Method, apparatus and computing device for determining formation property values

Abstract

The application discloses a stratum attribute value determination method, device and computing equipment. The method comprises the following steps: acquiring a first number of groups of adjacent well stratum attribute value distribution data of a target well logging, and generating a first number of attribute value reference sequences; using a random algorithm to generate a derived sequence set under the constraint of each attribute value reference sequence of any inversion stratum depth, each derived sequence set comprising a second number of attribute value derived sequences; taking any derived sequence set as an initial model vector, using a random gradient evolution hybrid optimization algorithm to perform inversion on the well logging data of the inversion stratum depth to obtain an inversion result set, each inversion result set comprising a third number of inversion results; and determining a target inversion result corresponding to the inversion stratum depth from the inversion results. The scheme can improve the inversion accuracy and efficiency of the stratum attribute value.

Description

Technical Field

[0001] This application relates to the field of drilling data processing technology, specifically to a method, apparatus, computing device, computer storage medium, and computer program product for determining formation attribute values. Background Technology

[0002] Formation attribute inversion is the process of inferring the attribute values ​​(such as resistivity, formation thickness, etc.) of an underground formation model from observation data such as well logging data.

[0003] Existing technologies typically employ a single inversion method, such as gradient algorithms, evolutionary algorithms, or artificial intelligence algorithms. However, gradient algorithms are sensitive to initial conditions; improper initial value selection can lead to slower convergence or even failure to converge, and they are prone to getting trapped in local optima. Evolutionary algorithms are global optimization algorithms, but they require extensive evaluation of the objective function, resulting in huge computational costs. Artificial intelligence algorithms often require large amounts of sample data for training, leading to high training costs and low efficiency. Furthermore, gradient algorithms and artificial intelligence algorithms often lack sufficient inversion accuracy in certain scenarios. Summary of the Invention

[0004] This application provides a method, apparatus, computing device, computer storage medium, and computer program product for determining stratigraphic attribute values ​​that overcomes or at least partially solves existing problems.

[0005] According to a first aspect of this application, a method for determining stratigraphic attribute values ​​is provided, comprising:

[0006] Obtain the distribution data of formation attribute values ​​of the first number of adjacent wells in the target well logging, and generate a first number of attribute value reference sequences based on the distribution data of formation attribute values ​​of the adjacent wells.

[0007] For any inverted stratigraphic depth, a random algorithm is used to generate a set of derived sequences for that inverted stratigraphic depth under the constraints of each attribute value reference sequence; wherein each set of derived sequences contains a second number of attribute value derived sequences;

[0008] For any derived sequence set of the inverted formation depth, the derived sequence set is used as the initial model vector, and the stochastic gradient evolution hybrid optimization algorithm is used to invert the well logging data of the inverted formation depth to obtain the inversion result set corresponding to the derived sequence set; wherein, each inversion result set contains a third number of inversion results;

[0009] The target inversion result corresponding to the inversion stratigraphic depth is determined from the inversion results.

[0010] In one optional implementation, the step of using the derived sequence set as the initial model vector and employing a stochastic gradient evolutionary hybrid optimization algorithm to invert the well logging data for the inverted formation depth to obtain the inversion result set corresponding to the derived sequence set includes:

[0011] Construct the inversion objective function;

[0012] The derived sequence set is used as the initial population, and each individual in the initial population is used as the initial model vector for iteration;

[0013] For any individual in any iteration step, the evolutionary solution for the current iteration step is generated based on the optimal solution of the previous iteration step, the optimal individual in the previous iteration step, the worst individual in the previous iteration step, and the first random number; wherein, the optimal individual in the previous iteration step and the worst individual in the previous iteration step are determined based on the objective function values ​​corresponding to the optimal solutions of each individual in the previous iteration step.

[0014] And for any individual in any iteration step, generate a random solution corresponding to the current iteration step based on the best individual in the previous iteration step, the worst individual in the previous iteration step, and the second random number;

[0015] A preset acceptance probability algorithm is used to select candidate solutions for the current iteration step from the evolutionary solutions and the random solutions;

[0016] For the same individual, if the objective function value of the candidate solution in the current iteration step is less than the optimal solution in the previous iteration step, then the candidate solution in the current iteration step is taken as the optimal solution in the current iteration step; if the objective function value of the candidate solution in the current iteration step is greater than or equal to the optimal solution in the previous iteration step, then the optimal solution in the previous iteration step is taken as the optimal solution in the current iteration step.

[0017] When the preset iteration termination condition is met, the iteration ends; the inversion result set is generated based on the optimal solution of the current iteration step for each individual after the iteration ends.

[0018] In one optional implementation, determining the target inversion result corresponding to the inversion stratigraphic depth from the inversion result includes:

[0019] For each set of inversion results, a fourth number of candidate inversion results are selected from the set based on the similarity between the inversion results in the set and the corresponding attribute value reference sequence.

[0020] Based on the sorting results of the inversion residuals corresponding to each candidate inversion result, the target inversion result corresponding to the inversion stratigraphic depth is determined from the candidate inversion results.

[0021] In an optional implementation, after determining the target inversion result corresponding to the inversion stratigraphic depth from the inversion results, the method further includes:

[0022] Anomaly inversion depths are determined from the target inversion results corresponding to multiple inversion stratigraphic depths;

[0023] The target inversion result corresponding to the abnormal inversion stratigraphic depth is re-determined from the inversion results corresponding to the abnormal inversion stratigraphic depth.

[0024] In one optional implementation, determining the anomalous inverted stratigraphic depth from the target inversion results corresponding to multiple inverted stratigraphic depths includes:

[0025] The similarity between the inversion results of any two targets is calculated using the dynamic time warping algorithm.

[0026] For any target inversion result, a similarity index value is generated based on the similarity between the target inversion result and other target inversion results.

[0027] Based on the similarity index values ​​of each target inversion result, the most similar target inversion result corresponding to the highest similarity index value is determined;

[0028] The inversion stratigraphic depth corresponding to the inversion result of the target whose similarity to the inversion result of the most similar target is lower than a preset threshold is taken as the abnormal inversion stratigraphic depth.

[0029] In an optional implementation, the method further includes:

[0030] The initial two-dimensional stratigraphic model is divided into uniformly sized pixel units;

[0031] Based on the target inversion results, determine multiple initial attribute values ​​corresponding to each pixel unit;

[0032] For any pixel unit, determine the optimized attribute value of the pixel unit based on multiple initial attribute values ​​of the pixel unit;

[0033] The final attribute value of the pixel unit is obtained by filtering the optimized attribute value of the pixel unit.

[0034] Coloring is performed based on the final attribute values ​​of the pixel units to generate a two-dimensional stratigraphic image.

