A method for identifying dominant flow channels based on fuzzy comprehensive evaluation

Abstract

The application discloses a kind of based on fuzzy comprehensive evaluation's advantage seepage channel identification method, it is related to oil reservoir development technical field.The method includes: obtaining oil reservoir related data, screening evaluation index and constructing evaluation index system;Using analytic hierarchy process to determine the weight of each index, and obtaining weight vector;According to the distribution characteristics of each index, membership function is constructed;Membership degree is calculated to each index observation value, and fuzzy relation matrix is formed;The weight vector and fuzzy relation matrix are carried out fuzzy matrix multiplication, and the fuzzy comprehensive evaluation result vector is obtained;Whether there is advantage seepage channel in oil reservoir area is determined according to the comparison result of evaluation value and preset threshold value, and dynamic data and / or numerical simulation result is verified.The application can effectively handle the fuzziness and uncertainty of oil reservoir data, improve the accuracy and reliability of advantage seepage channel identification, and optimize oil reservoir development decision.

Description

Technical Field

[0001] This invention relates to the field of reservoir development technology, specifically to a method for identifying dominant seepage channels based on fuzzy comprehensive evaluation. Background Technology

[0002] After oilfield development enters the medium-to-high water-cut stage, long-term water injection development can easily create dominant flow channels within the reservoir. Dominant flow channels refer to low-resistance flow paths formed in localized areas of the reservoir due to geological heterogeneity or inappropriate development measures. Their existence causes injected water to rush rapidly along these channels, resulting in serious problems such as reduced water drive efficiency, decreased oil recovery, and increased development costs. Therefore, accurately identifying the spatial distribution characteristics of dominant flow channels is crucial for optimizing water injection strategies, adjusting development plans, and improving the ultimate recovery rate of the oilfield.

[0003] Currently, the industry mainly uses dynamic analysis, static analysis, and numerical simulation methods to identify dominant flow channels. Dynamic analysis relies on dynamic well production data, such as changes in production, water cut trends, and pressure variations, to infer the existence of dominant flow channels by analyzing abnormal changes in these parameters. However, this method requires high data quality and accuracy; field data is often affected by measurement errors and interference factors, leading to unstable identification results. Furthermore, dynamic analysis struggles to quantitatively describe the specific morphology and scale of dominant flow channels, providing only qualitative judgments. Static analysis primarily relies on static geological data such as core analysis and well logging interpretation, identifying potential dominant flow channels by analyzing the spatial distribution characteristics of reservoir physical parameters. While this method reflects the inherent heterogeneity of the reservoir, it cannot capture the dynamic evolution characteristics caused by fluid flow and pressure changes during reservoir development, resulting in a disconnect between the identification results and actual development conditions. Numerical simulation, by establishing detailed geological and flow models, simulates the fluid migration process in the reservoir, thereby identifying dominant flow channels. Although this method can theoretically provide a relatively comprehensive analysis, it faces many challenges in practical applications: the model construction process is complex and requires a large number of accurate geological and engineering parameters; the computation is large and time-consuming; it is sensitive to the initial parameter settings, and small changes in parameters may lead to significant differences in the identification results; and model verification is difficult, making it hard to guarantee that the simulation results completely match the actual reservoir dynamics.

[0004] Fuzzy comprehensive evaluation, as an effective mathematical tool for handling fuzzy and uncertain problems, can comprehensively consider the influence of multiple factors and indicators, making it suitable for identifying dominant seepage channels, which are affected by various fuzzy factors. During the formation of dominant seepage channels, key parameters such as permeability, porosity, and water cut exhibit continuous changes and fuzzy boundaries, which traditional binary logic struggles to accurately describe. Theoretically, fuzzy comprehensive evaluation models can better handle these fuzzy characteristics, but current technologies lack a complete and systematic application method. Specifically, existing technologies fail to reasonably construct membership functions for different evaluation indicators and fail to accurately calculate fuzzy relation matrices, resulting in insufficient reliability of fuzzy comprehensive evaluation results and failing to meet the high-precision identification needs of oilfields. Existing methods have significant shortcomings in handling the fuzziness of evaluation indicators, making it difficult to simultaneously consider the comprehensive analysis of multi-source data and the reliability of identification results, thus limiting the accuracy of dominant seepage channel identification and affecting the scientific nature of subsequent development decisions. Summary of the Invention

[0005] The purpose of this invention is to provide a method for identifying dominant seepage channels based on fuzzy comprehensive evaluation, which has the advantages of effectively handling the fuzziness and uncertainty of reservoir data, improving the accuracy and reliability of dominant seepage channel identification, and thus optimizing reservoir development decisions.

[0006] This invention provides a method for identifying dominant seepage channels based on fuzzy comprehensive evaluation, comprising the following steps: S1. Acquire reservoir-related data; the reservoir-related data includes static data and dynamic data; S2. Based on the acquired reservoir-related data, select evaluation indicators and construct an evaluation indicator system for identifying dominant seepage channels; S3. Use the analytic hierarchy process (AHP) to determine the weights of each evaluation index and obtain the weight vector. S4. Based on the distribution characteristics of each evaluation index and the preset identification criteria for dominant seepage channels, construct the membership function for each evaluation index respectively; the membership function is used to characterize the degree to which each evaluation index belongs to dominant seepage channels and non-dominant seepage channels. S5. Based on the constructed membership function, calculate the membership degree of the actual observed values ​​of each evaluation index to obtain the membership degree value of each evaluation index belonging to the dominant seepage channel, and form a fuzzy relation matrix by combining the membership degree values ​​of all evaluation indicators; wherein, the actual observed values ​​come from the sample to be identified, and the sample to be identified is the reservoir area unit to be determined whether there is a dominant seepage channel. S6. Perform fuzzy matrix multiplication on the weight vector and the fuzzy relation matrix to obtain the fuzzy comprehensive evaluation result vector; S7. Based on the comparison results of the evaluation values ​​of each sample to be identified in the fuzzy comprehensive evaluation result vector with the preset evaluation threshold, determine whether there is a dominant seepage channel in the reservoir area corresponding to the target sample to be identified, and verify the identification results using reservoir dynamic data and / or numerical simulation results.

[0007] As can be seen from the above, the dominant seepage channel identification method based on fuzzy comprehensive evaluation provided in this application comprehensively processes multi-source data through steps such as acquiring reservoir-related data, constructing an evaluation index system, determining weights, calculating membership degrees, and performing fuzzy comprehensive evaluation. It effectively identifies dominant seepage channels and has the advantages of effectively handling the fuzziness and uncertainty of reservoir data, improving the accuracy and reliability of dominant seepage channel identification, thereby optimizing reservoir development decisions. Attached Figure Description

[0008] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0009] Figure 1 This invention provides a method for identifying advantageous seepage channels based on fuzzy comprehensive evaluation. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0011] After oilfield development enters the high water-cut stage, dominant seepage channels easily form within the reservoir, leading to inefficient circulation of injected water and a decline in oil recovery. Existing identification methods, such as dynamic analysis, static analysis, and numerical simulation, have limitations, including limited data accuracy, inability to reflect dynamic changes, complex models, and strong parameter dependencies. Particularly in handling the fuzziness and uncertainty of influencing factors, current technologies have not yet formed a complete and systematic method based on a fuzzy comprehensive evaluation model, resulting in a need to improve identification accuracy and reliability.

[0012] In this regard, such as Figure 1 As shown, this application proposes a method for identifying dominant seepage channels based on fuzzy comprehensive evaluation, including the following steps: S1. Obtain reservoir-related data; reservoir-related data includes static data and dynamic data.

