Intelligent while-drilling comprehensive fracturing dessert identification method

Abstract

The invention discloses an intelligent while-drilling comprehensive fracturing dessert identification method, which comprises the following steps of enabling logging data of different small layers in a target block reservoir to correspond to a dessert comprehensive evaluation coefficient and production well gas production rate data, and establishing an intelligent dessert identification model training database; and a neural network model is established, part of data in the intelligent fractured dessert recognition model training basic database serves as a training set to be input into the neural network model for training, establishment of a deep learning model is completed, and the intelligent fractured dessert recognition model is formed. In the drilling construction process, logging-while-drilling data are fed back in real time, the logging-while-drilling data serve as input parameters and are uploaded to the intelligent fracturing dessert recognition model, and corresponding dessert comprehensive evaluation coefficients and production well gas production rate data are output in real time; and intelligent while-drilling comprehensive fracturing dessert identification is realized according to the output dessert comprehensive evaluation coefficient and the production well gas production rate data. According to the method, real-time sweet spot identification in the while-drilling process is comprehensively realized.

Description

Technical Field

[0001] The present application relates to the field of fracturing sweet spot identification, and in particular to an intelligent while-drilling comprehensive fracturing sweet spot identification method. Background Art

[0002] In recent years, with the gradual deepening of oil and gas reservoir exploration and development, unconventional oil and gas reservoirs represented by shale and tight sandstone oil and gas have become important successor blocks for increasing reserves and production and tapping potential and stabilizing production in China. Such reservoirs can only effectively release production capacity through fracturing transformation. In addition, the development of conventional onshore oil and gas reservoirs and shallow sea oil and gas reservoirs has gradually entered the marginal area, and the reservoir quality has deteriorated, and fracturing transformation is also required to achieve increased and stable production. Therefore, the continuous advancement of fracturing technology is the key to promoting the deep development of oil and gas resources.

[0003] The purpose of fracturing is that oil and gas reservoirs are highly heterogeneous, and the physical properties and geomechanical properties of the reservoirs encountered in long horizontal sections vary greatly. In order to achieve accurate and low-cost fracturing and achieve the effect of increasing and stabilizing production, it is necessary to efficiently, quickly and accurately identify the "sweet spot" area before fracturing to improve the pertinence and effectiveness of the transformation. Conventional "sweet spot" identification usually requires the determination of various parameters such as physical properties through indoor nuclear magnetic resonance, scanning electron microscopy and other experiments, and qualitative judgment combined with seismic, logging, lithology analysis, etc., which is highly dependent on human experience, lacks reliability, takes a long time, and identifies more geological "sweet spots". However, a good geological sweet spot is not necessarily an engineering sweet spot with good fracturability. When identifying engineering sweet spots, it is necessary to collect, analyze, calculate and interpret a large amount of basic data such as reservoir physical properties, logging, rock mechanics, etc., which takes a long time and has a complicated workload. The model used for calculation is not pertinent to the fracturing block and cannot provide a fast and accurate sweet spot mark for the next fracturing. Therefore, this technology is still insufficient in applying the idea of ​​geological-engineering integration and comprehensively identifying geological "sweet spots" and engineering "sweet spots". It cannot well unify the gas-rich areas of the reservoir with the areas that are conducive to fracturing, affecting the overall production increase effect.

[0004] In addition, fracturing has long been widely used in unconventional oil and gas blocks. The amount of data such as logging data and fracturing construction data in each block is huge and diverse, but the application of these data is mainly limited to single well logging interpretation, post-fracturing analysis, etc., and the potential of data tapping and comprehensive utilization are far from enough. Summary of the invention

[0005] The present application provides an intelligent while-drilling comprehensive fracturing sweet spot identification method, which can realize real-time identification of "sweet spots" during the while-drilling process.

[0006] The embodiment of the present application provides an intelligent while-drilling comprehensive fracturing sweet spot identification method, comprising the following steps:

[0007] Collect reservoir characteristic data of the target block reservoir and gas production data of the production wells.

