Lithology sensitive logging curve analysis method and device, electronic equipment and medium
Through a data-driven supervised learning method, the distance and similarity of the logging curve are calculated, and the problem of low lithologic recognition accuracy in the existing technology is solved, and a more accurate lithologic sensitivity analysis is achieved.
Patent Information
- Application Number
- CN202311675035.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-07
- Publication Date
- 2025-06-10
AI Technical Summary
The prior art has problems in lithology identification with low accuracy, many qualitative analysis, and it is difficult to effectively distinguish transitional rocks or complex lithology.
Using a data-driven supervised learning method, the distance and similarity of the logging curve are calculated, the objective function is established and solved, and the weight is updated to determine the sensitivity of each logging parameter to lithology.
It improves the accuracy of lithologic identification, reduces errors, can more effectively distinguish different lithologic properties, and provides more accurate lithologic sensitivity analysis.
Smart Images
Figure CN120122141A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of oil and gas and coalbed methane exploration and development, and more specifically, to a method, device, electronic device and medium for analyzing lithology-sensitive logging curves. Background Art
[0002] Lithology identification is a very important and fundamental content in the fields of geology, resource exploration, etc. The identification and description of lithology is an important task throughout the entire process of oil and gas field exploration and development. Of course, it is also an essential key link in research such as logging data interpretation and reservoir prediction. At present, the main logging lithology interpretation methods include conventional qualitative interpretation methods, crossplot methods, curve overlay methods, discriminant analysis methods, mathematical statistics methods, and artificial intelligence methods such as fuzzy recognition, cluster analysis, and artificial neural networks. The conventional qualitative interpretation method mainly uses the anomalies and anomaly amplitudes of spontaneous potential and natural gamma logging curves for lithology identification. This method is suitable for the qualitative interpretation of conventional sandstone-shale profiles. The crossplot method generally uses theoretical interpretation charts such as neutron-density, neutron-sonic, sonic-density, M-N, etc. for qualitative and semi-quantitative interpretation. These theoretical interpretation charts are made for pure formations saturated with water and are only applicable to the case where the rock is composed of one or two minerals, and are easily affected by shale. Another commonly used crossplot method is to plot the logging information of different lithologies, such as spontaneous potential, natural gamma, three porosity logs, resistivity, nuclear magnetic resonance, etc., pairwise on a single graph based on lithology identification, in order to analyze the numerical values and ranges of the logging information possessed by different lithologies, and identify the lithology based on this numerical value and range. Since the logging characteristic values of different lithologies are often not a specific value but a range value, and the characteristic values of transitional lithologies or complex lithologies often cross and overlap, the ability of this method to distinguish lithologies is reduced. The curve overlay method is also a commonly used qualitative identification method. It commonly overlaps neutron porosity and density porosity, and has good applicability to pure lithology formations. The advantage of these interpretation methods is that they are fast and simple, but the accuracy is not high, and most are qualitative analyses.
[0003] With the continuous development of artificial intelligence technology, intelligent lithology identification methods based on technologies such as data mining and image recognition have begun to emerge continuously. The method of taking rock images or microscopic images as the research object and forming a mapping by establishing the relationship between image features and corresponding lithologies to achieve automatic lithology identification is automatic and efficient, but only using the single feature of the rock image will inevitably produce large errors in the process of lithology identification, resulting in problems such as deviation or even error in lithology interpretation. On the other hand, not all logging curves have a one-to-one correspondence with rock geological characteristics. In fact, the sensitivity of logging curve information to reflect reservoir lithology is different. Some logging curves obtained through rock physics calculations play an interfering role in lithology prediction and classification. Some logs are correlated or even repeated with each other, resulting in a large amount of redundant information.
[0004] There is still a need to develop a method for analyzing lithology-sensitive logging curves.
