n1 are the numbers of seismic attribute vectors without oil and gas and with oil and gas, respectively.

[0017] In step 2, set the non-hydrocarbon representation vector as the zero vector, the hydrocarbon target representation vector as the zero vector, the noise matrix as the zero vector, the Lagrange multiplier vector as the zero vector. Set the trade-off coefficients λ1, λ2, μ > 0 as positive real numbers, set the convergence coefficient ∈ > 0 as a positive real number, and set the judgment factor k > 0.

[0018] In step 2, set ∈ < 10 -3 , λ1 ∈ (2, 200), λ1 ∈ (0.2, 20), μ ∈ (20, 200), k ∈ (0.1, 10).

[0019] In step 2, set ∈ < 10 -4 , λ1 ∈ (1, 100), λ1 ∈ (0.1, 10), μ ∈ (5, 50), k ∈ (0.2, 5).

[0020] In step 2, set ∈ < 10 -1 , λ1 ∈ (1, 50), λ1 ∈ (1, 100), μ ∈ (0.2, 10), k ∈ (0.1, 10).

[0021] In step 3, for the seismic attribute vector of an unknown coordinate point, solving the following problem can obtain the optimal non-hydrocarbon representation vector and the hydrocarbon target representation vector That is:

[0022]

[0023] In step 3, the involved and The solution steps are as follows:

[0024] Step 301. Fix the values of W0, W1, and Θ and solve for E, that is, solve:

[0025]

[0026] Let Then the current optimal E can be solved.

[0027] Step 302: Fix the values of W1, E, and Θ to solve for W0, that is, solve

[0028]

[0029] The current optimal W0 can be solved by using proximal gradient descent.

[0030] Step 303: Fix the values of W0, E, and Θ to solve for W1, that is, solve

[0031]

[0032] Let Then the current optimal W1 can be solved.

[0033] Step 304: Fix the values of W0, W1, and E to solve for Θ, that is, solve

[0034]

[0035] Let Then the current optimal Θ can be solved.

[0036] Step 305: If Then jump to Step 301; otherwise, assign the current W0 to And assign the current W1 to And jump to Step 4.

[0037] In Step 4, first, calculate the of the reconstructed non-hydrocarbon information vector of the unknown point and the of the reconstructed hydrocarbon target information vector. Then, calculate the corresponding reconstruction error and Finally, if ε0 < kε1, then predict that the unknown point does not contain hydrocarbons; otherwise, when ε0 ≥ kε1, then predict that the unknown point contains hydrocarbons.

[0038] The object of the present invention can also be achieved by the following technical measures: a hydrocarbon identification system, which includes a seismic attribute data processing unit, a parameter initialization unit, a representation vector calculation unit, and an unknown point hydrocarbon prediction unit. The seismic attribute data processing unit acquires the seismic attribute data of the research area and processes the seismic attribute data; the parameter initialization unit performs parameter initialization, the representation vector calculation unit calculates the non-hydrocarbon representation vector and the hydrocarbon target representation vector, and the unknown point hydrocarbon prediction unit predicts the hydrocarbon information of the unknown point.

[0039] The object of the present invention can also be achieved by the following technical measures:

[0040] The seismic attribute data processing unit obtains the seismic attribute data of the research area, and marks the hydrocarbon-bearing property of the seismic attribute vector of a certain coordinate point according to the geodetic coordinates. If the logging information, oil testing results, and manual interpretation results of this coordinate point show the presence of oil and gas, it is marked as having oil and gas; otherwise, it is marked as having no oil and gas. Then, a set of seismic attribute vectors without oil and gas can be obtained. And the set of seismic attribute vectors with oil and gas where d is the dimension of the seismic attribute, and n0 and n1 are the numbers of the seismic attribute vectors without oil and gas and with oil and gas respectively.

[0041] The parameter initialization unit sets the non-hydrocarbon representation vector as a zero vector, the hydrocarbon target representation vector as a zero vector, the noise matrix as a zero vector, and the Lagrange multiplier vector as a zero vector. Appropriately set the trade-off coefficients λ1, λ2, μ > 0 as positive real numbers, and appropriately set the convergence coefficient ∈ > 0 as a positive real number.

