Methods, apparatuses, devices, media, and products for predicting lithofacies of a hydrocarbon reservoir

Abstract

The embodiments of the present disclosure provide a method, device, equipment, medium and product for predicting lithofacies of oil and gas reservoir. The method comprises: obtaining a seismic attribute data sequence and a statistical rock physics model based on a calibration well; precomputing and caching a global likelihood probability matrix; performing Markov chain Monte Carlo iteration based on the matrix, generating a candidate lithofacies sequence for local disturbance of a current lithofacies sequence, and reading a corresponding likelihood probability value from the matrix according to a sampling point index and a lithofacies state index; combining prior information in the model to calculate a posterior probability of the candidate and current lithofacies sequences; performing update on the current lithofacies sequence according to the posterior probability ratio, and capturing a global optimal lithofacies sequence based on the updated current sequence in the iteration process and outputting as a lithofacies prediction result of the oil and gas reservoir. By using the method of the present disclosure, the timeliness problem of large-scale reservoir lithofacies prediction can be effectively solved.

Description

Technical Field

[0001] This disclosure relates to the field of geophysical exploration and oil and gas reservoir prediction technology, specifically to a method, apparatus, equipment, medium and product for predicting the lithofacies of oil and gas reservoirs. Background Technology

[0002] Reservoir facies is a core parameter characterizing the lithology and fluid properties of subsurface reservoirs, and its accurate prediction is crucial for oil and gas reserve assessment and well location deployment. Given the complexity of subsurface media and the limited availability of observational data, Bayesian statistical inversion methods, with their ability to integrate prior information and quantify prediction uncertainties, have been widely applied in reservoir facies prediction. From early statistical rock physics frameworks (Mukerji et al., 2001) and one-dimensional Markov chain priors (Larsen et al., 2006), to facies classification based on Gaussian mixture models (Grana, 2018) and three-dimensional Markov random field priors (Fjeldstad et al., 2021), the Bayesian inversion framework has continuously evolved. For the aforementioned complex posterior probability models with high dimensions and nonlinearity, the Markov chain Monte Carlo (MCMC) algorithm is currently the most commonly used solution method (Mosegaard and Tarantola, 1995; Eidsvik et al., 2004). By constructing a Markov chain with the target distribution as the steady-state distribution and performing random sampling, it can not only infer the optimal lithofacies sequence, but also effectively quantify the uncertainty of the prediction results.

[0003] However, traditional MCMC algorithms face severe computational efficiency bottlenecks in practical applications. In the Bayesian inversion process, the likelihood probability term needs to be calculated in each iteration to evaluate the posterior probability of candidate lithofacies sequences. This likelihood probability is typically given by a multidimensional Gaussian probability density function, involving complex numerical operations such as determinant calculation of the high-dimensional covariance matrix, matrix inversion, vector-matrix multiplication, and exponential operations. In the traditional MCMC algorithm architecture, this complex numerical computation is highly coupled with the random sampling iteration process—that is, in each iteration, the likelihood function formula needs to be repeatedly called to calculate the relevant sampling points in order to evaluate newly proposed candidate lithofacies sequences. To ensure a steady-state distribution, the algorithm typically requires thousands to tens of thousands of iterations. This repeated, complex, and time-consuming computation accumulates into a significant computational burden, severely limiting the convergence speed and application efficiency of the MCMC algorithm in large-scale reservoir lithofacies prediction. Summary of the Invention

[0004] To address the aforementioned technical problems, the present disclosure provides a solution. Embodiments of this disclosure offer a method, apparatus, equipment, medium, and product for predicting the lithofacies of oil and gas reservoirs.

[0005] According to a first aspect of the present disclosure, a method for predicting the lithofacies of oil and gas reservoirs is provided, wherein the method includes: Seismic attribute data sequences of the target oil and gas reservoir and a statistical rock physics model constructed based on calibration wells are obtained; wherein, the statistical rock physics model characterizes the probability distribution characteristics of seismic attributes under finite discrete lithofacies states and provides prior information on finite discrete lithofacies states. Based on the earthquake attribute data sequence and the statistical rock physics model, a global likelihood probability matrix is ​​pre-calculated and cached; wherein, the global likelihood probability matrix is ​​used for joint addressing through sampling point index and lithofacies state index; Based on the global likelihood probability matrix, a Markov chain Monte Carlo iteration is performed to locally perturb the current lithofacies sequence to generate candidate lithofacies sequences, and the corresponding likelihood probability values ​​are read from the global likelihood probability matrix according to the sampling point index and the lithofacies state index. The likelihood probability value is read from the global likelihood probability matrix, and combined with the prior information in the statistical rock physics model, the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence are calculated. The current lithofacies sequence is updated based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence. During the Markov chain Monte Carlo iteration process, the globally optimal lithofacies sequence is captured based on the updated current lithofacies sequence and output as the oil and gas reservoir lithofacies prediction result.

[0006] According to a second aspect of the present disclosure, an apparatus for predicting the lithofacies of oil and gas reservoirs is provided, wherein the apparatus comprises: The acquisition unit is configured to: acquire the seismic attribute data sequence of the target oil and gas reservoir and the statistical rock physics model constructed based on the calibration well; wherein the statistical rock physics model characterizes the probability distribution characteristics of seismic attributes under finite discrete lithofacies states and provides prior information on finite discrete lithofacies states; The pre-computation unit is configured to: pre-compute and cache the global likelihood probability matrix based on the seismic attribute data sequence and the statistical rock physics model; wherein the global likelihood probability matrix is ​​used for joint addressing through the sampling point index and the lithofacies state index; The iteration unit is configured to: perform Markov chain Monte Carlo iteration based on the global likelihood probability matrix, locally perturb the current lithofacies sequence to generate candidate lithofacies sequences, and read the corresponding likelihood probability value from the global likelihood probability matrix according to the sampling point index and the lithofacies state index. The calculation unit is configured to: read the likelihood probability value from the global likelihood probability matrix, and combine it with the prior information in the statistical rock physics model to calculate the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence. The output unit is configured to update the current lithofacies sequence based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence, and capture the globally optimal lithofacies sequence based on the updated current lithofacies sequence and output it as the oil and gas reservoir lithofacies prediction result during the Markov chain Monte Carlo iteration process.

[0007] According to a third aspect of the present disclosure, an electronic device is provided, the electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the method for predicting the lithology of oil and gas reservoirs as described in the present disclosure.

[0008] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, the storage medium storing a computer program for executing the method for predicting the lithofacies of oil and gas reservoirs described in the present disclosure.

[0009] According to a fifth aspect of the present disclosure, a computer program product is provided, including a computer program, wherein the computer program, when executed by a processor, implements the method for predicting the lithofacies of oil and gas reservoirs as described in the present disclosure.