[0005] The information disclosed in the background section of the present invention is only intended to deepen the understanding of the general background of the present invention, and should not be regarded as an admission or any form of implication that this information constitutes the prior art known to those skilled in the art. Summary of the Invention
[0006] The present invention provides a method, device, electronic device and medium for analyzing lithology-sensitive logging curves, which realizes supervised learning based on data driving to determine the sensitivity of each logging parameter to lithology.
[0007] In a first aspect, an embodiment of the present disclosure provides a method for analyzing lithology-sensitive logging curves, including:
[0008] Determine known lithologies and their corresponding multiple logging curves, and assign initial weights to the multiple logging curves;
[0009] According to the initial weights, calculate the distances between pairwise logging curves, and then calculate the corresponding similarities;
[0010] Calculate the probability that two logging curves belong to the same class, and then establish and solve an objective function;
[0011] For each logging curve, calculate the gradient of the corresponding objective function;
[0012] Update the weights according to the gradient, and re-solve the objective function. If the solutions of the objective function twice are less than a set threshold, the updated weights are the sensitivities of the corresponding logging curves to lithology.
[0013] As a specific implementation manner of the embodiment of the present disclosure, the distance of the logging curve is:
[0014]
[0015] where X ir , X kr are logging curves, i = 1, 2,..., N represents the serial number of data points, r = 1, 2,..., P represents the serial number of variables of data at each data point, and w r is the weight.
[0016] As a specific implementation manner of the embodiment of the present disclosure, the similarity is:
[0017]
[0018] where σ is a parameter.
[0019] As a specific implementation manner of the embodiments of the present disclosure, the probability that two logging curves belong to the same category is:
[0020]
[0021] where Y i is the lithology.
[0022] As a specific implementation manner of the embodiments of the present disclosure, the objective function is:
[0023]
[0024] where λ is a parameter.
[0025] As a specific implementation manner of the embodiments of the present disclosure, the gradient of the objective function is:
[0026]
[0027] As a specific implementation manner of the embodiments of the present disclosure, updating the weight according to the gradient is:
[0028] w r = w r + α·Δ r
[0029] where α is a parameter.
[0030] In a second aspect, the embodiments of the present disclosure further provide a lithology-sensitive logging curve analysis device, including:
[0031] An assignment module, which determines known lithologies and their corresponding multiple logging curves, and assigns initial weights to the multiple logging curves;
[0032] A calculation module, which calculates the distance between two logging curves pairwise according to the initial weights, and further calculates the corresponding similarity;
[0033] An objective function solving module, which calculates the probability that two logging curves belong to the same category, and further establishes and solves the objective function;
[0034] A gradient calculation module, which calculates the gradient of the corresponding objective function for each logging curve;
[0035] An update module, which updates the weight according to the gradient, re-solves the objective function, and if the solutions of the objective function twice are less than a set threshold, the updated weight is the sensitivity of the corresponding logging curve to the lithology.
[0036] As a specific implementation manner of the embodiments of the present disclosure, the distance of the logging curve is:
[0037]
[0038] Among them, X ir and X kr are logging curves, i = 1, 2, …, N represents the data point serial number, r = 1, 2, …, P represents the variable serial number of the data at each data point, and w r is the weight.
[0039] As a specific implementation manner of the embodiment of the present disclosure, the similarity is:
[0040]
[0041] Among them, σ is a parameter.
[0042] As a specific implementation manner of the embodiment of the present disclosure, the probability that two logging curves belong to the same class is:
[0043]
[0044] Among them, Y i is the lithology.
[0045] As a specific implementation manner of the embodiment of the present disclosure, the objective function is:
[0046]
[0047] Among them, λ is a parameter.
[0048] As a specific implementation manner of the embodiment of the present disclosure, the gradient of the objective function is:
[0049]
[0050] As a specific implementation manner of the embodiment of the present disclosure, updating the weight according to the gradient is:
[0051] w r = w r + α·Δ r
[0052] Among them, α is a parameter.
[0053] In a third aspect, the embodiment of the present disclosure further provides an electronic device, and the electronic device includes:
[0054] A memory storing executable instructions;
[0055] A processor, and the processor runs the executable instructions in the memory to implement the lithology-sensitive logging curve analysis method.