[0013] S2. Based on the acquired reservoir-related data, evaluation indicators are selected, and an evaluation indicator system for identifying advantageous seepage channels is constructed.

[0014] S3. Use the analytic hierarchy process (AHP) to determine the weights of each evaluation index and obtain the weight vector.

[0015] S4. Based on the distribution characteristics of each evaluation index and the preset identification criteria for dominant seepage channels, construct the membership function for each evaluation index. The membership function is used to characterize the degree to which each evaluation index belongs to dominant seepage channels and non-dominant seepage channels.

[0016] S5. Based on the constructed membership function, calculate the membership degree of the actual observed values ​​of each evaluation index to obtain the membership degree value of each evaluation index belonging to the dominant seepage channel, and form a fuzzy relation matrix by combining the membership degree values ​​of all evaluation indicators; where the actual observed values ​​come from the sample to be identified, the sample to be identified is the reservoir area unit to be determined whether there is a dominant seepage channel.

[0017] S6. Perform fuzzy matrix multiplication on the weight vector and the fuzzy relation matrix to obtain the fuzzy comprehensive evaluation result vector.

[0018] S7. Based on the comparison results of the evaluation values ​​of each sample to be identified in the fuzzy comprehensive evaluation result vector with the preset evaluation threshold, determine whether there is a dominant seepage channel in the reservoir area corresponding to the target sample to be identified, and verify the identification results using reservoir dynamic data and / or numerical simulation results.

[0019] Reservoir-related data refers to information describing the geological characteristics and development dynamics of an oil reservoir. Static data reflects inherent reservoir properties, such as effective reservoir thickness, permeability, porosity, outburst coefficient, sedimentary microfacies, and clay content. Dynamic data reflects the dynamic response of the reservoir under injection and production, such as daily production, water cut, bottomhole flowing pressure, daily water injection volume, water-injected oil pressure, apparent water absorption index, pressure drop, production change rate, and water injection intensity.

[0020] Evaluation indicators refer to parameters used to measure the existence of dominant seepage channels in an oil reservoir area. The dominant seepage channel identification evaluation indicator system is a comprehensive evaluation set formed by organizing the selected evaluation indicators according to logical relationships.

[0021] The Analytic Hierarchy Process (AHP) is a multi-objective decision analysis method that decomposes complex problems into different levels and determines the relative importance of each element to the objective of the next higher level through pairwise comparisons. The weight vector, calculated using the AHP, represents the relative importance of the evaluation indicators to the final identification result.

[0022] Membership function is a concept in fuzzy mathematics used to describe the degree to which an element belongs to a fuzzy set. In this embodiment, the function quantifies the degree to which the actual observed values ​​of each evaluation index belong to a "dominant seepage channel" or a "non-dominant seepage channel".

[0023] The fuzzy relation matrix is ​​a matrix composed of the membership values ​​of various evaluation indicators. Its rows represent evaluation indicators, and its columns represent samples to be identified. Each element represents the degree to which a specific sample belongs to the dominant seepage channel under a specific evaluation indicator.

[0024] The sample to be identified refers to the reservoir area unit where it is necessary to determine whether there is a dominant flow channel.

[0025] Fuzzy matrix multiplication is a fuzzy mathematical operation rule that combines weight vectors with fuzzy relation matrices to obtain a comprehensive evaluation result.

[0026] The fuzzy comprehensive evaluation result vector is the output of the fuzzy matrix multiplication operation. It contains the comprehensive evaluation value of each sample to be identified, reflecting the comprehensive probability that there are dominant seepage channels in the sample area.

[0027] The preset evaluation threshold is a pre-set critical value used to binarize the evaluation values ​​in the fuzzy comprehensive evaluation result vector to determine whether there is a dominant seepage channel in the reservoir area.

[0028] Specifically, the first step is to acquire reservoir-related data. This data can be entered manually, for example, by manually reviewing paper archives or electronic documents to input static and dynamic data into the data processing system. Alternatively, the data can be directly exported from the oilfield database, for example, by writing database query scripts to extract static and dynamic data for a specified time period and area in batches.

[0029] Secondly, based on the acquired reservoir-related data, evaluation indicators are selected, and a system of evaluation indicators for identifying dominant seepage channels is constructed. When selecting evaluation indicators, an expert team can discuss and determine a set of indicators that influence the formation of dominant seepage channels, based on reservoir development experience. For example, permeability, porosity, water cut, and pressure drop can be initially selected as candidate indicators. Subsequently, these selected indicators are organized according to their inherent relationships and importance to the identification target, forming a structured evaluation indicator system.

[0030] Furthermore, the weights of each evaluation index are determined using the analytic hierarchy process (AHP), resulting in a weight vector. WThe process may include the following steps: First, establish a hierarchical model, with the identification of dominant seepage channels as the highest target layer and the selected evaluation indicators as the criterion layer. Then, compare the importance of each evaluation indicator pairwise using expert scoring or questionnaires, and construct a judgment matrix based on the comparison results. Next, calculate the largest eigenvalue and the corresponding eigenvector of the judgment matrix, and normalize the eigenvector to obtain the weight vector of each evaluation indicator.

[0031] Subsequently, based on the distribution characteristics of each evaluation indicator and the pre-defined criteria for identifying dominant seepage channels, membership functions for each evaluation indicator are constructed. These membership functions characterize the degree to which each evaluation indicator belongs to a dominant or non-dominant seepage channel. For example, for a given evaluation indicator, a threshold or interval can be manually set based on its historical data distribution, and a simple linear membership function can be constructed accordingly. Alternatively, a piecewise function can be directly specified based on experience to describe the relationship between the indicator value and the membership degree.

[0032] Based on this, according to the constructed membership function, the membership degree of each evaluation index is calculated based on the actual observed values, obtaining the membership degree value of each evaluation index belonging to the dominant seepage channel. The membership degree values ​​of all evaluation indices are then combined to form a fuzzy relation matrix. Specifically, for each sample to be identified, the actual observed values ​​of each evaluation index are substituted into the corresponding membership function to calculate its membership value. Arranging and combining the membership degree values ​​of all samples under all indices forms a fuzzy relation matrix. These actual observed values ​​come from the sample to be identified, which is the reservoir unit for which the existence of a dominant seepage channel is to be determined.

[0033] Next, a fuzzy matrix multiplication operation is performed between the weight vector and the fuzzy relation matrix to obtain the fuzzy comprehensive evaluation result vector. This fuzzy matrix multiplication operation can employ various fuzzy operators; for example, traditional matrix multiplication rules can be used, but multiplication is replaced with minimum value operations, and addition is replaced with maximum value operations. Through this operation, the weights of each evaluation index are combined with their corresponding membership values ​​to obtain the comprehensive evaluation value for each sample to be identified.

[0034] Finally, based on the comparison between the evaluation values ​​of each sample to be identified in the fuzzy comprehensive evaluation result vector and the preset evaluation threshold, it is determined whether a dominant seepage channel exists in the reservoir area corresponding to the target sample. The identification results are then verified using reservoir dynamic data and / or numerical simulation results. For example, a fixed threshold can be manually set; if the evaluation value of a sample is higher than this threshold, a dominant seepage channel is determined to exist; otherwise, it is determined not to exist. After identification, the accuracy of the identification results can be evaluated by comparing them with actual reservoir production dynamic data (such as water cut trends and pressure drop) or numerical simulation results. If the identification results do not match the actual situation, it may be necessary to manually adjust the evaluation indicators, weights, or membership functions to improve the accuracy of the identification.

