Comprehensive analysis method for main control factors of fracturing data aiming at yield

Abstract

The invention provides a comprehensive analysis method for main control factors of fracturing data for yield, and belongs to the technical field of unconventional oil and gas reservoir exploitation. According to the method, through parameter screening, modeling analysis, index evaluation, comprehensive analysis and effect verification, a relatively objective and accurate yield main control factor evaluation result is obtained, and a reference basis is provided for unconventional oil reservoir development. The invention provides a data-driven comprehensive analysis method flow for the main control factors of the fracturing data aiming at the yield, data internal relations are mined as much as possible, results of various analysis methods are fully considered through mathematical means, and comprehensive evaluation, judgment and verification of the main control factors of the yield are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of unconventional oil and gas reservoir exploitation, and in particular to a comprehensive analysis method of main control factors of fracturing data for production. Background Art

[0002] At present, unconventional oil reservoirs have developed micro-fractures, strong reservoir heterogeneity, narrow pore throats, complex seepage mechanisms, strong reservoir heterogeneity, unbalanced fracturing transformation, and poor production capacity prediction. Traditional physical drive models are difficult to accurately characterize the actual situation on site, and have high requirements for physical mechanisms, making it difficult to guide on-site fracturing operations. The development of unconventional oil and gas resources such as coalbed methane wells and conglomerate reservoirs requires the use of horizontal well volume fracturing technology to increase production capacity and recovery.

[0003] However, the fracturing effect is affected by a variety of geological and engineering factors, such as coal structure, reservoir parameters, perforation parameters, proppant dosage, etc. Therefore, how to fully mine fracturing data, accurately identify the main controlling factors affecting the fracturing effect, and optimize the fracturing plan is the key to improving fracturing efficiency and economy.

[0004] The analysis of the main controlling factors of fracturing data production refers to the use of fracturing data, through various mathematical methods and models, to identify the main parameters affecting production, and to evaluate and optimize them to improve the fracturing effect and production capacity of unconventional oil and gas wells.

[0005] At present, there are many methods for evaluating the importance of features, but the evaluation results are not uniform. For example, methods based on gray correlation degree analysis, linear correlation analysis, lasso regression, random forest model weights, etc., for the same analysis data set and target parameters, the importance rankings produced are different, and the emphasis of each importance is not the same. Subsequent analysis mainly relies on manual experience inspection and screening. In addition, the input parameters have not been filtered for collinearity and correlation tests, and there are errors with the actual main control factors, and there are errors in the analysis results. In the indicator weight evaluation method, fuzzy mathematics is often used for weight evaluation, or it is directly performed on the original data, and the parameter size and distribution are mistakenly directly linked to the degree of importance; or the indicator system is determined by qualitative analysis, which is greatly affected by manual experience. Finally, the main control factors obtained by analysis are often difficult to verify, and it is too costly to conduct field experiments. Summary of the invention

[0006] In order to solve the problems existing in the prior art, the present invention provides a comprehensive analysis method for the main controlling factors of fracturing data for production. Through parameter screening, modeling analysis, index evaluation, comprehensive analysis, and effect verification, a relatively objective and accurate evaluation result of the main controlling factors of production is obtained, providing a reference basis for the development of unconventional oil reservoirs. The present invention provides a data-driven, comprehensive analysis method process for the main controlling factors of fracturing data for production, which explores the internal connections of the data as much as possible, fully considers the results of multiple analysis methods through mathematical means, and realizes the comprehensive evaluation, judgment and verification of the main controlling factors of production.

[0007] The present invention provides a comprehensive analysis method for main control factors of fracturing data for production, comprising the following steps:

[0008] Step S1, preprocessing the fracturing data of the field production: performing data missing and data anomaly processing on the fracturing data of the field production to obtain processed fracturing data;

[0009] Step S2, feature selection and transformation of geotechnical parameters that affect fracturing data: calculating the collinearity of geotechnical parameters and the correlation between geotechnical parameters, screening geotechnical parameters, and extracting features of geotechnical parameters by data dimensionality reduction and data augmentation of processed fracturing data;

[0010] Step S3, using multiple evaluation methods to evaluate the importance of the fracturing parameters of the target production: using multiple importance evaluation methods to evaluate the importance of the characteristics of the target production parameters, and integrating the importance evaluation results of multiple importance evaluation methods to establish an importance evaluation matrix of geological engineering parameters, wherein the importance evaluation methods include grey correlation method, CatBoost ensemble learning method and SHAP evaluation;

[0011] Step S4, conduct a comprehensive evaluation of the main control factors of the evaluation matrix through fuzzy mathematics: determine the weight of the characteristic vector of each importance evaluation method in the importance evaluation matrix based on the entropy weight method, obtain the final characteristic importance score by weighted calculation, and sort to obtain the main control factors of production.

[0012] Preferably, in step S1, the missing data is filled using a KNN nearest neighbor prediction filling method.