[0035] According to a second aspect of this application, a stratigraphic property value determination apparatus is provided, comprising:

[0036] The reference sequence generation module is used to acquire the distribution data of formation attribute values ​​of the first number of adjacent wells of the target well logging, and generate a first number of attribute value reference sequences based on the distribution data of formation attribute values ​​of the adjacent wells.

[0037] The derived sequence generation module is used to generate a set of derived sequences for each attribute value reference sequence constraint for any inverted stratigraphic depth using a random algorithm; wherein each set of derived sequences contains a second number of attribute value derived sequences;

[0038] The inversion module is used to invert the well logging data at the inverted formation depth for any derived sequence set, using the derived sequence set as the initial model vector, and employing a stochastic gradient evolution hybrid optimization algorithm to obtain the inversion result set corresponding to the derived sequence set; wherein, each inversion result set contains a third number of inversion results;

[0039] The determination module is used to determine the target inversion result corresponding to the inversion stratum depth from the inversion results.

[0040] In one alternative implementation, the inversion module is used to: construct the inversion objective function;

[0041] The derived sequence set is used as the initial population, and each individual in the initial population is used as the initial model vector for iteration;

[0042] For any individual in any iteration step, the evolutionary solution for the current iteration step is generated based on the optimal solution of the previous iteration step, the optimal individual in the previous iteration step, the worst individual in the previous iteration step, and the first random number; wherein, the optimal individual in the previous iteration step and the worst individual in the previous iteration step are determined based on the objective function values ​​corresponding to the optimal solutions of each individual in the previous iteration step.

[0043] And for any individual in any iteration step, generate a random solution corresponding to the current iteration step based on the best individual in the previous iteration step, the worst individual in the previous iteration step, and the second random number;

[0044] A preset acceptance probability algorithm is used to select candidate solutions for the current iteration step from the evolutionary solutions and the random solutions;

[0045] For the same individual, if the objective function value of the candidate solution in the current iteration step is less than the optimal solution in the previous iteration step, then the candidate solution in the current iteration step is taken as the optimal solution in the current iteration step; if the objective function value of the candidate solution in the current iteration step is greater than or equal to the optimal solution in the previous iteration step, then the optimal solution in the previous iteration step is taken as the optimal solution in the current iteration step.

[0046] When the preset iteration termination condition is met, the iteration ends; the inversion result set is generated based on the optimal solution of the current iteration step for each individual after the iteration ends.

[0047] In one optional implementation, the determining module is used to: for each set of inversion results, select a fourth number of candidate inversion results from the set of inversion results based on the similarity between the inversion results in the set of inversion results and the corresponding attribute value reference sequence;

[0048] Based on the sorting results of the inversion residuals corresponding to each candidate inversion result, the target inversion result corresponding to the inversion stratigraphic depth is determined from the candidate inversion results.

[0049] In one optional implementation, the determining module is used to: determine the anomalous inverted stratigraphic depth from the target inversion results corresponding to multiple inverted stratigraphic depths;

[0050] The target inversion result corresponding to the abnormal inversion stratigraphic depth is re-determined from the inversion results corresponding to the abnormal inversion stratigraphic depth.

[0051] In one optional implementation, the determining module is used to: calculate the similarity between any two target inversion results using a dynamic time warping algorithm;

[0052] For any target inversion result, a similarity index value is generated based on the similarity between the target inversion result and other target inversion results.

[0053] Based on the similarity index values ​​of each target inversion result, the most similar target inversion result corresponding to the highest similarity index value is determined;

[0054] The inversion stratigraphic depth corresponding to the inversion result of the target whose similarity to the inversion result of the most similar target is lower than a preset threshold is taken as the abnormal inversion stratigraphic depth.

[0055] In one alternative embodiment, the device further includes an imaging module for dividing the initial two-dimensional stratigraphic model into uniformly sized pixel units.

[0056] Based on the target inversion results, determine multiple initial attribute values ​​corresponding to each pixel unit;

[0057] For any pixel unit, determine the optimized attribute value of the pixel unit based on multiple initial attribute values ​​of the pixel unit;

[0058] The final attribute value of the pixel unit is obtained by filtering the optimized attribute value of the pixel unit.

[0059] Coloring is performed based on the final attribute values ​​of the pixel units to generate a two-dimensional stratigraphic image.

[0060] According to a third aspect of this application, a computing device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0061] The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described method for determining formation attribute values.

[0062] According to a fourth aspect of this application, a computer storage medium is provided, wherein the storage medium stores at least one executable instruction that causes a processor to perform the operation corresponding to the above-described method for determining stratigraphic attribute values.

[0063] According to a fifth aspect of this application, a computer program product is provided, comprising at least one executable instruction that causes a processor to perform operations corresponding to the above-described method for determining stratigraphic attribute values.

[0064] The formation attribute value determination method, apparatus, computing device, computer storage medium, and computer program product provided in this application generate an attribute value reference sequence based on the distribution data of formation attribute values ​​from adjacent wells. Then, an attribute value derived sequence is generated through a random algorithm. The attribute value derived sequence is used as the initial value data for inversion, thereby reducing the parameter search space and improving inversion efficiency and accuracy. Moreover, the inversion is performed using a stochastic gradient evolutionary hybrid optimization algorithm that integrates stochastic, evolutionary, and gradient ideas. This ensures the global optimality of attribute value inversion while also possessing fast convergence, achieving global reliability and local efficiency.

[0065] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0066] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0067] Figure 1 A flowchart of a method for determining stratigraphic attribute values ​​provided in Embodiment 1 of this application is shown;

[0068] Figure 2 This illustration shows a schematic diagram of a target inversion result generation process provided in Embodiment 1 of this application;

[0069] Figure 3 A flowchart of a method for determining stratigraphic attribute values ​​provided in Embodiment 2 of this application is shown;

[0070] Figure 4 A flowchart of an anomaly inversion method for determining stratigraphic depth is shown in Embodiment 2 of this application;

[0071] Figure 5 This document shows a flowchart of a stratigraphic imaging method based on target inversion results, provided in Embodiment 3 of this application.

[0072] Figure 6 A structural diagram of a formation attribute value determination device provided in Embodiment 4 of this application is shown;

[0073] Figure 7 A structural diagram of a computing device provided in Embodiment 5 of this application is shown. Detailed Implementation

[0074] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0075] Example 1

[0076] Figure 1 A flowchart illustrating a method for determining stratigraphic attribute values ​​provided in Embodiment 1 of this application is shown. Figure 1 As shown, the method specifically includes the following steps:

[0077] Step S101: Obtain the distribution data of formation attribute values ​​of the first number of adjacent wells in the target well logging, and generate a first number of attribute value reference sequences based on the distribution data of formation attribute values ​​of the adjacent wells.