[0010] As described above, the method for predicting the lithology of oil and gas reservoirs disclosed in this disclosure is based on the characteristics that seismic attribute data remains unchanged during the inversion process and that the lithology state belongs to a finite discrete set. Before the Markov chain Monte Carlo sampling iteration, all sampling points and lithology state combinations are pre-traversed to construct and cache a global likelihood probability matrix that can be jointly addressed by the sampling point index and the lithology state index. During the iteration process, the likelihood probability value is directly retrieved from the matrix according to the index to replace the complex numerical calculation based on the formula. The posterior probability of the candidate lithology sequence and the current lithology sequence is calculated by combining the prior information in the statistical rock physics model. The sequence update method is determined according to the ratio of the two. Finally, the globally optimal lithology sequence after convergence is output as the prediction result. This method removes the likelihood probability calculation characterized by high-dimensional matrix operations from the random sampling iteration process in advance and eliminates the computational redundancy in the iteration loop with low-complexity matrix lookup operations. Thus, while ensuring that the prediction result is strictly consistent with the traditional Markov chain Monte Carlo method and without introducing any approximate calculation error, it achieves an order-of-magnitude improvement in computational efficiency and effectively solves the timeliness problem of large-scale reservoir lithology prediction. Attached Figure Description

[0011] The above and other objects, features, and advantages of this disclosure will become more apparent from the more detailed description of the embodiments thereof in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this disclosure and form part of the specification. They are used together with the embodiments of this disclosure to explain the disclosure and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0012] Figure 1 This is a schematic flowchart of a method for predicting the lithofacies of oil and gas reservoirs provided in an exemplary embodiment of this disclosure; Figure 2 This is a public announcement Figure 1 One of the exemplary process diagrams for predicting the lithofacies of oil and gas reservoirs provided in the embodiments; Figure 3 This is a public announcement Figure 1 Example flowchart 2 for a method for predicting the lithofacies of oil and gas reservoirs provided in the embodiments; Figure 4 This is a public announcement Figure 1 The third exemplary flowchart of the method for predicting the lithofacies of oil and gas reservoirs provided in the embodiments; Figure 5 This is a public announcement Figure 1 Example flowchart four of the embodiments for predicting the lithofacies of oil and gas reservoirs; Figure 6(a) is a schematic diagram of the lithofacies profile distribution predicted by the traditional MCMC method; Figure 6(b) is a schematic diagram of the lithofacies profile distribution predicted by the method of this disclosure; Figure 7(a) is a schematic diagram of the lithofacies prediction results for the W1 well bypass provided in an exemplary embodiment of this disclosure; Figure 7(b) is a schematic diagram of the lithofacies prediction results for the W2 well bypass provided in an exemplary embodiment of this disclosure; Figure 8 This is a schematic flowchart of a method for predicting the lithofacies of oil and gas reservoirs provided in another exemplary embodiment of this disclosure; Figure 9 This is a schematic diagram of a structure for predicting the lithology of oil and gas reservoirs provided in an exemplary embodiment of the present disclosure; Figure 10 This is a schematic diagram of the structure of an application embodiment of the electronic device disclosed herein. Detailed Implementation

[0013] The present disclosure will be further described below with reference to the embodiments shown in the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present disclosure, and not all embodiments of the present disclosure. It should be understood that the present disclosure is not limited to the exemplary embodiments described herein.

[0014] It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of this disclosure.

[0015] Those skilled in the art will understand that the terms "first," "second," etc., in the embodiments of this disclosure are only used to distinguish different steps, devices, or modules, and do not represent any specific technical meaning, nor do they indicate a necessary logical order between them.

[0016] It should also be understood that in the embodiments disclosed herein, "a plurality of" may refer to two or more, and "at least one" may refer to one, two or more.

[0017] It should also be understood that any component, data or structure mentioned in the embodiments of this disclosure can generally be understood as one or more unless expressly defined or given to the contrary in the context.

[0018] Furthermore, the term "and / or" in this disclosure is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this disclosure generally indicates that the preceding and following related objects have an "or" relationship.

[0019] It should also be understood that the description of the various embodiments in this disclosure emphasizes the differences between the various embodiments, and the similarities or similarities can be referred to each other. For the sake of brevity, they will not be described in detail.

[0020] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0021] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0022] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0023] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0024] Overview of the inventive concept The core inventive concept of this disclosed technical solution lies in addressing the problem that the Markov chain Monte Carlo algorithm in reservoir lithofacies sequence prediction suffers from severely limited computational efficiency due to the repeated invocation of complex multidimensional Gaussian likelihood functions in each iteration. Based on the characteristics that seismic attribute data remains unchanged during inversion and that lithofacies states belong to a finite discrete set, before the sampling iteration is implemented, all combinations of sampling points and all lithofacies states are pre-traversed, and a global likelihood probability matrix is ​​calculated and cached in one go. In subsequent iterations, the corresponding likelihood probability values ​​are directly retrieved from this matrix based on the sampling point index and the lithofacies state index. This low-cost matrix lookup operation completely replaces the formula-based likelihood function numerical calculation that needs to be re-executed in each iteration of the traditional method. Thus, while ensuring that the prediction results are strictly consistent with the traditional method and without introducing any approximation errors, computational redundancy within the iteration loop is completely eliminated, achieving an order-of-magnitude improvement in computational efficiency.

[0025] Based on the above-mentioned inventive concept, this disclosure proposes a scheme for predicting the lithology of oil and gas reservoirs as described in the following embodiments.

[0026] Example 1 Figure 1 This is a schematic flowchart of a method for predicting the lithofacies of oil and gas reservoirs provided in an exemplary embodiment of this disclosure. The method is executed by a server, which may be a locally deployed computer node or a cloud server connected via a high-speed communication network.

[0027] Specifically, refer to Figure 1 The method for predicting the lithofacies of oil and gas reservoirs includes: S110. Obtain the seismic attribute data sequence of the target oil and gas reservoir and the statistical rock physics model constructed based on the calibration well.

[0028] The statistical rock physics model characterizes the probability distribution of seismic properties under finite discrete lithofacies states and provides prior information about finite discrete lithofacies states.

[0029] Optionally, obtaining the seismic attribute data sequence of the target oil and gas reservoir specifically involves obtaining multidimensional seismic attribute vectors corresponding to multiple sampling points within the target work area. These seismic attribute vectors include one or more combinations of P-wave velocity, S-wave velocity, and density. The statistical rock physics model constructed based on calibration wells specifically involves using seismic attribute sample data labeled with lithofacies in the calibration wells to extract statistical characteristic parameters of the seismic attribute data for different lithofacies categories, including the mean vector and covariance matrix corresponding to each lithofacies category. Simultaneously, the frequency of occurrence of each lithofacies in the calibration wells is statistically analyzed to obtain the initial state probability, and the transition frequency of lithofacies states between adjacent sampling points is statistically analyzed to obtain the state transition probability matrix. The finite discrete lithofacies state set includes sandstone and mudstone facies, meaning the number of lithofacies types is generally 2 or 3.

[0030] S120. Based on the earthquake attribute data sequence and the statistical rock physics model, pre-calculate and cache the global likelihood probability matrix.

[0031] The global likelihood probability matrix is ​​used for joint addressing through sampling point index and lithofacies state index.

