Seismic Parameter Optimization Method and Device Based on Multivariate Nonlinear Stepwise Regression

Abstract

The present invention provides a method and apparatus for optimizing seismic parameters based on multivariate non - linear stepwise regression. The method for optimizing seismic parameters based on multivariate non - linear stepwise regression includes: establishing a plurality of first regression functions according to target predicted reservoir parameters and a plurality of seismic parameters; screening the plurality of regression functions to determine a second regression function; and performing a significance test on the second regression function to optimize the seismic parameters. The present invention can quickly and accurately select, from a large number of seismic parameters (seismic attributes) that are originally independent of each other, the seismic parameters that are sensitive to the reservoir and reflect the essential characteristics of the reservoir. Furthermore, the reservoir development situation in the target work area can be predicted using the optimized seismic parameters, providing a reliable basis for the exploration and development of oilfields.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas field exploration and development, in particular to the technical field of seismic data processing and interpretation, and in particular to a seismic parameter optimization method and device based on multivariate nonlinear stepwise regression. Background Art

[0002] It is understandable that reservoir prediction using seismic parameters is primarily based on spatial variations in reservoir physical properties and the properties of the fluids within them, which can cause changes in a series of geometric, kinematic, and dynamic seismic parameters, such as seismic reflection waveforms, amplitude, frequency, energy, and phase. By analyzing the longitudinal and lateral variations in seismic parameters, the distribution range and reservoir characteristics of the reservoir can be predicted. Reservoir prediction using seismic parameters includes techniques such as seismic parameter extraction, seismic parameter optimization, and reservoir parameter conversion. Seismic parameter optimization is the core and foundation of reservoir prediction, and its purpose is to select seismic parameter combinations with high correlation with reservoir prediction parameters.

[0003] Seismic parameter extraction involves using various analytical methods to extract information related to lithology and reservoir properties from seismic data. Seismic parameter extraction is divided into three aspects: 3D attribute extraction, along-layer seismic parameter extraction, and interlayer absorption attribute extraction. 3D attribute extraction is based on 3D seismic data and can be extracted using different spatial combination patterns of seismic traces, which reflect reservoir characteristics from different perspectives. Along-layer seismic parameter extraction requires opening a time window along the target interval and performing autocorrelation, power spectrum, Fourier spectrum, autoregression, and other statistical analysis on the records within the time window to extract relevant seismic parameters. Interlayer absorption attributes are average absorption parameters of formations measured using seismic reflection wave data. The amplitude ratio method is typically used, which examines the absorption coefficient determined by studying the attenuation of the reflection wave amplitude between two adjacent layers. Analysis methods include Fourier spectrum, power spectrum, and reciprocal spectrum.

[0004] The purpose of seismic parameter optimization is to analyze the correlation between seismic parameters and identify independent seismic parameters that reflect the essential characteristics of the reservoir. With the in-depth study of seismic parameters, the types of seismic parameters are increasing. Although the increase in attributes can provide us with more useful information, its unlimited increase also has an adverse impact on reservoir prediction. The relationship between seismic parameters and the predicted objects is complex. The geological conditions and reservoir conditions in different regions, different depths, and different reservoirs are also different. The sensitivity of reservoirs to seismic parameters is not exactly the same. Even in the same work area and the same set of reservoirs, the corresponding sensitive attributes are significantly different depending on the observed objects. Therefore, it is necessary to optimize the large number of extracted attributes.

[0005] At present, there are mainly four methods for optimizing seismic parameters, namely the expert experience method, the mathematical theory method, the combination method of expert experience and mathematical theory, and the forward modeling determination method. The expert experience method includes two ways: determining attributes by experts selecting attribute combinations and experts specifying optimization attribute criteria. This method requires analyzing all attributes in combination with data in aspects such as geology and logging. Its advantages are high credibility, and the optimized attributes generally have relatively clear geological significance; the disadvantages are that it requires a relatively in-depth understanding of the work area and the meanings of various seismic parameters, and the subjectivity is large. The mathematical theory method uses mathematical methods to optimize some attributes for reservoir prediction, such as the genetic BP neural network method, the correlation method, the neural network method, the seismic parameter effectiveness method calculated based on seismic parameter values beside wells and logging characteristic values, the method of simultaneously controlling with multiple mathematical means, etc. This method does not require analyzing all attributes, but directly selects some ideal attributes using attribute methods, then conducts simple analysis, eliminates some unreasonable attributes, and finally uses the remaining few attributes for prediction. The advantages are that it reduces the workload of researchers, does not require an in-depth understanding of the work area and the meanings of all seismic parameters, and is relatively objective; the disadvantages are low credibility, and the optimized seismic parameters sometimes do not have clear geological significance. The combination method of expert experience and mathematical theory combines expert experience with advanced mathematical theory for attribute optimization. The advantages of this type of method are high credibility, and the optimized seismic parameters generally have relatively clear geological significance and are relatively objective; the disadvantages are that it requires an in-depth understanding of the work area and the meanings of various seismic parameters, the workload is large, and the subjectivity is large. The forward modeling determination method uses data such as logging and core to establish a geological model under certain conditions, then studies the responses of various seismic parameters to this model, classifies the seismic parameters that can reflect various reservoir characteristics, and thus uses the classified seismic parameters to predict reservoir characteristics. Its main advantages are a certain degree of credibility, and the optimized seismic parameters often have clear geological significance; the disadvantages are that it requires collecting a large amount of data such as logging and core for reservoir simulation, the workload is large, and there is a large deviation between the reservoir simulation and the actual reservoir situation, and the simulation results are very different from the actual reservoir prediction results. Summary of the Invention

[0006] Aiming at the problems in the prior art, the seismic parameter optimization method and device based on multivariate nonlinear stepwise regression provided by the present invention can quickly and accurately optimize the seismic parameters (seismic attributes) that are sensitive to the reservoir and reflect the essential characteristics of the reservoir from a large number of seismic parameters that are originally independent of each other, and then the optimized seismic parameters can be used to predict the reservoir development situation in the target work area, providing a reliable basis for the exploration and development of oilfields.