[0056] Fourthly, an embodiment of the present disclosure further provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the above-described lithology-sensitive log curve analysis method.
[0057] The methods and apparatuses of the present invention have other characteristics and advantages, which will be apparent from the accompanying drawings incorporated herein and the subsequent detailed description, or will be described in detail in the accompanying drawings incorporated herein and the subsequent detailed description, and these accompanying drawings and detailed description are used together to explain the specific principles of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] By describing the exemplary embodiments of the present invention in more detail in conjunction with the accompanying drawings, the above and other objects, features, and advantages of the present invention will become more apparent, wherein, in the exemplary embodiments of the present invention, the same reference numerals generally represent the same components.
[0059] Figure 1 FIG. shows a flowchart of steps of a lithology-sensitive log curve analysis method according to an embodiment of the present invention.
[0060] Figure 2 FIG. shows a schematic diagram of a sensitivity histogram according to an embodiment of the present invention.
[0061] Figure 3 FIG. shows a schematic diagram of a lithology-log curve sensitivity matrix according to an embodiment of the present invention.
[0062] Figure 4 FIG. shows a block diagram of a lithology-sensitive log curve analysis apparatus according to an embodiment of the present invention.
[0063] DESCRIPTION OF REFERENCE NUMERALS:
[0064] 201, assignment module; 202, calculation module; 203, objective function solving module; 204, gradient calculation module; 205, update module. DETAILED DESCRIPTION
[0065] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein.
[0066] To facilitate understanding of the solutions and effects of the embodiments of the present invention, six specific application examples are given below. Those skilled in the art should understand that this example is only for facilitating understanding of the present invention, and any specific details are not intended to limit the present invention in any way.
[0067] Example 1
[0068] Figure 1 The flowchart shows the steps of a lithology-sensitive log curve analysis method according to an embodiment of the present invention.
[0069] As Figure 1 shown, the lithology-sensitive log curve analysis method includes: Step 101, determining known lithologies and their corresponding multiple log curves, and assigning initial weights to the multiple log curves; Step 102, calculating the distances between pairwise log curves according to the initial weights, and then calculating the corresponding similarities; Step 103, calculating the probability that two log curves belong to the same class, and then establishing and solving an objective function; Step 104, calculating the gradient of the corresponding objective function for each log curve; Step 105, updating the weights according to the gradient and re-solving the objective function. If the solutions of the two objective functions are less than a set threshold, the updated weights are the sensitivities of the corresponding log curves to the lithology.
[0070] In one example, the distance of the log curve is:
[0071]
[0072] where X ir 、X kr are log curves, i = 1, 2, …, N represents the data point serial number, r = 1, 2, …, P represents the variable serial number of the data at each data point, and w r is the weight.
[0073] In one example, the similarity is:
[0074]
[0075] where σ is a parameter.
[0076] In one example, the probability that two log curves belong to the same class is:
[0077]
[0078] where Y i is the lithology.
[0079] In one example, the objective function is:
[0080]
[0081] where λ is a parameter.
[0082] In one example, the gradient of the objective function is:
[0083]
[0084] In one example, the weights are updated according to the gradient as follows:
[0085] w r = w r + α·Δ r
[0086] where α is a parameter.
[0087] Specifically, known lithologies and their corresponding multiple logging curves are determined. For the multiple logging curves, initial weights are assigned as 1; according to the initial weights, the distances between every two logging curves are calculated as:
[0088]
[0089] Furthermore, the corresponding similarity is calculated as:
[0090]
[0091] The probability that two logging curves belong to the same class is calculated as:
[0092]
[0093] Furthermore, an objective function is established and solved:
[0094]
[0095] For each logging curve, the gradient of the corresponding objective function is calculated as::
[0096]
[0097] The weights are updated according to the gradient:
[0098] w r = w r + α·Δ r
[0099] The objective function is solved again. If the solutions of the objective function in two times are less than the set threshold value, the updated weights are the sensitivity degrees of the corresponding logging curves to the lithology; otherwise, the above steps are repeated until the solutions of the objective function in the front and back two times are less than the set threshold value.