[0078] To facilitate the inversion of formation attribute values, simplify the inversion process, and improve inversion efficiency, the initial formation model used for formation attribute value inversion in this application embodiment is a two-dimensional formation model. In specific implementation, well logging data from the target well can be obtained, and an initial three-dimensional formation body can be constructed using geophysical exploration data or other formation data. The design well trajectory of the target well can be obtained, and bedding plane information in the three-dimensional formation body can be extracted along the vertical plane of the design well trajectory, thereby obtaining the two-dimensional formation model.

[0079] A first number of neighboring wells are selected for the target logging, for example, N neighboring wells are identified. For each neighboring well, logging data obtained by measuring the neighboring well using various logging instruments is acquired. Then, the neighboring well logging data is inverted to obtain the distribution data of formation attribute values ​​for the neighboring wells. For example, a two-dimensional true resistivity inversion is performed on the resistivity logging data of the neighboring well to obtain the true resistivity value of each layer and the vertical thickness of the formation, thereby obtaining the distribution data of different attribute values ​​of the neighboring well, i.e., the distribution data of formation attribute values ​​for the neighboring wells. A set of neighboring well formation attribute value distribution data corresponds to a formation model. The attribute value distribution data of the neighboring wells reflects the distribution of at least one attribute value of the neighboring well in the formation. The attribute value distribution data of one neighboring well constitutes a set of neighboring well formation attribute value distribution data; thus, this application obtains a first number of sets of neighboring well formation attribute value distribution data.

[0080] The attribute values ​​described in the embodiments of this application may include one or more of the following attribute parameters: resistivity, formation depth thickness, density, natural gamma, PE value, etc.

[0081] After obtaining the distribution data of formation attribute values ​​of the first number of adjacent wells in the target well logging, for each group of adjacent well formation attribute value distribution data, a corresponding attribute value reference sequence is generated based on the distribution data of formation attribute values ​​of the adjacent wells. The attribute value reference sequence corresponds one-to-one with the distribution data of formation attribute values ​​of the adjacent wells, thereby obtaining the first number of attribute value reference sequences.

[0082] Specifically, the distribution data of formation attribute values ​​from adjacent wells typically consists of attribute values ​​at different formation depths. A corresponding attribute value reference sequence is obtained by serializing the distribution data of formation attribute values ​​from adjacent wells (e.g., sorting by depth). Optionally, during the serialization process, attribute values ​​can be normalized, and / or a factoring algorithm can be used to amplify the attribute value range, thereby obtaining the attribute value reference sequence. The attribute value reference sequence can be represented as a vector. For example, the initial model vector for 5 formation layers can be represented as [R1, R2, R3, R4, R5, H2, H3, H4, dip], where R1 to R5 are the resistivity of the 5 formation layers, H2, H3, and H4 are the layer thicknesses of the 2nd, 3rd, and 4th layers, the 1st and 5th layers are infinitely thick, and dip is the relative dip angle between the instrument and the formation.

[0083] In one optional implementation, to improve inversion efficiency and accuracy, this implementation extracts the distribution data of formation attribute values ​​from adjacent wells within a preset range, and then generates a corresponding attribute value reference sequence. The preset range can be determined as follows: First, the effective detection depth D of the drilling azimuth electromagnetic wave instrument is determined. The effective detection depth D is the maximum detection distance at which the drilling azimuth electromagnetic wave instrument can effectively identify formation interfaces or changes in physical properties. Specifically, a two-layer medium model is constructed (i.e., the model includes a high-resistivity target layer and a low-resistivity surrounding rock layer). The effective detection depth D is determined comprehensively based on the resistivity of the target layer and the surrounding rock layer, as well as a set signal threshold. For example, to acquire well logging data for the target well (specifically including logging while drilling resistivity data, logging while drilling azimuth electromagnetic wave data, logging while drilling natural gamma ray data, etc.), the maximum resistivity value in the formation is used as the resistivity value of the formation where the instrument is located, and the minimum resistivity value is used as the resistivity value of the surrounding rock. The geological signal of the logging while drilling azimuth electromagnetic wave instrument is calculated in a two-layer medium model, and the distance between the instrument and the formation corresponding to the signal threshold (the set minimum value) is used as the effective detection depth D. Then, the vertical depth range ± D meters, with the midpoint of the target formation as the midpoint, is used as the preset range.

[0084] Step S102: For any inverted stratigraphic depth, a random algorithm is used to generate a set of derived sequences for each attribute value reference sequence constraint for that inverted stratigraphic depth; wherein each set of derived sequences contains a second number of attribute value derived sequences.

[0085] This step uses a random algorithm to generate a set of derived sequences corresponding to any inverted stratigraphic depth. Each set of derived sequences contains a second set of attribute value derived sequences. These attribute value derived sequences within the same set are generated under the same attribute value reference sequence constraint. The constraint refers to the inherent magnitude relationship of the reference sequence, such as the relative magnitude of the resistivity attribute values ​​of each layer or the relative magnitude of the thickness of each layer. The relative magnitude relationship of the attribute values ​​in the derived sequence set must be consistent with the reference sequence. Different derived sequence sets correspond to different attribute value reference sequences, thus ensuring a one-to-one correspondence between the derived sequence set and the attribute value reference sequence. Therefore, at any inverted stratigraphic depth, a first set of derived sequences can be obtained.

[0086] like Figure 2 As shown, the first quantity is n, and the second quantity is m. Step S101 generates the attribute value reference sequence N1, N2, ..., N n In step S102, for the inverted stratigraphic depth D... i A random algorithm is used to generate the attribute value derived sequence M under the constraint of the attribute value reference sequence N1. 1_1 M 1_2 M 1_m Attribute value derived sequence M 1_1M 1_2 M 1_m There are m elements in total, forming a derived sequence set M1; under the constraint of the attribute value reference sequence N2, an attribute value derived sequence M is generated. 2_1 M 2_2 M 2_m Attribute value derived sequence M 2_1 M 2_2 M 2_m There are m items in total, forming a derived sequence set M2; and so on, in the attribute value reference sequence N. n Generate attribute value derived sequence M under constraints n_1 M n_2 M n_m Attribute value derived sequence M n_1 M n_2 M n_m There are m sequences in total, forming a set M of derived sequences. n .

[0087] In one optional implementation, the attribute value derived sequence is generated as follows: using the true resistivity value and formation thickness in the attribute value reference sequence as averages, a uniform distribution or normal distribution method is used to generate the corresponding attribute value derived sequence. For example, a truncated normal distribution method can be used to remove random values ​​that are more than two standard deviations away from the average value and then regenerate new random values.