[0042] For the seismic attribute vector of an unknown coordinate point, the representation vector obtaining unit can obtain the optimal non-hydrocarbon representation vector and the hydrocarbon target representation vector of this unknown point by solving the following problem of the unknown point, that is: Namely:

[0043]

[0044] The hydrocarbon prediction unit of the unknown point first calculates the of the reconstructed non-hydrocarbon information vector and the of the reconstructed hydrocarbon target information vector of the unknown point. Then, calculate the corresponding reconstruction errors and Finally, if ε0 < kε1, it is predicted that the unknown point does not contain oil and gas; otherwise, when ε0 ≥ kε1, it is predicted that the unknown point contains oil and gas.

[0045] The oil and gas identification method in the present invention includes the following steps: seismic attribute data processing, parameter initialization, solving the non-oil and gas representation vector and the oil and gas target representation vector, and predicting the oil and gas information at unknown points. Compared with the prior art, starting from the perspective of pixel reconstruction of the seismic attribute image, the present invention reconstructs the information of unknown points using a dictionary structure, which has strong interpretability. At the same time, the sparse expression of the non-oil and gas sample dictionary and the dense expression of the oil and gas sample dictionary are introduced, which can improve the expression ability and discrimination ability of features simultaneously, thereby improving the recognition accuracy. Compared with the prior art, starting from the perspective of pixel reconstruction of the seismic attribute image, the present invention reconstructs the information of unknown points using a dictionary structure, which has strong interpretability. At the same time, the sparse expression of the non-oil and gas sample dictionary and the dense expression of the oil and gas sample dictionary are introduced, which can improve the expression ability and discrimination ability of features simultaneously, thereby improving the recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flowchart of a specific embodiment of the oil and gas identification method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] It should be noted that the following detailed description is exemplary and is intended to provide further illustration of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0048] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, and / or combinations thereof.

[0049] As Figure 1 shown, Figure 1 It is a flowchart of the oil and gas identification method of the present invention. The oil and gas identification method includes the following steps:

[0050] Step 1, seismic attribute data processing

[0051] Obtain the seismic attribute data of the research area, and label the seismic attribute vector of a certain coordinate point according to the geodetic coordinates for its oil and gas content. If the logging information, oil testing results, and manual interpretation results of this coordinate point show the presence of oil and gas, it is labeled as having oil and gas; otherwise, it is labeled as having no oil and gas. Then, a certain number of non-oil and gas seismic attribute vector sets and oil and gas seismic attribute vector sets Where d is the dimension of seismic attributes, and n0 and n1 are the numbers of non-hydrocarbon seismic attribute vectors and hydrocarbon seismic attribute vectors, respectively.

[0052] Step 2: Parameter initialization

[0053] Set the non-hydrocarbon representation vector as a zero vector, and the hydrocarbon target representation vector as a zero vector, the noise matrix as a zero vector, and the Lagrange multiplier vector as a zero vector. Appropriately set the trade-off coefficients λ1, λ2, and μ > 0 as positive real numbers, and appropriately set the convergence coefficient ∈ > 0 as a positive real number.

[0054] Step 3: Solve for the non-hydrocarbon representation vector and the hydrocarbon target representation vector

[0055] For the seismic attribute vector at a certain unknown coordinate point Solving the following problem can obtain the optimal non-hydrocarbon representation vector of this unknown point and the hydrocarbon target representation vector That is:

[0056]

[0057] Step 4: Predict the hydrocarbon information of the unknown point

[0058] First, calculate the of the reconstructed non-hydrocarbon information vector of the unknown point and the of the reconstructed hydrocarbon target information vector Finally, if ε0 < kε1, then predict that the unknown point does not contain hydrocarbons; otherwise, when ε0 ≥ kε1, then predict that the unknown point contains hydrocarbons.

[0059] The and involved in Step 3 are solved as follows:

[0060] Step 301: Fix the values of W0, W1, and Θ and solve for E, that is, solve:

[0061]

[0062] Let Then the current optimal E can be solved;

[0063] Step 302: Fix the values of W1, E, and Θ and solve for W0, that is, solve

[0064]

[0065] The current optimal W0 can be obtained by using proximal gradient descent;

[0066] Step 303: Fix the values of W0, E, and Θ and solve for W1, that is, solve

[0067]

[0068] Let The current optimal W1 can be obtained;

[0069] Step 304: Fix the values of W0, W1, and E and solve for Θ, that is, solve

[0070]

[0071] Let The current optimal Θ can be obtained;

[0072] Step 305: If Then jump to step 301, otherwise let And And jump to step 4.