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a method for optimizing seismic parameters based on multivariate non - linear stepwise regression, including:

[0009] Establish a plurality of first regression functions based on the target predicted reservoir parameters and a plurality of seismic parameters;

[0010] Screen the plurality of regression functions to determine a second regression function;

[0011] Conduct a significance test on the second regression function to optimize the seismic parameters.

[0012] In one embodiment, the establishing a plurality of first regression functions based on the target predicted reservoir parameters and a plurality of seismic parameters includes:

[0013] Using the non - linear regression analysis method, establish a plurality of multivariate non - linear regression functions between the target predicted reservoir parameters and the plurality of seismic parameters.

[0014] In one embodiment, the screening the plurality of regression functions to determine a second regression function includes:

[0015] Calculate the error between the function values and the true values of the plurality of multivariate non - linear regression functions;

[0016] Determine the screening of the plurality of multivariate non - linear regression functions according to a plurality of errors to determine the second regression function.

[0017] In one embodiment, the conducting a significance test on the second regression function to optimize the seismic parameters includes:

[0018] Perform an iterative operation:

[0019] Delete the seismic parameters included in the second regression function from the plurality of seismic parameters to generate a preferred set of seismic parameters;

[0020] Substitute the seismic parameters in the preferred set of seismic parameters into the regression function in the previous iterative operation;

[0021] Using the non - linear regression analysis method, calculate the coefficients of the seismic parameters substituted in the current iterative operation in the regression function after substitution;

[0022] Determine the regression function of the current iterative operation according to the coefficients until the number of seismic parameters in the preferred set of seismic parameters is 0.

[0023] In one embodiment, the substituting the seismic parameters in the preferred set of seismic parameters into the regression function in the previous iterative operation includes:

[0024] Multiply the seismic parameters in the optimized set of seismic parameters by the seismic parameters in the regression function in the previous iteration operation;

[0025] Generate the independent variable of the regression function in the current iteration operation based on the multiplied seismic parameters.

[0026] In one embodiment, the seismic parameter optimization method based on multiple non - linear stepwise regression further includes:

[0027] Establish a prediction function for the target predicted reservoir parameters based on the optimized seismic parameters;

[0028] Predict the reservoir of the target work area according to the prediction function.

[0029] In a second aspect, the present invention provides a device for optimizing seismic parameters based on multiple non - linear stepwise regression, and the device includes:

[0030] A first function establishment unit, configured to establish a plurality of first regression functions according to the target predicted reservoir parameters and a plurality of seismic parameters;

[0031] A second function determination unit, configured to screen the plurality of regression functions to determine a second regression function;

[0032] A seismic parameter optimization unit, configured to perform a significance test on the second regression function to optimize the seismic parameters.

[0033] In one embodiment, the first function establishment unit is specifically configured to use a non - linear regression analysis method to establish a plurality of multiple non - linear regression functions between the target predicted reservoir parameters and the plurality of seismic parameters.

[0034] In one embodiment, the second function determination unit includes:

[0035] An error calculation module, configured to calculate the error between the function values of the plurality of multiple non - linear regression functions and the true values;

[0036] A second function determination module, configured to screen the plurality of multiple non - linear regression functions according to a plurality of errors to determine the second regression function.

[0037] In one embodiment, the seismic parameter optimization unit includes:

[0038] An iteration module, configured to perform an iteration operation:

[0039] Delete the seismic parameters included in the second regression function from the plurality of seismic parameters to generate an optimized set of seismic parameters;

[0040] Substitute the seismic parameters in the optimized set of seismic parameters into the regression function in the previous iteration operation;

[0041] Using the non - linear regression analysis method, calculate the coefficients of the seismic parameters substituted in the current iteration operation in the regression function after substitution.

[0042] Determine the regression function of the current iteration operation according to the coefficients until the number of seismic parameters in the preferred set of seismic parameters is 0.

[0043] In one embodiment, the iteration module includes:

[0044] A parameter multiplication module, configured to multiply the seismic parameters in the preferred set of seismic parameters by the seismic parameters in the regression function of the previous iteration operation.

[0045] A current independent variable generation module, configured to generate the independent variable of the regression function in the current iteration operation according to the multiplied seismic parameters.

[0046] In one embodiment, the seismic parameter optimization device based on multiple - variable non - linear stepwise regression further includes:

[0047] A prediction function establishment unit, configured to establish a prediction function for the target predicted reservoir parameters according to the optimized seismic parameters.

[0048] A reservoir prediction unit, configured to predict the reservoir of the target work area according to the prediction function.

[0049] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the seismic parameter optimization method based on multiple - variable non - linear stepwise regression are implemented.

[0050] In a fourth aspect, the present invention provides a computer - readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the seismic parameter optimization method based on multiple - variable non - linear stepwise regression are implemented.