[0088] In one optional implementation, the presence of formation resistivity anisotropy is determined based on the anisotropic images generated by the edge resistivity instrument during drilling. For example, if the image values ​​of two sectors differing by 90 degrees are different, anisotropy is determined to exist in the formation where the instrument is located; otherwise, anisotropy is determined not to exist. For measurement depths without anisotropy, a derived sequence of attribute values ​​corresponding to formation thickness and horizontal resistivity is generated, meaning no additional vertical resistivity attribute value is added at that logging depth. For measurement depths with anisotropy, a derived sequence of attribute values ​​corresponding to formation thickness, horizontal resistivity, and vertical resistivity is generated, thereby adding an additional vertical resistivity attribute value at that logging depth. Using this method, vertical resistivity attribute values ​​are added only to formations with anisotropy, thereby reducing data processing volume and improving overall inversion efficiency.

[0089] Step S103: For any derived sequence set of the inverted formation depth, using the derived sequence set as the initial model vector, the stochastic gradient evolution hybrid optimization algorithm is used to invert the well logging data of the inverted formation depth to obtain the inversion result set corresponding to the derived sequence set; wherein, each inversion result set contains a third number of inversion results.

[0090] The well logging data at the inverted formation depth is acquired, including measured data from various measurement points of the target well within the inverted formation depth range. This well logging data is then combined with each derived sequence set to perform inversion, resulting in a corresponding inversion result set. Each derived sequence set corresponds one-to-one with the inversion result set, thus yielding a first set of inversion result sets for each inverted formation depth. Each inversion result set contains a third set of inversion results. The inversion processes corresponding to different derived sequence sets can be executed concurrently, thereby improving inversion efficiency.

[0091] In the inversion process corresponding to any set of derived sequences, the attribute value derived sequences in the derived sequence set are used as the initial model vector, and the stochastic gradient evolutionary hybrid optimization algorithm is used to invert the well logging data. The stochastic gradient evolutionary hybrid optimization algorithm is a hybrid algorithm that combines stochastic gradient and evolutionary ideas. It can achieve global reliability and local efficiency, and using the attribute value derived sequences in the derived sequence set as the initial data can improve the inversion efficiency.

[0092] In one optional implementation, using the derived sequence set as the initial model vector, the stochastic gradient evolutionary hybrid optimization algorithm is used to invert the well logging data for the inverted formation depth to obtain the inversion result set corresponding to the derived sequence set. Specifically, this includes the following steps:

[0093] S1: Construct the inversion objective function.

[0094] Specifically, the inversion objective function can be constructed based on the inversion residuals, or it can be constructed based on the measured data and the specified solution (such as preset formation parameters).

[0095] The inversion objective function can be specifically shown in Formula 1:

[0096]

[0097] Where C(x) represents the inversion objective function; r(x) represents the inversion residual; D represents the measured data (well logging data of this well); and S(x) represents the forward model response of the formation model when the formation parameter is x.

[0098] S2: Use the derived sequence set as the initial population, and use each individual in the initial population as the initial model vector for iteration.

[0099] In this embodiment, each individual is iterated continuously using an iterative algorithm. During the iteration process, in order to improve the iteration accuracy and efficiency, the set of derived sequences determined in the previous steps is used as the initial population, and each individual in the initial population is used as the iteration initial model vector of the corresponding individual.

[0100] S3: For any individual in any iteration step, generate the evolutionary solution for the current iteration step based on the optimal solution of the previous iteration step, the optimal individual in the previous iteration step, the worst individual in the previous iteration step, and the first random number; wherein, the optimal individual in the previous iteration step and the worst individual in the previous iteration step are determined based on the objective function values ​​corresponding to the optimal solutions of each individual in the previous iteration step.

[0101] After obtaining the optimal solution for each individual in the previous iteration, the optimal solution for each individual in the previous iteration is substituted into the variable x in the inversion objective function to obtain the objective function value for each individual in the previous iteration. In the previous iteration, the optimal solution of the individual with the largest objective function value is taken as the worst individual, and the optimal solution of the individual with the smallest objective function value is taken as the best individual.

[0102] This step calculates the evolutionary solution for any individual in the current iteration step. This evolutionary solution is obtained from the optimal solution of the individual in the previous iteration step, the best individual in the previous iteration step, the worst individual in the previous iteration step, and the first random number generated in this iteration, thereby determining the direction of gradient evolution.

[0103] Specifically, the evolutionary solution for the current iteration step of any individual can be calculated using the following formula 2:

[0104]

[0105] in, This represents the evolutionary solution of the i-th individual in the (j+1)-th iteration step (the current iteration step); X_j represents the optimal solution for the i-th individual in the j-th iteration step (the previous iteration step); rand1 represents the first random number in the range [0,1]; best X_j represents the optimal individual in the j-th iteration step. worst This represents the worst individual in the j-th iteration step.

[0106] This evolutionary solution combines population information (the optimal individual guides the direction of improvement) and randomness (the first random number introduces a perturbation to prevent all individuals from moving in the same direction, which would lead to premature convergence).

[0107] S4: For any individual in any iteration step, generate a random solution corresponding to the current iteration step based on the best individual in the previous iteration step, the worst individual in the previous iteration step, and the second random number.

[0108] In each iteration step, a random solution is generated for each individual in the current iteration step. Specifically, the random solution can be calculated using the following formula 3:

[0109]

[0110] in, X_j represents the random solution of the i-th individual in the (j+1)-th iteration step (the current iteration step); rand2 represents the second random number in the range [0,1]; best X_j represents the optimal individual in the j-th iteration step. worst This represents the worst individual in the j-th iteration step.

[0111] S5: Select candidate solutions for the current iteration step from evolutionary solutions and random solutions using a preset acceptance probability algorithm.

[0112] For any individual in any iteration step, a candidate solution for the current iteration step is selected from the evolutionary solution and random solution of that individual in the current iteration step using a preset acceptance probability algorithm.

[0113] Specifically, candidate solutions can be selected using the following formula 4:

[0114]

[0115] in, This represents a candidate solution for the i-th individual in the (j+1)-th iteration step; This represents the random solution for the i-th individual in the (j+1)-th iteration step; represents the evolutionary solution of the i-th individual in the (j+1)-th iteration step; rand3 represents the third random number in the range [0,1]; CR is a constant in the range [0,1].

[0116] This method can effectively avoid getting trapped in local optima and ensure the accuracy of the inversion.

[0117] S6: For the same entity, if the objective function value of the candidate solution in the current iteration step is less than the optimal solution in the previous iteration step, then the candidate solution in the current iteration step is taken as the optimal solution in the current iteration step; if the objective function value of the candidate solution in the current iteration step is greater than or equal to the optimal solution in the previous iteration step, then the optimal solution in the previous iteration step is taken as the optimal solution in the current iteration step.