[0073] The present invention also provides an oil and gas identification system, including a seismic attribute data processing unit, a parameter initialization unit, a representation vector calculation unit, and an unknown point oil and gas prediction unit, wherein:

[0074] The seismic attribute data processing unit processes seismic attribute data;

[0075] Obtain the seismic attribute data of the research area, label the oil and gas content of the seismic attribute vector of a certain coordinate point according to the geodetic coordinates. If the logging information, oil testing results, and manual interpretation results of this coordinate point show the presence of oil and gas, it is labeled as having oil and gas, otherwise it is labeled as having no oil and gas; then a certain number of non-oil-and-gas seismic attribute vector sets And the oil-and-gas seismic attribute vector set where d is the dimension of the seismic attribute, and n0 and n1 are the numbers of non-oil-and-gas seismic attribute vectors and oil-and-gas seismic attribute vectors respectively

[0076] The parameter initialization unit initializes parameters.

[0077] Set the non-oil-and-gas representation vector As a zero vector, the oil and gas target representation vector As a zero vector, the noise matrix As a zero vector, the Lagrange multiplier vector As a zero vector, appropriately set the trade-off coefficients λ1, λ2, μ > 0 as positive real numbers, and appropriately set the convergence coefficient ∈ > 0 as a positive real number.

[0078] The vector representation obtaining unit solves the non-hydrocarbon representation vector and the hydrocarbon target representation vector.

[0079] For the seismic attribute vector of a certain unknown coordinate point Solving the following problem can obtain the optimal non-hydrocarbon representation vector of the unknown point and the hydrocarbon target representation vector That is:

[0080]

[0081] The unknown point hydrocarbon prediction unit predicts the hydrocarbon information of the unknown point.

[0082] First, calculate the of the reconstructed non-hydrocarbon information vector of the unknown point and the of the reconstructed hydrocarbon target information vector Finally, if ε0 < kε1, it is predicted that the unknown point contains no hydrocarbons, otherwise when ε0 ≥ kε1, it is predicted that the unknown point contains hydrocarbons.

[0083] The following are several specific embodiments of applying the present invention.

[0084] Embodiment 1

[0085] In a specific Embodiment 1 of applying the present invention, the hydrocarbon identification method includes the following steps:

[0086] Step 1: Seismic attribute data processing

[0087] Obtain the seismic attribute data of the research area, mark the hydrocarbon-bearing property of the seismic attribute vector of a certain coordinate point according to the geodetic coordinates. If the logging information, oil testing results, and manual interpretation results of this coordinate point show hydrocarbons, it is marked as having hydrocarbons, otherwise it is marked as having no hydrocarbons; then a certain number of non-hydrocarbon seismic attribute vector sets and hydrocarbon-bearing seismic attribute vector sets can be obtained, where d is the dimension of the seismic attribute, and n0 and n1 are the numbers of non-hydrocarbon seismic attribute vectors and hydrocarbon-bearing seismic attribute vectors respectively

[0088] Step 2: Parameter initialization

[0089] Set the non-hydrocarbon representation vector as the zero vector, and the hydrocarbon target representation vector as the zero vector, the noise matrix as the zero vector, and the Lagrange multiplier vector is a zero vector, appropriately set the trade-off coefficients λ1, λ2, μ > 0 as positive real numbers, and appropriately set the convergence coefficient ∈ > 0 as a positive real number; ∈ < 10 -3 , λ1 ∈ (2, 200), λ1 ∈ (0.2, 20), μ ∈ (20, 200).

