Dynamic evaluation method, system, medium and equipment for fracturing effect of deep coal bed gas

Abstract

The invention relates to the field of oil and gas field development, and discloses a dynamic evaluation method, system, medium and equipment for the fracturing effect of deep coal bed gas, and the method comprises the steps: collecting basic data of a fractured well of a target deep coal bed gas development block, and carrying out the data preprocessing of the basic data, so as to determine the dynamic evaluation characteristic parameters of the fracturing effect of the deep coal bed gas; based on the preprocessed data, the production dynamic state of the fractured well is analyzed, and production index stage quantification is conducted on the average daily gas production rate and the cumulative gas production rate of a single well in multiple different periods; based on the preprocessed data, the feature parameters and the quantized production indexes of the multiple different stages, establishing a non-linear multi-model fusion interpretable machine learning model between the feature parameters and the production indexes of the different stages, and sequentially analyzing the influence of the feature parameters on the production indexes of the different stages; determining a main control factor from each characteristic parameter; and quantitatively evaluating the influence degree and the change rule of the main control factors on the production indexes in different stages.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas field development, and particularly to a method, system, medium and device for dynamic evaluation of deep coalbed methane fracturing effect based on preprocessing of geological-engineering-production dynamic big data and construction of machine learning models, as well as construction of a machine learning model. Background Art

[0002] Evaluation of fracturing productivity effect and analysis of main productivity control factors are crucial for efficient fracturing development of deep coalbed methane, and are important bases for guiding the formulation of fracturing process strategies and the optimization of fracturing process parameters. Currently, with the rapid development of computer science and the explosive accumulation of oilfield data, machine learning has been widely applied in the petroleum industry due to its powerful non-linear approximation ability. The design of fracturing process parameters for on-site deep coalbed methane requires a large amount of geological modeling, fracture simulation and production simulation, which is time-consuming and laborious. Big data and artificial intelligence algorithms help to mine a large amount of useful information from massive data, thus greatly improving the level and efficiency of fracturing design.

[0003] Therefore, how to use data analysis and mining to better solve the problems such as difficult evaluation of deep coalbed methane fracturing effect, fuzzy main productivity control factors of fracturing, and time-consuming optimization calculation of fracturing process parameter design has become a technical problem to be solved urgently at present. Summary of the Invention

[0004] In view of the above problems, the present invention aims to provide a method, system, medium and device for dynamic evaluation of deep coalbed methane fracturing effect, which integrates the characteristics of deep coalbed methane fracturing development, and based on the analysis of geological-engineering-production dynamic big data, solves the problem that the current evaluation method of deep coalbed methane development effect is not applicable, and further fills the technical gap in the evaluation of deep coalbed methane fracturing development effect.

[0005] To achieve the above object, in a first aspect, the technical solution adopted by the present invention is: a method for dynamic evaluation of deep coalbed methane fracturing effect, which includes: collecting basic data of fractured wells in a target deep coalbed methane development block, and performing data preprocessing on the basic data to determine dynamic evaluation characteristic parameters of deep coalbed methane fracturing effect; based on the preprocessed data, analyzing the production dynamics of the fractured wells, and performing stage quantification of production indexes such as the peak average daily gas production of a single well and the cumulative gas production of a single well in multiple different periods; based on the preprocessed data, various characteristic parameters and the quantified production indexes in multiple different stages, establishing an interpretable machine learning model with non-linear multi-model fusion between various characteristic parameters and production indexes in different stages, sequentially analyzing the influence of various characteristic parameters on production indexes in different stages, and determining the main control factors from various characteristic parameters; quantitatively evaluating the influence degree and change law of the main control factors on production indexes in different stages.

[0006] Further, collect the basic data of the fractured wells in the target deep coalbed methane development block, and perform data preprocessing on the basic data to determine the dynamic evaluation characteristic parameters of the fracturing effect of deep coalbed methane, including:

[0007] Collect the basic data of the fractured wells in the target deep coalbed methane development block, and perform data preprocessing on the basic data; the basic data includes the geological reservoir basic data, drilling, completion and fracturing construction data, and post-fracturing production dynamic data of the fractured wells in the target deep coalbed methane development block;

[0008] Determine the dynamic evaluation characteristic parameters of the fracturing effect of deep coalbed methane according to the primary characteristic parameters and secondary characteristic parameters from the preprocessed basic data; the primary characteristic parameters include geological reservoir parameters, engineering construction parameters, and production dynamic parameters; the secondary characteristic parameters are detailed parameters determined according to each primary characteristic parameter.

[0009] Further, perform data preprocessing on the basic data, including: data cleaning and feature screening.