[0013] Preferably, in step S2, feature selection and transformation of geological engineering parameters affecting fracturing data are performed, including the following steps:

[0014] The VIF analysis method was used to conduct collinearity analysis on the geo-engineering parameters, and the principal component analysis method was used to conduct correlation analysis between the geo-engineering parameters;

[0015] The VIF analysis method is used to reduce the dimension of geoengineering parameters and remove collinearity while retaining the interpretability of the data.

[0016] Preferably, the variance inflation factor VIF is used in pairs to determine the collinearity between the geological engineering parameters. The VIF calculation method is as follows:

[0017]

[0018] For parameter combinations with multicollinearity and variance inflation factor VIF greater than 10, principal component replacement is performed, and dimension reduction is performed using principal component analysis, retaining one principal component. The principal component calculation method is as follows:

[0019] And meet

[0020] in,

[0021] R j 2 is the determination coefficient obtained by linear regression of the other k-1 geo-engineering parameters when the j-th geo-engineering parameter is used as the explained variable, and k is the total number of geo-engineering parameters involved in dimensionality reduction;

[0022] Among them, Z1 is the first principal component, X j is the jth geoengineering parameter dataset, φ 1j is the first principal component weight of the jth geoengineering parameter;

[0023] n is the total number of sample wells;

[0024] m is the total number of geo-engineering parameters;

[0025] j is the serial number of the geological engineering parameter;

[0026] i is the sample well number;

[0027] x ij is the jth geological engineering parameter of the ith well.

[0028] PCA (Principal Component Analysis) is a commonly used data analysis method that can reduce high-dimensional data to a low-dimensional space while retaining the main characteristic components of the data. The mathematical principle of PCA is to find the direction with the largest data variance as the new coordinate axis through covariance matrix and eigenvalue analysis, and then project the data onto these coordinate axes to achieve the purpose of dimensionality reduction. The advantage of PCA is that it can reduce the complexity of the data, reduce computational overhead, remove noise, and extract important information. The disadvantage of PCA is that it may not be applicable to all data, there may be a loss of useful information, and it can only process numerical data. Figure 4As shown in Figure 1, PCA maps two-dimensional data points into a one-dimensional axis and preserves the distribution information of the points as much as possible.

[0029] Preferably, for parameters with weak collinearity, indicating that they are more independent, the data is augmented by cross-multiplication to increase data complexity, and the two-dimensional transformation converts X1, X2 into X1, X2, X1 2 ,X2 2 ,X1·X2 five vectors.

[0030] The dataset after dimension reduction and augmentation is represented as {X 改造段长 ,Z 主成分1 ,…,X 用液强度}.

[0031] Preferably, in step S3, the geological engineering parameter data set after dimensionality reduction and augmentation is used to evaluate the importance of the characteristics of the target production parameters using a variety of importance evaluation methods, and the importance evaluation results of the multiple importance evaluation methods are combined to establish an importance evaluation matrix of the geological engineering parameters. The target production parameter is used as the dependent variable, and other geological engineering parameters are used as independent variables. The importance score of each geological engineering parameter to the characteristic of the target production parameter is calculated, and finally the importance evaluation matrix E is constructed.

[0032] Preferably, in step S4, the analysis results of the importance evaluation matrix E are comprehensively evaluated by fuzzy mathematics, the final parameter importance scores of the target parameters are calculated and ranked, and the main controlling factors of the output are analyzed and determined.

[0033] Preferably, in step S4, the importance evaluation matrix E is standardized to obtain a fuzzy evaluation matrix, and the entropy weight method in fuzzy mathematics is used to calculate the weights of each importance evaluation method and the weights of each geological engineering parameter to obtain a comprehensive fuzzy evaluation vector, and the final importance score of each geological engineering parameter is obtained, and the scores are sorted from high to low to determine the main controlling factors of production.

[0034] Preferably, when the importance evaluation matrix E is standardized, for any geological engineering parameter, one of positive indicator normalization or negative indicator normalization is selected to normalize the importance evaluation matrix E based on the principle of whether it is closer to the actual situation of the parameter.

[0035] Preferably, the positive indicator normalization process is as follows:

[0036]

[0037] in,

[0038] E b is the data vector of evaluation index b in the evaluation matrix, e ab The value of evaluation parameter a in evaluation index b in the evaluation matrix, nab It is the normalized score value of evaluation parameter a in evaluation index b.

[0039] Preferably, the negative indicator normalization process is as follows:

[0040]

[0041] in,

[0042] E b is the data vector of evaluation index b in the evaluation matrix, e ab The value of evaluation parameter a in evaluation index b in the evaluation matrix, n ab It is the normalized score value of evaluation parameter a in evaluation index b.