[0032] Optionally, the pre-calculation and caching of the global likelihood probability matrix specifically involves: establishing a two-dimensional matrix with the number of rows equal to the total number of sampling points and the number of columns equal to the number of lithofacies types; traversing the location of each sampling point and each lithofacies state, calculating the likelihood probability value of the sampling point under that lithofacies state using the statistical rock physics model, and caching the calculation results according to the row-column correspondence of "sampling point index - lithofacies state index". In a preferred embodiment, the specific calculation method of the likelihood probability assumes that the seismic attribute data for each lithofacies state follows a multivariate Gaussian distribution, and the corresponding likelihood probability value is given by the multivariate Gaussian probability density function. This calculation process involves solving the determinant of the covariance matrix, matrix inversion, vector-matrix multiplication, and exponential operations. By calculating and storing the likelihood probability values ​​of all sampling points and all lithofacies states all at once before iteration, the complex calculations mentioned above do not need to be repeated in subsequent iterations.

[0033] S130. Based on the global likelihood probability matrix, perform Markov chain Monte Carlo iteration to locally perturb the current lithofacies sequence to generate candidate lithofacies sequences, and read the corresponding likelihood probability values ​​from the global likelihood probability matrix according to the sampling point index and lithofacies state index.

[0034] Optionally, the execution of Markov chain Monte Carlo iteration specifically involves: first, randomly generating an initial lithofacies sequence as the current lithofacies sequence and entering an iterative loop; in each iteration, locally perturbing the current lithofacies sequence to generate candidate lithofacies sequences, wherein the local perturbation, in a preferred embodiment, specifically involves randomly changing the lithofacies state of one or more sampling points in the candidate lithofacies sequence; subsequently, for each sampling point in the candidate lithofacies sequence, directly reading the corresponding likelihood probability value from the global likelihood probability matrix based on the index of the sampling point and the index of the lithofacies state at that sampling point, thereby replacing the numerical calculation that requires re-calling the complex likelihood function formula in each iteration of the traditional method with a single matrix lookup operation. The likelihood probability value corresponding to each sampling point in the current lithofacies sequence is updated to the likelihood probability value of the candidate lithofacies sequence read in the current iteration if the candidate lithofacies sequence is accepted, and inherited from the likelihood probability value of the current lithofacies sequence retained in the previous iteration if the candidate lithofacies sequence is rejected.

[0035] S140. Read the likelihood probability value from the global likelihood probability matrix, and calculate the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence by combining the prior information in the statistical rock physics model.

[0036] Optionally, calculating the posterior probability of the candidate lithofacies sequence specifically involves: reading the likelihood probability value corresponding to each sampling point in the candidate lithofacies sequence from the global likelihood probability matrix, obtaining the initial state probability and the state transition probability between adjacent sampling points from the prior information in the statistical rock physics model, and multiplying the initial state probability, the state transition probability, and the likelihood probability value corresponding to each sampling point to obtain the posterior probability of the candidate lithofacies sequence. Calculating the posterior probability of the current lithofacies sequence specifically involves: reading the likelihood probability value corresponding to each sampling point in the current lithofacies sequence from the global likelihood probability matrix, obtaining the initial state probability and the state transition probability between adjacent sampling points from the prior information in the statistical rock physics model, and multiplying the aforementioned terms to obtain the posterior probability of the current lithofacies sequence. In a preferred embodiment, to avoid numerical underflow and simplify calculation, the posterior probability is calculated in logarithmic form, i.e., the logarithm of each term is taken and then summed.

[0037] S150. Based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence, update the current lithofacies sequence. During the Markov chain Monte Carlo iteration process, capture the globally optimal lithofacies sequence based on the updated current lithofacies sequence and output it as the oil and gas reservoir lithofacies prediction result.

[0038] Optionally, updating the current lithofacies sequence specifically involves: calculating an acceptance rate based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence, where the acceptance rate is the smaller value between the ratio and 1; subsequently generating a random number between 0 and 1, and determining whether to replace the current lithofacies sequence with the candidate lithofacies sequence based on the numerical relationship between the random number and the acceptance rate. Simultaneously, a globally optimal lithofacies sequence and its corresponding posterior probability value are recorded in each iteration. When the posterior probability of the updated current lithofacies sequence is greater than the posterior probability of the globally optimal lithofacies sequence, the globally optimal lithofacies sequence is updated to the current lithofacies sequence, thereby continuously capturing the globally optimal lithofacies sequence during the iteration process. The Markov chain Monte Carlo iteration process has a termination condition; the iteration terminates when a preset number of iterations or a convergence condition is reached. The globally optimal lithofacies sequence captured during the iteration process is the predicted lithofacies result of the oil and gas reservoir. In a specific application case, the initial stabilization stage is set to iterate 2000 times, the basic stabilization stage to iterate 5000 times, and the fully stabilization stage to iterate 10000 times. Finally, the globally optimal lithofacies sequence captured in the fully stabilization stage is output.

[0039] As described above, the lithofacies method for predicting oil and gas reservoirs disclosed in this disclosure is based on the characteristics that seismic attribute data remains unchanged during the inversion process and that lithofacies states belong to a finite discrete set. Before the Markov chain Monte Carlo sampling iteration, all sampling points and lithofacies state combinations are pre-traversed to construct and cache a global likelihood probability matrix that can be jointly addressed by the sampling point index and the lithofacies state index. During the iteration process, the likelihood probability value is directly retrieved from the matrix based on the index to replace the complex numerical calculation based on formulas. Furthermore, the posterior probability of the candidate lithofacies sequence and the current lithofacies sequence is calculated by combining prior information from the statistical rock physics model. The ratio of the two probabilities is then used to predict the lithofacies of the reservoir. The current lithofacies sequence is updated, and the globally optimal lithofacies sequence is captured based on the updated current lithofacies sequence during the Markov chain Monte Carlo iteration process and output as the oil and gas reservoir lithofacies prediction result. This method removes the likelihood probability calculation, which is characterized by high-dimensional matrix operations, from the random sampling iteration process in advance, and eliminates the computational redundancy in the iteration loop with low-complexity matrix lookup operations. Thus, while ensuring that the prediction results are strictly consistent with the traditional Markov chain Monte Carlo method and without introducing any approximate calculation errors, it achieves an order-of-magnitude improvement in computational efficiency and effectively solves the timeliness problem of large-scale reservoir lithofacies prediction.

[0040] Example 2 Based on the above embodiment 1, referring to Figure 2 As an optional implementation, the step of "pre-calculating and caching the global likelihood probability matrix based on the seismic attribute data sequence and the statistical rock physics model" in S120 can be achieved in the following way: S1210. Traverse each sampling point in the seismic attribute data sequence, and for each sampling point, use the statistical rock physics model to calculate the likelihood probability value of the sampling point under each lithofacies state.