[0035] This embodiment effectively addresses the fuzziness and uncertainty of multiple factors and indicators in the identification of dominant seepage channels in oil reservoirs through a systematic fuzzy comprehensive evaluation process. By acquiring static and dynamic data of the reservoir, constructing an evaluation index system and determining weights, and then constructing membership functions and fuzzy relation matrices, fuzzy comprehensive evaluation and verification are finally performed. This achieves accurate identification of dominant seepage channels in the oil reservoir area, overcoming the limitations of traditional methods in terms of data accuracy, dynamic reflection, and model complexity, improving the reliability of identification, and providing technical support for the optimization of reservoir development plans.

[0036] In one optional implementation, after acquiring reservoir-related data, the data is further preprocessed, including data cleaning, data standardization, and data normalization.

[0037] Specifically, data preprocessing is a crucial step in the data analysis process, aiming to improve data quality and make it more suitable for subsequent analysis and modeling. Raw reservoir data often contains missing values, outliers, noise, redundancy, and different dimensions. If these problems are not addressed, they will severely affect the accuracy of evaluation indicators and model performance. Data cleaning aims to identify and handle errors, inconsistencies, and missing values ​​in the data. For example, for missing data, methods such as mean imputation, median imputation, regression imputation, or record deletion can be used; for outliers, statistical methods (such as the 3σ criterion) or model-based methods can be used for detection and correction. Data standardization transforms data of different dimensions and orders of magnitude to a unified scale, eliminating the influence of dimensions. Common standardization methods include Z-score standardization and Min-Max standardization. Data normalization is a special type of standardization, typically referring to scaling data to a specific interval, such as [0,1]. This helps eliminate differences in dimensions and numerical ranges between different indicators, ensuring that all indicators have the same weight or influence in subsequent calculations.

[0038] Based on this, the specific steps for screening evaluation indicators include: using grey relational analysis to screen evaluation indicators that affect the formation and development of dominant seepage channels from the preprocessed data, based on reservoir development theory and the formation mechanism of dominant seepage channels, and constructing an evaluation indicator system for identifying dominant seepage channels.

[0039] Evaluation indicators include static and dynamic indicators. Static indicators include permeability, porosity, sedimentary microfacies, and clay content; dynamic indicators include water cut, pressure drop, production change rate, and injection intensity.

[0040] Specifically, grey relational analysis is an analytical method based on grey system theory, used to measure the degree of correlation between factors. It judges the tightness of the correlation between factors by calculating the similarity of the geometric shapes of sequence curves; the closer the curves are, the greater the correlation. In the identification of dominant seepage channels, this method can screen out key evaluation indicators with a high degree of correlation with the formation and development of dominant seepage channels from numerous reservoir characteristic parameters, thereby constructing a more representative and effective indicator system. Its implementation typically includes determining reference and comparison sequences, performing dimensionless processing on the sequences, calculating the correlation coefficient, and obtaining the grey relational degree. Reservoir development theory and the formation mechanism of dominant seepage channels are the theoretical foundation guiding the selection of evaluation indicators. Reservoir development theory covers multiple aspects such as reservoir physics, fluid flow, and displacement mechanisms, providing a macroscopic framework for understanding reservoir behavior; the formation mechanism of dominant seepage channels delves into how factors such as geological heterogeneity, injection-production well patterns, and development methods lead to the formation of local dominant flow paths in the reservoir. By combining these theoretical knowledge, we can initially define the range of candidate indicators that may be related to the dominant seepage channels, and provide a reasonable interpretation and verification of the results of the grey relational analysis.

[0041] Static indicators reflect the geological attributes and reservoir characteristics of an oil reservoir. These characteristics are relatively stable or change slowly during reservoir development and are the basic geological conditions for the formation of dominant flow channels. For example, permeability measures the ability of fluids to pass through porous media, porosity measures the proportion of pore space in the total volume of the reservoir, sedimentary microfacies reflect the sedimentary environment and geological characteristics during reservoir formation, and clay content affects permeability and porosity. Dynamic indicators reflect the fluid flow and production response characteristics of an oil reservoir during development. These characteristics change over time and can directly or indirectly reveal the dynamic evolution of dominant flow channels. For example, water cut is the proportion of water in the produced fluid, pressure drop reflects the resistance or driving force of fluid flow, production change rate reflects the trend of production changes in wells or areas, and water injection intensity reflects the displacement effect of water injection on the reservoir.

[0042] By preprocessing reservoir-related data through data cleaning, standardization, and normalization, problems such as noise, redundancy, and dimensional inconsistencies in the original data were effectively eliminated, significantly improving data quality and reliability. Based on this, and combining reservoir development theory and the formation mechanism of dominant seepage channels, a grey relational analysis method was used to screen static indicators such as permeability, porosity, sedimentary microfacies, and clay content, as well as dynamic indicators such as water cut, pressure drop, production change rate, and water injection intensity, from the preprocessed data, constructing a more scientific and targeted evaluation index system. This method ensures that the selected indicators can comprehensively and accurately reflect the essential characteristics and dynamic evolution process of dominant seepage channels, avoiding subjective assumptions and interference from irrelevant indicators. This makes the subsequent fuzzy comprehensive evaluation results more accurate and reliable, significantly improving the accuracy and efficiency of dominant seepage channel identification.

[0043] In one optional implementation, step S3 of determining the weight vector of each evaluation index specifically includes: First, in step S31, a hierarchical model is established, with the identification of dominant seepage channels as the target layer and the various evaluation indicators as the criterion layer. A hierarchical model is a structured method that decomposes complex problems into several levels. In this step, the ultimate goal of "identifying dominant seepage channels" is first set as the highest level, i.e., the target layer. Subsequently, the various evaluation indicators used to evaluate dominant seepage channels, such as permeability, porosity, and water content, are set as the next level, i.e., the criterion layer. This hierarchical structure helps to clearly define the scope of the problem, clarify the logical relationship between the evaluation objectives and evaluation elements, and lay the foundation for subsequent quantitative analysis. In this way, the complex identification problem is decomposed into components that are easier to manage and analyze.

[0044] Secondly, in step S32, a judgment matrix is ​​constructed to compare the importance of each evaluation indicator pairwise, using a 1-9 scale to determine the elements of the judgment matrix. The judgment matrix is ​​a core tool in the Analytic Hierarchy Process (AHP) used to quantify the relative importance of each evaluation indicator. When constructing the judgment matrix, domain experts or historical data analysis are required to compare any two evaluation indicators in the criteria layer pairwise, assessing their relative importance to achieving the target layer (identification of advantageous seepage channels). This comparison is quantified using a 1-9 scale, where 1 indicates that the two indicators are equally important, 9 indicates that one indicator is extremely more important than the other, 2, 4, 6, and 8 represent intermediate levels, and the reciprocal indicates a reverse comparison. For example, if indicator A is 3 times more important than indicator B, the corresponding element in the judgment matrix is ​​3, while indicator B is 1 / 3 more important than indicator A. Through this pairwise comparison, the subjective judgment of experts or decision-makers is transformed into a numerical form, forming a positive-reciprocal matrix.

[0045] Next, in step S33, the largest eigenvalue and corresponding eigenvector of the judgment matrix are calculated, and the eigenvectors are normalized to obtain the weight vectors of each evaluation index. After the judgment matrix is ​​constructed, it needs to be mathematically processed to extract the weight information. Specifically, this is done by calculating the largest eigenvalue of the judgment matrix. The eigenvectors are the result of the evaluation matrix and their corresponding eigenvalues. The largest eigenvalue reflects the consistency of the judgment matrix, while the eigenvectors contain the relative importance information of each evaluation indicator. To make these relative importance values ​​meaningful and convenient for subsequent calculations, the obtained eigenvectors need to be normalized, that is, the sum of all its elements is adjusted to 1. The normalized eigenvectors are the weight vectors of each evaluation indicator, where each element represents the relative weight of the corresponding evaluation indicator in the identification of dominant seepage channels.