[0010] Further, the process of feature screening includes:

[0011] Calculate the Spearman correlation coefficient between the characteristic factors of dynamic development indicators at multiple different stages and the target production. If the absolute value of the correlation coefficient is greater than the first set value, the first feature subset is obtained;

[0012] Calculate the mutual information between the characteristic factors of dynamic development indicators at multiple different stages and the target production. If the mutual information value is greater than the second set value, the second feature subset is obtained;

[0013] Perform a union operation on the first feature subset and the second feature subset to obtain a feature union;

[0014] Calculate the Spearman correlation coefficient between two features in the feature union. If the absolute value of the correlation coefficient is greater than the third set value, delete the features with a correlation higher than the third set value, and output the features to form the third feature subset; if the absolute value of the correlation coefficient is less than the third set value, output the features to form the third feature subset.

[0015] Further, perform stage quantification on the peak average daily gas production and cumulative gas production of a single well in multiple different cycles, including:

[0016]

[0017] In the formula, q Bit is the peak average daily gas production of a single well in the i-th cycle; Q total is the cumulative gas production of a single well; t B is the peak gas appearance time of a single well; t total is the cumulative gas production time of a single well; qt is the gas production of a single well at time t.

[0018] Furthermore, based on the preprocessed data, various characteristic parameters, and the quantified production indicators at multiple different stages, an interpretable machine learning model with non-linear multi-model fusion between various characteristic parameters and production indicators at different stages is established. Analyze the influence of each characteristic parameter on the production indicators at different stages in turn, and determine the main control factors from each characteristic parameter, including:

[0019] Calculate the feature importance of the third feature subset obtained after feature screening using the recursive elimination method, SHAP method, and Boruta method respectively to obtain the feature importance scores;

[0020] Normalize and weighted average the three feature importance scores to determine the feature importance ranking, and select the third feature subset with the greatest importance as the feature subset with the best performance;

[0021] Use the random forest and recursive feature elimination methods to comprehensively evaluate the influence characteristics of each characteristic parameter in the feature subset with the best performance on the production indicators of the peak average daily gas production and cumulative gas production of a single well in different dynamic cycles, and determine the main control factors affecting the post-fracturing development effect of deep coalbed methane.

[0022] Furthermore, quantitatively evaluate the influence degree and change law of the main control factors on the production indicators at different stages, including:

[0023] Normalize and weighted average the importance scores of the production capacity main control factors under the production indicators at different stages respectively, and construct a comparison chart to quantitatively evaluate the influence degree and change law of the production capacity main control factors under the production indicators at different stages.

[0024] In the second aspect, the technical solution adopted by the present invention is: a dynamic evaluation system for the fracturing effect of deep coalbed methane, which includes: an acquisition module that collects the basic data of the fractured wells in the target deep coalbed methane development block, and preprocesses the basic data to determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect; a processing module that analyzes the production dynamics of the fractured wells and quantifies the production indicators of the peak average daily gas production and cumulative gas production of a single well in multiple different cycles; a training module that, based on the preprocessed data, various characteristic parameters, and the quantified production indicators at multiple different stages, establishes an interpretable machine learning model with non-linear multi-model fusion between various characteristic parameters and production indicators at different stages, analyzes the influence of each characteristic parameter on the production indicators at different stages in turn, and determines the main control factors from each characteristic parameter to quantitatively evaluate the influence degree and change law of the main control factors on the production indicators at different stages.

[0025] In a third aspect, the technical solution adopted by the present invention is: a computer-readable storage medium storing one or more programs, where the one or more programs include instructions that, when executed by a computing device, cause the computing device to execute any one of the above methods.

[0026] In a fourth aspect, the technical solution adopted by the present invention is: a computing device, which includes: one or more processors, a memory, and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the above methods.

[0027] Due to the above technical solution adopted by the present invention, it has the following advantages:

[0028] The present invention can effectively avoid large-scale fracture-seepage numerical simulation of hydraulic fracturing, tap the value from a large amount of real data, and at the same time, the present invention breaks the previous analysis of only the influencing factors of single-stage production capacity, realizes the dynamic evaluation of the main controlling factors of hydraulic fracturing production capacity, helps to clarify the influencing factors and change trend laws of deep coal seam gas fracturing production capacity in the short term, medium term, and long term, so as to provide technical support for the customization of hydraulic fracturing process strategies, the iterative optimization design of hydraulic fracturing process parameters, and the reasonable decision-making of on-site construction adjustments. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following briefly introduces the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are Some embodiments of the present invention For those of ordinary skill in the art, on the premise of not paying creative labor Under such circumstances, other drawings can also be obtained according to these drawings.