[0041] Optionally, traversing each sampling point in the seismic attribute data sequence specifically involves following the sampling point index. The sampling point is visited sequentially in the following order; for each sampling point, the likelihood probability value of that sampling point under each lithofacies state is calculated using the statistical rock physics model, specifically for the current sampling point. z traverse the finite discrete set of lithofacies states in sequence Each lithofacies state The seismic attribute data vector observed at the sampling point is calculated using the multivariate Gaussian probability density function given by the statistical rock physics model. d z In lithofacies state S i Likelihood probability value at time In a preferred embodiment, the statistical rock physics model assumes that the seismic attribute data for each lithofacies state follows a multivariate Gaussian distribution, the likelihood of which is given by the following formula:

[0042] in, n For the dimensions of earthquake attribute data, and Representing lithofacies S i The mean vector and covariance matrix of the corresponding seismic attribute data, Represents the covariance matrix The determinant, This is the inverse of the covariance matrix. The calculation process involves solving for the determinant of the covariance matrix, matrix inversion, vector-matrix multiplication, and exponentiation. In a specific application case, the seismic attribute data includes three dimensions: P-wave velocity, S-wave velocity, and density, i.e., n=3, and the finite discrete lithofacies set includes two lithofacies: sandstone and mudstone. The mean vector Covariance Matrix The results were obtained from statistical analysis of elastic parameter samples with labeled lithofacies in the calibration wells.

[0043] S1220. Store the calculated likelihood probability values ​​into the global likelihood probability matrix according to the correspondence between the sampling point index and the lithofacies state index.

[0044] Optionally, storing the calculated likelihood probability values ​​into the global likelihood probability matrix according to the correspondence between the sampling point index and the lithofacies state index specifically involves: establishing a matrix of size... Two-dimensional likelihood probability matrix L Index the sampling points z As row index of the matrix, lithofacies state index i (Right now , indicating lithofacies f z State in discrete set F Using the sequence number in the matrix as the column index, the likelihood probability value calculated in step S1210 is used... Fill the matrix L The z Line 1 i Column position, i.e.:

[0045] Therefore, the matrix L Each element in the matrix uniquely corresponds to a likelihood probability value under the combination of "sampling point - lithofacies state". After completing the entire traversal, the global likelihood probability matrix is... L It is fully constructed and cached in the server's memory or storage device, for use in subsequent Markov chain Monte Carlo iterations via row indexes. z With column index i Direct joint addressing reads eliminate the need to repeatedly call the aforementioned multivariate Gaussian probability density function for recalculation during iterations.

[0046] Based on the above embodiment 1, referring to Figure 3 As an optional implementation, in S130, "based on the global likelihood probability matrix, performing Markov chain Monte Carlo iteration to locally perturb the current lithofacies sequence to generate candidate lithofacies sequences, and reading the corresponding likelihood probability values ​​from the global likelihood probability matrix according to the sampling point index and lithofacies state index" can be implemented in the following way: S1310. Randomly generate an initial lithofacies sequence as the current lithofacies sequence and enter the Markov chain Monte Carlo iteration loop.

[0047] Optionally, the random generation of the initial lithofacies sequence is specifically performed according to the sampling point index. z =1,2,…, N z The initial lithofacies sequence is obtained by randomly assigning a lithofacies state to each sampling point in the order of sampling. Each of them After generating the initial sequence, the server calculates the posterior probability of the initial sequence, specifically by analyzing the global likelihood probability matrix. LRead the likelihood probability value corresponding to each sampling point. Combined with the initial state probability obtained from the statistical rock physics model and the state transition probability between adjacent sampling points Calculate using the following multiplication formula:

[0048] Meanwhile, this initial sequence is tentatively designated as the current optimal lithofacies sequence. In a preferred embodiment, to avoid numerical underflow, the likelihood function and posterior probability are calculated in logarithmic form. Subsequently, the Markov chain Monte Carlo iteration loop is entered, and subsequent steps S1320 to S1330 are repeated according to a preset number of iterations (e.g., 2000 iterations for the initial stable phase, 5000 iterations for the basic stable phase, and 10000 iterations for the fully stable phase). S1320. In each iteration of the iterative loop, the current lithofacies sequence is locally perturbed to generate a candidate lithofacies sequence.

[0049] Optionally, the local perturbation of the current lithofacies sequence specifically refers to, in the t-th iteration, adjusting the current lithofacies sequence... By randomly changing the lithofacies state of one or more sampling points, a new candidate lithofacies sequence can be obtained. In a specific application case, the finite discrete lithofacies set includes two lithofacies: sandstone and mudstone. The local perturbation involves randomly changing the lithofacies state at a sampling point in the candidate lithofacies sequence from sandstone to mudstone, or vice versa, while keeping the lithofacies states of the remaining sampling points unchanged. After generating the candidate lithofacies sequence, the server calculates the posterior probability of the candidate lithofacies sequence according to the following formula:

[0050] S1330. Based on the sampling point index and the lithofacies state index, read the likelihood probability value corresponding to each sampling point in the candidate lithofacies sequence from the global likelihood probability matrix.

[0051] Optionally, the step of reading the likelihood probability value corresponding to each sampling point in the candidate lithofacies sequence from the global likelihood probability matrix based on the sampling point index and the lithofacies state index specifically involves: for the candidate lithofacies sequence Each sampling point in To determine the lithofacies state at the sampling point In a finite discrete set of lithofacies states F Index in Then indexed by sampling point z Indexed by row, indexed by lithofacies state For column indexing, from the global likelihood probability matrix L Read the corresponding element value directly This element value represents the likelihood probability of the sampling point under its current lithofacies condition. Therefore, all candidate lithofacies sequences... The likelihood probability value of each sampling point is obtained through a single matrix lookup operation, eliminating the need for repeated calculations by calling the multivariate Gaussian probability density function during the iteration process.

[0052] The likelihood probability value corresponding to each sampling point in the current lithofacies sequence is obtained by updating the likelihood probability value of the candidate lithofacies sequence read in this iteration, or by inheriting the likelihood probability value of the current lithofacies sequence retained in the previous iteration.

[0053] Optionally, the update or inheritance mechanism specifically involves: in the t-th iteration, the server calculates the acceptance rate based on the posterior probability ratio of the candidate lithofacies sequence to the current lithofacies sequence.

[0054] Generate a random number between 0 and 1 u ,like u≤α If so, then accept the candidate lithofacies sequence, that is, update the current lithofacies sequence to the candidate lithofacies sequence. At this point, the likelihood probability value corresponding to each sampling point in the current lithofacies sequence is the likelihood probability value of each sampling point in the candidate lithofacies sequence obtained from the matrix in this iteration; if u>α If the condition is not met, then the update will be refused and the original sequence will be preserved. At this point, the likelihood probability value corresponding to each sampling point in the current lithofacies sequence is inherited from the likelihood probability value of the current lithofacies sequence calculated and retained in the previous iteration. Simultaneously, after each iteration, if... Then the globally optimal lithofacies sequence will be updated to Therefore, through repeated iterations, the server finally outputs the globally optimal lithofacies sequence. As a result of oil and gas reservoir lithofacies prediction.