[0046] Finally, in step S34, a consistency check is performed. If the check passes, the weight vector is valid; if it fails, the judgment matrix is ​​reconstructed until the check passes. Since the judgment matrix is ​​constructed based on subjective judgment, it may contain a certain degree of logical inconsistency. Therefore, a consistency check must be performed after obtaining the weight vector. The consistency check aims to evaluate the logical rationality of the expert judgment, usually by calculating the consistency ratio (CR). If the CR is less than a preset threshold (e.g., 0.1), the judgment matrix is ​​considered to have satisfactory consistency, and the obtained weight vector is valid and can be used for subsequent fuzzy comprehensive evaluation. Conversely, if the CR exceeds the threshold, it indicates serious inconsistency in the expert judgment, requiring a re-examination and adjustment of the pairwise comparison values ​​in the judgment matrix until the consistency check passes. This iterative process ensures the scientific validity and reliability of the weight vector, avoiding identification errors caused by subjective judgment bias.

[0047] Through the above technical solution, this application provides a systematic, quantitative, and verifiable method for determining the weights of evaluation indicators. By establishing a hierarchical model, the complex problem of identifying dominant seepage channels is decomposed into clear target and criterion layers, thus structuring the problem. By constructing a judgment matrix and using the 1-9 scaling method for pairwise comparisons, the subjective experience of experts or decision-makers can be transformed into objective numerical values, thereby quantifying the relative importance of each evaluation indicator. Furthermore, by calculating the maximum eigenvalue and eigenvector of the judgment matrix and performing normalization, the weight vectors of each evaluation indicator can be scientifically derived. More importantly, the introduction of a consistency check mechanism can effectively identify and correct logical inconsistencies in the judgment process, ensuring that the determined weight vectors have high reliability and rationality. This significantly improves the scientificity and accuracy of weight assignment in fuzzy comprehensive evaluation, thereby improving the accuracy of dominant seepage channel identification results and providing a more solid data foundation for reservoir development decisions.

[0048] In an optional implementation, in step S4, the distribution characteristics of each evaluation index are determined by statistical analysis, and the parameters of the membership function are determined based on the distribution characteristics. Specifically, this includes: obtaining historical sample data of reservoir areas with known dominant seepage channels and without dominant seepage channels; performing statistical analysis on the historical sample data to determine the distribution range of each evaluation index under the two conditions of dominant seepage channels and without dominant seepage channels; and determining the inflection point and interval parameters of the membership function based on the overlapping area or boundary point of the distribution ranges of the two types of samples.

[0049] Specifically, the distribution characteristics of each evaluation index refer to the statistical performance of its values ​​in different reservoir regions (e.g., regions with and without dominant seepage channels), including central tendency, dispersion, skewness, and kurtosis. Determining these characteristics through statistical analysis aims to objectively and quantitatively understand the actual behavioral patterns of the indicators, providing data support for the subsequent construction of membership functions. Statistical analysis methods may include, but are not limited to, histogram analysis, probability density function estimation, box plot analysis, and descriptive statistics calculations (such as mean, median, standard deviation, and quartiles). These methods can reveal the numerical range, concentration areas, and outliers of data across different categories.

[0050] Based on this, the parameters of the membership function (such as inflection points, interval boundaries, slopes, etc.) directly determine how much the function maps a specific index value to its membership in a fuzzy set (such as "dominant seepage channels"). Determining these parameters based on the distribution characteristics obtained from statistical analysis ensures that the membership function more accurately and objectively reflects the fuzzy relationship between the index and the dominant seepage channels, avoiding subjective assumptions. Parameter determination can be based on statistical analysis results; for example, the mean or median of a certain index in the "dominant seepage channel" sample can be used as a key inflection point of the membership function; the boundary of the intersection or overlapping region of the two types of sample distributions can be used as the transition interval of the membership function.

[0051] To achieve the above objectives, it is first necessary to obtain historical sample data of reservoir areas with known dominant flow channels and those without. Historical sample data forms the basis for statistical analysis; these data are collections of reservoir area units whose existence of dominant flow channels has been clearly identified through actual exploration, development verification, or long-term production dynamic monitoring. Obtaining these samples with clear classification labels is crucial for supervised learning and pattern recognition, providing a reliable training set for subsequent statistical analysis. Historical sample data can be derived from production reports, well logging data, core analysis data, geological research reports, numerical simulation results, etc., from developed oilfields. During acquisition, it is essential to ensure the completeness, accuracy, and representativeness of the data, and to classify and label the samples.

[0052] Subsequently, statistical analysis was performed on historical sample data to determine the distribution range of each evaluation indicator under two conditions: the presence and absence of dominant seepage channels. The statistical analysis aimed to quantify the numerical characteristics of each evaluation indicator under different categories (presence / absence of dominant seepage channels). Determining the distribution range, i.e., clarifying the minimum, maximum, and main concentration intervals of the indicators under each condition, helps to intuitively understand the performance differences of the indicators under different states and provides a quantitative basis for the subsequent construction of membership functions. Descriptive statistics such as the maximum, minimum, mean, standard deviation, and quartiles of each indicator in the two types of samples can be calculated. Simultaneously, histograms or kernel density estimation plots can be drawn to visually display the numerical distribution of the indicators in the two types of samples, observing the degree of overlap and separation.

[0053] Finally, the inflection point and interval parameters of the membership function are determined based on the overlapping region or boundary point of the two types of sample distributions. The overlapping region or boundary point is the key area distinguishing between the two types of samples. In fuzzy logic, the membership function typically transitions from 0 to 1 or from 1 to 0 in these regions. Using these statistically significant points as the inflection point and interval parameters of the membership function allows it to more accurately capture the fuzzy boundary of the indicator's transition from a "non-dominant seepage channel" to a "dominant seepage channel," thereby improving the accuracy of fuzzy evaluation. For example, for an indicator, if its maximum value in the "no dominant seepage channel" sample overlaps with its minimum value in the "dominant seepage channel" sample, then the boundary point of this overlapping region or a specific value within the overlapping region (such as the midpoint or weighted average) can be used as the inflection point of the membership function. If there is a clear boundary point between the two types of sample distributions, this boundary point can be directly used as the critical value for the membership function to change from 0 to 1 or from 1 to 0.

[0054] Through the above technical solution, this application, when constructing the membership function, no longer relies on subjective experience or pre-set fuzzy standards, but is based on statistical analysis of a large amount of historical sample data. By acquiring historical sample data of known classifications and conducting in-depth statistical analysis, the true distribution patterns and ranges of each evaluation index under both the presence and absence of dominant seepage channels can be objectively revealed. Based on this, the inflection points and interval parameters of the membership function are precisely determined by utilizing the overlapping areas or boundary points of the two types of sample distributions. This allows the constructed membership function to more accurately and objectively reflect the degree to which each evaluation index belongs to the dominant seepage channel. This significantly improves the scientific rigor and accuracy of fuzzy comprehensive evaluation, thus making the finally identified dominant seepage channels more consistent with the actual reservoir conditions, providing a more reliable basis for the refined development and management of the reservoir.