[0090] Step 3, Solve the non-oil-gas representation vector and the oil-gas target representation vector

[0091] For the seismic attribute vector of a certain unknown coordinate point Solving the following problem can obtain the optimal non-oil-gas representation vector of this unknown point and the oil-gas target representation vector That is:[[ID=z15]]

[0092]

[0093] Step 4, Predict the oil-gas information of the unknown point

[0094] First, calculate the of the reconstructed non-oil-gas information vector of the unknown point and the of the reconstructed oil-gas target information vector Finally, if ε0 < kε1, then predict that this unknown point does not contain oil and gas, otherwise when ε0 ≥ kε1, then predict that this unknown point contains oil and gas.

[0095] Involved in Step 3 and The solution steps are as follows:

[0096] Step 301, Fix the values of W0, W1, Θ and solve for E, that is, solve:

[0097]

[0098] Let Then the current optimal E can be solved;

[0099] Step 302, Fix the values of W1, E, Θ and solve for W0, that is, solve

[0100]

[0101] The proximal gradient descent can be used to solve the current optimal W0;

[0102] Step 303, Fix the values of W0, E, Θ and solve for W1, that is, solve

[0103]

[0104] Let The current optimal W1 can be obtained by solving the equation.

[0105] Step 304: Fix the values of W0, W1, and E and solve for Θ, that is, solve the equation

[0106]

[0107] Let The current optimal Θ can be obtained by solving the equation.

[0108] Step 305: If Jump to Step 301; otherwise, let And Jump to Step 4.

[0109] Embodiment 2

[0110] In the specific Embodiment 2 of applying the present invention, the oil and gas identification method includes the following steps:

[0111] Step 1: Seismic attribute data processing

[0112] Obtain the seismic attribute data of the research area. According to the geodetic coordinates, label the oil and gas content of the seismic attribute vector of a certain coordinate point. If the logging information, oil testing results, and manual interpretation results of this coordinate point show the presence of oil and gas, label it as having oil and gas; otherwise, label it as having no oil and gas. Then, a certain number of seismic attribute vector sets without oil and gas can be obtained And the seismic attribute vector set with oil and gas where d is the dimension of the seismic attribute, and n0 and n1 are the numbers of seismic attribute vectors without oil and gas and with oil and gas respectively

[0113] Step 2: Parameter initialization

[0114] Set the non - oil - gas representation vector as the zero vector, the oil and gas target representation vector as the zero vector, the noise matrix as the zero vector, the Lagrange multiplier vector as the zero vector. Appropriately set the trade - off coefficients λ1, λ2, μ > 0 as positive real numbers, and appropriately set the convergence coefficient ∈ > 0 as a positive real number; ∈ < 10 -4 , λ1 ∈ (1, 100), λ1 ∈ (0.1, 10), μ ∈ (5, 50), k ∈ (0.2, 5).

[0115] Step 3: Solve for the non - oil - gas representation vector and the oil and gas target representation vector

[0116] [[ID=�0]]For the seismic attribute vector of an unknown coordinate point The following problem can be solved to obtain the optimal non - oil - gas representation vector of this unknown point With the oil and gas target representation vector That is:

[0117]

[0118] Step 4. Predict the oil and gas information of unknown points

[0119] First, calculate the of the reconstructed non - oil and gas information vector of the unknown point and the Then, calculate the corresponding reconstruction error and Finally, if ε0 < kε1, then predict that the unknown point does not contain oil and gas; otherwise, when ε0 ≥ kε1, then predict that the unknown point contains oil and gas.

[0120] Involved in Step 3 and The solution steps are as follows:

[0121] Step 301. Fix the values of W0, W1, Θ and solve for E, that is, solve:

[0122]

[0123] Let Then the current optimal E can be solved;

[0124] Step 302. Fix the values of W1, E, Θ and solve for W0, that is, solve

[0125]

[0126] Using proximal gradient descent, the current optimal W0 can be solved;

[0127] Step 303. Fix the values of W0, E, Θ and solve for W1, that is, solve

[0128]

[0129] Let Then the current optimal W1 can be solved;

[0130] Step 304. Fix the values of W0, W1, E and solve for Θ, that is, solve

[0131]

[0132] Let Then the current optimal Θ can be solved;

[0133] Step 305. If Then jump to Step 301; otherwise, let and And jump to step 4.