[0008] According to the reservoir characteristic data, characteristic parameters for evaluating reservoir physical quality and characteristic parameters for evaluating completion quality are screened to form sweet spot evaluation characteristic parameters, and a corresponding relationship is established with the corresponding production well gas production data to form a basic database.

[0009] The grey correlation between the sweet spot evaluation characteristic parameters and the gas production data of the production wells in the basic database is obtained, and the main controlling factors for sweet spot identification are determined according to the size of the grey correlation.

[0010] The main controlling factors of dessert identification are analyzed and the weights of the main controlling factors of dessert identification are determined.

[0011] According to the main controlling factors of dessert identification and the corresponding weights, the comprehensive coefficient of dessert evaluation is calculated.

[0012] The logging data of different sub-layers in the target block reservoir are matched with the comprehensive evaluation coefficient of the sweet spot and the gas production data of the production wells to establish an intelligent sweet spot identification model training database.

[0013] A neural network model is established, and part of the data in the basic training database of the intelligent fracturing sweet spot identification model is used as a training set to input the neural network model for training, completing the establishment of the deep learning model, and the remaining data in the basic training database of the intelligent fracturing sweet spot identification model is used as a test set to test the deep learning model to verify that the accuracy of the sweet spot comprehensive evaluation coefficient and the gas production data of the production wells reaches the preset effect, thereby forming an intelligent fracturing sweet spot identification model.

[0014] During the drilling construction process, the logging while drilling data is fed back in real time and uploaded to the intelligent fracturing sweet spot identification model as input parameters. The corresponding sweet spot comprehensive evaluation coefficient and production well gas production data are output in real time. Based on the output sweet spot comprehensive evaluation coefficient and production well gas production data, intelligent comprehensive fracturing sweet spot identification while drilling is realized.

[0015] In some of the embodiments, during the drilling construction, the reservoir physical property characteristic parameters and geomechanical parameters output by the intelligent fracturing sweet spot identification model are transmitted in real time to the target block well surrounding geological model to update the target block well surrounding geological model.

[0016] In some of the embodiments, after the drilling construction is completed, the intelligent fracturing sweet spot identification model outputs a geomechanical parameter profile along the wellbore trajectory and a geological and engineering integrated sweet spot profile to guide the next step of fracturing layer selection.

[0017] In some embodiments, obtaining the grey correlation between the sweet spot evaluation characteristic parameters and the gas production data of the production well in the basic database includes:

[0018] ① Establish evaluation matrix X iand the reference matrix X0, corresponding to the sweet spot evaluation characteristic parameters and the gas production of the production wells;

[0019]

[0020] X0=[X0(1),X2(2)…,X m (n)]

[0021] Where: m is the number of characteristic parameters for sweet spot evaluation, pieces;

[0022] n——The number of data contained in each dessert evaluation feature parameter, pieces;

[0023] X i ——Numerical value of sweet spot identification characteristic parameter, pcs;

[0024] X0——gas production of production well, cubic meters / day.

[0025] ②Use mean standardization to normalize the matrix [X0, X1, X2…, X n ] T Non-dimensionalization.

[0026] ③Calculate the correlation coefficient ξ i ;

[0027]

[0028] Where: i ——correlation coefficient, dimensionless;

[0029] P——resolution coefficient, P=0.5;

[0030] k——the number of rows in the evaluation matrix, rows;

[0031] i——The number of columns in the evaluation matrix, columns.

[0032] ④Calculate the grey correlation degree;

[0033]

[0034] Where: R i ——Grey correlation.

[0035] In some embodiments, analyzing the main controlling factors of dessert identification to determine the weights of the main controlling factors of dessert identification includes:

[0036] ① Use the range method to identify the main control factor X of the sweet spot as an evaluation index ij To standardize;

[0037] When X ij When it is a positive indicator:

[0038]

[0039] When X ij When it is a negative indicator:

[0040]

[0041] Where, X ij Y is the value of the jth evaluation index in the i-th group of dessert evaluation feature parameters; ij is the standardized evaluation index value; max(X ij ) and min(X ij ) represent X ij The maximum and minimum values ​​of .