[0055] In one optional implementation, in step S4, different forms of membership functions are used for different types of evaluation indicators: for indicators that are larger and more conducive to the formation of dominant seepage channels, a semi-trapezoidal membership function is used; for indicators that are smaller and more conducive to the formation of dominant seepage channels, an inverse semi-trapezoidal membership function is used; and for indicators that are conducive to the formation of dominant seepage channels within a certain interval, a triangular membership function is used.

[0056] Specifically, the evaluation indicators are diverse, and their mechanisms of influence on the formation of dominant seepage channels vary. For example, higher values ​​for some indicators are more conducive to the formation of dominant seepage channels, while lower values ​​for others are more conducive, and some indicators are only conducive to formation within specific ranges. Therefore, to accurately quantify the fuzzy relationship between these different types of indicators and dominant seepage channels, it is necessary to select the function form that best reflects the degree of membership based on the inherent characteristics of the indicators.

[0057] For indices where larger values ​​are more conducive to the formation of dominant seepage channels, the characteristic is that the larger the value, the stronger its membership in the dominant seepage channel. A semi-trapezoidal membership function can well simulate this monotonically increasing relationship. Specifically, when the index value is below a certain threshold, the membership degree may be 0 or close to 0; when the index value is above another threshold, the membership degree may be 1 or close to 1; between the two thresholds, the membership degree value increases linearly or non-linearly. For example, high permeability usually means that fluid flows more easily, thus being more conducive to the formation of dominant seepage channels.

[0058] For indices where smaller values ​​are more conducive to the formation of dominant seepage channels, the characteristic is that the smaller the value, the higher the degree to which it belongs to the dominant seepage channel. The membership function of an inverse semi-trapezoidal distribution can simulate this monotonically decreasing relationship. Specifically, when the index value is below a certain threshold, the membership degree may be 1 or close to 1; when the index value is above another threshold, the membership degree may be 0 or close to 0; between the two thresholds, the membership degree value decreases linearly or non-linearly. For example, low clay content usually means a purer rock skeleton and better pore-throat connectivity, thus being more conducive to the formation of dominant seepage channels.

[0059] For an index that favors the formation of dominant seepage channels within a certain range, the characteristic is that its value is most conducive to the formation of dominant seepage channels when it is within an optimal range, while deviations from this range are detrimental. A triangular membership function can simulate this "moderate is best" relationship. Specifically, the membership degree reaches its maximum value of 1 at the center point of the optimal range; as the index value deviates to either side, the membership degree gradually decreases until it reaches 0 outside the range. For example, excessively high or low porosity may be detrimental to the formation of dominant seepage channels; only within a moderate porosity range can good storage and seepage conditions be guaranteed.

[0060] By employing the above technical solutions, membership function forms that align with the characteristics of different types of evaluation indicators can more accurately characterize the fuzzy relationships between each evaluation indicator and the formation of dominant seepage channels. For example, semi-trapezoidal, inverse semi-trapezoidal, and triangular membership functions are suitable for monotonically increasing, monotonically decreasing, and interval-optimal indicators, respectively, thus avoiding membership calculation biases caused by inappropriate function form selection. This makes the membership values ​​of each evaluation indicator belonging to dominant seepage channels calculated in step S5 more realistic and reliable, and the constructed fuzzy relationship matrix can more accurately reflect the actual situation of the sample to be identified. Finally, the fuzzy comprehensive evaluation performed in steps S6 and S7 will be based on more precise input, significantly improving the accuracy and reliability of dominant seepage channel identification results, effectively solving the problem that a single membership function form cannot fully adapt to the diverse characteristics of indicators.

[0061] In an optional implementation, in step S5, the expression for the fuzzy relation matrix is: Among them, the fuzzy relation matrix R for matrix, m The number of evaluation indicators, n The number of samples to be identified, matrix elements r ij Indicates the first j The sample to be identified in the first... i The membership degree value of a dominant seepage channel under each evaluation index. , .

[0062] Specifically, the expression for the fuzzy relation matrix refers to the expression used to clearly and formally represent the fuzzy relation matrix. R The mathematical form of the fuzzy relation matrix is ​​typically represented using the general notation of matrices, which uses the arrangement of rows and columns to show the positions and relationships of the elements within the matrix. Its purpose is to provide an intuitive and easily understood structure, enabling subsequent fuzzy calculations to be performed accurately and avoiding computational errors caused by unclear matrix structures. The fuzzy relation matrix R is a core component of the fuzzy comprehensive evaluation method; it quantifies the degree to which the sample to be identified belongs to a specific fuzzy concept (here, the dominant seepage channel) under various evaluation indicators. The construction of this matrix is ​​a crucial step in connecting the membership degree calculation results of the evaluation indicators with the final comprehensive evaluation result. Matrix representation of fuzzy relation matrix R The dimension, in which, m Represents the number of rows in the matrix. nThis represents the number of columns in the matrix. This dimension definition clearly defines the size of the matrix, ensuring that the matrix can correctly accommodate all relevant data when dealing with multiple evaluation metrics and multiple samples to be identified. m The number of evaluation indicators represents the total number of evaluation indicators used to identify dominant seepage channels. These evaluation indicators can be static or dynamic, and together they constitute the evaluation indicator system. m The determination of the value directly affects the number of rows in the fuzzy relation matrix, that is, each row in the matrix corresponds to an evaluation index. n The number of samples to be identified represents the total number of reservoir units for which dominant flow channels need to be determined. These samples are discrete regions in the actual reservoir that need to be analyzed, and each sample will be evaluated using fuzzy comprehensive criteria based on evaluation indicators. n The determination of the number of columns in the fuzzy relation matrix directly affects the number of columns in the matrix; that is, each column in the matrix corresponds to a sample to be identified. Matrix elements r ij It is a fuzzy relation matrix R A specific value, located at the [number]th position in the matrix. i Line number j This element is one of the core data points in fuzzy comprehensive evaluation, directly reflecting the degree of fuzzy membership of a single evaluation index to a single sample to be identified. j The nth sample to be identified refers to the nth sample among all samples to be identified, arranged in a certain order. j Each reservoir unit will be analyzed using a fuzzy comprehensive evaluation method to determine whether a dominant flow channel exists. i The evaluation indicator refers to the first evaluation indicator in a certain order among all evaluation indicators. i Several indicators are used to identify dominant seepage channels. Each evaluation indicator calculates the membership degree for all samples to be identified, reflecting their contribution to the dominant seepage channel. The membership degree value for belonging to a dominant seepage channel is a value between 0 and 1, indicating the degree to which a sample belongs to the fuzzy concept of "dominant seepage channel" under a certain evaluation indicator. The closer the value is to 1, the higher the degree of membership; the closer the value is to 0, the lower the degree of membership. These membership degree values ​​are the basic data for fuzzy comprehensive evaluation and are calculated using a membership function.

[0063] This explicit matrix structure allows the membership values ​​of the evaluation indicators calculated in step S5 to be systematically organized, providing standardized input for the fuzzy matrix multiplication operation in the subsequent step S6. Because the dimensions of the matrix and the meaning of each element are precisely defined, data confusion and calculation errors are effectively avoided, ensuring the accuracy and reliability of the fuzzy comprehensive evaluation results, thereby improving the accuracy of identifying dominant seepage channels.

[0064] In an optional implementation, in step S6, the fuzzy matrix multiplication operation employs... The operator, namely the rule of taking the smaller value first and then the larger value, is specifically: Let the fuzzy comprehensive evaluation result vector be... , For the first j The evaluation value of each sample to be identified is then: in, For the first i The weights of each evaluation indicator are all arranged into a weight vector. ; For the first j The sample to be identified in the first... i The membership degree values ​​of dominant seepage channels under each evaluation index, i.e., the fuzzy relation matrix. R The Middle i Line number j The elements of the column.