[0134] Embodiment 3:

[0135] In the specific Embodiment 3 of the application of the present invention, the oil and gas identification method includes the following steps:

[0136] Step 1, seismic attribute data processing

[0137] Obtain the seismic attribute data of the research area, and label the oil and gas content of the seismic attribute vector of a certain coordinate point according to the geodetic coordinates. If the logging information, oil testing results, and manual interpretation results of this coordinate point show oil and gas, it is labeled as having oil and gas, otherwise it is labeled as having no oil and gas; then a certain number of non-oil-and-gas seismic attribute vector sets can be obtained And the oil-and-gas seismic attribute vector set Where d is the dimension of the seismic attribute, and n0 and n1 are the numbers of non-oil-and-gas seismic attribute vectors and oil-and-gas seismic attribute vectors respectively.

[0138] Step 2, parameter initialization

[0139] Set the non-oil-and-gas representation vector As a zero vector, and the oil-and-gas target representation vector As a zero vector, the noise matrix As a zero vector, the Lagrange multiplier vector As a zero vector, appropriately set the trade-off coefficients λ1, λ2, μ > 0 as positive real numbers, and appropriately set the convergence coefficient ∈ > 0 as a positive real number; set ∈ < 10 -1 , λ1 ∈ (1, 50), λ1 ∈ (1, 100), μ ∈ (0.2, 10), k ∈ (0.1, 10).

[0140] Step 3, solve the non-oil-and-gas representation vector and the oil-and-gas target representation vector

[0141] For the seismic attribute vector of an unknown coordinate point Solving the following problem can obtain the optimal non-oil-and-gas representation vector of this unknown point And the oil-and-gas target representation vector That is:

[0142]

[0143] Step 4, predict the oil and gas information of the unknown point

[0144] First, calculate the Of the reconstructed non-oil-and-gas information vector of the unknown point And the Of the reconstructed oil-and-gas target information vector. Then, calculate the corresponding reconstruction error And Finally, if ε0 < kε1, it is predicted that the unknown point does not contain oil and gas; otherwise, when ε0 ≥ kε1, it is predicted that the unknown point contains oil and gas.

[0145] Those involved in step 3 and The solution steps are as follows:

[0146] Step 301: Fix the values of W0, W1, and Θ and solve for E, that is, solve:

[0147]

[0148] Let The current optimal E can be solved;

[0149] Step 302: Fix the values of W1, E, and Θ and solve for W0, that is, solve

[0150]

[0151] The proximal gradient descent can be used to solve the current optimal W0;

[0152] Step 303: Fix the values of W0, E, and Θ and solve for W1, that is, solve

[0153]

[0154] Let The current optimal W1 can be solved;

[0155] Step 304: Fix the values of W0, W1, and E and solve for Θ, that is, solve

[0156]

[0157] Let The current optimal Θ can be solved;

[0158] Step 305: If Jump to step 301; otherwise, let and Jump to step 4.

[0159] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0160] Except for the technical features described in the specification, all are well-known technologies to those skilled in the art.

Claims

1. An oil and gas identification method, characterized in that, The oil and gas identification method includes: Step 1: Obtain seismic attribute data of the research area and process the seismic attribute data; Step 2: Initialize the parameters; Step 3: Solve the non-oil and gas representation vector and the oil and gas target representation vector; Step 4: Predict the oil and gas information of unknown points; In step 1, seismic attribute data of the research area is obtained. According to the geodetic coordinates, the hydrocarbon-bearing property of the seismic attribute vector at a certain coordinate point is labeled. If the logging information, oil testing results, and manual interpretation results of this coordinate point show the presence of oil and gas, it is labeled as having oil and gas; otherwise, it is labeled as having no oil and gas. A set of seismic attribute vectors with no oil and gas of a certain quantity is obtained and the set of seismic attribute vectors with oil and gas where d is the dimension of the seismic attribute, and n0 and n1 are the quantities of the seismic attribute vectors with no oil and gas and with oil and gas respectively; In step 2, set the non-oil-gas representation vector to be a zero vector, the oil-gas target representation vector to be a zero vector, the noise matrix to be a zero vector, the Lagrange multiplier vector to be a zero vector, set the trade-off coefficients λ1, λ2, μ > 0 as positive real numbers, and set the convergence coefficient ε > 0 as a positive real number; In step 3, for the seismic attribute vector of a certain unknown coordinate point The following problem is solved to obtain the optimal non-hydrocarbon representation vector of the unknown point And the hydrocarbon target representation vector That is:

2. The oil and gas identification method according to claim 1, wherein In step 2, set ε < 10 -3 , λ1 ∈ (2, 2000), λ1 ∈ (0.2, 20), μ ∈ (20, 200).