[0055] Based on the above embodiment 1, referring to Figure 4 As an optional implementation, the step S140, "reading the likelihood probability value from the global likelihood probability matrix and calculating the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence by combining the prior information in the statistical rock physics model," can be achieved in the following way: S1410. Read the likelihood probability value corresponding to each sampling point in the candidate lithofacies sequence from the global likelihood probability matrix, and obtain the initial state probability of the candidate lithofacies sequence and the state transition probability between adjacent sampling points from the prior information in the statistical rock physics model.

[0056] Optionally, the step of reading the likelihood probability value corresponding to each sampling point in the candidate lithofacies sequence from the global likelihood probability matrix specifically involves: for the candidate lithofacies sequence Each sampling point in Based on the index of the sampling point z and the lithofacies state at the sampling point Index in the finite discrete set of lithofacies states F Read the corresponding element value from the global likelihood probability matrix L. The step of obtaining the initial state probability of the candidate lithofacies sequence and the state transition probability between adjacent sampling points from the prior information in the statistical rock physics model specifically involves: obtaining the initial state probability of the first sampling point of the candidate lithofacies sequence. This probability is directly given by the prior information in the statistical rock physics model; meanwhile, for z =2 to N z For each sampling point, obtain the state transition probability between adjacent sampling points. This probability is determined by the state transition matrix in the statistical rock physics model. Provided.

[0057] S1420. Multiply the initial state probability, state transition probability, and likelihood probability value corresponding to each sampling point of the candidate lithofacies sequence to obtain the posterior probability of the candidate lithofacies sequence.

[0058] Optionally, the posterior probability of the candidate lithofacies sequence obtained by the multiplication is specifically calculated according to the following formula.

[0059] Initial state probability State transition probability between all adjacent sampling points The product of all samples and the likelihood probability values ​​corresponding to all sampling points. The product of these three factors is multiplied together to obtain the joint probability of the candidate lithofacies sequence, which is proportional to the posterior probability of the candidate lithofacies sequence.

[0060] In a preferred embodiment, to avoid numerical underflow and simplify the calculation process, the present invention uses logarithmic form to operate on the likelihood function and posterior probability, that is, taking the natural logarithm of each term in the above formula and converting it into an addition operation:

[0061] In actual calculations, the server can use the above logarithmic form for cumulative comparison to avoid numerical underflow caused by the original probability value being too small.

[0062] S1430. Read the likelihood probability value corresponding to each sampling point in the current lithofacies sequence from the global likelihood probability matrix, and obtain the initial state probability of the current lithofacies sequence and the state transition probability between adjacent sampling points from the prior information in the statistical rock physics model.

[0063] Optionally, the step of reading the likelihood probability value corresponding to each sampling point in the current lithofacies sequence from the global likelihood probability matrix specifically involves: for the current lithofacies sequence Each sampling point in z Based on the index of the sampling point z and the lithofacies state at the sampling point index From the global likelihood probability matrix L Read the corresponding element value The step of obtaining the initial state probability of the current lithofacies sequence and the state transition probability between adjacent sampling points from the prior information in the statistical rock physics model specifically involves: obtaining the initial state probability of the first sampling point of the current lithofacies sequence. and for z =2 to N z For each sampling point, obtain the state transition probability. The data source is the same as the acquisition method of the candidate lithofacies sequence, both of which are given by the prior information in the statistical rock physics model.

[0064] S1440. Multiply the initial state probability, state transition probability, and likelihood probability value corresponding to each sampling point of the current lithofacies sequence to obtain the posterior probability of the current lithofacies sequence.

[0065] Optionally, the posterior probability of the current lithofacies sequence obtained by the multiplication is specifically calculated according to the following formula:

[0066] The joint probability of the current lithofacies sequence is obtained by multiplying the initial state probability, the state transition probability, and the likelihood probability of each sampling point. This joint probability is proportional to the posterior probability of the current lithofacies sequence. In a preferred embodiment, the calculation can also be performed logarithmically, i.e.:

[0067] After the server has calculated the posterior probability (or log-posterior probability) of the candidate lithofacies sequence and the current lithofacies sequence, it can be used for the calculation of the acceptance rate and the sequence update judgment in the subsequent step S150.

[0068] Based on the above embodiment 1, referring to Figure 5 As an optional implementation, in S150, "updating the current lithofacies sequence based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence, and capturing the globally optimal lithofacies sequence and outputting it as an oil and gas reservoir lithofacies prediction result based on the updated current lithofacies sequence" can be achieved in the following way: S1510. Calculate the acceptance rate based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence.

[0069] The acceptance rate is the smaller value between the ratio and 1.

[0070] The acceptance rate is calculated by the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence, specifically according to the following formula:

[0071] in, This represents the posterior probability of the candidate lithofacies sequence calculated in S1420. This is the posterior probability of the current lithofacies sequence calculated in S1440. This acceptance rate represents the upper bound of the probability that a candidate lithofacies sequence will be accepted as the new current lithofacies sequence.

[0072] S1520. Generate a random number between 0 and 1, and update the current lithofacies sequence based on the numerical comparison between the random number and the acceptance rate and the candidate lithofacies sequence.

[0073] Optionally, the server generates a random number that is uniformly distributed between 0 and 1. u ;like u≤α If so, then accept the candidate lithofacies sequence, that is, update the current lithofacies sequence to the candidate lithofacies sequence. ;like u>α If the current lithofacies sequence is not updated, the current lithofacies sequence will be retained. Therefore, by using random numbers and the acceptance rate... α The numerical values ​​were compared, and the update result of the current lithofacies sequence was determined in a machine-executable manner.

[0074] S1530. During the Markov chain Monte Carlo iteration process, the globally optimal lithofacies sequence and its corresponding posterior probability value are recorded. Based on the comparison between the posterior probability of the updated current lithofacies sequence and the posterior probability of the globally optimal lithofacies sequence, the globally optimal lithofacies sequence is updated and output as the oil and gas reservoir lithofacies prediction result.

[0075] Optionally, the recording of the globally optimal lithofacies sequence and its corresponding posterior probability value specifically involves: during the iterative initialization phase (S1310), the initial lithofacies sequence is temporarily set as the globally optimal lithofacies sequence. The corresponding posterior probability value is recorded. After each iteration completes the sequence update, the updated current lithofacies sequence is... posterior probability The posterior probability of the globally optimal lithofacies sequence Compare; if Then the globally optimal lithofacies sequence will be updated to Therefore, the globally optimal lithofacies sequence is continuously maintained and conditionally updated during the iteration process, ensuring that the lithofacies sequence with the highest posterior probability is captured as the globally optimal lithofacies sequence throughout the entire iteration process. The globally optimal lithofacies sequence captured during the iteration process is the lithofacies prediction result of the oil and gas reservoir.

[0076] It should be noted that the Markov chain Monte Carlo iteration process has a termination condition, and the iteration terminates when the preset number of iterations or the convergence condition is reached.

[0077] Example 3 To verify the technical effectiveness of the method proposed in this disclosure in practical applications, a specific implementation verification was conducted in an oilfield in eastern my country. This comparative verification example comprehensively compares and analyzes the method of this disclosure (hereinafter referred to as "this method") with the traditional Markov chain Monte Carlo method under the same data and parameter settings. The specific implementation process and verification results are as follows.