[0118] For the same individual, compare the objective function values ​​of the candidate solution in the current iteration step with the optimal solution in the previous iteration step. If the objective function value of the candidate solution in the current iteration step is less than that of the optimal solution in the previous iteration step, then the candidate solution in the current iteration step is better than the optimal solution in the previous iteration step, and the candidate solution in the current iteration step for that individual is finally taken as the optimal solution in the current iteration step; otherwise, the optimal solution in the previous iteration step is used as the optimal solution in the current iteration step.

[0119] S7. When the preset iteration termination condition is met, the iteration ends; and an inversion result set is generated based on the optimal solution of the current iteration step for each individual after the iteration ends.

[0120] When the maximum number of iterations is reached or the error reaches a predetermined threshold, the current iteration ends if the preset termination condition is met.

[0121] After the iteration ends, an inversion result set is generated based on the optimal solution of the current iteration step for each individual.

[0122] The inversion result set contains r individuals, where r is the third quantity, and these r inversion results constitute an inversion result set. Here, r can be equal to m.

[0123] This implementation method, through an initial value generation method constrained by prior information stratigraphic attribute sequences, defines gradient evolution factors and random factors to balance global optimization with the evolutionary acceleration process, ensuring the global optimality of attribute value inversion while possessing fast convergence. It solves the problem of gradient-based methods easily getting trapped in local optima and improves the time-consuming nature of conventional evolutionary searches, thereby enhancing the accuracy and efficiency of stratigraphic attribute value inversion.

[0124] like Figure 2 As shown, using the attribute value derived sequences in the derived sequence set M1 as the initial population, and combining them with the well logging data of this well for inverting the formation depth, a stochastic gradient evolution hybrid optimization algorithm is used for inversion, resulting in r (third number) inversion results R. 1_1 R 1_2 ..., R 1_r The r (third quantity) inversion results R 1_1 R 1_2 ..., R 1_r The inversion result set R1 is formed; where r = m.

[0125] Accordingly, using the attribute value derived sequences in the derived sequence set M2 as the initial population, and combining them with the well logging data of this well for inverting the formation depth, a stochastic gradient evolution hybrid optimization algorithm is used for inversion, resulting in r inversion results R. 2_1 R 2_2 ..., R 2_r The r inversion results R 2_1 R 2_2 ..., R 2_r The inversion result set R2 is constructed; the derived sequence set M is constructed. n The attribute value-derived sequence in the data is used as the initial population. Combined with the well logging data of this well for inverting the formation depth, a stochastic gradient evolution hybrid optimization algorithm is used for inversion to obtain r inversion results R. n_1 R n_2 ..., R n_r The r inversion results R n_1 R n_2 ..., R n_r The inversion result set R n .

[0126] Step S104: Determine the target inversion result corresponding to the inversion stratigraphic depth from the inversion results.

[0127] Through the implementation of steps S101-S103 above, a first number of inversion result sets corresponding to the inverted stratigraphic depth can be obtained. Each inversion result set contains a third number of inversion results, thus the inverted stratigraphic depth corresponds to a first number × a third number of inversion results. Further, based on the obtained inversion results, the final target inversion result corresponding to the inverted stratigraphic depth is determined. For example, the target inversion result can be selected from the inversion results based on the inversion residuals, or the target inversion result can be obtained by fusing the inversion results.

[0128] In one optional implementation, to improve the accuracy of the target inversion results, this implementation can specifically determine the target inversion results in the following manner:

[0129] First, a screening process is performed within the inversion result set: For each inversion result set, based on the similarity between the inversion results in that set and the corresponding attribute value reference sequence, a fourth number of candidate inversion results are selected from the set, where the fourth number is less than the third number. Specifically, for each inversion result set, the inversion results in that set are normalized to obtain an inversion result sequence, and the corresponding attribute value reference sequence is determined. This inversion result set is generated based on the corresponding attribute value reference sequence. Algorithms such as dynamic time warping are used to calculate the similarity between each inversion result sequence in the set and the corresponding attribute reference sequence, and then the top q (fourth number) inversion results with the highest similarity are selected as candidate inversion results. Figure 2 As shown, the inversion result set R1 corresponds to the attribute value reference sequence N1. Therefore, calculate the R values ​​in the inversion result set R1 respectively. 1_1 Similarity to N1, R 1_2 Similarity to N1...R 1_r The similarity to N1 is used to sort the inversion results from high to low, and the inversion results with the highest similarity (q) are selected as candidate inversion results. Figure 2 China Q 1_1 Q 1_2 ...Q 1_q The candidate inversion results selected from the inversion result set R1 constitute the candidate inversion result set Q1; correspondingly, the candidate inversion result set Q2 (containing candidate inversion results Q1) is obtained from the inversion result set R2 after similarity screening. 2_1 Q 2_2 ...Q 2_q From the inversion result set R n The candidate inversion result set Q is obtained after similarity screening. n (Including candidate inversion results Q) n_1 Qn_2 ...Q n_q ).

[0130] Further inter-assembly screening is performed: based on the ranking of the inversion residuals corresponding to each candidate inversion result, the target inversion result corresponding to the inverted stratigraphic depth is determined from the candidate inversion results. Specifically, each candidate inversion result is screened out by similarity, the inversion residual of each candidate inversion result is calculated, and they are sorted in ascending order of inversion residuals. One or more candidate inversion results with the lowest inversion residuals are selected as the target inversion result.

[0131] Therefore, the formation attribute value determination method provided in this application generates an attribute value reference sequence based on the distribution data of formation attribute values ​​from adjacent wells, and then generates an attribute value derived sequence through a random algorithm. The attribute value derived sequence is used as the initial value data for inversion, thereby reducing the parameter search space and improving inversion efficiency and accuracy. Moreover, the use of a stochastic gradient evolutionary hybrid optimization algorithm that integrates random, evolutionary, and gradient ideas for inversion can ensure the global optimality of attribute value inversion while possessing fast convergence, achieving global reliability and local efficiency.

[0132] Example 2

[0133] Figure 3 A flowchart illustrating a method for determining stratigraphic attribute values ​​provided in Embodiment 2 of this application is shown. Figure 3 As shown, the method specifically includes the following steps:

[0134] Step S301: Obtain the distribution data of formation attribute values ​​of the first number of adjacent wells in the target well logging, and generate a first number of attribute value reference sequences based on the distribution data of formation attribute values ​​of the adjacent wells.

[0135] Step S302: For any inverted stratigraphic depth, determine the target inversion result corresponding to that inversion stratigraphic depth.

[0136] If there are currently k inversion stratigraphic depths (D1, D2...D...), i ...D k Then, for each inversion stratigraphic depth, the corresponding target inversion result is determined. The process for determining the target inversion result for different inversion stratigraphic depths can be referred to the relevant description in Example 1, and will not be repeated here. The process for determining the target inversion result for different inversion stratigraphic depths can be executed concurrently.