[0065] The M(∧,∨) operator, also known as the max-min composition operator, is a commonly used fuzzy composition operation rule in fuzzy mathematics. The core idea of ​​this operator is to first use the "minimum" (∧) operation to adjust the weights of each evaluation index. Its corresponding membership value The contributions of each indicator are combined to determine its effective contribution to a specific sample to be identified. The "smallest contribution" operation here reflects the "weakest link effect," meaning that the contribution of an indicator will not exceed the smaller of its weight or membership value, ensuring the rigor of the evaluation. Subsequently, the "largest contribution" (∨) operation selects the maximum value from the effective contributions of all indicators as the final evaluation value for the sample to be identified. This "take the largest" operation reflects the principle of "highlighting the advantages", that is, as long as one or a few key indicators perform well in terms of weight and membership, it is enough to make the sample to be identified receive a high evaluation.

[0066] The above technical solution clarifies the specific operator for fuzzy matrix multiplication, namely, using the M(∧,∨) operator to perform the operation rule of taking the smaller value first and then the larger value. This gives the fuzzy comprehensive evaluation process a clear mathematical foundation and physical meaning, avoiding uncertainty in the evaluation results caused by inappropriate operator selection. This operator can effectively integrate the weight and membership information of various evaluation indicators, and is particularly good at capturing and highlighting key indicators that perform well in both weight and membership, thus ensuring the final evaluation value... This allows for a more accurate reflection of the likelihood of a dominant flow channel existing in the sample to be identified. This precise and interpretable evaluation mechanism significantly improves the reliability and accuracy of the dominant flow channel identification results, providing a solid foundation for the subsequent threshold-based judgment in step S7, and thus contributing to the optimization of reservoir development decisions.

[0067] In an optional implementation, in step S7, the preset evaluation threshold is set using the receiver operating characteristic (ROC) curve method. Specifically, this method includes plotting an ROC curve, calculating the area under the curve, and selecting a value that optimizes both sensitivity and specificity as the preset evaluation threshold λ. Subsequently, if the evaluation value b ≥ λ for a sample to be identified in the fuzzy comprehensive evaluation result vector, it is determined that the reservoir area corresponding to the identified sample has a dominant flow channel; if b < λ, it is determined that the reservoir area corresponding to the identified sample does not have a dominant flow channel.

[0068] The Receiver Operating Characteristic (ROC) curve method is a statistical method widely used for evaluating the performance of binary classification models. Its core lies in visually demonstrating the model's performance at different classification thresholds by plotting a curve showing the relationship between the true positive rate and the false positive rate. In this application, this method aims to provide an objective and systematic way to determine the classification threshold of the fuzzy comprehensive evaluation result, thereby maximizing the overall performance of the identification method—that is, effectively identifying dominant seepage channels while minimizing false positives. In practice, a set of historical sample data with known true classification results (i.e., the presence or absence of dominant seepage channels) is typically required as a training set. The fuzzy comprehensive evaluation method is applied to these sample data to obtain their evaluation values, which are then compared with the true classification results to construct the ROC curve.

[0069] Plotting the ROC curve is a fundamental step in the ROC curve method. It iterates through a series of possible evaluation thresholds, calculates the true positive rate (Sensitivity) and false positive rate (1-Specificity) at each threshold, and plots these points on a two-dimensional coordinate system. The true positive rate represents the proportion of samples with actual dominant seepage channels that are correctly identified, while the false positive rate represents the proportion of samples with actual non-dominant seepage channels that are incorrectly identified. The area under the curve (AUC) is a commonly used metric for measuring the overall performance of a classifier; a larger AUC value indicates a stronger discriminative ability of the classifier.

[0070] Sensitivity refers to the proportion of samples where a dominant seepage channel actually exists that are correctly identified as having one; specificity refers to the proportion of samples where a dominant seepage channel does not actually exist that are correctly identified as not having one. In dominant seepage channel identification, it is often necessary to find an optimal balance between these two metrics. The preset evaluation threshold is chosen to achieve the optimal values ​​for both sensitivity and specificity. This can be achieved through various methods. For example, the threshold corresponding to the point on the ROC curve closest to the top-left corner (0,1) can be selected, representing the ideal state with the lowest false positive rate and the highest true positive rate. Alternatively, the Youden Index can be used, which selects the threshold that maximizes (sensitivity + specificity - 1). These methods aim to ensure that the identification method identifies as many true dominant seepage channels as possible while minimizing the misclassification of non-dominant seepage channels as dominant seepage channels.

[0071] Once the optimal preset evaluation threshold is determined Thus, the final decision rule is clarified. For any sample to be identified in the fuzzy comprehensive evaluation result vector, if its evaluation value... b Greater than or equal to If the value is positive, the corresponding reservoir area is determined to have a dominant flow channel. Conversely, if the evaluation value is negative... b Less than If the condition is met, it is determined that there is no dominant flow channel in the reservoir area. This logical judgment transforms the continuous evaluation results into a clear binary classification conclusion, thus completing the identification process of the dominant flow channel.

[0072] By employing the receiver operating characteristic (ROC) curve method to determine the preset evaluation threshold, this application overcomes the problems of strong subjectivity and insufficient accuracy in threshold setting found in traditional methods. This method systematically analyzes the sensitivity and specificity of the identification method under different thresholds and selects the optimal value for both as the evaluation threshold, thereby ensuring the objectivity and reliability of the identification results. This allows for a more accurate balance between identifying true dominant flow channels (high sensitivity) and avoiding misidentification of non-dominant flow channels (high specificity) when identifying dominant flow channels. Therefore, this application significantly improves the accuracy and practicality of dominant flow channel identification, providing a more reliable decision-making basis for reservoir development and effectively guiding subsequent injection-production adjustments and production enhancement measures.

[0073] In one optional implementation, in step S7, if the identification result matches the actual reservoir development dynamics, the identification of the dominant seepage channel is completed; if the identification result does not match the actual reservoir development dynamics, the evaluation index system, weight vector, or membership function for the identification of the dominant seepage channel is adjusted, and the calculations of S2 to S7 are repeated until the identification result matches the actual reservoir development dynamics.

[0074] Specifically, after verifying the identification results using reservoir dynamic data and / or numerical simulation results, if the verification results show a high degree of consistency between the identified dominant seepage channel distribution and the actual reservoir production performance (such as injection-production connectivity, waterflooding patterns, production changes, etc.), then the current identification model and parameter settings are considered accurate and effective, and the identification process can be terminated. This consistency can be judged in various ways, such as by comparing whether the identified dominant seepage channel regions are consistent with areas with fast advance rates at the waterflood front in actual production, or whether they match the streamline distribution predicted by numerical simulation.

[0075] However, when the identification results differ significantly from the actual reservoir development dynamics, it indicates that the current fuzzy comprehensive evaluation model needs improvement. In this case, the model parameters need to be adjusted. Adjusting the evaluation index system for identifying dominant seepage channels means re-examining the evaluation indicators selected in step S2. It may be necessary to introduce new indicators that better reflect the characteristics of dominant seepage channels, or to remove those indicators that do not contribute significantly to the identification effect. For example, if it is found that the existing indicators fail to fully capture the influence of micropore throat structure on seepage, micropore structure parameters can be considered as new evaluation indicators. Adjusting the weight vector involves the importance of each evaluation indicator determined in step S3. If some indicators are found to have a greater impact on the formation of dominant seepage channels in practice, but their weights are low, their weights need to be readjusted to more accurately reflect their relative importance. This can be done through expert scoring, recalculating using the analytic hierarchy process, or combining it with sensitivity analysis. Adjusting the membership function refers to the function constructed in step S4 used to characterize the degree to which each evaluation indicator belongs to the dominant seepage channel. If the membership function of a certain indicator is found to be set too broadly or too strictly, resulting in inaccurate calculation of the membership degree for actual observations, then its shape, inflection point, or interval parameters need to be corrected. For example, by analyzing more historical sample data or combining expert experience, the threshold of the membership function can be fine-tuned to make it more accurately reflect the relationship between the indicator value and the dominant seepage channel.