[0051] As can be seen from the above description, for the seismic parameter optimization method and device based on multivariate non - linear step - by - step regression provided by the embodiments of the present invention, first, a plurality of first regression functions are established according to the target predicted reservoir parameters and a plurality of seismic parameters; then, the plurality of regression functions are screened to determine the second regression function; finally, a significance test is performed on the second regression function to optimize the seismic parameters. The present invention, based on the reservoir prediction parameters and the actual observed values of seismic attributes, uses a non - linear regression analysis method of "step - by - step introduction" and "step - by - step elimination" to perform a significance test on the correlation of seismic attributes, selects the seismic attributes with a large correlation with the reservoir prediction parameters, and on this basis, establishes an approximate multivariate non - linear regression model between the two. Then, during the model establishment process, a statistical analogy graph of the regression predicted value and the actual value is drawn to judge the correctness of the regression equation and the accuracy of the prediction result. Finally, based on the established reliable reservoir prediction model, the effective prediction of the target reservoir parameters is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0053] Figure 1 Flow chart of the seismic parameter optimization method based on multivariate non - linear step - by - step regression in the embodiments of the present invention Figure 1 ;

[0054] Figure 2 Flow chart of step 100 in the embodiments of the present invention;

[0055] Figure 3 Flow chart of step 200 in the embodiments of the present invention;

[0056] Figure 4 [[ID=SS]] Flow chart of step 300 in the embodiments of the present invention;

[0057] Figure 5 Flow chart of step 301 in the embodiments of the present invention;

[0058] Figure 6 Flow chart of the seismic parameter optimization method based on multivariate non - linear step - by - step regression in the embodiments of the present invention Figure 2 ;

[0059] Figure 7 Flow chart of the seismic parameter optimization method based on multivariate non - linear step - by - step regression in the specific application example of the present invention;

[0060] Figure 8 Mind map of the seismic parameter optimization method based on multivariate non - linear stepwise regression in the specific application example of the present invention;

[0061] Figure 9 Statistical analogy diagram of the measured value and regression value of the reservoir thickness of the F1S oil layer in the specific application example of the present invention;

[0062] Figure 10 Prediction result diagram of the reservoir thickness of the oil layer in the specific application example of the present invention;

[0063] Figure 11 Statistical analogy diagram of the measured value and regression value of the reservoir thickness of the F1Z oil layer in the specific application example of the present invention;

[0064] Figure 12 Prediction result diagram of the reservoir thickness of the F1Z oil layer in the specific application example of the present invention;

[0065] Figure 13 Statistical analogy diagram of the measured value and regression value of the reservoir thickness of the F1X oil layer in the specific application example of the present invention;

[0066] Figure 14 Prediction result diagram of the reservoir thickness of the F1X oil layer in the specific application example of the present invention;

[0067] Figure 15 Structural schematic of the seismic parameter optimization device based on multivariate non - linear stepwise regression in the embodiment of the present invention Figure 1 ;

[0068] Figure 16 Structural schematic diagram of the second function determination unit in the embodiment of the present invention;

[0069] Figure 17 Structural schematic diagram of the seismic parameter optimization unit in the embodiment of the present invention;

[0070] Figure 18 Structural schematic diagram of the iteration module in the embodiment of the present invention;

[0071] Figure 19 Structural schematic of the seismic parameter optimization device based on multivariate non - linear stepwise regression in the embodiment of the present invention Figure 2 ;

[0072] Figure 20 Structural schematic diagram of the electronic device in the embodiment of the present invention. Detailed implementation manners

[0073] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0074] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products, or devices.

[0075] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0076] An embodiment of the present invention provides a specific implementation manner of a seismic parameter optimization method based on multivariate nonlinear stepwise regression. See Figure 1 , and the method specifically includes the following content:

[0077] Step 100: Establish a plurality of first regression functions according to the target predicted reservoir parameters and a plurality of seismic parameters.

[0078] Specifically, s seismic attributes are selected and denoted as x i (i = 1, 2,..., s); through n sets of observed values obtained in actual work:

[0079] (x 1k , x 2k ,..., x sk , y k )(k = 1, 2,..., n)

[0080] Nonlinear stepwise regression analysis is used to establish an approximate multivariate nonlinear functional relationship between the reservoir parameters and the significantly correlated seismic attributes.

[0081] Step 200: Screen the plurality of regression functions to determine a second regression function.

[0082] Calculate the multiple correlation coefficients of the plurality of regression functions in sequence, and determine the second regression function according to the magnitudes of the negative correlation coefficients. Specifically, the regression function corresponding to the largest negative correlation coefficient is the second regression function.

[0083] Step 300: Conduct a significance test on the second regression function to optimize the seismic parameters.

[0084] Specifically, a non - linear regression analysis method of "step - by - step introduction" and "step - by - step elimination" is adopted to conduct a correlation significance test on seismic attributes, and seismic attributes with high correlation with reservoir prediction parameters are optimized. Further, according to the magnitude of the influence of variable x i (i = 1, 2,..., s) on reservoir parameter y, they are successively introduced into the regression equation. At the same time, each variable introduced into the regression equation also needs to be tested one by one, and variables with insignificant influence on y are promptly eliminated. Proceed in this way until there are no variables with significant influence on y that can be introduced into the regression equation, and there are no variables with insignificant influence on y in the regression equation that are eliminated. At this time, the regression equation only contains variables with significant influence on y.

[0085] As can be seen from the above description, for the seismic parameter optimization method based on multiple non - linear step - by - step regression provided by the embodiments of the present invention, first, multiple first regression functions are established according to the target predicted reservoir parameters and multiple seismic parameters; then, multiple regression functions are screened to determine the second regression function; finally, a significance test is conducted on the second regression function to optimize the seismic parameters. The present invention, based on the reservoir prediction parameters and the actual observed values of seismic attributes, adopts a non - linear regression analysis method of "step - by - step introduction" and "step - by - step elimination" to conduct a correlation significance test on seismic attributes, optimizes seismic attributes with high correlation with reservoir prediction parameters, and on this basis, establishes an approximate multiple non - linear regression model between the two. Then, during the model establishment process, a statistical analogy graph of regression predicted values and actual values is drawn to judge the correctness of the regression equation and the accuracy of the prediction result. Finally, based on the established reliable reservoir prediction model, effective prediction of the target reservoir parameters is realized.