[0078] I. Experimental data and parameter settings.

[0079] The data for this application case comes from an actual seismic exploration area in an oilfield in eastern my country. Using seismic attribute data (including P-wave velocity, S-wave velocity, and density) and corresponding lithofacies labels from calibration wells, statistical feature parameters required for the algorithm are extracted: for elastic parameter samples under different lithofacies categories, the mean vector and covariance matrix required for the Gaussian mixture model are calculated; simultaneously, the frequency of each lithofacies in the calibration well data is statistically analyzed to obtain the initial state probability, and the transition frequencies of all adjacent lithofacies are statistically analyzed to obtain the transition probability matrix. The finite discrete lithofacies state set includes two lithofacies: sandstone and mudstone.

[0080] To avoid numerical underflow and simplify the calculation process, this disclosure uses logarithmic forms for the likelihood function and posterior probability. To comprehensively verify the convergence performance and computational efficiency of the method, multiple observation nodes representing different convergence states were set up in the experiment: 2000 iterations for the initial stable stage, 5000 iterations for the basic stable stage, and 10000 iterations for the fully stable stage. While maintaining parameter consistency, the lithofacies prediction results of this method and the traditional Markov chain Monte Carlo method at the above different stages were output, and the corresponding running times were recorded for comparative analysis.

[0081] II. Verification results in Figure 6.

[0082] Referring to Figure 6, which shows the comparison results of the predicted lithofacies profile distribution, Figure 6(a) is the lithofacies predicted profile obtained by the traditional Markov chain Monte Carlo method when it reaches the fully stable stage, and Figure 6(b) is the lithofacies predicted profile obtained by this method when it reaches the fully stable stage.

[0083] Comparing Figures 6(a) and 6(b), it is clear that the lithofacies profiles predicted by the two methods exhibit a high degree of consistency in the spatial distribution morphology of sandstone and mudstone, local lithofacies boundaries, and lateral extension. This comparison demonstrates that our method, by pre-calculating the global likelihood probability matrix and replacing formula-based numerical calculations of the likelihood function with table lookup operations, achieves an efficient replacement for the traditional Markov chain Monte Carlo method without introducing any computational errors, and the prediction accuracy is strictly consistent with the traditional method.

[0084] III. Verification results in Figure 7.

[0085] Referring to Figure 7, Figure 7 shows a comparison of the lithofacies prediction results of the well-side passage. Figure 7(a) shows the lithofacies prediction results of the calibration well W1 obtained by the two methods, and Figure 7(b) shows the lithofacies prediction results of the blind well W2 obtained by the two methods.

[0086] In Figure 7(a), it can be clearly observed that at the calibration well W1, when in the same stable stage, the lithofacies results predicted by this method are completely consistent with the results obtained by the traditional Markov chain Monte Carlo method in terms of sequence structure and sandstone-mudstone boundary location. In Figure 7(b), at the blind well W2, this method and the traditional Markov chain Monte Carlo method also show a high degree of agreement in the characterization of sandstone-mudstone boundaries.

[0087] Once fully stable, the predictions from both methods closely matched the reference lithofacies, accurately reconstructing the lithofacies distribution characteristics of the subsurface reservoir. This result further demonstrates that our method is strictly consistent with the traditional Markov chain Monte Carlo method in terms of prediction accuracy and convergence characteristics, without introducing any prediction bias.

[0088] IV. Verification results in Table 1.

[0089] Referring to Table 1, which details the comparison of computation time required to predict the entire profile at different stabilization stages between the traditional Markov chain Monte Carlo method and the proposed method.

[0090] Table 1 Comparison of Calculation Time

[0091] According to the data in Table 1, the traditional Markov chain Monte Carlo method takes 3.67 × 10⁻⁶ times to reach complete stability (10,000 iterations) when predicting the same profile. 4 The traditional method takes only 24.45 seconds, while the current method takes approximately 1500 times longer, representing a significant improvement in computational efficiency. In the stable phase (5000 iterations), the traditional method takes 1.85 × 10⁻⁶ seconds. 4 In terms of computation time, the traditional method takes only 13.05 seconds, which is about 1400 times more efficient. In the initial stabilization stage (2000 iterations), the traditional method takes 7.43 × 10³ seconds, while the traditional method takes 5.33 seconds. Moreover, the computation time of the traditional method in the initial stabilization stage is still much longer than the time required for the traditional method to reach the fully stable stage, which is about 300 times longer.

[0092] The data above clearly demonstrates that this method transforms the originally high-frequency and repetitive multivariate numerical calculations into matrix lookup operations with extremely low complexity, completely eliminating computational redundancy in iterative loops. While achieving lossless prediction accuracy, this method can significantly shorten the computation time, achieving extremely high inversion accuracy with extremely low computational cost, and has strong industrial application value.

[0093] V. Overall Conclusion

[0094] Based on the verification results in Figures 6 and 7 and Table 1, it can be seen that the method proposed in this disclosure for predicting the lithology of oil and gas reservoirs successfully breaks the strong coupling between the numerical computation of the likelihood function and the random sampling iteration process in the traditional Markov chain Monte Carlo algorithm by pre-calculating the global likelihood probability matrix and retrieving the likelihood values ​​from the table during iteration. This method achieves an order-of-magnitude improvement in computational efficiency while ensuring complete consistency of the prediction results with the traditional Markov chain Monte Carlo method, effectively solving the timeliness problem of large-scale reservoir lithology prediction.

[0095] Example 4 To facilitate understanding of the complete process of the solution described in the foregoing embodiments of this disclosure, please refer to... Figure 8 The complete process is as follows: Step 1: Construct the global likelihood probability matrix L。

[0096] Traverse all sampling points and each lithofacies state The likelihood probability values ​​for each combination were calculated using a statistical rock physics model. And cache it as a global matrix. L .in, dz For the first z Seismic attribute data vector of each sampling point f z This represents the lithofacies state corresponding to this sampling point. N f This represents the total number of finite discrete lithofacies states. L z,j For the matrix of the first z Line 1 j The elements of the column.

[0097] Step 2, Initialization.

[0098] Randomly generate initial lithofacies sequence According to the matrix L Calculate their joint probability:

[0099] This sequence is tentatively designated as the current optimal lithofacies sequence. .in Let be the initial state probability. This represents the state transition probability between adjacent sampling points. lithofacies state Index in a discrete set.

[0100] Step 3, iterative loop (t=1,2,…,T).

[0101] State perturbation: for the current lithofacies sequence Local perturbation is performed to randomly change the lithofacies state of one or more sampling points, generating candidate lithofacies sequences. ; Likelihood values ​​are retrieved from a table: For each sampling point in the candidate lithofacies sequence, the likelihood value is retrieved based on the sampling point index. z Index of Lithofacies From the matrix L Read the corresponding likelihood probability value directly. ; Calculate the posterior probability of candidate lithofacies sequences:

[0102] Calculate the acceptance rate:

[0103] Sequence Update: Generate a random number u between 0 and 1. If u ≤ α, then let... Otherwise If the posterior probability of the updated current lithofacies sequence is greater than The posterior probability will then be Update to the current lithofacies sequence.