[0030] Figure 1 It is a schematic flow chart of the dynamic evaluation method for deep coal seam gas fracturing effect based on geological engineering big data analysis in an embodiment of the present invention;

[0031] Figure 2 It is an example diagram of secondary characteristic parameters in an embodiment of the present invention;

[0032] Figure 3 It is a work flow chart of data preprocessing in an embodiment of the present invention;

[0033] Figure 4a It is a flow chart of feature screening in an embodiment of the present invention;

[0034] Figure 4b It is a flow chart of determining the main controlling factors from each characteristic parameter in an embodiment of the present invention;

[0035] Figure 5Schematic diagram of a quantitative evaluation method of main control factors based on SHAP analysis in an embodiment of the present invention;

[0036] Figure 6a This is a graph showing the calculation results of the influence of each characteristic parameter on the dynamic development indicators at different stages based on the random forest method in an embodiment of the present invention;

[0037] Figure 6b This is a calculation result diagram of the influence of each characteristic parameter on the dynamic development index at different stages based on the recursive elimination method in an embodiment of the present invention;

[0038] Figure 7 The importance scores and change trend diagrams of the comprehensive impact of geological and engineering factors on dynamic development indicators at different stages based on the machine learning model in the embodiment of the present invention;

[0039] Figure 8 It is a structural schematic diagram of an embodiment of a deep coalbed methane fracturing effect dynamic evaluation system based on geological engineering big data analysis in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In view of the current problems of difficult evaluation of deep coalbed methane fracturing effect, fuzzy main control factors of fracturing capacity, and time-consuming calculation of fracturing process parameter design optimization, it is necessary to carry out big data analysis and mining research based on geological gas reservoirs, completion fracturing, and production data, clarify the key geological and engineering factors affecting the fracturing effect, and establish a prediction model for development indicators under different geological gas reservoirs and fracturing process parameters. The application of big data analysis and mining methods can effectively avoid the complex mechanism research of deep coalbed methane fracturing, tap the potential value from a large amount of real data, and thus provide technical support for the customization of fracturing process strategies, iterative optimization design of fracturing process parameters, and reasonable decision-making for on-site construction adjustments. The present invention proposes a new prediction framework based on geological-engineering-production dynamic big data analysis and machine learning. Starting from the post-fracturing capacity dynamics of deep coalbed methane wells at different stages of production, based on multi-source heterogeneous data, the framework systematically and quantitatively evaluates the fracturing development effect and clarifies the dynamic main control factors affecting fracturing capacity, providing guidance for the customization of fracturing strategies and iterative optimization of fracturing process parameters.

[0041] Compared with existing technologies, including traditional intersection chart method, multivariate linear regression, predetermined interval method, experiment and numerical simulation methods, machine learning can effectively avoid the complex mechanism research of unconventional reservoir fracturing with its powerful nonlinear mapping ability and tap the potential value from a large amount of real data. There are obvious differences between the quantitative evaluation modeling of development effects based on oil and gas big data and traditional machine learning modeling, and it is necessary to establish a systematic and targeted big data analysis process and method.

[0042] The technical concept of the present invention lies in: combining the characteristics of oil and gas big data and the modeling requirements, establishing a non-linear multi-model fusion interpretable machine between each characteristic parameter and each production index at different stages from oil and gas dataA system framework from data collection, parameter extraction, data preprocessing, feature selection to model optimization and prediction. Specifically, the present invention provides a method, system, medium and device for dynamically evaluating the fracturing effect of deep coalbed methane based on big data analysis of geological engineering, including: collecting the basic data of the fractured wells in the target deep coalbed methane development block, and performing data preprocessing on the basic data to determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect. Data preprocessing is carried out for data missing, noise and outliers to improve the data quality of the whole cycle of geology-engineering-production in the target block. Based on the preprocessed data, analyze the production dynamics of the fractured wells, and quantitatively quantify the production indicators of the peak average daily gas production per well and the cumulative gas production per well in multiple different periods; based on the preprocessed data, each characteristic parameter and the quantified production indicators in multiple different stages, establish an interpretable machine learning model with non-linear multi-model fusion between each characteristic parameter and each production indicator in different stages, analyze the influence of each characteristic parameter on the production indicators in different stages in turn, and determine the main control factors from each characteristic parameter; quantitatively evaluate the influence degree and change law of the main control factors on the production indicators in different stages. The present invention can effectively avoid large-scale fracturing-seepage numerical simulation, tap the value from a large amount of real data, so as to optimize the fracturing plan design and make reasonable decisions on on-site construction adjustment, solve the problem that the current evaluation methods for the development effect of deep coalbed methane reservoirs are not applicable in the field of deep coalbed methane reservoirs, and fill the technical gap in the evaluation of the development effect of deep coalbed methane reservoirs.

[0043] 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 of the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the described embodiments of the present invention fall within the scope of protection of the present invention.