[0086] In one embodiment, referring to Figure 2 , step 100 specifically includes:

[0087] Step 101: Use the non - linear regression analysis method to establish multiple multiple non - linear regression functions between the target predicted reservoir parameters and the multiple seismic parameters.

[0088] Separate regression equations of x i (i = 1, 2,..., m) (independent variable, equivalent to seismic parameter) on y (dependent variable, equivalent to target predicted reservoir parameter) are established, denoted as:

[0089]

[0090] In one embodiment, referring to Figure 3 , step 200 specifically includes:

[0091] Step 201: Calculate the error between the function values of the multiple multivariate non-linear regression functions and the true values;

[0092] It can be understood that in multiple regression analysis, the regression equation is not a linear function. The dependent variable in the equation is a random variable, and the relationship between it and other variables (ordinary variables) in the equation is called a multiple non-linear regression relationship.

[0093] Step 202: Determine the screening of the multiple multivariate non-linear regression functions according to multiple errors to determine the second regression function.

[0094] In Step 201 and Step 202, based on Step 101, calculate R (multiple correlation coefficient) one by one to test the significance of each multiple non-linear regression function, select the most significant regression equation from them, and introduce the corresponding independent variable (seismic parameter) into the regression equation. The formula for calculating the multiple correlation coefficient is as follows:

[0095]

[0096] In the formula, is the sum of squared deviations between n sets of observed values y k and the regression values , mainly reflecting the deviation caused by the regression model. It is the sum of squared differences between n sets of regression values and the average value of y , reflecting the fluctuation of y caused by the change of the independent variable x i (i = 1, 2,..., r), called the regression sum of squares. The value range of the multiple correlation coefficient R is [0, 1]. The closer R is to 1, the higher the significance of the regression equation; conversely, the worse the significance of the regression equation, and the regression equation obtained from the observed values has no practical significance.

[0097] In one embodiment, referring to Figure 4 , Step 300 specifically includes:

[0098] Step 301: Perform iterative operations:

[0099] Delete the seismic parameters included in the second regression function from the multiple seismic parameters to generate a preferred set of seismic parameters;

[0100] Substitute the seismic parameters in the preferred set of seismic parameters into the regression function in the previous iterative operation;

[0101] Using the non-linear regression analysis method, calculate the coefficients of the seismic parameters substituted in this iterative operation in the regression function after substitution;

[0102] Determine the regression function for the current iteration operation according to the coefficients until the number of seismic parameters in the preferred set of seismic parameters is 0.

[0103] Specifically, assume that among s variables, m (m < s) variables are introduced into the regression equation, and the following regression equation can be established:

[0104]

[0105] The above equation contains linear terms and quadratic terms. If the quadratic terms are regarded as new variables, then this equation is a linear regression equation, and this equation contains independent variables, which can be denoted as x i (i = 1, 2,..., r). To determine whether this equation is representative and the influence of each linear term and quadratic term on the dependent variable, a significance test can be performed on the regression equation, linear terms, and quadratic terms. The significance of the regression equation can be reflected by the multiple correlation coefficient R:

[0106]

[0107] In the formula, is the sum of squared deviations between n sets of observed values y k and the regression values , which mainly reflects the deviation caused by the regression model. It is the sum of squared differences between n sets of regression values and the average value of y , which reflects the fluctuation of y caused by the change of the independent variable x i (i = 1, 2,..., r), and is called the regression sum of squares. The value range of the multiple correlation coefficient R is [0, 1]. The closer R is to 1, the higher the significance of the regression equation; conversely, the worse the significance of the regression equation, and the regression equation obtained from the observed values has no practical significance.

[0108] For each independent variable, the statistic F r can be used for significance testing.

[0109]

[0110] Among them For a given test level F α , if F i > F α , then it is considered that the variable x i has a significant effect on y. If F i ≤F α , then it is considered that the variable x i has an insignificant effect on y, and the variable x i should not enter the regression equation. For the variable x iCheck one by one for (i = 1, 2,..., r), screen out the variables that have a significant effect on y, and re - perform regression analysis to establish a simpler and more effective regression equation.

[0111] In one embodiment, referring to Figure 5 , substituting the seismic parameters in the optimized set of seismic parameters into the regression function in the previous iteration operation in step 301 includes:

[0112] Step 3011: Multiply the seismic parameters in the optimized set of seismic parameters by the seismic parameters in the regression function in the previous iteration operation;

[0113] Step 3012: Generate the independent variable of the regression function in this iteration operation according to the multiplied seismic parameters.

[0114] Here, the quadratic term can be regarded as a new variable, that is, there are r independent variables in the regression equation. That is, the multiplied seismic parameters are used as the independent variable of the regression function in this iteration operation.

[0115] In one embodiment, referring to Figure 6 , the method for optimizing seismic parameters based on multivariate non - linear step - by - step regression further includes:

[0116] Step 400: Establish a prediction function for the target predicted reservoir parameters according to the optimized seismic parameters;

[0117] Step 500: Predict the reservoir of the target work area according to the prediction function.

[0118] For step 400 and step 500, during the model establishment process, draw a statistical analogy graph of the regression predicted value and the actual value to judge the correctness of the prediction function and the accuracy of the prediction result. Finally, based on the established reliable reservoir prediction model, effectively predict the target reservoir parameters.