[0043] Preferably, the weights of each importance evaluation method and the weights of each geological engineering parameter are calculated by the entropy weight method, and the steps are as follows:

[0044] Calculate the value e of evaluation parameter a in evaluation index b ab The proportion of evaluation index b ab :

[0045]

[0046] Among them, h is the number of evaluation parameters;

[0047] Calculate the information entropy value r of the evaluation index b b :

[0048]

[0049] Calculate the weight ω of each evaluation index b :

[0050]

[0051] Among them, u is the number of evaluation indicators;

[0052] Finally, the evaluation matrix E is multiplied by the index weight vector to obtain the fuzzy comprehensive evaluation matrix S:

[0053]

[0054] in,

[0055]

[0056] in,

[0057] [ξ], [J] and are the feature importance score vectors obtained by grey correlation method, CatBoost ensemble learning method and SHAP evaluation respectively.

[0058] Each element in is an element of the comprehensive evaluation matrix E, which represents the importance score of the geological engineering parameters obtained by using different importance evaluation methods;

[0059] ω is the evaluation index vector,

[0060] The weight calculation is performed on the principal component vector, and the abstract dimensionality reduction and augmented geological engineering parameters in the comprehensive fuzzy evaluation vector S are restored to the original geological engineering parameters to obtain the importance score of the original geological engineering parameters. The relationship between the collinear geological engineering parameter combination and its dimensionality reduction principal component is shown as follows:

[0061]

[0062] in,

[0063] S Z is the comprehensive evaluation score of the principal component evaluation parameters;

[0064] j is the serial number of the geological engineering parameter;

[0065] k is the total number of geo-engineering parameters involved in dimensionality reduction;

[0066] S Zj is the comprehensive evaluation score of the jth geo-engineering parameter after principal component reduction;

[0067] φ zj is the weight of the jth geoengineering parameter of the principal component.

[0068] Preferably, step S5 is also included, establishing a regression prediction model of geological engineering parameters and target production parameters, introducing the final feature importance score of each importance rating method as the basis for initializing the regression prediction model, and verifying the rationality of the main control factor analysis results by comparing the training results of the regression prediction model and the accuracy of the regression prediction model.

[0069] Preferably, in step S5, the rationality of the main control factor analysis result is evaluated by using a prediction accuracy index combined with the accuracy requirement:

[0070]

[0071] Among them, err is the minimum error requirement, e true is the true value, e pred is the predicted value, n is the total number of sample wells, acc err To meet the accuracy requirements of the prediction, acc err If it is 1, it means that the prediction accuracy of all samples meets the requirements.

[0072] Preferably, the calculation method for feature importance scoring using the grey correlation method is as follows:

[0073]

[0074] Among them, ξ j is the grey correlation coefficient between geological engineering parameter j and production parameter y, y i is the yield of sample well i, x ij is the fracturing data of geological engineering parameter j of sample well i, n is the total number of sample wells, and ρ is the resolution coefficient, which is generally 0.5. The larger the grey correlation coefficient is, the closer the relationship with the target parameter is, which is a positive indicator.

[0075] Preferably, when using the CatBoost ensemble learning evaluation method for importance evaluation, the geological engineering parameters are used as input and the production parameters are used as output, the CatBoost ensemble learning model is trained, and the calculation method for deriving the geological engineering parameter feature importance score is as follows:

[0076]

[0077]

[0078] in, is the importance evaluation of the geological engineering parameter j in the integrated learning model, M is the number of trees of the geological engineering parameter j in the generative model; T is the tree, T q is the qth tree, L is the number of leaf nodes in each tree, L-1 is the number of non-leaf nodes in the tree, v t is the feature associated with node t, I t 2 is the reduction in squared loss after node t is split. For example, before the split, the node has 100 data points and the MSE is 1; after the split, it is divided into two nodes, one with 30 data points and the MSE is 0.9, and the other with 70 data points and the MSE is 0.8. Considering the weighting, I t 2 It is equal to (100 1-300.9-70 0.8) / 100=0.17. t 2 The larger the value, the greater the impact of the node on the accuracy of the final predicted output. Therefore, the greater the importance of the node feature, the greater the impact on the prediction of the target parameter, which is a positive indicator.

[0079] The CatBoost model uses a fully symmetric decision tree as the base learner, and uses combined category features and target feature statistics to enhance feature dimensions. At the same time, an unbiased gradient estimate is obtained through the sorting boosting algorithm, and the performance of the model is improved by increasing the computational complexity, which is more in line with the current situation of small sample data on site. The feature importance ranking is determined by extracting the internal feature weights of the CatBoost model.

[0080] Preferably, based on the trained CatBoost ensemble learning model, the SHAP value is used to analyze the feature importance. The SHAP importance score calculation method is as follows:

[0081]

[0082] in, is the SHAP importance score of geotechnical parameter j, {X 改造段长 ,Z 主成分1 ,…,X 用液强度} is the geological engineering dataset after dimension reduction and augmentation, X j is the fracturing data of geological engineering parameter j; h is the number of geological engineering parameters after dimension reduction and augmentation, {X 改造段长 ,Z 主成分1 ,…,X 用液强度}\{X j} does not include {X j} is the possible set of all input geotechnical parameter characteristics, f x (S) is the prediction of the characteristic subset S of geo-engineering parameters. is the proportion of characteristic combinations of the geological engineering parameter characteristic subset S, and the sum of the proportions of characteristic combinations of all possible geological engineering parameter characteristic subsets S is equal to 1. x The () function is the prediction function of the CatBoost model. The larger the SHAP value, the greater the impact on the target parameter, which is a positive indicator.