3. The oil and gas identification method according to claim 1, characterized in that In step 2, set ε < 10 -4 , λ1 ∈ (1, 100), λ1 ∈ (0.1, 10), μ ∈ (5, 50), k ∈ (0.2, 5).

4. The oil and gas identification method according to claim 1, wherein In step 2, set ε < 10 -1 , λ1 ∈ (1, 50), λ1 ∈ (1, 100), μ ∈ (0.2, 10), k ∈ (0.1, 10).

5. The oil and gas identification method according to claim 1, wherein In step 3, those involved and The solution steps are as follows: Step 301: Fix the values of W0, W1, and Θ to solve for E, that is, solve: Let the current optimal E can be obtained by solving Step 302: Fix the values of W1, E, and Θ to solve for W0, that is, solve The current optimal w0 can be solved by using the proximal gradient descent; Step 303: Fix the values of W0, E, and Θ to solve for W1, that is, solve Let Then the current optimal w1 can be obtained by solving Step 304: Fix the values of W0, W1, and E to solve for Θ, that is, solve Let the current optimal Θ can be solved and obtained; Step 305. If ο > ε, then jump to Step 301; otherwise, let and and jump to Step 4.

6. The oil and gas identification method according to claim 4, characterized in that In step 4, first, calculate the of the reconstructed non-hydrocarbon information vector of the unknown point and the of the reconstructed hydrocarbon target information vector. Then, calculate the corresponding reconstruction error and Finally, if ε0 < kε1, predict that the unknown point does not contain hydrocarbons; otherwise, when ε0 ≥ kε1, predict that the unknown point contains hydrocarbons.

7. An oil and gas identification system, characterized in that, The oil and gas identification system includes a seismic attribute data processing unit, a parameter initialization unit, a representation vector calculation unit, and an unknown point oil and gas prediction unit. The seismic attribute data processing unit obtains the seismic attribute data of the research area and processes the seismic attribute data; the parameter initialization unit initializes the parameters, the representation vector calculation unit solves the non-oil and gas representation vector and the oil and gas target representation vector, and the unknown point oil and gas prediction unit predicts the oil and gas information of unknown points; The seismic attribute data processing unit obtains the seismic attribute data of the research area, and marks the hydrocarbon-bearing property of the seismic attribute vector of a certain coordinate point according to the geodetic coordinates. If the logging information, oil testing results, and manual interpretation results of this coordinate point show the presence of oil and gas, it is marked as having oil and gas; otherwise, it is marked as having no oil and gas. Then, a set of seismic attribute vectors without oil and gas with a certain quantity is obtained. and the set of seismic attribute vectors with oil and gas where d is the dimension of the seismic attribute, and n0 and n1 are the quantities of the seismic attribute vectors without oil and gas and with oil and gas respectively. The parameter initialization unit sets the non-oil / gas representation vector to a zero vector, and the oil / gas target representation vector to a zero vector, the noise matrix to a zero vector, and the Lagrange multiplier vector to a zero vector. Appropriately set the trade-off coefficients λ1, λ2, μ > 0 as positive real numbers, and appropriately set the convergence coefficient ε > 0 as a positive real number; The representation vector obtaining unit calculates the seismic attribute vector of a certain unknown coordinate point By solving the following problem, the optimal non-hydrocarbon representation vector of the unknown point can be obtained and the hydrocarbon target representation vector W1 * , that is:

8. The oil and gas identification system according to claim 7, wherein This unknown point oil and gas prediction unit first calculates the of the reconstructed non-oil and gas information vector of the unknown point and the of the reconstructed oil and gas target information vector. Then, it calculates the corresponding reconstruction error and Finally, if ε0 < kε1, it is predicted that the unknown point does not contain oil and gas; otherwise, when ε0 ≥ kε1, it is predicted that the unknown point contains oil and gas.

Images