[0119] As can be seen from the above description, the embodiment of the present invention provides a method for optimizing seismic parameters based on multivariate non - linear step - by - step regression. Aiming at the problem in the background technology of lacking comprehensive multi - parameter and quantitative division methods for paleokarst micro - landforms, referring to Ruhe's classic slope position classification system in geomorphology, four parameters of slope, profile curvature, relative elevation, and aspect ratio are optimized, the quantitative division criteria for each parameter are determined, and combined with the gully automatic extraction algorithm, eight types of paleokarst micro - landform units are divided, thus realizing the quantitative division of comprehensive multi - parameters and micro - landform units, making the division of paleokarst micro - landform units more systematic and reasonable, and having a large applicable range.

[0120] To further illustrate the solution, the present invention also provides a specific application example of the method for optimizing seismic parameters based on multivariate non - linear step - by - step regression.

[0121] It can be understood that in reservoir prediction, the relationship between reservoir parameters and seismic attributes is complex and there is a certain degree of correlation, but there is no definite functional relationship. Such variables with an uncertain relationship are called correlated variables. This specific application example studies the establishment of an approximate functional relationship between reservoir parameters and seismic attributes using the non-linear stepwise regression analysis method. This functional relationship can reflect the correlation degree of seismic attributes with reservoir parameters and can also judge the accuracy of prediction results. Specifically, this specific application example includes the following contents. See Figure 7 and Figure 8 .

[0122] S1: Load seismic result data.

[0123] S2: Select the seismic attribute prediction object y.

[0124] S3: Select seismic attributes.

[0125] Specifically, select r seismic attributes, which can be regarded as r variables. The quadratic combination of all attributes is regarded as a new variable, then the number of variables is denoted as x i (i = 1, 2,..., m).

[0126] S4: Extract observation values.

[0127] Extract n groups of observation values of y and x i (i = 1, 2,..., m) (x 1k , x 2k ,..., x mk , y k )(k = 1, 2,..., n).

[0128] S5: Conduct stepwise regression analysis based on the observation values to establish the regression equation of y and x i (i = 1, 2,..., m).

[0129] Furthermore, step S5 also includes the following steps:

[0130] S51: Respectively establish the regression equations of x i (i = 1, 2,..., m) for y, denoted as:

[0131]

[0132] Then calculate the significance of R i test one by one, select the most significant regression equation from them and introduce the variable x corresponding to i into the regression equation. Without loss of generality, assume x i is x1.

[0133] S52: Compare the values of (x1, x2), (x1, x3), ..., (x1, x3) one by one. m ) regression equation, add variable x to the test equation i (i≠1) after that, x i Is the coefficient of significantly different from 0? Then select the most significant equation among the regression equations that are significantly different from 0. And The corresponding variable x i Then introduce the regression equation, and let x introduced in the second step be i is x2.

[0134] S53: Based on the introduction of x1 and x2 in the regression equation, add variables x one by one i (i = 3, 4, ..., m), and check whether the regression equation after introducing the new variable is significantly improved compared with the regression equation containing variables x1 and x2. If so, introduce new variables and calculate the F of each variable. i , eliminating insignificant variables. Repeat this process until there are no more variables that can be introduced and eliminated, and then the final regression equation is established.

[0135] S6: Calculate the predicted value based on the stepwise regression equation;

[0136] S7: Draw a graph comparing the regression prediction value and the actual value;

[0137] S8: Judgment on the rationality and accuracy of seismic attribute reservoir prediction results;

[0138] S9: Conduct reservoir prediction for the entire area.

[0139] Based on the above method, this specific application example also provides a seismic parameter optimization method based on multivariate nonlinear stepwise regression with a specific work area as an example.

[0140] There are 11 exploration wells in the work area. Based on the work area data, the reservoir thickness of multiple sets of oil-producing layers is predicted.

[0141] (1) F1S oil reservoir prediction.

[0142] For the F1S reservoir, the average instantaneous amplitude, average frequency and total absolute amplitude are selected for prediction. By performing quadratic regression analysis on the selected attributes, the test level F is given. α =0.004, see Table 1, and obtain the multivariate quadratic stepwise regression equation with a multiple correlation coefficient of R=0.960:

[0143] y = -865.621 + 1441.568x2 - 0.005x3 - 0.076x1x2

[0144] where y is the reservoir thickness, x1 is the average instantaneous amplitude, x2 is the average frequency, and x3 is the total absolute amplitude.

[0145] Table 1 Data Sheet of F1S Oil Reservoir Samples

[0146]

[0147]

[0148] Figure 9 It is the statistical analogy graph of the measured value and the regression value of the reservoir thickness of the F1S oil reservoir. The blue diagonal line is the 45° line. The closer all coordinate points are to the diagonal line, the closer the predicted regression value is to the measured value, and the more representative the regression equation is. Figure 10 It is the prediction result graph of the reservoir thickness of the F1S oil reservoir.

[0149] (2) Prediction of the F1Z oil reservoir.

[0150] For the F1Z oil reservoir, we select three attributes: average instantaneous phase, root mean square amplitude, and average energy. Given the test level F α = 0.01. Referring to Table 2, we obtain the following multiple quadratic stepwise regression equation with a multiple correlation coefficient of R = 0.960:

[0151]

[0152] where y is the reservoir thickness, x1 is the average instantaneous phase, x2 is the root mean square amplitude, and x3 is the average energy.