[0044] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they specify the presence of features, steps, operations, devices, components and / or combinations thereof.

[0045] In an embodiment of the present invention, a method for dynamically evaluating the fracturing effect of deep coalbed methane is provided. In this embodiment, as Figure 1 shown, the method includes the following steps:

[0046] Step S101: Collect the basic data of the fractured wells in the target deep coalbed methane development block, and perform data preprocessing on the basic data for data missing, noise, and outliers to determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect, so as to improve the data quality of the whole process of geology-engineering-production in the target block.

[0047] Step S102: Based on the preprocessed data, analyze the production dynamics of the fractured wells, and conduct stage quantification of production indicators for the average daily gas production at the peak of a single well and the cumulative gas production of a single well in multiple different periods.

[0048] For example, the average daily gas production at the peak of a single well in multiple different periods includes: the average daily production of a single well at the peak for 30 days, the average daily production of a single well at the peak for 60 days, the average daily production of a single well at the peak for 90 days, and the average daily production of a single well at the peak for 180 days.

[0049] Step S103, based on the preprocessed data, each characteristic parameter and the quantified production indicators at multiple different stages Based on the determined characteristic parameters and production indicators, establish an interpretable machine learning model with non-linear multi-model fusion between each characteristic parameter and each production indicator at different stages, analyze the influence of each characteristic parameter on the production indicators at different stages in turn, and determine the main control factors from each characteristic parameter.

[0050] Step S104: Quantitatively evaluate the influence degree and change law of the main control factors on the production indicators at different stages.

[0051] In the above step S101, in this embodiment, by modifying the Python source program, collect, analyze, and sort out the big data related materials of the fractured wells in the large-scale developed blocks in the region, such as parameters of geology, gas reservoir, fracturing construction, rock mechanics, logging, and oil and gas production dynamics.

[0052] Specifically, collecting the basic data of the fractured wells in the target deep coalbed methane development block and performing data preprocessing on the basic data to determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect includes the following steps:

[0053] Step S1011: Collect the basic data of the fractured wells in the target deep coalbed methane development block and perform data preprocessing on the basic data; the basic data includes the geological gas reservoir basic data, drilling, completion and fracturing construction data, and post-fracture production dynamic data of the fractured wells in the target deep coalbed methane development block.

[0054] Step S1012: According to the preprocessed basic data, determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect according to the first-level characteristic parameters and the second-level characteristic parameters; the first-level characteristic parameters include geological gas reservoir parameters, engineering construction parameters, and production dynamic parameters; the second-level characteristic parameters are detailed parameters determined according to each first-level characteristic parameter respectively.

[0055] For example, as Figure 2As shown in the figure, the secondary characteristic parameters are detailed parameters determined according to the geological gas reservoir parameters in the primary characteristic parameters, including coal seam vertical depth, pure coal vertical thickness, interlayer vertical thickness, gas content, coal rock density, reservoir pressure, GSI value, minimum horizontal in-situ stress, elastic modulus, Poisson's ratio, and stress difference between reservoir and barrier layer; the secondary characteristic parameters are detailed parameters determined according to the engineering construction parameters in the primary characteristic parameters, including perforation thickness, perforation degree, actual net fracturing fluid volume, actual sand addition volume in fracturing, net fracturing fluid intensity, sand addition intensity in fracturing, whether to use temporary plugging, quartz sand volume (70 - 140 mesh), quartz sand volume (40 - 70 mesh), quartz sand volume (30 - 50 mesh), quartz sand volume (20 - 40 mesh), fracturing construction displacement, preflush volume, and preflush ratio; the secondary characteristic parameters are detailed parameters determined according to the production dynamic parameters in the primary characteristic parameters, including average daily gas production in the peak 30 days, average daily gas production in the peak 60 days, average daily gas production in the peak 90 days, and average daily gas production in the peak 180 days.

[0056] In the above step S101, data preprocessing is performed on the basic data, including data cleaning and feature screening. As Figure 3 shown, specifically: the input basic data is respectively processed by box plot, isolation forest, and dimensionality reduction clustering, and outliers are identified and judged from the perspectives of statistical features and feature correlation, and the outliers are processed in combination with professional experience, and the processed data is output. By comprehensively using the three methods of box plot, isolation forest, and clustering, outliers in the dataset can be effectively identified, and the model performance can be improved.

[0057] In this embodiment, dealing with outliers in combination with professional experience includes:

[0058] (1) Correct or delete data errors according to the data source file.

[0059] (2) Retain abnormal engineering parameters for experimental wells.

[0060] (3) For extremely good and extremely poor wells, retain abnormal geological parameters.