[0137] Step S303: Determine the abnormal inverted stratigraphic depth from the target inversion results corresponding to multiple inverted stratigraphic depths.

[0138] By implementing steps S301-S302, multiple target inversion results corresponding to inversion stratigraphic depths can be obtained. To improve the accuracy of the final inversion results, this embodiment further performs continuity detection on the target inversion results corresponding to multiple inversion stratigraphic depths to identify abnormal inversion stratigraphic depths that clearly do not conform to stratigraphic characteristics.

[0139] In one alternative implementation, the anomaly inversion depth can be obtained by... Figure 4 The steps shown are as follows:

[0140] S3031 uses the dynamic time warping algorithm to calculate the similarity between the inversion results of any two targets.

[0141] The dynamic time warping algorithm is used to calculate the similarity between every two target inversion results, thereby obtaining the pairwise similarity between the target inversion results.

[0142] S3032, For any target inversion result, generate a similarity index value for the target inversion result based on the similarity between the target inversion results and other target inversion results.

[0143] For each target inversion result, the similarity between it and other target inversion results is obtained. The obtained similarities are then statistically calculated using methods such as averaging and / or summing to obtain a similarity index value for that target inversion result. The similarity index value reflects the overall similarity between the target inversion result and other target inversion results.

[0144] S3033, determine the most similar target inversion result corresponding to the highest similarity index value based on the similarity index value of each target inversion result.

[0145] The maximum value among the similarity index values ​​is determined; this maximum value is the highest similarity index value. The target inversion result corresponding to the highest similarity index value is taken as the most similar target inversion result.

[0146] S3034, the inversion stratigraphic depth corresponding to the target inversion result whose similarity to the most similar target inversion result is lower than a preset threshold is taken as the abnormal inversion stratigraphic depth.

[0147] The similarity between the inversion results of other targets and the inversion results of the most similar target is obtained, and the inversion results of targets with similarity below a preset threshold are identified. The inversion stratigraphic depth corresponding to the identified target inversion results is taken as the abnormal inversion stratigraphic depth.

[0148] Step S304: Re-determine the target inversion result corresponding to the abnormal inversion stratigraphic depth from the inversion results corresponding to the abnormal inversion stratigraphic depth.

[0149] For any anomalous inversion stratigraphic depth, step S104 in Example 1 is re-executed to reselect a target inversion result from the inversion results for that anomalous inversion stratigraphic depth. For example, an inversion result different from the original target inversion result is selected from the candidate inversion results based on the inversion residual as a new target inversion result. The generated target inversion result re-enters step S303 to verify whether the updated anomalous inversion stratigraphic depth is still an anomalous inversion stratigraphic depth. If so, step S104 is continued, and this process is repeated until no anomalous inversion stratigraphic depth exists.

[0150] Therefore, the stratigraphic attribute value determination method provided in this application embodiment obtains multiple target inversion results corresponding to different inversion stratigraphic depths, thereby improving the accuracy of subsequent stratigraphic analysis based on the target inversion results; furthermore, it identifies abnormal inversion stratigraphic depths and reselects target inversion results with abnormal inversion stratigraphic depths, thereby further improving the final inversion accuracy.

[0151] Example 3

[0152] Figure 5 A flowchart of a stratigraphic imaging method based on target inversion results, provided in Embodiment 3 of this application, is shown. The process of obtaining the target inversion results can be referred to the descriptions in Embodiments 1 and 2.

[0153] Specifically, such as Figure 5 As shown, the method specifically includes the following steps:

[0154] Step S501: Divide the initial two-dimensional stratigraphic model into uniformly sized pixel units.

[0155] A preset range (target stratum center depth ± D) is determined based on the effective detection depth D, and an initial two-dimensional stratigraphic model within this range is constructed. The initial two-dimensional stratigraphic model is then divided into uniform pixel units (such as square pixel blocks), thereby converting the model into a uniformly sized pixel network.

[0156] Step S502: Determine multiple initial attribute values ​​corresponding to each pixel unit based on the target inversion results.

[0157] Through Example 1 and / or Example 2, inversion results for each target can be obtained. These results contain attribute values ​​(such as formation thickness, resistivity, etc.) at different locations. Based on the mapping relationship between pixel units and model locations, the attribute values ​​corresponding to each pixel unit can be quickly determined; these attribute values ​​are called the initial attribute values ​​corresponding to the pixel unit. A single pixel unit typically corresponds to multiple initial attribute values.

[0158] Step S503: For any pixel unit, determine the optimized attribute value of the pixel unit based on multiple initial attribute values ​​of the pixel unit.

[0159] For each pixel unit, mathematical statistics and / or cluster analysis are performed on multiple initial attribute values ​​of the pixel unit, such as determining statistical parameters such as the mean and variance of multiple initial attribute values, and / or determining the cluster center values ​​of multiple initial attribute values. Then, based on the statistical results (such as using the mean) and / or the cluster analysis results (such as the cluster center values), the optimized attribute values ​​of the pixel element are obtained, thereby improving the reliability of the attribute values ​​of the pixel element.

[0160] Step S504: After filtering the optimized attribute values ​​of the pixel unit, the final attribute values ​​of the pixel unit are obtained.

[0161] Specifically, the optimized attribute values ​​of pixel units can be subjected to vertical and / or horizontal filtering to eliminate inversion noise and enhance formation continuity. These filtering methods include, but are not limited to, multi-point smoothing filtering, Gaussian filtering, and Kalman filtering. The attribute values ​​obtained after filtering are called the final attribute values.

[0162] Step S505: Color processing is performed based on the final attribute values ​​of the pixel units to generate a two-dimensional stratigraphic image.

[0163] Based on the mapping relationship between attribute values ​​and colors, the color value of each pixel unit is determined, and then color processing is performed to obtain a two-dimensional stratigraphic image, thereby intuitively displaying stratigraphic characteristics.

[0164] Therefore, the formation imaging method provided in this application converts the initial two-dimensional formation model into a uniformly sized pixel network, and obtains the optimized attribute values ​​of the pixel units based on multiple initial attribute values ​​of the pixel units, thereby improving the reliability of the optimized attribute values; and performs longitudinal and lateral filtering on the optimized attribute values ​​to reduce image noise and ensure formation continuity, thereby improving the image quality of the final generated formation image.

[0165] Example 4

[0166] Figure 6 A structural diagram of a formation attribute value determination device provided in Embodiment 4 of this application is shown. Figure 6 As shown, the device 600 includes: a reference sequence generation module 610, a derived sequence generation module 620, an inversion module 630, and a determination module 640.

[0167] The reference sequence generation module 610 is used to acquire the distribution data of formation attribute values ​​of the first number of adjacent wells of the target well logging, and generate a first number of attribute value reference sequences based on the distribution data of formation attribute values ​​of the adjacent wells.