[0153] Table 2 Data Sheet of F1Z Oil Reservoir Samples

[0154]

[0155] Figure 11 It is the statistical analogy graph of the measured value and the regression value of the reservoir thickness of the F1Z oil reservoir, Figure 12 It is the prediction result graph of the reservoir thickness of the F1Z oil reservoir.

[0156] (3) Prediction of the F1X oil reservoir.

[0157] For the F1X oil reservoir, we select three attributes: average instantaneous phase, average absolute amplitude, and total energy for reservoir prediction. Referring to Table 3, given the test level F α = 0.08, we obtain a quadratic regression equation with a correlation coefficient of 0.992:

[0158]

[0159] where x1 is the average instantaneous phase, x2 is the average absolute amplitude, and x3 is the overall energy.

[0160] Table 3 Data Sheet of F1X Oil Reservoir Samples

[0161]

[0162] Figure 13 is the statistical analogy graph of the measured value and the regression value of the reservoir thickness of the F1X oil reservoir, Figure 14 is the prediction result graph of the reservoir thickness of the F1X oil reservoir.

[0163] In summary, for the 3 oil reservoirs in this work area, the method of stepwise regression analysis of seismic attributes is used to predict the reservoir thickness. Judging from the multiple correlation coefficient of the multiple nonlinear regression equation, the conformity degree between the predicted value and the measured value of the work area samples, and the statistical analogy graph, this method has good reservoir prediction effect and high prediction accuracy. The present invention is applicable to the prediction objects of different work areas and different reservoirs, has universal applicability, has achieved good economic benefits in actual production applications, and has very good practical value.

[0164] In summary, the seismic parameter optimization method based on multiple nonlinear stepwise regression provided by this specific application example has the following beneficial effects:

[0165] (1) Through parameters R and F i judge whether there is a correlation between the reservoir prediction object y and the seismic attributes and quadratic combinations x i (i = 1, 2,..., p; p ≤ m), and if so, a multiple nonlinear function expression representing the correlation between them can be established;

[0166] (2) Through the established multiple nonlinear regression equation model, the reservoir parameter y can be predicted, and the accuracy of the prediction result can be known;

[0167] (3) Through the stepwise regression analysis process, it can be determined which seismic attributes and quadratic combinations have a significant effect on reservoir prediction and the object y, and which have an insignificant effect on y, thereby simplifying the research model.

[0168] Based on the same inventive concept, an embodiment of the present application further provides a seismic parameter optimization device based on multiple nonlinear stepwise regression, which can be used to implement the method described in the above embodiment, as described in the following embodiment. Since the principle of solving problems by the seismic parameter optimization device based on multiple nonlinear stepwise regression is similar to that of the seismic parameter optimization method based on multiple nonlinear stepwise regression, the implementation of the seismic parameter optimization device based on multiple nonlinear stepwise regression can refer to the implementation of the seismic parameter optimization method based on multiple nonlinear stepwise regression, and the repeated parts will not be elaborated. Hereinafter, the term "unit" or "module" can be a combination of software and / or hardware that can achieve a predetermined function. Although the systems described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0169] An embodiment of the present invention provides a specific implementation manner of a seismic parameter optimization device based on multiple nonlinear stepwise regression that can implement the seismic parameter optimization method based on multiple nonlinear stepwise regression, see Figure 15 , the seismic parameter optimization device based on multiple nonlinear stepwise regression specifically includes the following contents:

[0170] The first function establishment unit 10 is used to establish a plurality of first regression functions according to the target predicted reservoir parameters and a plurality of seismic parameters;

[0171] The second function determination unit 20 is used to screen the plurality of regression functions to determine the second regression function;

[0172] The seismic parameter optimization unit 30 is used to perform a significance test on the second regression function to optimize the seismic parameters.

[0173] In one embodiment, the first function establishment unit 10 is specifically used to establish a plurality of multiple nonlinear regression functions between the target predicted reservoir parameters and the plurality of seismic parameters by using a nonlinear regression analysis method.

[0174] In one embodiment, see Figure 16 , the second function determination unit 20 includes:

[0175] The error calculation module 201 is used to calculate the error between the function values of the plurality of multiple nonlinear regression functions and the true values;

[0176] The second function determination module 202 is used to screen the plurality of multiple nonlinear regression functions according to a plurality of errors to determine the second regression function.

[0177] In one embodiment, see Figure 17 , the seismic parameter optimization unit 30 includes:

[0178] Iterative module 301, for performing iterative operations:

[0179] Delete the seismic parameters included in the second regression function from the multiple seismic parameters to generate a seismic parameter priority set;

[0180] Substitute the seismic parameters in the seismic parameter priority set into the regression function in the previous iterative operation;

[0181] Using the non-linear regression analysis method, calculate the coefficients of the seismic parameters substituted in this iterative operation in the regression function after substitution;

[0182] Determine the regression function of this iterative operation according to the coefficients until the number of seismic parameters in the seismic parameter priority set is 0.

[0183] In one embodiment, refer to Figure 18 , the iterative module 301 includes:

[0184] Parameter multiplication module 3011, for multiplying the seismic parameters in the seismic parameter priority set by the seismic parameters in the regression function in the previous iterative operation;

[0185] This independent variable generation module 3012, for generating the independent variable of the regression function in this iterative operation according to the multiplied seismic parameters.

[0186] In one embodiment, refer to Figure 19 , the seismic parameter optimization device based on multivariate non-linear stepwise regression further includes:

[0187] Prediction function establishment unit 40, for establishing a prediction function for the target predicted reservoir parameters according to the optimized seismic parameters;

[0188] Reservoir prediction unit 50, for predicting the reservoir of the target work area according to the prediction function.