[0083] SHAP (Shapley Additive exPlanations) is a game theory-based method that can explain the output of any machine learning model. The core idea of ​​SHAP is to calculate the marginal contribution of each feature to the model output, that is, the SHAP value of each feature. The SHAP value satisfies additivity, that is, the model output is equal to the sum of the SHAP values ​​of all features plus a baseline value. SHAP supports many types of models, including deep learning, tree models, linear models, etc., and can handle high-dimensional and time series data. Explain the feature importance of the CatBoost model through SHAP.

[0084] Fuzzy mathematics evaluation is a comprehensive evaluation method based on fuzzy mathematics, which can handle evaluation problems involving fuzzy concepts. The basic steps of fuzzy mathematics comprehensive evaluation are as follows: determine the evaluation target and evaluation index system, that is, the object to be evaluated and the various factors that affect the object; determine the factor set and comment set, that is, use ordinary sets to represent each factor and possible evaluation results; determine the weight of each factor, that is, use fuzzy sets to represent the importance of each factor to the evaluation target; determine the membership function of each factor, that is, use functions to represent the membership of each factor to each comment, that is, the degree of belonging to a certain comment; construct a fuzzy comprehensive evaluation matrix, that is, use a matrix to represent the membership of each factor to each comment; perform fuzzy comprehensive operations, that is, use weights and membership to calculate and obtain the final evaluation results.

[0085] Preferably, KNN nearest neighbor prediction to fill missing data specifically includes: considering the sample as a group of vectors, measuring the distance between sample vectors by Euclidean geometric distance, and filling sample vectors with missing values ​​using the average value of known values ​​of multiple groups of sample vectors closest to them.

[0086] Preferably, the padded data set is represented as {X 改造段长 ,X 加砂强度 ,…,X 用液强度 ,Y}, where Y is a column vector of production parameter data. The geological engineering parameters come from n sample wells, with a total of m items of geological engineering parameters. The data set {X 改造段长 ,X 加砂强度 ,…,X 用液强度 The matrix formed by the matrix is ​​n×m, where the data of the geological engineering parameter j of the sample well i is represented by x ij .

[0087] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0088] 1. The present invention realizes a comprehensive evaluation method of main control factors that combines modeling analysis with fuzzy mathematics, overcomes the problem that multiple main control schemes are difficult to comprehensively consider in the modeling analysis method, as well as the influence of human factors on weights and the difficulty in obtaining the indicator matrix in fuzzy mathematics, and provides a verification method for the rationality of the combination of main control factors.

[0089] 2. The present invention proposes a comprehensive analysis method for the main controlling factors of fracturing data for production, which explores the internal connections of the data as much as possible, fully considers the results of multiple analysis methods, and comprehensively evaluates and judges the main controlling factors of production. Effectively utilize fracturing data, extract the main features related to production, and integrate multiple analysis methods and models to obtain more objective and accurate evaluation results of the main controlling factors of production, providing a reference for the development of unconventional oil reservoirs. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1A schematic diagram of a geological engineering parameter characteristic transformation process according to an embodiment of the present invention;

[0091] Figure 2 A schematic diagram of a process for constructing a geological engineering parameter importance evaluation matrix according to an embodiment of the present invention;

[0092] Figure 3 A schematic diagram of a fuzzy comprehensive evaluation process of geological engineering parameter importance according to an embodiment of the present invention;

[0093] Figure 4 Schematic diagram of reducing two-dimensional data to one dimension using principal component analysis (large information loss);

[0094] Figure 5 Schematic diagram of reducing two-dimensional data to one dimension using principal component analysis (minimum information loss);

[0095] Figure 6 The flowchart is a method for comprehensive analysis of main controlling factors of fracturing data for production according to an embodiment of the present invention. DETAILED DESCRIPTION

[0096] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings.

[0097] The present invention provides a comprehensive analysis method for main control factors of fracturing data for production, comprising the following steps:

[0098] Step S1, preprocessing the fracturing data of the field production: performing data missing and data anomaly processing on the fracturing data of the field production to obtain processed fracturing data;

[0099] Step S2, feature selection and transformation of geotechnical parameters that affect fracturing data: calculating the collinearity of geotechnical parameters and the correlation between geotechnical parameters, screening geotechnical parameters, and extracting features of geotechnical parameters by data dimensionality reduction and data augmentation of processed fracturing data;

[0100] Step S3, using multiple evaluation methods to evaluate the importance of the fracturing parameters of the target production: using multiple importance evaluation methods to evaluate the importance of the characteristics of the target production parameters, and integrating the importance evaluation results of multiple importance evaluation methods to establish an importance evaluation matrix of geological engineering parameters, wherein the importance evaluation methods include grey correlation method, CatBoost ensemble learning method and SHAP evaluation;

[0101] Step S4, conduct a comprehensive evaluation of the main control factors of the evaluation matrix through fuzzy mathematics: determine the weight of the characteristic vector of each importance evaluation method in the importance evaluation matrix based on the entropy weight method, obtain the final characteristic importance score by weighted calculation, and sort to obtain the main control factors of production.