[0061] In this embodiment, box plot and isolation forest are used to process data outliers. Box plot: It is a graphical description formed by the quartiles of a dataset, which can describe the data dispersion. It is a very simple and effective statistical method for visualizing outliers. This method does not make assumptions about the data distribution and is widely used. Denote the distance between the lower quartile (Q1) and the upper quartile (Q3) as IQR. The upper and lower whiskers are the boundaries of the data distribution. Any data point above the upper whisker (Q3 + 1.5IQR) or below the lower whisker (Q1 - 1.5IQR) can be considered an outlier.

[0062] Isolation Forest: An outlier is defined as "a point that is easily isolated", which can be understood as a point with sparse distribution and far from the group with high density. The isolation forest algorithm recursively and randomly partitions the dataset until all sample points are isolated. Under this random partitioning strategy, outliers usually have shorter paths. Intuitively, those clusters with high density need to be cut many times to be isolated, but those points with low density can be easily isolated.

[0063] In this embodiment, the initial characteristic variable parameters collected from geology, development, fracturing construction, rock mechanics, and logging data contain "irrelevant features" and "redundant features". Before determining the main control factors, correlation filtering and mutual information method are used for correlation analysis to reduce data complexity and noise, eliminate the linear relationship between data, and increase the readability of the model. Therefore, the present invention uses correlation filtering and mutual information method for correlation analysis, and the method of removing multicollinearity for feature screening. Specifically, as Figure 4a shown, the process of feature screening includes the following steps:

[0064] Step S10111: Calculate the Spearman correlation coefficient between the influencing factor features of dynamic development indicators at multiple different stages and the target production. If the absolute value of the correlation coefficient is greater than the first set value, the first feature subset S1 is obtained;

[0065] In this embodiment, the first set value is 0.2.

[0066] Step S10112: Calculate the mutual information between the influencing factor features of dynamic development indicators at multiple different stages and the target production. If the mutual information value is greater than the second set value, the second feature subset S2 is obtained; In this embodiment, the second set value is 0.01.

[0067] Step S10113: Perform a union operation on the first feature subset and the second feature subset to obtain the feature union S3.

[0068] Step S10114: Calculate the Spearman correlation coefficient between pairwise features in the feature union. If the absolute value of the correlation coefficient is greater than the third set value, delete the features with a correlation higher than the third set value, and output the features to form the third feature subset S4; If the absolute value of the correlation coefficient is less than the third set value, output the features to form the third feature subset S4; In this embodiment, the third set value is 0.8.

[0069] In this embodiment, the correlation analysis method is used to calculate the correlation between the influencing factor features of dynamic development indicators at multiple different stages and the target production; among them, the Spearman correlation coefficient and the mutual information method are used to comprehensively calculate the correlation. Feature filtering is performed based on the correlation, and irrelevant or weakly correlated features are removed.

[0070] In this embodiment, the Spearman correlation coefficient is a rank correlation coefficient, that is, it is calculated based on the correlation of the sorting positions of the data in two features (the formula is shown in Equation (1)). As long as the corresponding values of the two features are sorted within each group The order is Similarly, there is a significant correlation, so the Spearman correlation coefficient can not only measure linear correlation It can also evaluate non-linear correlation.

[0071]

[0072] The mutual information method is used to measure the correlation between features and production. Mutual information is a measure in information theory used to evaluate the degree of dependence between two random variables, and the calculation formula is shown in Equation (2). The larger the mutual information value between a certain feature and production, the more the uncertainty of the target variable is reduced after knowing a certain feature, that is, the stronger the correlation between the two.

[0073]

[0074] In the above step S102, the production index stage quantization of the peak average daily gas production and cumulative gas production of a single well in multiple different periods is specifically as follows:

[0075]

[0076] In the formula, q Bit is the peak average daily gas production of a single well in the i-th period. For example, i = 30 days, 60 days, 90 days, 180 days; Q total is the cumulative gas production of a single well; t B is the peak gas appearance time of a single well; t total is the cumulative gas production time of a single well; q t is the gas production of a single well at time t.

[0077] Specifically, the peak 30-day average daily gas production q B30 of a single well, the peak 60-day average daily gas production q B60 of a single well, the peak 90-day average daily gas production q B90 of a single well, and the peak 180-day average daily gas production q B180 of a single well are respectively:

[0078]

[0079] In the above step S103, as Figure 5 shown, based on the preprocessed data, each feature parameter, and the production indexes of multiple different quantized stages, an interpretable machine learning model of non-linear multi-model fusion between each feature parameter and each different stage production index is established, and the influence of each feature parameter on the production indexes of different stages is analyzed in turn, and the main control factors are determined from each feature parameter. As Figure 4b shown, it includes the following steps:

[0080] Step S1031: Calculate the feature importance of the third feature subset obtained after feature screening by using the recursive elimination method, SHAP method, and Boruta method respectively, to obtain the feature importance scores.