[0189] As can be seen from the above description, the seismic parameter optimization device based on multivariate non-linear stepwise regression provided by the embodiments of the present invention first establishes a plurality of first regression functions according to the target predicted reservoir parameters and a plurality of seismic parameters; then, screens the plurality of regression functions to determine the second regression function; and finally performs a significance test on the second regression function to optimize the seismic parameters. The present invention uses a non-linear regression analysis method of "stepwise introduction" and "stepwise elimination" to perform a significance test on the correlation of seismic attributes based on the reservoir prediction parameters and the actual observed values of seismic attributes, optimizes the seismic attributes with a large correlation with the reservoir prediction parameters, and on this basis, establishes an approximate multivariate non-linear regression model between the two. Then, during the model establishment process, a statistical analogy graph of the regression predicted value and the actual value is drawn to judge the correctness of the regression equation and the accuracy of the prediction result. Finally, based on the established reliable reservoir prediction model, the effective prediction of the target reservoir parameters is realized.

[0190] The devices, modules or units illustrated in the above embodiments may be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is an electronic device. Specifically, the electronic device may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0191] In a typical example, the electronic device specifically includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the above dynamic buried point method based on the front-end framework are implemented. The steps include:

[0192] Step 100: Establish a plurality of first regression functions according to the target predicted reservoir parameters and a plurality of seismic parameters;

[0193] Step 200: Screen the plurality of regression functions to determine the second regression function;

[0194] Step 300: Perform a significance test on the second regression function to optimize the seismic parameters.

[0195] Next, refer to Figure 20 , which shows a schematic structural diagram of an electronic device 600 suitable for implementing the embodiments of the present application.

[0196] As Figure 20As shown, the electronic device 600 includes a central processing unit (CPU) 601, which can perform various appropriate operations and processes according to programs stored in a read-only memory (ROM) 602 or programs loaded from a storage section 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the system 600 are also stored. The CPU 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0197] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as required. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as required so that a computer program read therefrom can be installed in the storage section 608 as required.

[0198] In such an embodiment, the computer program can be downloaded and installed from the network via the communication section 609, and / or installed from the removable medium 611.

[0199] Computer-readable media includes both permanent and non-permanent, removable and non-removable media and can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transitory medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0200] For convenience of description, the above-described apparatus is described by functionally dividing it into various units. Of course, when implementing the present application, the functions of the various units can be implemented in one or more software and / or hardware.

[0201] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to work in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device, the instruction device implementing the functions specified in one process Figure 1 or more processes and / or blocks Figure 1 or more blocks.

[0202] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, such that a series of operation steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one process Figure 1 or more processes and / or blocks Figure 1 or more blocks.

[0203] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can take the form of an all-hardware embodiment, an all-software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0204] The above are only the embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for optimizing earthquake parameters based on multivariate nonlinear stepwise regression, characterized in that: include: Establishing a plurality of first regression functions according to target predicted reservoir parameters and a plurality of seismic parameters; screening the plurality of regression functions to determine a second regression function; performing a significance test on the second regression function to optimize the earthquake parameters; The establishing of a plurality of first regression functions according to target predicted reservoir parameters and a plurality of seismic parameters comprises: Using a nonlinear regression analysis method, establishing a plurality of multivariate nonlinear regression functions between the target predicted reservoir parameters and the plurality of seismic parameters; The screening of the plurality of regression functions to determine the second regression function comprises: Calculating errors between function values of the multiple multivariate nonlinear regression functions and true values; The multiple multivariate nonlinear regression functions are screened according to the multiple errors to determine the second regression function, specifically: Calculate the multiple correlation coefficients R of multiple regression functions one by one to test the significance of each multivariate nonlinear regression function, select the most significant regression equation, and assign the corresponding independent variable x to the regression equation. i , i=1, 2, …, m is introduced into the regression equation, where the independent variable is the earthquake parameter, and the complex correlation coefficient R is calculated as follows: In the formula, is the sum of squared deviations between n groups of observed values y k and the regression values , which is used to reflect the deviation caused by the regression model; y is the dependent variable and is the target predicted reservoir parameter; is the sum of squared differences between n groups of regression values and the mean value of y , which is used to reflect the fluctuation of y caused by the change of independent variables x i , i = 1, 2, …, r; the value range of the multiple correlation coefficient R is [0, 1]. The closer R is to 1, the higher the significance of the regression equation; on the contrary, the worse the significance of the regression equation, and the regression equation obtained from the observed values has no practical significance; The performing a significance test on the second regression function to optimize the earthquake parameters includes: Perform iterative operations: Deleting the seismic parameters included in the second regression function from the plurality of seismic parameters to generate a priority set of seismic parameters; Substituting the seismic parameters in the seismic parameter optimization set into the regression function in the last iterative operation; Using nonlinear regression analysis methods, the coefficients of the earthquake parameters substituted in this iteration operation are calculated in the regression function after substitution; The regression function of this iterative operation is determined according to the coefficient until the number of seismic parameters in the seismic parameter priority set is 0, specifically: Suppose among s variables, there are m, m <s个变量引入到回归方程中,建立下面的回归方程: The above equation contains a linear term and a quadratic term. If the quadratic term is regarded as a new variable, then the equation is a linear regression equation. The equation contains An independent variable, denoted as x i , i = 1, 2, ..., r, in order to determine whether the equation is representative and the effect of each linear term and quadratic term on the dependent variable, a significance test is performed on the regression equation and the linear term and quadratic term; the significance of the regression equation is reflected by the multiple correlation coefficient R: In the formula, is the sum of squared deviations between n sets of observed values y k and the regression values , which reflects the deviation caused by the regression model; It is the sum of squared differences between n sets of regression values and the mean value of y , which reflects the fluctuation of y caused by the change of the independent variable x i , i = 1, 2, …, r, and is called the regression sum of squares; Use the statistic F for each independent variable r to conduct a significance test: in For a given test level F α , if F i >F α , then the variable x i The effect on y is significant. If F i ≤F α , then the variable x i The effect on y is not significant, the variable x i Should not enter the regression equation; for variable x i , i=1,2,…,r are tested one by one, and the variables with significant effects on y are screened out and re-regression analysis is performed to establish a simpler and more effective regression equation.