[0168] The derived sequence generation module 620 is used to generate a set of derived sequences for each attribute value reference sequence constraint for any inverted stratigraphic depth using a random algorithm; wherein each set of derived sequences contains a second number of attribute value derived sequences;

[0169] The inversion module 630 is used to invert the well logging data at the inverted formation depth for any derived sequence set, using the derived sequence set as the initial model vector, and employing a stochastic gradient evolution hybrid optimization algorithm to obtain the inversion result set corresponding to the derived sequence set; wherein, each inversion result set contains a third number of inversion results;

[0170] The determination module 640 is used to determine the target inversion result corresponding to the inversion stratum depth from the inversion result.

[0171] In one alternative implementation, the inversion module 630 is used to: construct an inversion objective function;

[0172] The derived sequence set is used as the initial population, and each individual in the initial population is used as the initial model vector for iteration;

[0173] For any individual in any iteration step, the evolutionary solution for the current iteration step is generated based on the optimal solution of the previous iteration step, the optimal individual in the previous iteration step, the worst individual in the previous iteration step, and the first random number; wherein, the optimal individual in the previous iteration step and the worst individual in the previous iteration step are determined based on the objective function values ​​corresponding to the optimal solutions of each individual in the previous iteration step.

[0174] And for any individual in any iteration step, generate a random solution corresponding to the current iteration step based on the best individual in the previous iteration step, the worst individual in the previous iteration step, and the second random number;

[0175] A preset acceptance probability algorithm is used to select candidate solutions for the current iteration step from the evolutionary solutions and the random solutions;

[0176] For the same individual, if the objective function value of the candidate solution in the current iteration step is less than the optimal solution in the previous iteration step, then the candidate solution in the current iteration step is taken as the optimal solution in the current iteration step; if the objective function value of the candidate solution in the current iteration step is greater than or equal to the optimal solution in the previous iteration step, then the optimal solution in the previous iteration step is taken as the optimal solution in the current iteration step.

[0177] When the preset iteration termination condition is met, the iteration ends; the inversion result set is generated based on the optimal solution of the current iteration step for each individual after the iteration ends.

[0178] In an optional implementation, the determining module 640 is configured to: for each set of inversion results, select a fourth number of candidate inversion results from the set of inversion results based on the similarity between the inversion results in the set of inversion results and the corresponding attribute value reference sequence;

[0179] Based on the sorting results of the inversion residuals corresponding to each candidate inversion result, the target inversion result corresponding to the inversion stratigraphic depth is determined from the candidate inversion results.

[0180] In one alternative implementation, the determining module 640 is used to: determine the anomalous inverted stratigraphic depth from the target inversion results corresponding to multiple inverted stratigraphic depths;

[0181] The target inversion result corresponding to the abnormal inversion stratigraphic depth is re-determined from the inversion results corresponding to the abnormal inversion stratigraphic depth.

[0182] In one optional implementation, the determining module 640 is used to: calculate the similarity between any two target inversion results using a dynamic time warping algorithm;

[0183] For any target inversion result, a similarity index value is generated based on the similarity between the target inversion result and other target inversion results.

[0184] Based on the similarity index values ​​of each target inversion result, the most similar target inversion result corresponding to the highest similarity index value is determined;

[0185] The inversion stratigraphic depth corresponding to the inversion result of the target whose similarity to the inversion result of the most similar target is lower than a preset threshold is taken as the abnormal inversion stratigraphic depth.

[0186] In an alternative embodiment, the device further includes an imaging module (not shown) for dividing the initial two-dimensional stratigraphic model into uniformly sized pixel units.

[0187] Based on the target inversion results, determine multiple initial attribute values ​​corresponding to each pixel unit;

[0188] For any pixel unit, determine the optimized attribute value of the pixel unit based on multiple initial attribute values ​​of the pixel unit;

[0189] The final attribute value of the pixel unit is obtained by filtering the optimized attribute value of the pixel unit.

[0190] Coloring is performed based on the final attribute values ​​of the pixel units to generate a two-dimensional stratigraphic image.

[0191] Therefore, the formation attribute value determination device provided in this application generates an attribute value reference sequence based on the distribution data of formation attribute values ​​of adjacent wells, and then generates an attribute value derived sequence through a random algorithm. The attribute value derived sequence is used as the initial value data for inversion, thereby improving inversion efficiency and inversion accuracy. Moreover, the use of a stochastic gradient evolution hybrid optimization algorithm for inversion can achieve global reliability and local efficiency.

[0192] Example 5

[0193] Figure 7 A structural diagram of a computing device according to Embodiment 5 of this application is shown. The specific embodiments of this application do not limit the specific implementation of the computing device.

[0194] like Figure 7 As shown, the computing device may include: a processor 702, a communication interface 704, a memory 706, and a communication bus 708.

[0195] The processor 702, communication interface 704, and memory 706 communicate with each other via communication bus 708. Communication interface 704 is used to communicate with other network elements such as clients or other servers. Processor 702 executes program 710, specifically performing the relevant steps in the above-described embodiment of the method for determining the stratigraphic attribute values ​​of the computing device.

[0196] Specifically, program 710 may include program code that includes computer operation instructions.

[0197] The processor 702 may be a central processing unit (CPU), a specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. The computing device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0198] Memory 706 is used to store program 710. Memory 706 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device. Program 710 can specifically be used to cause processor 702 to perform the operations of any of the above embodiments.

[0199] Example 6

[0200] Embodiment 6 of this application provides a non-volatile computer storage medium storing at least one executable instruction or computer program that enables a processor to perform the operation corresponding to the formation attribute value determination method in any of the above method embodiments.

[0201] Example 7

[0202] Embodiment 7 of this application provides a computer program product, which includes at least one executable instruction or computer program that enables a processor to perform the operation corresponding to the formation attribute value determination method in any of the above method embodiments.

[0203] In summary, based on the computing device, computer storage medium, and computer program product provided in this embodiment, an attribute value reference sequence is generated based on the distribution data of formation attribute values ​​from adjacent wells. Then, an attribute value derived sequence is generated through a random algorithm, and the attribute value derived sequence is used as the initial value data for inversion, thereby improving inversion efficiency and accuracy. Moreover, the use of a stochastic gradient evolution hybrid optimization algorithm for inversion can achieve global reliability and local efficiency.

[0204] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, the embodiments of this application are not directed to any particular programming language. It should be understood that the content of this application described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of this application.

[0205] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0206] Similarly, it should be understood that, in order to streamline this application and aid in understanding one or more of the various inventive aspects, features of the embodiments of this application are sometimes grouped together in a single embodiment, figure, or description thereof in the above description of exemplary embodiments of this application. However, this method of disclosure should not be construed as reflecting an intention that the claimed application requires more features than expressly recited in each claim. Rather, as reflected in the claims, the inventive aspect lies in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of this application.