[0081] Step S1032: Normalize and perform weighted averaging on the three feature importance scores, determine the feature importance ranking, and select the third feature subset with the greatest importance as the feature subset with the best performance.

[0082] Step S1033: Use the random forest and recursive feature elimination methods to comprehensively evaluate the influence characteristics of each feature parameter in the feature subset with the best performance on the production indexes of the average daily gas production at the peak of a single well and the cumulative gas production of a single well in different periods, and determine the main controlling factors affecting the post-development effect of deep coalbed methane, such as Figure 6a 、 Figure 6b as shown.

[0083] Specifically, starting from the development effects such as the control and utilization of geological reserves, cumulative gas production, and peak stable gas production, use the random forest and recursive feature elimination methods to comprehensively evaluate the influence characteristics of each feature parameter on the development indexes of the average daily gas production at the peak of a single well and the cumulative gas production of a single well in different periods, and determine the main controlling factors affecting the post-development effect of deep coalbed methane.

[0084] In this embodiment, the average daily gas production in 30 days at the peak of a single well, the average daily gas production in 60 days at the peak of a single well, the average daily gas production in 90 days at the peak of a single well, and the average daily gas production in 180 days at the peak of a single well are used as the development indexes. Through a voting method, comprehensively evaluate the influence characteristics of each parameter on the development indexes, and determine the main controlling factors affecting the development effect of deep coalbed methane.

[0085] The machine learning model in this embodiment is based on the Keras library, coded in Python 3.7, and runs on a computer equipped with a 3.40GHz CPU.

[0086] In the above step S104, quantitatively evaluate the influence degree and change law of the main controlling factors on the production indexes at different stages, specifically: respectively perform normalized weighted averaging on the importance scores of the production capacity main controlling factors under the production indexes at different stages, and construct a comparison chart to quantitatively evaluate the influence degree and change law of the production capacity main controlling factors under the production indexes at different stages, such as Figure 7 as shown.

[0087] The present invention uses machine learning and data mining algorithms such as random forest, gradient boosting, and linear regression to establish an interpretable machine learning model of the main controlling factors and multi-development indexes, comprehensively evaluate the model accuracy, and combine with the SHAP interpretation method to quantitatively evaluate the influence degree and change law of the main controlling factors on the development indexes.

[0088] As Figure 8 shown, in an embodiment of the present invention, a dynamic evaluation system 400 for the fracturing effect of deep coalbed methane is provided, which includes:

[0089] An acquisition module 401, which collects the basic data of the fractured wells in the target deep coalbed methane development block, preprocesses the basic data to determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect;

[0090] A processing module 402, which analyzes the production dynamics of the fractured wells, and quantifies the production indexes of the average daily gas production at the peak of a single well and the cumulative gas production of a single well in multiple different periods;

[0091] A training module 403, based on the preprocessed data, each characteristic parameter, and the production Production Index, establish an interpretable machine for non-linear multi-model fusion between each characteristic parameter and each production index at different stages of multiple different stages, analyzes the influence of each characteristic parameter on the production indexes of different stages in turn, and determines the main control factors from each characteristic parameter to quantitatively evaluate the influence degree and variation law of the main control factors on the production indexes of different stages.

[0092] In the above embodiment, collecting the basic data of the fractured wells in the target deep coalbed methane development block, preprocessing the basic data to determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect includes:

[0093] Collecting the basic data of the fractured wells in the target deep coalbed methane development block and preprocessing the basic data; the basic data includes the geological reservoir basic data, drilling, completion and fracturing construction data, and post-fracturing production dynamic data of the fractured wells in the target deep coalbed methane development block;

[0094] According to the preprocessed basic data, determining the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect according to the first-level characteristic parameters and the second-level characteristic parameters; the first-level characteristic parameters include geological reservoir parameters, engineering construction parameters, and production dynamic parameters; the second-level characteristic parameters are detailed parameters determined according to each first-level characteristic parameter.

[0095] In the above embodiment, preprocessing the basic data includes: data cleaning and feature screening.

[0096] In the above embodiment, the process of feature screening includes:

[0097] Calculating the Spearman correlation coefficient between the characteristic factors of the dynamic development indexes in multiple different stages and the target production. If the absolute value of the correlation coefficient is greater than the first set value, a first feature subset is obtained;

[0098] Calculating the mutual information between the characteristic factors of the dynamic development indexes in multiple different stages and the target production. If the mutual information value is greater than the second set value, a second feature subset is obtained;

[0099] Perform a union operation on the first feature subset and the second feature subset to obtain a feature union;

[0100] Calculate the Spearman correlation coefficient between every two features in the feature union. If the absolute value of the correlation coefficient is greater than the third set value, delete the features with a correlation higher than the third set value, and output the features to form a third feature subset; if the absolute value of the correlation coefficient is less than the third set value, output the features to form a third feature subset.