[0042] ② Calculate the weight of the evaluation index;

[0043] i=1,2,···,n; j=1,2,···,m

[0044] Where P ij Y is the weight of the jth evaluation index in the i-th group of dessert evaluation feature parameters; ij is the standardized evaluation index value.

[0045] ③ Calculate the information entropy of the evaluation index;

[0046]

[0047] In the formula, E j is the information entropy; n is the number of data contained in each dessert evaluation feature parameter; P ij is the weight of the jth evaluation index in the i-th group of dessert evaluation feature parameters; if P ij If is 0, then define

[0048]

[0049] ④ Calculate the weight of evaluation indicators based on information entropy;

[0050]

[0051] Where W j is the weight of evaluation index j; E j is the information entropy; m is the number of sweet spot evaluation feature parameters.

[0052] In some embodiments, the dessert evaluation comprehensive coefficient is calculated based on the dessert identification main control factors and corresponding weights, including:

[0053]

[0054] Where: δ——comprehensive coefficient of dessert evaluation;

[0055] G i ——The main control factors of dessert identification are normalized and dimensionless;

[0056] W i ——The weight of the main controlling factors for dessert identification.

[0057] In some of the embodiments, the well logging data is normalized before establishing the neural network model.

[0058] In some of the embodiments, the well logging data include: natural gamma, resistivity, sonic transit time, shale content, density, porosity, and wellbore diameter.

[0059] In some of the embodiments, characteristic parameters for evaluating reservoir physical quality include permeability, porosity, and saturation.

[0060] In some of the embodiments, characteristic parameters for evaluating completion quality include Young's modulus, Poisson's ratio, horizontal principal stress, brittleness index, and horizontal stress difference coefficient.

[0061] According to an embodiment of the present application, an intelligent while-drilling integrated fracturing sweet spot identification method is provided, comprising the following steps: collecting reservoir characteristic data of the target block reservoir and gas production data of the production well. According to the reservoir characteristic data, characteristic parameters for evaluating the physical quality of the reservoir and characteristic parameters for evaluating the completion quality are screened to form sweet spot evaluation characteristic parameters, and a corresponding relationship is established with the corresponding production well gas production data to form a basic database. The grey correlation between the sweet spot evaluation characteristic parameters and the production well gas production data in the basic database is obtained, and the main controlling factors for sweet spot identification are determined according to the size of the grey correlation. The main controlling factors for sweet spot identification are analyzed to determine the weights of the main controlling factors for sweet spot identification. According to the main controlling factors for sweet spot identification and the corresponding weights, the comprehensive coefficient for sweet spot evaluation is calculated. The logging data of different sub-layers in the target block reservoir are matched with the sweet spot comprehensive evaluation coefficient and the production well gas production data to establish an intelligent sweet spot identification model training database. Establish a neural network model, input part of the data in the basic database of the intelligent fracturing sweet spot identification model training as a training set into the neural network model for training, complete the establishment of the deep learning model, and use the remaining part of the data in the basic database of the intelligent fracturing sweet spot identification model training as a test set to test the deep learning model, verify that the accuracy of the sweet spot comprehensive evaluation coefficient and the production well gas production data reaches the preset effect, and form an intelligent fracturing sweet spot identification model. During the drilling construction process, the logging data while drilling is fed back in real time, and the logging data while drilling is uploaded as an input parameter to the intelligent fracturing sweet spot identification model, and the corresponding sweet spot comprehensive evaluation coefficient and production well gas production data are output in real time. According to the output sweet spot comprehensive evaluation coefficient and production well gas production data, the intelligent comprehensive fracturing sweet spot identification while drilling is realized. An intelligent comprehensive fracturing sweet spot identification method while drilling in the present application can deeply mine multi-dimensional data such as well logging, reservoir physical properties, geomechanics, and post-fracturing production, organically combine engineering sweet spots with geological sweet spots, and use machine learning methods. Taking expert analysis results as a reference, it comprehensively realizes the real-time identification of "sweet spots" in the drilling process, solves the problems of excessive influence of human factors and experience and insufficient reliability in the process of fracturing section and layer selection, reduces the impact of decision-making results differences caused by human factors on the final fracturing effect, improves fracturing design and construction efficiency, reduces costs, and achieves the purpose of increasing production in a more targeted manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 Schematic diagram of the process of the intelligent while-drilling comprehensive fracturing sweet spot identification method in the embodiment of the present application. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0065] See also Figure 1 The embodiment of the present application provides an intelligent while-drilling comprehensive fracturing sweet spot identification method, comprising the following steps:

[0066] Step 1: Collect reservoir characteristic data of the target block reservoir and gas production data of the production wells.

[0067] In the above steps, the current logging while drilling (LWD) can basically cover most of the logging information obtained by traditional cable logging, including: natural gamma, resistivity, acoustic time difference, shale content, density, porosity, well diameter and even imaging logging data. We select a certain amount of LWD data and conventional logging data from different layers in the target block, combined with the gas production of the corresponding layer, as the data basis for subsequent data processing and analysis. Therefore, the reservoir characteristic data can include reservoir logging data and reservoir physical property experimental analysis data. Logging data can include: natural gamma, resistivity, acoustic time difference, shale content, density, porosity, well diameter, etc.

[0068] Step 2: Based on the reservoir characteristic data, select the characteristic parameters for evaluating the physical quality of the reservoir and the characteristic parameters for evaluating the completion quality, form the sweet spot evaluation characteristic parameters, and establish a corresponding relationship with the gas production data of the production wells in the corresponding target layer to form a basic database.

[0069] In the above steps, the sweet spot evaluation should follow the concept of geological engineering integration, comprehensively consider geological sweet spots and engineering sweet spots, and include two main aspects, namely reservoir physical quality (RQ) and completion quality (CQ). Among them, the characteristic parameters of reservoir physical quality are mainly reservoir physical parameters such as porosity, permeability k, saturation Sg or So, and the above parameters can be directly obtained through logging interpretation. The characteristic parameters of completion quality are mainly geomechanical parameters such as Young's modulus, Poisson's ratio, horizontal principal stress, brittleness, horizontal stress difference coefficient, etc., and the above parameters need to be calculated through logging data. Therefore, the logging data and reservoir physical property experimental analysis data can be combined to screen permeability, porosity, saturation, etc. as characteristic parameters for evaluating reservoir physical quality. Young's modulus, Poisson's ratio, horizontal principal stress, brittleness index, horizontal stress difference coefficient, etc. can be calculated through logging data as characteristic parameters for evaluating completion quality.

[0070] in,

[0071] (1) Young's modulus: The higher the Young's modulus, the better the completion quality of the reservoir. The calculation method is as follows:

[0072]

[0073] Where, E is Young's modulus, GPa; is the formation shear wave time difference, μs / m; is the formation P-wave time difference, μs / m; ρ b is the rock bulk density, g / cm 3 ; β is the conversion coefficient.

[0074] (2) Poisson's ratio: The lower the Poisson's ratio, the better the completion quality of the reservoir. The calculation method is as follows:

[0075]

[0076] Among them, Δt s is the shear wave time difference, μs / m; Δt c is the longitudinal wave time difference, μs / m; v is the Poisson's ratio, dimensionless.

[0077] (3) Horizontal principal stress: The lower the horizontal principal stress, the better the completion quality of the reservoir. The calculation method is as follows:

[0078]

[0079]

[0080] Among them, σ H is the maximum horizontal principal stress, MPa; σ h is the minimum horizontal principal stress, MPa; σ v is the overburden pressure, MPa; α is the Biot coefficient, dimensionless; is the maximum and minimum well diameter, cm; D g is the rock density, g / cm 3 .