[0076] After adjusting the evaluation index system, weight vector, or membership function, these modified parameters need to be re-introduced into the fuzzy comprehensive evaluation process. Starting from step S2 (screening evaluation indicators, if the index system has been adjusted), or S3 (determining weights, if the weight vector has been adjusted), or S4 (constructing membership functions, if the membership functions have been adjusted), and continuing until step S7, new identification results are obtained through recalculation. This process is a closed-loop iterative optimization process. After each adjustment, identification and verification are performed again until the identified dominant seepage channel distribution matches the expected production dynamics of the actual reservoir.

[0077] Through the above technical solution, this application provides an adaptive method for identifying dominant seepage channels. When the initial identification results do not match the actual reservoir development dynamics, it can provide timely feedback and make targeted adjustments to the evaluation index system, weight vector, or membership function, and then recalculate and verify. This iterative optimization mechanism effectively avoids identification deviations caused by improper initial parameter settings or reservoir complexity, significantly improving the accuracy and reliability of dominant seepage channel identification. Through continuous correction and verification, it ensures a high degree of consistency between the final identification results and the actual production status of the reservoir, thereby providing a more accurate and reliable basis for the refined development and management of the reservoir, reducing decision-making risks, and improving development efficiency.

[0078] The following example will provide a more detailed explanation of the above technical solution: In an oil field X This block, belonging to a sandstone reservoir, is currently in a high water-cut development stage. Low injected water circulation efficiency has been observed, suggesting the potential existence of dominant seepage channels. To accurately identify these channels, optimize the water injection strategy, and improve oil recovery, this method is employed for dominant seepage channel identification.

[0079] First, reservoir-related data were acquired and preprocessed. Static data for the block were collected, such as effective reservoir thickness, permeability, porosity, sedimentary microfacies, and clay content. Simultaneously, dynamic data were collected, including daily production, water cut, bottomhole flowing pressure, daily water injection volume, water-injected oil pressure, apparent water absorption index, pressure drop, production change rate, and water injection intensity. To ensure data quality, these raw data underwent preprocessing, including data cleaning to remove outliers and missing values, data standardization to eliminate the influence of different dimensions, and data normalization to map the numerical range to the [0,1] interval. This preprocessing step effectively improved the data foundation for subsequent analysis, avoiding the significant impact of data accuracy on the results in traditional dynamic analysis methods.

[0080] Next, based on reservoir development theory and the formation mechanism of dominant seepage channels, grey relational analysis was used to screen key evaluation indicators affecting the formation and development of dominant seepage channels from the preprocessed data. The constructed evaluation index system for identifying dominant seepage channels includes eight indicators: permeability (X1), porosity (X2), clay content (X3), water cut (X4), pressure drop (X5), production change rate (X6), water injection intensity (X7), and sedimentary microfacies (X8). Among them, sedimentary microfacies were quantified into numerical values; for example, channel sand bodies were assigned a value of 1, distributary channels 0.8, and floodplains 0.2. This systematic index screening and quantification overcomes the limitation of traditional static analysis methods in reflecting the dynamic changes of reservoirs.

[0081] Subsequently, the analytic hierarchy process is used to determine the weights of each evaluation index. First, a hierarchical structure model is established, with "identification of dominant seepage channels" as the target layer and the above 8 evaluation indexes as the criterion layer. Then, several experts in the field of reservoir development are invited to compare the importance of each evaluation index pairwise, and a judgment matrix is constructed using the 1-9 scale method. By calculating the maximum eigenvalue and the corresponding eigenvector of the judgment matrix, and normalizing the eigenvector, the weight vector of each evaluation index is obtained. W , for example: .

[0082] Finally, a consistency test is performed to ensure the validity of the weight vector. If the test fails, the judgment matrix is reconstructed until it passes. This way of determining weights by combining expert experience and mathematical methods makes the weight assignment more reasonable and reliable, reducing the excessive influence of subjective factors on the identification results.

[0083] After determining the weights, according to the distribution characteristics of each evaluation index and the preset identification criteria for dominant seepage channels, the membership functions of each evaluation index are constructed respectively. By statistically analyzing the historical sample data of reservoir areas where dominant seepage channels are known to exist and not exist, the distribution ranges of each evaluation index in the two cases are determined, and then the inflection points and interval parameters of the membership function are determined. For different types of indexes, membership functions of different forms are used: for indexes such as permeability (X1), water cut (X4), and water injection intensity (X7) where the larger the value, the more conducive to the formation of dominant seepage channels, a membership function with a semi-trapezoidal distribution is used; for indexes such as shale content (X3) where the smaller the value, the more conducive to the formation of dominant seepage channels, an inverse semi-trapezoidal distribution membership function is used; for indexes such as porosity (X2) where being in a certain interval is conducive to the formation of dominant seepage channels, a triangular distribution membership function is used. For example: For permeability (X1), a membership function with a semi-trapezoidal distribution is used: When X1 ≥ 100 mD, the membership degree is 1; When X1 ≤ 50 mD, the membership degree is 0; When 50 mD < X1 < 100 mD, the membership degree is μ(X1) = (X1 - 50) / (100 - 50).

[0084] For shale content (X3), an inverse semi-trapezoidal distribution membership function is used.

[0085] When X3 ≤ 5%, the membership degree is 1; When X3 ≥ 15%, the membership degree is 0; When 5% < X3 < 15%, the membership degree is μ(X3) = (15 - X3) / (15 - 5).

[0086] For porosity (X2): A triangular distribution membership function is adopted. When X2 = 25%, the membership degree is 1; When X2 ≤ 15% or X2 ≥ 35%, the membership degree is 0; When 15% < X2 < 25%, the membership degree is μ(X2)=(X2 - 15) / (25 - 15); When 25% < X2 < 35%, the membership degree is μ(X2)=(35 - X2) / (35 - 25).

[0087] For other indicators (X4 - X8), corresponding semi - trapezoidal or triangular membership functions are constructed according to their distribution characteristics.

[0088] This refined construction of membership functions can effectively handle the fuzziness and uncertainty inherent in various factors affecting the formation of dominant seepage channels, which are difficult to directly handle by traditional numerical simulation methods.

[0089] Next, calculate the fuzzy relation matrix. Twenty samples to be identified (reservoir area units) in this block are selected, and the membership degrees of the actual observed values of the 8 evaluation indicators for each sample are calculated to obtain the membership degree values of each indicator belonging to the dominant seepage channel. These membership degree values form an 8×20 fuzzy relation matrix R . For example, the membership degree values of each indicator for the sample to be identified 1 are [0.8, 0.9, 0.7, 0.85, 0.75, 0.6, 0.82, 0.9], and the membership degree values of each indicator for the sample to be identified 2 are [0.3, 0.4, 0.6, 0.25, 0.3, 0.2, 0.35, 0.8].