[0101] In this embodiment, the production index stage quantization of the single-well peak average daily gas production and single-well cumulative gas production in multiple different periods includes:

[0102]

[0103] where q Bit is the single-well peak average daily gas production in the i-th period; Q total is the single-well cumulative gas production; t B is the single-well peak gas appearance time; t total is the single-well cumulative gas production time; q t is the single-well gas production at time t.

[0104] In the above embodiment, based on the preprocessed data, each feature parameter, and the quantified production indexes in multiple different stages, establish an interpretable machine learning model of nonlinear multi-model fusion between each feature parameter and each different-stage production index, and analyze the influence of each feature parameter on the production indexes in different stages in turn, and determine the main control factors from each feature parameter, including:

[0105] Calculate the feature importance of the features in the third feature subset obtained after feature screening by using the recursive elimination method, SHAP method, and Boruta method respectively to obtain the feature importance score;

[0106] Normalize and weighted average the three feature importance scores to determine the feature importance ranking, and select the third feature subset with the greatest importance as the feature subset with the best performance;

[0107] Use the random forest and recursive feature elimination methods to comprehensively evaluate the influence characteristics of each feature parameter in the feature subset with the best performance on the production indexes of the dynamic single-well peak average daily gas production and single-well cumulative gas production in different periods, and determine the main control factors affecting the post-development effect of deep coal seam gas.

[0108] In the above embodiment, quantitatively evaluate the influence degree and change law of the main control factors on the production indexes in each different stage, including:

[0109] Normalize and perform weighted average on the importance scores of the main factors affecting production capacity under different production indicators at each stage, and construct a comparison chart to quantitatively evaluate the influence degree and variation law of the main factors affecting production capacity under different production indicators at each stage.

[0110] The system provided in this embodiment is used to execute the above method embodiments. For the specific process and detailed content, please refer to the above embodiments and will not be elaborated here.

[0111] In an embodiment of the present invention, a computing device is provided. The computing device may be a terminal, and it may include: a processor, a communications interface, a memory, a display screen, and an input device. Among them, the processor, the communications interface, and the memory communicate with each other through a communication bus. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. When the computer program is executed by the processor, it is used to implement the methods in the above embodiments; the internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communications interface is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be implemented through WIFI, a carrier network, NFC (Near Field Communication), or other technologies. The display screen may be a liquid crystal display screen or an electronic ink display screen. The input device may be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computing device, or an external keyboard, a touchpad, or a mouse, etc. The processor can call the logical instructions in the memory.

[0112] In addition, when the logical instructions in the above memory are implemented in the form of a software functional unit and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. And the foregoing storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), magnetic disk or optical disk, etc., various media that can Store program code.

[0113] In an embodiment of the present invention, a computer program product is provided. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions that, when executed by a computer, enable the computer to execute the methods provided in the above method embodiments.

[0114] In an embodiment of the present invention, a non-transitory computer-readable storage medium is provided. The non-transitory computer-readable storage medium stores server instructions that cause a computer to execute the methods provided in the above embodiments.

[0115] For the computer-readable storage medium provided in the above embodiments, its implementation principle and technical effects are similar to those of the above method embodiments, and will not be elaborated here.

[0116] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0117] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0118] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 block or multiple blocks.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dynamic evaluation method for the fracturing effect of deep coal seam gas, characterized in that, Including: Collect the basic data of the fractured wells in the target deep coalbed methane development block, and perform data preprocessing on the basic data to determine the dynamic evaluation characteristic parameters of the fracturing effect of deep coalbed methane; Based on the preprocessed data, analyze the production dynamics of the fractured wells, and perform stage quantification of the production indicators such as the peak average daily gas production per well and the cumulative gas production per well in multiple different cycles; Based on the preprocessed data, various characteristic parameters, and the quantified production indicators in multiple different stages, establish an interpretable machine learning model with non-linear multi-model fusion between various characteristic parameters and production indicators in different stages, and analyze the influence of each characteristic parameter on the production indicators in different stages in turn, and determine the main control factors from each characteristic parameter; Quantitatively evaluate the influence degree and variation law of the main control factors on the production indicators in different stages.