[0104] Step 4, Termination Check: Determine if the current iteration count has reached the preset iteration count. T If the target is not reached, return to step 3 to continue the iteration; if the target is reached, proceed to step 5.

[0105] Step 5, Output: The globally optimal lithofacies sequence captured during the Markov chain Monte Carlo iteration process. The output is the predicted lithology of oil and gas reservoirs.

[0106] As described above, the method for predicting oil and gas reservoir facies provided in the embodiments of this disclosure, based on the characteristics that seismic attribute data remains unchanged during the inversion process and that the facies state belongs to a finite discrete set, pre-constructs and caches the global likelihood probability matrix before the Markov chain Monte Carlo iteration. This pre-deploys the complex likelihood function numerical calculation that was originally highly coupled with random sampling iteration, allowing the likelihood probability value to be directly retrieved from the matrix based on the sampling point index and the facies state index during the iteration process. The low-complexity matrix lookup operation replaces the high-frequency repetitive formula calculation. Thus, while ensuring that the prediction accuracy is strictly consistent with traditional methods and without introducing any approximate calculation errors, the computation time is significantly shortened, computational redundancy in the iteration loop is effectively eliminated, and computational efficiency is improved by orders of magnitude. This solves the timeliness problem of large-scale reservoir facies prediction and has strong industrial application value.

[0107] Example 5 It should be understood that the methods described in the foregoing embodiments of this document for predicting the lithology of oil and gas reservoirs can also be similarly applied to the following apparatus for predicting the lithology of oil and gas reservoirs for similar extensions. For simplicity, they are not described in detail.

[0108] Figure 9 This is a schematic diagram of a device for predicting the lithology of oil and gas reservoirs provided in an exemplary embodiment of this disclosure. (Refer to...) Figure 9 The device includes: The acquisition unit 110 is configured to: acquire the seismic attribute data sequence of the target oil and gas reservoir and the statistical rock physics model constructed based on the calibration well; wherein the statistical rock physics model characterizes the probability distribution characteristics of seismic attributes under finite discrete lithofacies state and provides prior information on finite discrete lithofacies state.

[0109] The pre-computation unit 120 is configured to pre-compute and cache the global likelihood probability matrix based on the seismic attribute data sequence and the statistical rock physics model; wherein the global likelihood probability matrix is ​​used for joint addressing through the sampling point index and the lithofacies state index.

[0110] Iteration unit 130 is configured to: perform Markov chain Monte Carlo iteration based on the global likelihood probability matrix, locally perturb the current lithofacies sequence to generate candidate lithofacies sequences, and read the corresponding likelihood probability values ​​from the global likelihood probability matrix according to the sampling point index and the lithofacies state index.

[0111] The calculation unit 140 is configured to: read the likelihood probability value from the global likelihood probability matrix, and combine it with the prior information in the statistical rock physics model to calculate the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence.

[0112] Output unit 150 is configured to: update the current lithofacies sequence based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence; and capture the globally optimal lithofacies sequence based on the updated current lithofacies sequence and output it as the oil and gas reservoir lithofacies prediction result during the Markov chain Monte Carlo iteration process.

[0113] Optionally, the pre-computation unit 120 is further configured to: traverse each sampling point in the seismic attribute data sequence, and for each sampling point, calculate the likelihood probability value of the sampling point in each lithofacies state using the statistical rock physics model; and store the calculated likelihood probability values ​​in the global likelihood probability matrix according to the correspondence between the sampling point index and the lithofacies state index.

[0114] As described above, the device for predicting oil and gas reservoir facies provided in the embodiments of this disclosure, based on the characteristics that seismic attribute data remains unchanged during the inversion process and that the facies state belongs to a finite discrete set, pre-constructs and caches the global likelihood probability matrix before the Markov chain Monte Carlo iteration. This pre-deploys the complex likelihood function numerical calculation that was originally highly coupled with random sampling iteration, allowing the likelihood probability value to be directly retrieved from the matrix based on the sampling point index and the facies state index during the iteration process. The low-complexity matrix lookup operation replaces the high-frequency repetitive formula calculation. Thus, while ensuring that the prediction accuracy is strictly consistent with traditional methods and without introducing any approximate calculation errors, the computation time is significantly shortened, computational redundancy in the iteration loop is effectively eliminated, and computational efficiency is improved by orders of magnitude. This solves the timeliness problem of large-scale reservoir facies prediction and has extremely strong industrial application value.

[0115] Example 6 In addition, this disclosure also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program stored in the memory, wherein when the computer program is executed, it implements the method for predicting the lithology of oil and gas reservoirs described in any of the above embodiments of this disclosure.

[0116] Figure 10 This is a schematic diagram of the structure of an application embodiment of the electronic device disclosed herein. Below, reference is made to… Figure 10 This describes an electronic device according to embodiments of the present disclosure. The electronic device may be either or both of a first device and a second device, or a standalone device independent of them, which may communicate with the first device and the second device to receive acquired input signals from them.

[0117] like Figure 10 As shown, the electronic device includes one or more processors and a memory. The processor may be a central processing unit (CPU) or other processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. The memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor may execute the program instructions to implement the methods for predicting oil and gas reservoir lithofacies and / or other desired functions described in the various embodiments of this disclosure above.

[0118] In one example, the electronic device may further include input and output devices, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown). Furthermore, the input device may include, for example, a keyboard, a mouse, etc. The output device can output various information to the outside. The output device may include, for example, a monitor, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0119] Of course, for the sake of simplicity, Figure 10 Only some of the components of the electronic device relevant to this disclosure are shown, omitting components such as buses, input / output interfaces, etc. In addition, the electronic device may include any other suitable components depending on the specific application.

[0120] In addition to the methods and apparatus described above, embodiments of this disclosure may also be computer program products comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods for predicting the lithofacies of oil and gas reservoirs according to various embodiments of this disclosure as described in the foregoing portion of this specification.

[0121] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this disclosure. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on a user's computing device, partially on a user's computing device, as a standalone software package, partially on a user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0122] Furthermore, embodiments of this disclosure may also be computer-readable storage media storing computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods for predicting the lithofacies of oil and gas reservoirs according to various embodiments of this disclosure as described in the foregoing portion of this specification.

[0123] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0124] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as ROM, RAM, magnetic disk, or optical disk.

[0125] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0126] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system embodiments, since they largely correspond to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0127] The block diagrams of devices, apparatuses, devices, and systems disclosed herein are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0128] The methods and apparatus of this disclosure may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the methods is for illustrative purposes only, and the steps of the methods of this disclosure are not limited to the order specifically described above unless otherwise specifically stated. Furthermore, in some embodiments, this disclosure may also be implemented as a program recorded on a recording medium, the program including machine-readable instructions for implementing the methods according to this disclosure. Thus, this disclosure also covers recording media storing programs for performing the methods according to this disclosure.