[0090] Then, perform the fuzzy matrix multiplication operation on the weight vector W and the fuzzy relation matrix R to obtain the fuzzy comprehensive evaluation result vector B . This operation uses the M(∧,∨) operator, that is, the operation rule of taking the minimum first and then the maximum. For each sample to be identified j , the calculation formula for its evaluation value bj is: . For example, the evaluation value of the sample to be identified 1: ; The evaluation value of the sample to be identified 2 . Through this fuzzy comprehensive evaluation, a comprehensive evaluation of multiple factors and multiple indicators is achieved, comprehensively reflecting various factors affecting the formation of dominant seepage channels, and improving the accuracy and reliability of identification.

[0091] Finally, the dominant seepage channels were identified and verified. The receiver operating characteristic (ROC) curve method was used to determine the preset evaluation threshold. By plotting ROC curves and calculating the area under the curve, the value that optimizes both sensitivity and specificity is selected as the evaluation threshold, for example... The evaluation values ​​of each sample to be identified in the fuzzy comprehensive evaluation result vector are compared with... Comparison: If the evaluation value of sample 1 to be identified is... If the evaluation value of sample 2 is positive, it is determined that there is a dominant seepage channel in the reservoir area corresponding to the sample; if the evaluation value of sample 2 to be identified is negative, it is determined that there is a dominant seepage channel in the reservoir area corresponding to the sample. If the value is 0.15 < 0.16, then the reservoir area corresponding to this sample is determined to lack a dominant flow channel. To verify the accuracy of the identification results, the production dynamic data of this block were analyzed.

[0092] The results showed that the well group corresponding to sample 1 exhibited a rapid increase in water cut and a significant decrease in production, consistent with the characteristics of dominant seepage channels. In contrast, the well group corresponding to sample 2 showed stable production with gradual changes in water cut and production, validating the accuracy of the identification results. If the identification results do not match the actual reservoir development dynamics, the dominant seepage channel identification evaluation index system, weight vector, or membership function is adjusted, and the calculations for S2 to S7 are repeated until the identification results conform to the actual reservoir development dynamics. This verification and iterative adjustment mechanism further ensures the accuracy of the identification results and overcomes the problems of traditional numerical simulation methods, such as strong parameter dependence and susceptibility to initial parameter settings affecting identification accuracy.

[0093] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for identifying dominant seepage channels based on fuzzy comprehensive evaluation, characterized in that, Includes the following steps: S1. Acquire reservoir-related data; the reservoir-related data includes static data and dynamic data; S2. Based on the acquired reservoir-related data, select evaluation indicators and construct an evaluation indicator system for identifying dominant seepage channels; S3. Use the analytic hierarchy process (AHP) to determine the weights of each evaluation index and obtain the weight vector. S3 specifically includes: S31. Establish a hierarchical model, with the identification of dominant seepage channels as the target layer and the evaluation indicators as the criteria layer. S32. Construct a judgment matrix, compare the importance of each evaluation index pairwise, and use the 1-9 scale method to determine the elements of the judgment matrix. S33. Calculate the largest eigenvalue and the corresponding eigenvector of the judgment matrix, normalize the eigenvector, and obtain the weight vector of each evaluation index. S34. Perform a consistency check. If the check passes, the weight vector is valid. If the check fails, reconstruct the judgment matrix until the check passes. S4. Based on the distribution characteristics of each evaluation index and the preset identification criteria for dominant seepage channels, construct the membership function for each evaluation index respectively; the membership function is used to characterize the degree to which each evaluation index belongs to dominant seepage channels and non-dominant seepage channels. S5. Based on the constructed membership function, calculate the membership degree of the actual observed values ​​of each evaluation index to obtain the membership degree value of each evaluation index belonging to the dominant seepage channel, and form a fuzzy relation matrix by combining the membership degree values ​​of all evaluation indicators; wherein, the actual observed values ​​come from the sample to be identified, and the sample to be identified is the reservoir area unit to be determined whether there is a dominant seepage channel. S6. Perform fuzzy matrix multiplication on the weight vector and the fuzzy relation matrix to obtain the fuzzy comprehensive evaluation result vector; S7. Based on the comparison results of the evaluation values ​​of each sample to be identified in the fuzzy comprehensive evaluation result vector with the preset evaluation threshold, determine whether there is a dominant seepage channel in the reservoir area corresponding to the target sample to be identified, and verify the identification results using reservoir dynamic data and / or numerical simulation results.

2. The method according to claim 1, characterized in that, S1 further includes: preprocessing the reservoir-related data; the preprocessing includes data cleaning, data standardization, and data normalization. S2 specifically includes: based on reservoir development theory and the formation mechanism of dominant seepage channels, using grey relational analysis to screen out evaluation indicators that affect the formation and development of dominant seepage channels from the preprocessed data, and constructing a dominant seepage channel identification and evaluation index system; The evaluation indicators include static indicators and dynamic indicators; among them, static indicators include permeability, porosity, sedimentary microfacies, and mud content; dynamic indicators include water content, pressure drop, production change rate, and water injection intensity.

3. The method according to claim 1, characterized in that, In S4, the distribution characteristics of each evaluation index are determined through statistical analysis, and the parameters of the membership function are determined based on the distribution characteristics, specifically including: Historical sample data of reservoir areas with known dominant seepage channels and those without are obtained; statistical analysis is performed on the historical sample data to determine the distribution range of each evaluation index under the two conditions of dominant seepage channels and non-dominant seepage channels; based on the overlapping area or boundary point of the distribution range of the two types of samples, the inflection point and interval parameters of the membership function are determined.

4. The method according to claim 3, characterized in that, In S4, different membership functions are used for different types of evaluation indicators: for indicators that are larger and more conducive to the formation of dominant seepage channels, a semi-trapezoidal membership function is used; for indicators that are smaller and more conducive to the formation of dominant seepage channels, an inverse semi-trapezoidal membership function is used; and for indicators that are conducive to the formation of dominant seepage channels within a certain interval, a triangular membership function is used.

5. The method according to claim 1, characterized in that, In S5, the expression for the fuzzy relation matrix is: Among them, the fuzzy relation matrix R for matrix, m The number of evaluation indicators, n The number of samples to be identified, matrix elements r ij Indicates the first j The sample to be identified in the first... i The membership degree value of a dominant seepage channel under each evaluation index. , .

6. The method according to claim 5, characterized in that, The fuzzy matrix multiplication operation in S6 adopts The operator, namely the rule of taking the smaller value first and then the larger value, is specifically: Let the fuzzy comprehensive evaluation result vector be... , For the first j The evaluation value of each sample to be identified is then: in, For the first i The weights of each evaluation indicator are all arranged into a weight vector. ; For the first j The sample to be identified in the first... i The membership degree values ​​of dominant seepage channels under each evaluation index, i.e., the fuzzy relation matrix. R The Middle i Line number j The elements of the column.

7. The method according to claim 6, characterized in that, In S7, the preset evaluation threshold adopts the receiver operating characteristic curve method, specifically including: Plot the ROC curve, calculate the area under the curve, and select the value that optimizes both sensitivity and specificity as the preset evaluation threshold. ; If the evaluation value of a certain sample to be identified in the fuzzy comprehensive evaluation result vector If so, it is determined that the reservoir area corresponding to the identified sample has a dominant seepage channel; if If the identification sample corresponds to an oil reservoir area with a dominant seepage channel, then it is determined that the oil reservoir area has a dominant seepage channel.

8. The method according to claim 1, characterized in that, In step S7, if the identification result matches the actual reservoir development dynamics, the identification of the dominant seepage channel is completed; if the identification result does not match the actual reservoir development dynamics, the evaluation index system, weight vector, or membership function for the identification of the dominant seepage channel is adjusted, and the calculations of steps S2 to S7 are repeated until the identification result matches the actual reservoir development dynamics.

Images