2. The dynamic evaluation method for the fracturing effect of deep coal seam gas as described in claim 1, wherein, Collect the basic data of the fractured wells in the target deep coalbed methane development block, and perform data preprocessing on the basic data to determine the dynamic evaluation characteristic parameters of the fracturing effect of deep coalbed methane, including: Collect the basic data of the fractured wells in the target deep coalbed methane development block, and perform preprocessing on the basic data; the basic data includes the geological reservoir basic data, drilling, completion and fracturing construction data, and post-fracturing production dynamic data of the fractured wells in the target deep coalbed methane development block; Based on the preprocessed basic data, determine the dynamic evaluation characteristic parameters of the fracturing effect of deep coalbed methane according to the primary characteristic parameters and secondary characteristic parameters; the primary characteristic parameters include geological reservoir parameters, engineering construction parameters, and production dynamic parameters; the secondary characteristic parameters are detailed parameters determined according to each primary characteristic parameter.

3. The dynamic evaluation method for the fracturing effect of deep coal seam gas according to claim 2, characterized in that, Perform data preprocessing on the basic data, including: data cleaning and feature screening.

4. The dynamic evaluation method for the fracturing effect of deep coal seam gas according to claim 3, characterized in that, The process of feature screening includes: Calculate the Spearman correlation coefficient between the influencing factor characteristics of dynamic development indicators in multiple different stages and the target production. If the absolute value of the correlation coefficient is greater than the first set value, the first feature subset is obtained; Calculate the mutual information between the influencing factor characteristics of dynamic development indicators in multiple different stages and the target production. If the mutual information value is greater than the second set value, the second feature subset is obtained; Perform a union operation on the first feature subset and the second feature subset to obtain a feature union; Calculate the Spearman correlation coefficient between two features in the feature union. If the absolute value of the correlation coefficient is greater than the third set value, delete the features with a correlation higher than the third set value, and output the features to form the third feature subset; if the absolute value of the correlation coefficient is less than the third set value, output the features to form the third feature subset.

5. The dynamic evaluation method for the fracturing effect of deep coal seam gas as described in claim 1, wherein, Perform stage quantification of the peak average daily gas production per well and the cumulative gas production per well in multiple different cycles, including: where q Bit is the average daily gas production at the peak of the i-th cycle for a single well; Q total is the cumulative gas production of a single well; t B is the peak gas breakthrough time of a single well; t total is the cumulative gas production time of a single well; q t is the gas production of a single well at time t.

6. The dynamic evaluation method for deep coal seam gas fracturing effect according to claim 1, characterized in that Based on the preprocessed data, various characteristic parameters, and the quantified production indicators in multiple different stages, establish an interpretable machine learning model with non-linear multi-model fusion between various characteristic parameters and production indicators in different stages, and analyze the influence of each characteristic parameter on the production indicators in different stages in turn, and determine the main control factors from each characteristic parameter, including: Calculate the feature importance of the features in the third feature subset obtained after feature screening by using the recursive elimination method, SHAP method, and Boruta method respectively to obtain the feature importance score; Normalize and perform weighted averaging on the three feature importance scores, determine the feature importance ranking, and select the third feature subset with the greatest importance as the feature subset with the best performance; Use the random forest and recursive feature elimination methods to comprehensively evaluate the influence characteristics of each feature parameter in the feature subset with the best performance on the production indicators of the peak average daily gas production and cumulative gas production of a single well in different periods of the dynamic, and determine the main control factors affecting the post-development effect of deep coalbed methane.

7. The dynamic evaluation method for the fracturing effect of deep coal seam gas as described in claim 1, wherein, Quantitatively evaluate the influence degree and change law of the main control factors on the production indicators at different stages, including: Normalize and perform weighted averaging on the importance scores of the production capacity main control factors under the production indicators at different stages respectively, and construct a comparison chart to quantitatively evaluate the influence degree and change law of the production capacity main control factors under the production indicators at different stages.

8. A dynamic evaluation system for fracturing effect of deep coal seam gas, characterized in that, Including: An acquisition module, which collects the basic data of the fractured wells in the target deep coalbed methane development block, and performs data preprocessing on the basic data to determine the dynamic evaluation characteristic parameters of the deep coalbed methane fracturing effect; A processing module, which analyzes the production dynamics of the fractured wells and quantifies the production indicator stages of the peak average daily gas production and cumulative gas production of a single well in multiple different periods; A training module, based on the preprocessed data, each feature parameter, and the quantified production indicators at multiple different stages, establishes an interpretable machine learning model with non-linear multi-model fusion between each feature parameter and the production indicators at different stages, analyzes the influence of each feature parameter on the production indicators at different stages in turn, and determines the main control factors from each feature parameter to quantitatively evaluate the influence degree and change law of the main control factors on the production indicators at different stages.

9. A computer-readable storage medium storing one or more programs, characterized in that, The one or more A program includes instructions that, when executed by a computing device, cause the computing device to perform any of the methods described in claims 1 to 7.

10. A computing device, characterized in that, Including: One or more processors, a memory, and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the methods described in claims 1 to 7.

Images