2. The earthquake parameter optimization method according to claim 1, characterized in that: Substituting the seismic parameters in the preferred set of seismic parameters into the regression function in the last iterative operation includes: multiplying the seismic parameters in the seismic parameter optimization set by the seismic parameters in the regression function in the last iterative operation; The independent variable of the regression function in this iterative operation is generated according to the multiplied earthquake parameters.

3. The seismic parameter optimization method according to claim 1, characterized in that Also includes: Establishing a prediction function for the target predicted reservoir parameters according to the optimized seismic parameters; The reservoir of the target work area is predicted according to the prediction function.

4. An earthquake parameter optimization device based on multivariate non-linear stepwise regression, characterized in that, include: A first function establishing unit, configured to establish a plurality of first regression functions according to target predicted reservoir parameters and a plurality of seismic parameters; a second function determining unit, configured to screen the plurality of regression functions to determine a second regression function; an earthquake parameter optimization unit, configured to perform a significance test on the second regression function to optimize the earthquake parameters; The first function establishing unit is specifically configured to establish a plurality of multivariate nonlinear regression functions between the target predicted reservoir parameters and the plurality of seismic parameters using a nonlinear regression analysis method; The second function determining unit includes: An error calculation module, used to calculate the error between the function values of the multiple nonlinear regression functions and the true values; The second function determination module is configured to screen the multiple multivariate nonlinear regression functions according to the multiple errors to determine the second regression function, specifically: Calculate the multiple correlation coefficients R of multiple regression functions one by one to test the significance of each multivariate nonlinear regression function, select the most significant regression equation, and assign the corresponding independent variable x to the regression equation. i , i=1,2,...,m are introduced into the regression equation, wherein the independent variable is the earthquake parameter, and the complex correlation coefficient R is calculated as follows: In the formula, is n sets of observations y k and regression value The sum of squared deviations between is used to reflect the deviation caused by the regression model; y is the dependent variable, which is the target predicted reservoir parameter; is n sets of regression values and the mean of y The sum of squares of the differences between i , i = 1, 2, ..., the fluctuation of y caused by the change of r; the value range of the multiple correlation coefficient R is [0, 1]. The closer R is to 1, the higher the significance of the regression equation; conversely, the lower the significance of the regression equation, the less practical significance the regression equation obtained from the observed values; The earthquake parameter optimization unit includes: Iteration module, used for iterative operations: Deleting the seismic parameters included in the second regression function from the plurality of seismic parameters to generate a priority set of seismic parameters; Substituting the seismic parameters in the seismic parameter optimization set into the regression function in the last iterative operation; Using nonlinear regression analysis methods, the coefficients of the earthquake parameters substituted in this iteration operation are calculated in the regression function after substitution; The regression function of this iterative operation is determined according to the coefficient until the number of seismic parameters in the seismic parameter priority set is 0, specifically: Suppose among s variables, there are m,m <s个变量引入到回归方程中,建立下面的回归方程: The above equation contains linear terms and quadratic terms. Regarding the quadratic terms as new variables, the equation becomes a linear regression equation, which contains independent variables, denoted as x i , where i = 1, 2,..., r. To determine whether the equation is representative and the effects of each linear term and quadratic term on the dependent variable, significance tests are performed on the regression equation, linear terms, and quadratic terms. The significance of the regression equation is reflected by the multiple correlation coefficient R: In the formula, is n sets of observations y k and regression value The sum of squared deviations between and reflects the deviation caused by the regression model; It is n sets of regression values and the mean of y The sum of the squares of the differences between the independent variable x i , i = 1, 2, ..., the fluctuation of y caused by the change of r is called the regression sum of squares; Use the statistic F for each independent variable r Conduct a significance test: Among them For a given test level F α , if F i > F α , it is considered that the effect of variable x i on y is significant. If F i ≤F α , it is considered that the effect of variable x i on y is not significant, and variable x i should not enter the regression equation; for each variable x i , i = 1, 2,..., r, conduct tests one by one, select the variables with significant effects on y, and re - conduct regression analysis to establish a simpler and more effective regression equation.

5. The earthquake parameter optimization device according to claim 4, characterized in that The iteration module includes: a parameter multiplication module, configured to multiply the seismic parameters in the seismic parameter optimization set by the seismic parameters in the regression function in the last iterative operation; The current independent variable generation module is used to generate the independent variable of the regression function in the current iterative operation according to the multiplied earthquake parameters.

6. The seismic parameter optimization device according to claim 4, wherein Also includes: A prediction function establishing unit, configured to establish a prediction function for the target predicted reservoir parameters according to the optimized seismic parameters; The reservoir prediction unit is used to predict the reservoir of the target work area according to the prediction function.

7. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, the steps of the seismic parameter optimization method based on multivariate nonlinear stepwise regression described in any one of claims 1 to 3 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the earthquake parameter optimization method based on multivariate nonlinear stepwise regression described in any one of claims 1 to 3 are implemented.

Images