[0100] Example 2
[0101] The present invention also provides a lithology-sensitive logging curve analysis device, including:
[0102] An assignment module, which determines known lithologies and their corresponding multiple logging curves, and assigns initial weights to the multiple logging curves;
[0103] A calculation module that calculates the distance between pairwise logging curves based on the initialized weights and then calculates the corresponding similarity;
[0104] An objective function solving module that calculates the probability that two logging curves belong to the same class, and then establishes and solves the objective function;
[0105] A gradient calculation module that calculates the gradient of the corresponding objective function for each logging curve;
[0106] An update module that updates the weights according to the gradient, re-solves the objective function, and if the solutions of the two objective functions are less than the set threshold, the updated weights are the sensitivity of the corresponding logging curves to lithology.
[0107] In one example, the distance of the logging curve is:
[0108]
[0109] where X ir and X kr are logging curves, i = 1, 2,..., N represents the data point serial number, r = 1, 2,..., P represents the variable serial number of the data at each data point, and w r is the weight.
[0110] In one example, the similarity is:
[0111]
[0112] where σ is a parameter.
[0113] In one example, the probability that two logging curves belong to the same class is:
[0114]
[0115] where Y i is the lithology.
[0116] In one example, the objective function is:
[0117]
[0118] where λ is a parameter.
[0119] In one example, the gradient of the objective function is:
[0120]
[0121] In one example, the weights are updated according to the gradient as:
[0122] w r = w r+α·Δ r
[0123] Among them, α is a parameter.
[0124] Specifically, determine the known lithologies and their corresponding multiple logging curves. For the multiple logging curves, assign an initial weight of 1; according to the initial weight, calculate the distance between every two logging curves as:
[0125]
[0126] Furthermore, calculate the corresponding similarity as:
[0127]
[0128] Calculate the probability that two logging curves belong to the same class as:
[0129]
[0130] Furthermore, establish an objective function and solve it:
[0131]
[0132] For each logging curve, calculate the gradient of the corresponding objective function as:
[0133]
[0134] Update the weight according to the gradient:
[0135] w r = w r +α·Δ r
[0136] Solve the objective function again. If the solutions of the objective function in two times are less than the set threshold, the updated weight is the sensitivity of the corresponding logging curve to the lithology; otherwise, repeat the above steps until the solutions of the objective function in the previous and subsequent times are less than the set threshold.
[0137] Example 3
[0138] Select the well logging data calibrated by cores, and perform normalization and standardization preprocessing on the well logging data (it is best to convert lithology into mineral component curves and hydrocarbon-bearing properties into fluid saturation curves). At the same time, use rock physics to calculate some other elastic parameter curves, such as shear wave impedance, etc.; then use principal component analysis (PCA) to select the principal components with large cumulative variance contribution rates to represent the input well logging information (for example, select well logging curves such as acoustic travel time, natural gamma, resistivity, shale content, spontaneous potential, effective porosity, water saturation, density, deep lateral resistivity, and shallow lateral resistivity); use the core calibration method to establish a learning sample with the well logging data and lithology of the core well.
[0139] The above well logging data (including elastic parameters calculated from well logging data), their interpretation results (porosity, saturation, lithology, etc.), and the lithology interpretation results after calibration together constitute the training dataset D(X ij ,Y i ), where X ij represents various well logging data, where i = 1, 2, …, N represents the data point serial number, and j = 1, 2, …, P represents the variable serial number (data dimension) of the data at each data point. Y i represents lithology.