[0102] Furthermore, in step S1, the missing data is filled using the KNN nearest neighbor prediction filling method.

[0103] Furthermore, in step S2, feature selection and transformation are performed on the geological engineering parameters that affect the fracturing data, including the following steps:

[0104] The VIF analysis method was used to conduct collinearity analysis on the geo-engineering parameters, and the principal component analysis method was used to conduct correlation analysis between the geo-engineering parameters;

[0105] The VIF analysis method is used to reduce the dimension of geoengineering parameters and remove collinearity while retaining the interpretability of the data.

[0106] Furthermore, the variance inflation factor (VIF) is used to determine the collinearity between the geological engineering parameters. The VIF calculation method is as follows:

[0107]

[0108] For parameter combinations with multicollinearity and variance inflation factor VIF greater than 10, principal component replacement is performed, and dimension reduction is performed using principal component analysis, retaining one principal component. The principal component calculation method is as follows:

[0109] And meet

[0110] in,

[0111] R j 2 is the determination coefficient obtained by linear regression of the other k-1 geo-engineering parameters when the j-th geo-engineering parameter is used as the explained variable, and k is the total number of geo-engineering parameters involved in dimensionality reduction;

[0112] Among them, Z1 is the first principal component, X j is the jth geoengineering parameter dataset, φ 1j is the first principal component weight of the jth geoengineering parameter;

[0113] n is the total number of sample wells;

[0114] m is the total number of geo-engineering parameters;

[0115] j is the serial number of the geological engineering parameter;

[0116] i is the sample well number;

[0117] x ij is the jth geological engineering parameter of the ith well.

[0118] Furthermore, for parameters with weak collinearity, it means that they are more independent. By cross-multiplying, the data is expanded to increase the data complexity. The two-dimensional transformation converts X1, X2 into X1, X2, X1 2 ,X2 2 ,X1·X2 five vectors.

[0119] The dataset after dimension reduction and augmentation is represented as {X 改造段长 ,Z 主成分1 ,…,X 用液强度}.

[0120] Furthermore, in step S3, the geological engineering parameter data set after dimensionality reduction and augmentation is used to evaluate the importance of the characteristics of the target production parameters using a variety of importance evaluation methods, and the importance evaluation results of a variety of importance evaluation methods are combined to establish an importance evaluation matrix of the geological engineering parameters. The target production parameter is used as the dependent variable, and other geological engineering parameters are used as independent variables. The importance score of each geological engineering parameter to the characteristic of the target production parameter is calculated, and finally the importance evaluation matrix E is constructed.

[0121] Furthermore, in step S4, the analysis results of the importance evaluation matrix E are comprehensively evaluated by fuzzy mathematics, the final parameter importance scores of the target parameters are calculated and ranked, and the main control factors of the output are analyzed and determined.

[0122] Furthermore, in step S4, the importance evaluation matrix E is standardized to obtain a fuzzy evaluation matrix, and the entropy weight method in fuzzy mathematics is used to calculate the weights of each importance evaluation method and the weights of each geological engineering parameter to obtain a comprehensive fuzzy evaluation vector, and the final importance score of each geological engineering parameter is obtained, and the parameters are sorted from high to low according to the score to determine the main controlling factors of production.

[0123] Furthermore, when the importance evaluation matrix E is standardized, for any geological engineering parameter, one of the positive indicator normalization or the negative indicator normalization is selected to normalize the importance evaluation matrix E according to the principle of whether it is closer to the actual situation of the parameter.

[0124] Furthermore, the normalization of positive indicators is as follows:

[0125]

[0126] in,

[0127] E b is the data vector of evaluation index b in the evaluation matrix, e ab The value of evaluation parameter a in evaluation index b in the evaluation matrix, n ab It is the normalized score value of evaluation parameter a in evaluation index b.

[0128] Furthermore, the negative indicators are normalized as follows:

[0129]

[0130] in,

[0131] E b is the data vector of evaluation index b in the evaluation matrix, e ab The value of evaluation parameter a in evaluation index b in the evaluation matrix, n ab It is the normalized score value of evaluation parameter a in evaluation index b.