[0207] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0208] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of this application and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0209] The various component embodiments of this application can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of this application. This application can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such an implementation of this application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0210] It should be noted that the above embodiments are illustrative of this application and not restrictive, and those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A method for determining stratigraphic attribute values, characterized in that, include: Obtain the distribution data of formation attribute values ​​of the first number of adjacent wells in the target well logging, and generate a first number of attribute value reference sequences based on the distribution data of formation attribute values ​​of the adjacent wells. For any inverted stratigraphic depth, a random algorithm is used to generate a set of derived sequences for that inverted stratigraphic depth under the constraints of each attribute value reference sequence; wherein each set of derived sequences contains a second number of attribute value derived sequences; For any derived sequence set of the inverted formation depth, the derived sequence set is used as the initial model vector, and the stochastic gradient evolution hybrid optimization algorithm is used to invert the well logging data of the inverted formation depth to obtain the inversion result set corresponding to the derived sequence set; wherein, each inversion result set contains a third number of inversion results; The target inversion result corresponding to the inversion stratigraphic depth is determined from the inversion result; The process of using the derived sequence set as the initial model vector and employing a stochastic gradient evolutionary hybrid optimization algorithm to invert the well logging data for the inverted formation depth to obtain the inversion result set corresponding to the derived sequence set includes: Construct the inversion objective function; The derived sequence set is used as the initial population, and each individual in the initial population is used as the initial model vector for iteration; For any individual in any iteration step, the evolutionary solution for the current iteration step is generated based on the optimal solution of the previous iteration step, the best individual in the previous iteration step, the worst individual in the previous iteration step, and a first random number. The best individual and the worst individual in the previous iteration step are determined based on the objective function values ​​corresponding to the optimal solutions of each individual in the previous iteration step. The evolutionary solution for the current iteration step for any individual is calculated using the following formula: in, Let represent the evolutionary solution of the i-th individual in the (j+1)-th iteration step; This represents the optimal solution for the i-th individual in the j-th iteration step, where the (j+1)-th iteration step is the current iteration step and the j-th iteration step is the previous iteration step; rand1 represents the first random number in the range [0,1]. This represents the optimal individual in the j-th iteration step; This represents the worst-performing individual in the j-th iteration step; And for any individual in any iteration step, generate a random solution corresponding to the current iteration step based on the best individual in the previous iteration step, the worst individual in the previous iteration step, and the second random number; A preset acceptance probability algorithm is used to select candidate solutions for the current iteration step from the evolutionary solutions and the random solutions; For the same individual, if the objective function value of the candidate solution in the current iteration step is less than the optimal solution in the previous iteration step, then the candidate solution in the current iteration step is taken as the optimal solution in the current iteration step; if the objective function value of the candidate solution in the current iteration step is greater than or equal to the optimal solution in the previous iteration step, then the optimal solution in the previous iteration step is taken as the optimal solution in the current iteration step. When the preset iteration termination condition is met, the iteration ends; the inversion result set is generated based on the optimal solution of the current iteration step for each individual after the iteration ends.

2. The method according to claim 1, characterized in that, The step of determining the target inversion result corresponding to the inversion stratigraphic depth from the inversion result includes: For each set of inversion results, a fourth number of candidate inversion results are selected from the set based on the similarity between the inversion results in the set and the corresponding attribute value reference sequence. Based on the sorting results of the inversion residuals corresponding to each candidate inversion result, the target inversion result corresponding to the inversion stratigraphic depth is determined from the candidate inversion results; The inversion objective function is shown in the following equation: Where C(x) represents the inversion objective function; r(x) represents the inversion residual; D represents the measured data; and S(x) represents the forward model response of the formation model when the formation parameter is x.

3. The method according to any one of claims 1-2, characterized in that, After determining the target inversion result corresponding to the inversion stratigraphic depth from the inversion results, the method further includes: Anomaly inversion depths are determined from the target inversion results corresponding to multiple inversion stratigraphic depths; The target inversion result corresponding to the abnormal inversion stratigraphic depth is re-determined from the inversion results corresponding to the abnormal inversion stratigraphic depth.

4. The method according to claim 3, characterized in that, The step of determining the anomalous inverted stratigraphic depth from the target inversion results corresponding to multiple inverted stratigraphic depths includes: The similarity between the inversion results of any two targets is calculated using the dynamic time warping algorithm. For any target inversion result, a similarity index value is generated based on the similarity between the target inversion results and other target inversion results. Based on the similarity index values ​​of each target inversion result, the most similar target inversion result corresponding to the highest similarity index value is determined; The inversion stratigraphic depth corresponding to the inversion result of the target whose similarity to the inversion result of the most similar target is lower than a preset threshold is taken as the abnormal inversion stratigraphic depth.

5. The method according to claim 3, characterized in that, The method further includes: The initial two-dimensional stratigraphic model is divided into uniformly sized pixel units; Based on the target inversion results, determine multiple initial attribute values ​​corresponding to each pixel unit; For any pixel unit, determine the optimized attribute value of the pixel unit based on multiple initial attribute values ​​of the pixel unit; The final attribute value of the pixel unit is obtained by filtering the optimized attribute value of the pixel unit. Coloring is performed based on the final attribute values ​​of the pixel units to generate a two-dimensional stratigraphic image.

6. A stratigraphic attribute value determination apparatus for performing the stratigraphic attribute value determination method according to claim 1, characterized in that, include: The reference sequence generation module is used to acquire the distribution data of formation attribute values ​​of the first number of adjacent wells of the target well logging, and generate a first number of attribute value reference sequences based on the distribution data of formation attribute values ​​of the adjacent wells. The derived sequence generation module is used to generate a set of derived sequences for each attribute value reference sequence constraint for any inverted stratigraphic depth using a random algorithm; wherein each set of derived sequences contains a second number of attribute value derived sequences; The inversion module is used to invert the well logging data at the inverted formation depth for any derived sequence set, using the derived sequence set as the initial model vector, and employing a stochastic gradient evolution hybrid optimization algorithm to obtain the inversion result set corresponding to the derived sequence set; wherein, each inversion result set contains a third number of inversion results; The determination module is used to determine the target inversion result corresponding to the inversion stratum depth from the inversion result.

7. A computing device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the method for determining stratigraphic attribute values ​​as described in any one of claims 1-5.

8. A computer storage medium, characterized in that, The storage medium stores at least one executable instruction that causes the processor to perform the operation corresponding to the formation attribute value determination method as described in any one of claims 1-5.

9. A computer program product, characterized in that, It includes at least one executable instruction that causes the processor to perform the operation corresponding to the formation attribute value determination method as described in any one of claims 1-5.

Images