[0140] The process of performing supervised learning based on the above data to analyze the sensitivity of well logging curves is as follows:
[0141] S1. For P well logging curves, initialize their weights w r to 1, r = 1, 2, …, P; or initialize w r to a random number between (0, 1);
[0142] S2. For all X ij , calculate the distance between sample X i. and X k. : Form a distance matrix D w (i, k);
[0143] S3. For all X ij calculate the similarity between sample X i. and X k. :
[0144]
[0145] where: σ is a parameter that can be specified as 1 or 2;
[0146] S4. For all X ij and its target lithology Y i , calculate sample Xi. and X k . Probability of belonging to the same class:
[0147] S5. Form the initial objective function: λ is a parameter and can be any number between (0, 1);
[0148] S6. For each logging curve, calculate the gradient of its objective function respectively:
[0149]
[0150] S7. Update w r : w r = w r + α·Δ r ; α is a parameter and can be any number between (0.4, 1);
[0151] S8. Recalculate the objective function with the updated w r :
[0152] S9. Compare the recalculated objective function with the objective function before update. If |ξ 0 (w) - ξ 1 (w)| < ε, then output w r , otherwise ξ 0 (w) = ξ 1 (w), and repeat S6 - S8; ε is a parameter and can be any number between (0, 0.05);
[0153] S10. The magnitude of the finally output w r value reflects the sensitivity of the logging curve r to a certain lithology.
[0154] Figure 2 shows a schematic diagram of a sensitivity histogram according to an embodiment of the present invention.
[0155] Figure 3 shows a schematic diagram of a lithology - logging curve sensitivity matrix according to an embodiment of the present invention.
[0156] For each lithology, calculate the sensitivity of each logging in sequence according to the above method and display it in the form of a histogram, as Figure 2 shown; the logging curve sensitivities of all lithologies form a lithology - logging curve sensitivity matrix and are displayed in the form of a matrix diagram, as Figure 3 shown.
[0157] The present invention selects the most appropriate logging curves, optimizes useful information from numerous logging curves, and thus more effectively recognizes, explores, and utilizes existing multi-type information.
[0158] Example 4
[0159] Figure 4 The block diagram of a lithology-sensitive logging curve analysis device according to an embodiment of the present invention is shown.
[0160] As Figure 4 shown, the lithology-sensitive logging curve analysis device includes:
[0161] An assignment module 201 determines known lithologies and their corresponding multiple logging curves, and assigns initial weights to the multiple logging curves.
[0162] A calculation module 202 calculates the distance between every two logging curves according to the initial weights, and further calculates the corresponding similarity.
[0163] An objective function solving module 203 calculates the probability that two logging curves belong to the same class, and further establishes and solves the objective function.
[0164] A gradient calculation module 204 calculates the gradient of the corresponding objective function for each logging curve.
[0165] An update module 205 updates the weights according to the gradient and re-solves the objective function. If the solutions of the objective function twice are less than a set threshold, the updated weights are the sensitivity degrees of the corresponding logging curves to the lithology.
[0166] As an optional solution, the distance between logging curves is:
[0167]
[0168] where X ir 、X kr are logging curves, i = 1, 2, …, N represents the serial number of data points, r = 1, 2, …, P represents the variable serial number of data at each data point, and w r is the weight.
[0169] As an optional solution, the similarity is:
[0170]
[0171] where σ is a parameter.
[0172] As an optional solution, the probability that two logging curves belong to the same class is:
[0173]
[0174] where Y i is the lithology.
[0175] As an alternative, the objective function is:
[0176]
[0177] where λ is a parameter.
[0178] As an alternative, the gradient of the objective function is:
[0179]
[0180] As an alternative, the weights are updated according to the gradient as:
[0181] w r = w r + α·Δ r
[0182] where α is a parameter.
[0183] Example 5
[0184] This embodiment provides an electronic device, which includes: a memory storing executable instructions; and a processor that runs the executable instructions in the memory to implement the above lithology-sensitive log curve analysis method.
[0185] The electronic device according to an embodiment of the present disclosure includes a memory and a processor.
[0186] The memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, and the computer program products may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory, etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.
[0187] The processor may be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In an embodiment of the present disclosure, the processor is used to run the computer-readable instructions stored in the memory.