[0129] It should also be noted that in the apparatus, devices, and methods of this disclosure, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions to this disclosure.

[0130] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0131] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for predicting the lithofacies of oil and gas reservoirs, characterized in that, The method includes: Seismic attribute data sequences of the target oil and gas reservoir and a statistical rock physics model constructed based on calibration wells are obtained; wherein, the statistical rock physics model characterizes the probability distribution characteristics of seismic attributes under finite discrete lithofacies states and provides prior information on finite discrete lithofacies states. Based on the earthquake attribute data sequence and the statistical rock physics model, a global likelihood probability matrix is ​​pre-calculated and cached; wherein, the global likelihood probability matrix is ​​used for joint addressing through sampling point index and lithofacies state index; Based on the global likelihood probability matrix, a Markov chain Monte Carlo iteration is performed to locally perturb the current lithofacies sequence to generate candidate lithofacies sequences, and the corresponding likelihood probability values ​​are read from the global likelihood probability matrix according to the sampling point index and the lithofacies state index. The likelihood probability value is read from the global likelihood probability matrix, and combined with the prior information in the statistical rock physics model, the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence are calculated. The current lithofacies sequence is updated based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence. During the Markov chain Monte Carlo iteration process, the globally optimal lithofacies sequence is captured based on the updated current lithofacies sequence and output as the oil and gas reservoir lithofacies prediction result.

2. The method according to claim 1, characterized in that, Based on the earthquake attribute data sequence and the statistical rock physics model, the global likelihood probability matrix is ​​pre-calculated and cached, including: Iterate through each sampling point in the seismic attribute data sequence, and for each sampling point, use the statistical rock physics model to calculate the likelihood probability value of the sampling point in each lithofacies state; The calculated likelihood probability values ​​are stored in the global likelihood probability matrix according to the correspondence between the sampling point index and the lithofacies state index.

3. The method according to claim 1, characterized in that, Based on the global likelihood probability matrix, a Markov chain Monte Carlo iteration is performed to locally perturb the current lithofacies sequence to generate candidate lithofacies sequences. The corresponding likelihood probability values ​​are then read from the global likelihood probability matrix according to the sampling point index and the lithofacies state index, including: An initial lithofacies sequence is randomly generated as the current lithofacies sequence, and the process enters a Markov chain Monte Carlo iteration loop. In each iteration of the iterative loop, the current lithofacies sequence is locally perturbed to generate candidate lithofacies sequences; Based on the sampling point index and the lithofacies state index, the likelihood probability value corresponding to each sampling point in the candidate lithofacies sequence is read from the global likelihood probability matrix; The likelihood probability value corresponding to each sampling point in the current lithofacies sequence is obtained by updating the likelihood probability value of the candidate lithofacies sequence read in this iteration, or by inheriting the likelihood probability value of the current lithofacies sequence retained in the previous iteration.

4. The method according to claim 1, characterized in that, The likelihood probability values ​​are read from the global likelihood probability matrix, and combined with the prior information in the statistical rock physics model, the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence are calculated, including: The likelihood probability value corresponding to each sampling point in the candidate lithofacies sequence is read from the global likelihood probability matrix, and the initial state probability and state transition probability between adjacent sampling points are obtained from the prior information in the statistical rock physics model. The posterior probability of the candidate lithofacies sequence is obtained by multiplying the initial state probability, state transition probability, and likelihood probability value corresponding to each sampling point. Read the likelihood probability value corresponding to each sampling point in the current lithofacies sequence from the global likelihood probability matrix, and obtain the initial state probability of the current lithofacies sequence and the state transition probability between adjacent sampling points from the prior information in the statistical rock physics model. The posterior probability of the current lithofacies sequence is obtained by multiplying the initial state probability, the state transition probability, and the likelihood probability value corresponding to each sampling point.

5. The method according to claim 1, characterized in that, The current lithofacies sequence is updated based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence. During the Markov chain Monte Carlo iteration process, the globally optimal lithofacies sequence is captured based on the updated current lithofacies sequence and output as the oil and gas reservoir lithofacies prediction result, including: The acceptance rate is calculated based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence; wherein the acceptance rate is the smaller value between the ratio and 1. Generate a random number between 0 and 1, and update the current lithofacies sequence based on the numerical comparison between the random number and the acceptance rate, as well as the candidate lithofacies sequence; During the Markov chain Monte Carlo iteration process, the globally optimal lithofacies sequence and its corresponding posterior probability value are recorded, as well as the comparison result between the posterior probability of the updated current lithofacies sequence and the posterior probability of the globally optimal lithofacies sequence. The globally optimal lithofacies sequence is then updated and output as the lithofacies prediction result of the oil and gas reservoir.

6. An apparatus for predicting the lithofacies of oil and gas reservoirs, characterized in that, The device includes: The acquisition unit is configured to: acquire the seismic attribute data sequence of the target oil and gas reservoir and the statistical rock physics model constructed based on the calibration well; wherein the statistical rock physics model characterizes the probability distribution characteristics of seismic attributes under finite discrete lithofacies states and provides prior information on finite discrete lithofacies states; The pre-computation unit is configured to: pre-compute and cache the global likelihood probability matrix based on the seismic attribute data sequence and the statistical rock physics model; wherein the global likelihood probability matrix is ​​used for joint addressing through the sampling point index and the lithofacies state index; The iteration unit is configured to: perform Markov chain Monte Carlo iteration based on the global likelihood probability matrix, locally perturb the current lithofacies sequence to generate candidate lithofacies sequences, and read the corresponding likelihood probability value from the global likelihood probability matrix according to the sampling point index and the lithofacies state index. The calculation unit is configured to: read the likelihood probability value from the global likelihood probability matrix, and combine it with the prior information in the statistical rock physics model to calculate the posterior probability of the candidate lithofacies sequence and the posterior probability of the current lithofacies sequence. The output unit is configured to update the current lithofacies sequence based on the ratio of the posterior probability of the candidate lithofacies sequence to the posterior probability of the current lithofacies sequence, and capture the globally optimal lithofacies sequence based on the updated current lithofacies sequence and output it as the oil and gas reservoir lithofacies prediction result during the Markov chain Monte Carlo iteration process.

7. The apparatus according to claim 6, characterized in that, The pre-computation unit is further configured as follows: Iterate through each sampling point in the seismic attribute data sequence, and for each sampling point, use the statistical rock physics model to calculate the likelihood probability value of the sampling point in each lithofacies state; The calculated likelihood probability values ​​are stored in the global likelihood probability matrix according to the correspondence between the sampling point index and the lithofacies state index.

8. An electronic device, characterized in that, The electronic device includes: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the method for predicting the lithology of oil and gas reservoirs as described in any one of claims 1 to 5.

9. A computer-readable storage medium storing a computer program, characterized in that, The computer program is used to execute the method for predicting the lithofacies of oil and gas reservoirs as described in any one of claims 1 to 5.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for predicting the lithofacies of oil and gas reservoirs as described in any one of claims 1 to 5.

Images