[0132] Furthermore, the entropy weight method is used to calculate the weights of each importance evaluation method and the weights of each geological engineering parameter. The steps are as follows:

[0133] Calculate the value e of evaluation parameter a in evaluation index b ab The proportion of evaluation index b ab :

[0134]

[0135] Among them, h is the number of evaluation parameters;

[0136] Calculate the information entropy value r of the evaluation index b b :

[0137]

[0138] Calculate the weight ω of each evaluation index b :

[0139]

[0140] Among them, u is the number of evaluation indicators;

[0141] Finally, the evaluation matrix E is multiplied by the index weight vector to obtain the fuzzy comprehensive evaluation matrix S:

[0142]

[0143] in,

[0144]

[0145] in,

[0146] [ξ], [J] and are the feature importance score vectors obtained by grey correlation method, CatBoost ensemble learning method and SHAP evaluation respectively.

[0147] Each element in is an element of the comprehensive evaluation matrix E, which represents the importance score of the geological engineering parameters obtained by using different importance evaluation methods;

[0148] ω is the evaluation index vector,

[0149] The weight calculation is performed on the principal component vector, and the abstract dimensionality reduction and augmented geological engineering parameters in the comprehensive fuzzy evaluation vector S are restored to the original geological engineering parameters to obtain the importance score of the original geological engineering parameters. The relationship between the collinear geological engineering parameter combination and its dimensionality reduction principal component is shown as follows:

[0150]

[0151] in,

[0152] S Z is the comprehensive evaluation score of the principal component evaluation parameters;

[0153] j is the serial number of the geological engineering parameter;

[0154] k is the total number of geo-engineering parameters involved in dimensionality reduction;

[0155] S Zj is the comprehensive evaluation score of the jth geo-engineering parameter after principal component reduction;

[0156] φ zj is the weight of the jth geoengineering parameter of the principal component.

[0157] Furthermore, it also includes step S5, establishing a regression prediction model of geological engineering parameters and target production parameters, introducing the final feature importance score of each importance rating method as the basis for initializing the regression prediction model, and verifying the rationality of the main control factor analysis results by comparing the training results of the regression prediction model and the accuracy of the regression prediction model.

[0158] Furthermore, in step S5, the rationality of the main control factor analysis results is evaluated using the prediction accuracy index combined with the accuracy requirement:

[0159]

[0160] Among them, err is the minimum error requirement, e true is the true value, e pred is the predicted value, n is the total number of sample wells, acc err To meet the accuracy requirements of the prediction, acc err If it is 1, it means that the prediction accuracy of all samples meets the requirements.

[0161] Furthermore, the calculation method of feature importance score using grey correlation method is as follows:

[0162]

[0163] Among them, ξ j is the grey correlation coefficient between geological engineering parameter j and production parameter y, y i is the yield of sample well i, x ij is the fracturing data of geological engineering parameter j of sample well i, n is the total number of sample wells, and ρ is the resolution coefficient, which is generally 0.5. The larger the grey correlation coefficient is, the closer the relationship with the target parameter is, which is a positive indicator.

[0164] Furthermore, when using the CatBoost ensemble learning evaluation method for importance evaluation, the geological engineering parameters are used as input and the production parameters are used as output. The CatBoost ensemble learning model is trained and the calculation method for deriving the geological engineering parameter feature importance score is as follows:

[0165]

[0166]

[0167] in, is the importance evaluation of the geological engineering parameter j in the integrated learning model, M is the number of trees of the geological engineering parameter j in the generative model; T is the tree, T q is the qth tree, L is the number of leaf nodes in each tree, L-1 is the number of non-leaf nodes in the tree, v t is the feature associated with node t, I t 2 is the reduction in squared loss after node t is split. For example, before the split, the node has 100 data points and the MSE is 1; after the split, it is divided into two nodes, one with 30 data points and the MSE is 0.9, and the other with 70 data points and the MSE is 0.8. Considering the weighting, I t 2 It is equal to (100 1-300.9-70 0.8) / 100=0.17. t 2 The larger the value, the greater the impact of the node on the accuracy of the final predicted output. Therefore, the greater the importance of the node feature, the greater the impact on the prediction of the target parameter, which is a positive indicator.

[0168] Furthermore, based on the trained CatBoost ensemble learning model, the SHAP value is used to analyze the feature importance. The SHAP importance score calculation method is as follows:

[0169]

[0170] in, is the SHAP importance score of geotechnical parameter j, {X 改造段长 ,Z 主成分1 ,…,X 用液强度} is the geological engineering dataset after dimension reduction and augmentation, X j is the fracturing data of geological engineering parameter j; h is the number of geological engineering parameters after dimension reduction and augmentation, {X 改造段长 ,Z 主成分1 ,…,X 用液强度}\{X j} does not include {X j} is the possible set of all input geotechnical parameter characteristics, f x (S) is the prediction of the characteristic subset S of geo-engineering parameters. is the proportion of characteristic combinations of the geological engineering parameter characteristic subset S, and the sum of the proportions of characteristic combinations of all possible geological engineering parameter characteristic subsets S is equal to 1. x The () function is the prediction function of the CatBoost model. The larger the SHAP value, the greater the impact on the target parameter, which is a positive indicator.