[0188] Those skilled in the art should understand that, in order to solve the technical problem of how to obtain good user experience effects, well-known structures such as communication buses and interfaces may also be included in this embodiment, and these well-known structures should also be included in the protection scope of this disclosure.
[0189] For the detailed description of this embodiment, reference may be made to the corresponding descriptions in the foregoing embodiments, which will not be elaborated herein.
[0190] Example 6
[0191] This embodiment provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the lithology-sensitive log curve analysis method described above is implemented.
[0192] According to the computer-readable storage medium of the embodiments of the present disclosure, non-temporary computer-readable instructions are stored thereon. When the non-temporary computer-readable instructions are run by a processor, all or part of the steps of the methods of the foregoing embodiments of the present disclosure are executed.
[0193] The above computer-readable storage medium includes but is not limited to: optical storage media (such as CD-ROMs and DVDs), magneto-optical storage media (such as MOs), magnetic storage media (such as magnetic tapes or external hard drives), media with built-in rewritable non-volatile memories (such as memory cards), and media with built-in ROMs (such as ROM cartridges).
[0194] Those skilled in the art should understand that the purpose of the above description of the embodiments of the present invention is only to exemplarily illustrate the beneficial effects of the embodiments of the present invention, and is not intended to limit the embodiments of the present invention to any of the examples given.
[0195] The above embodiments of the present invention have been described. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments.
Claims
1. A method for analyzing lithology-sensitive logging curves, characterized in that, it includes: Determine known lithologies and their corresponding multiple logging curves, and assign initial weights to the multiple logging curves; According to the initial weights, calculate the distances between pairwise logging curves, and then calculate the corresponding similarities; Calculate the probability that two logging curves belong to the same class, and then establish and solve an objective function; For each logging curve, calculate the gradient of the corresponding objective function; Update the weights according to the gradient, and re-solve the objective function. If the solutions of the two objective functions are less than a set threshold, the updated weights are the sensitivities of the corresponding logging curves to lithology.
2. The method for analyzing lithology-sensitive logging curves according to claim 1, wherein, the distance of the logging curve is: Among them, X ir , X kr are logging curves, i = 1, 2, …, N represents the data point serial number, r = 1, 2, …, P represents the variable serial number of the data at each data point, and w r is the weight.
3. The method for analyzing lithology-sensitive logging curves according to claim 1, wherein, the similarity is: Among them, σ is a parameter.
4. The method for analyzing lithology-sensitive logging curves according to claim 1, wherein, the probability that two logging curves belong to the same class is: Among them, Y i is lithology.
5. The method for analyzing lithology-sensitive logging curves according to claim 1, wherein, the objective function is: where λ is a parameter.
6. The method for analyzing lithology-sensitive logging curves according to claim 1, wherein, the gradient of the objective function is:
7. The method for analyzing lithology-sensitive logging curves according to claim 1, wherein, updating the weights according to the gradient is: w r = w r + α·Δ r where α is a parameter.
8. A device for analyzing lithology-sensitive logging curves, characterized in that, it includes: An assignment module that determines known lithologies and their corresponding multiple logging curves, and assigns initial weights to the multiple logging curves; A calculation module that calculates the distances between pairwise logging curves according to the initial weights, and then calculates the corresponding similarities; An objective function solving module that calculates the probability that two logging curves belong to the same class, and then establishes and solves an objective function; A gradient calculation module that calculates the gradient of the corresponding objective function for each logging curve; An update module that updates the weights according to the gradient, re-solves the objective function, and if the solutions of the two objective functions are less than a set threshold, the updated weights are the sensitivities of the corresponding logging curves to lithology.
9. An electronic device, characterized in that, the electronic device includes: A memory that stores executable instructions; A processor that runs the executable instructions in the memory to implement the method for analyzing lithology-sensitive logging curves according to any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the method for analyzing lithology-sensitive logging curves according to any one of claims 1-7.