[0171] Furthermore, KNN nearest neighbor prediction to fill missing data specifically includes: treating the sample as a group of vectors, measuring the distance between sample vectors by Euclidean geometric distance, and filling the sample vectors with missing values ​​using the average value of the known values ​​of the nearest multiple groups of sample vectors.

[0172] Furthermore, the padded dataset is represented as {X 改造段长 ,X 加砂强度 ,…,X 用液强度 ,Y}, where Y is a column vector of production parameter data. The geological engineering parameters come from n sample wells, with a total of m items of geological engineering parameters. The data set {X 改造段长 ,X 加砂强度 ,…,X 用液强度 The matrix formed by the matrix is ​​n×m, where the data of the geological engineering parameter j of the sample well i is represented by x ij .

[0173] Further, in one embodiment, the data set includes the following geological engineering parameters: inclination point, A sounding, A vertical depth, B sounding, B vertical depth, artificial well bottom / m, horizontal section length / m, operable well section length / m, designed measure section length / m, oil layer drilling rate / %, oil layer thickness / m, Class I oil layer / m, Class II oil layer / m, Class III oil layer / m, Φ_min / %, Φ_max / %, Φ_avg / %, K_min / mD, K_max / mD, K_avg / mD, So_min / %, So_max / %, So_avg / %, Slug_quartz sand (70 / 140 mesh), main_quartz sand (70 / 140 mesh), Slug_quartz sand (40 / 70 mesh), main_quartz sand (40 / 70 mesh), Slug_quartz sand (30 / 50 mesh), main_quartz sand (30 / 50 mesh), Slug_quartz sand (20 / 40 mesh), main_quartz sand (20 / 40 mesh), Slug_ceramsite (70 / 140 mesh), Slug_ceramsite (40 / 70 mesh), Slug_ceramsite (20 / 40 mesh), ceramsite (20 / 40 mesh), ceramsite (30 / 50 mesh), total construction liquid volume / m 3 , total construction sand volume / m 3 , actual length of the reconstruction section, actual number of sections, actual number of clusters, number of clusters above 6 in a section, number of clusters above 6 in a section, acid level, construction liquid-sand ratio, construction liquid strength / m 3 / m, construction sand adding strength m 3 / m, construction slippery water volume / m 3 , construction slippery water ratio / %, construction pre-liquid ratio / %, construction average sand ratio / %, construction maximum sand ratio / kg / m 3 , average pump stop pressure / MPa, maximum construction pressure / MPa, average bursting pressure / MPa, maximum bursting pressure / MPa, construction span (d), pure operation time (d), maximum number of construction sections per day, average number of construction sections per day, sand addition completion rate / %, minimum actual construction displacement / m3 / min, maximum actual construction displacement / m 3 / min, outer diameter of oil layer casing / mm, wall thickness of oil layer casing / mm, etc. The production index is selected as the early cumulative production index, such as the cumulative production in 30 days.

[0174] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A comprehensive analysis method for the main controlling factors of fracturing data for production, characterized in that: The steps include: Step S1, preprocessing the fracturing data of the field production: performing data missing and data anomaly processing on the fracturing data of the field production to obtain processed fracturing data; Step S2, feature selection and transformation of geotechnical parameters that affect fracturing data: calculating the collinearity of geotechnical parameters and the correlation between geotechnical parameters, screening geotechnical parameters, and extracting features of geotechnical parameters by data dimensionality reduction and data augmentation of processed fracturing data; Step S3, using multiple evaluation methods to evaluate the importance of the fracturing parameters of the target production: using multiple importance evaluation methods to evaluate the importance of the characteristics of the target production parameters, and integrating the importance evaluation results of multiple importance evaluation methods to establish an importance evaluation matrix of geological engineering parameters, wherein the importance evaluation methods include grey correlation method, CatBoost ensemble learning method and SHAP evaluation; Step S4, conduct a comprehensive evaluation of the main control factors of the evaluation matrix through fuzzy mathematics: determine the weight of the characteristic vector of each importance evaluation method in the importance evaluation matrix based on the entropy weight method, obtain the final characteristic importance score by weighted calculation, and sort to obtain the main control factors of production.

2. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 1 is characterized in that: In step S1, the missing data is filled using the KNN nearest neighbor prediction filling method.

3. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 1, characterized in that: In step S2, feature selection and transformation of geological engineering parameters affecting fracturing data are performed, including the following steps: The VIF analysis method was used to conduct collinearity analysis on the geo-engineering parameters, and the principal component analysis method was used to conduct correlation analysis between the geo-engineering parameters; The VIF analysis method is used to reduce the dimension of geoengineering parameters and remove collinearity while retaining the interpretability of the data.

4. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 3 is characterized in that: Variance inflation factor (VIF) is used to determine the collinearity between geological engineering parameters. The VIF calculation method is as follows: For parameter combinations with multicollinearity and variance inflation factor VIF greater than 10, principal component replacement is performed, and dimension reduction is performed using principal component analysis, retaining one principal component. The principal component calculation method is as follows: And meet in, is the determination coefficient obtained by linear regression of the other k-1 geo-engineering parameters when the j-th geo-engineering parameter is used as the explained variable, and k is the total number of geo-engineering parameters involved in dimensionality reduction; Among them, Z1 is the first principal component, X j is the jth geoengineering parameter dataset, φ 1j is the first principal component weight of the jth geoengineering parameter; n is the total number of sample wells; m is the total number of geo-engineering parameters; j is the serial number of the geological engineering parameter; i is the sample well number; x ij is the jth geological engineering parameter of the ith well.

5. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 4 is characterized in that: In step S3, the geological engineering parameter data set after dimensionality reduction and augmentation is used to evaluate the importance of the characteristics of the target production parameters using a variety of importance evaluation methods. The importance evaluation results of a variety of importance evaluation methods are combined to establish an importance evaluation matrix of the geological engineering parameters. The target production parameter is used as the dependent variable and other geological engineering parameters are used as independent variables. The importance score of each geological engineering parameter to the characteristic of the target production parameter is calculated, and finally the importance evaluation matrix E is constructed.

6. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 1, characterized in that: In step S4, the analysis results of the importance evaluation matrix E are comprehensively evaluated through fuzzy mathematics, the final parameter importance scores of the target parameters are calculated and ranked, and the main control factors of production are analyzed and determined.

7. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 6 is characterized in that: In step S4, the importance evaluation matrix E is standardized to obtain a fuzzy evaluation matrix. The entropy weight method in fuzzy mathematics is used to calculate the weights of each importance evaluation method and the weights of each geological engineering parameter to obtain a comprehensive fuzzy evaluation vector, and the final importance score of each geological engineering parameter is obtained. The parameters are sorted from high to low according to the score to determine the main controlling factors of production.

8. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 7, characterized in that: When normalizing the importance evaluation matrix E, for any geological engineering parameter, according to the principle of whether it is closer to the actual situation of the parameter, one of the positive indicator normalization or negative indicator normalization is selected to normalize the importance evaluation matrix E.

9. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 8, characterized in that: The normalization of positive indicators is as follows: in, E b is the data vector of evaluation index b in the evaluation matrix, e ab The value of evaluation parameter a in evaluation index b in the evaluation matrix, n ab It is the normalized score value of evaluation parameter a in evaluation index b.

10. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 8, characterized in that: The normalization of negative indicators is as follows: in, E b is the data vector of evaluation index b in the evaluation matrix, e ab The value of evaluation parameter a in evaluation index b in the evaluation matrix, n ab It is the normalized score value of evaluation parameter a in evaluation index b.

11. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 7, characterized in that: The entropy weight method is used to calculate the weights of each importance evaluation method and the weights of each geological engineering parameter. The steps are as follows: Calculate the value e of evaluation parameter a in evaluation index b ab The proportion of evaluation index b ab : ∈P ab ＝1 Among them, h is the number of evaluation parameters; Calculate the information entropy value r of the evaluation index b b : Calculate the weight ω of each evaluation index b : Among them, u is the number of evaluation indicators; Finally, the evaluation matrix E is multiplied by the index weight vector to obtain the fuzzy comprehensive evaluation matrix S: in, in, [ξ], [J] and are the feature importance score vectors obtained by the grey correlation method, CatBoost ensemble learning method and SHAP evaluation method, respectively. Each element in is an element of the comprehensive evaluation matrix E, which represents the importance score of the geological engineering parameters obtained by using different importance evaluation methods; ω is the evaluation index vector, The weight calculation is performed on the principal component vector, and the abstract dimensionality reduction and augmented geological engineering parameters in the comprehensive fuzzy evaluation vector S are restored to the original geological engineering parameters to obtain the importance score of the original geological engineering parameters. The relationship between the collinear geological engineering parameter combination and its dimensionality reduction principal component is shown as follows: in, S Z is the comprehensive evaluation score of the principal component evaluation parameters; j is the serial number of the geological engineering parameter; k is the total number of geo-engineering parameters involved in dimensionality reduction; S Zj is the comprehensive evaluation score of the jth geo-engineering parameter after principal component reduction; φ Zj is the weight of the jth geoengineering parameter of the principal component.

12. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 1, characterized in that: The method also includes step S5, establishing a regression prediction model of geological engineering parameters and target production parameters, introducing the final feature importance score of each importance rating method as the basis for initializing the regression prediction model, and verifying the rationality of the main control factor analysis results by comparing the training results of the regression prediction model and the accuracy of the regression prediction model.

13. The comprehensive analysis method of main controlling factors of fracturing data for production according to claim 12, characterized in that: In step S5, the prediction accuracy index combined with the accuracy requirement is used to evaluate the rationality of the main control factor analysis results: Among them, err is the minimum error requirement, e true is the true value, e pred is the predicted value, n is the total number of sample wells, acc err To meet the accuracy requirements of the prediction, acc err If it is 1, it means that the prediction accuracy of all samples meets the requirements.

Images