[0081] (4) Brittleness index: The brittleness index is an important indicator for determining the compressibility of the reservoir. The larger the brittleness index, the better the completion quality. The calculation method is as follows:

[0082] BRIT=(YM_BRIT+PR_BRIT) / 2

[0083] YM_BRIT=(EE min ) / (E max -E min )×100%

[0084] PR_BRIT=(v-υ min) / (v max -v min )×100%

[0085] Among them, BRIT is the rock brittleness index; YM_BRIT is the normalized Young's modulus; PR_BRIT is the normalized Poisson's ratio; Emax is the maximum Young's modulus; Emin is the minimum Young's modulus; υ max is the maximum Poisson's ratio; υmin is the minimum Poisson's ratio, E is the Young's modulus; υ is the Poisson's ratio.

[0086] (5) Horizontal stress difference coefficient: The horizontal stress difference coefficient is an important indicator for the formation of complex fractures. The larger the coefficient, the better the completion quality. The calculation method is as follows:

[0087]

[0088] Where ψ is the horizontal stress difference coefficient and is dimensionless.

[0089] The above steps 1 and 2 realize the establishment of a database of block physical properties and completion quality.

[0090] Step 3: Use grey correlation analysis to obtain the grey correlation between the sweet spot evaluation characteristic parameters in the basic database and the gas production data of the production wells, and determine the main controlling factors for sweet spot identification according to the size of the grey correlation.

[0091] In the above steps, obtaining the grey correlation between the sweet spot evaluation characteristic parameters in the basic database and the gas production data of the production wells may include:

[0092] ① Establish evaluation matrix X i and the reference matrix X0, corresponding to the sweet spot evaluation characteristic parameters and the gas production of the production wells;

[0093]

[0094] X0=[X0(1),X2(2)…,X m (n)]

[0095] Where: m is the number of characteristic parameters for sweet spot evaluation, pieces;

[0096] n——The number of data contained in each dessert evaluation feature parameter, pieces;

[0097] X i ——The value of the sweet spot identification characteristic parameter, pcs;

[0098] X0——gas production of production well, cubic meters / day.

[0099] ②Use mean standardization to normalize the matrix [X0, X1, X2…, X n ] TNon-dimensionalization.

[0100] ③Calculate the correlation coefficient ξ i ;

[0101]

[0102] Where: i ——correlation coefficient, dimensionless;

[0103] P——resolution coefficient, P=0.5;

[0104] k——the number of rows in the evaluation matrix, rows;

[0105] i——The number of columns in the evaluation matrix, columns.

[0106] ④Calculate the grey correlation degree;

[0107]

[0108] Where: R i ——Grey correlation.

[0109] Determining the main control factor for dessert identification according to the grey correlation degree may include: selecting a dessert evaluation characteristic parameter with a correlation coefficient greater than 0.8 as the main control factor for dessert evaluation.

[0110] The above steps use the grey correlation analysis method to study each characteristic parameter. The greater the correlation, the greater its influence on the evaluation of desserts.

[0111] Step 4: Analyze the main control factors of dessert identification through the entropy weight method to determine the weight of each main control factor of dessert identification.

[0112] In the above steps, analyzing the main control factors of dessert identification and determining the weights of the main control factors of dessert identification may include:

[0113] ① Use the range method to identify the main control factor X of the sweet spot as an evaluation index ij To standardize;

[0114] When X ij When it is a positive indicator:

[0115]

[0116] When X ij When it is a negative indicator:

[0117]

[0118] In the formula, X ij Y is the value of the jth evaluation index in the i-th group of dessert evaluation feature parameters; ijis the standardized evaluation index value; max(X ij ) and min(X ij ) represent X ij The maximum and minimum values ​​of .

[0119] ② Calculate the weight of the evaluation index;

[0120] i=1,2,···,n; j=1,2,···,m

[0121] Where P ij Y is the weight of the jth evaluation index in the i-th group of dessert evaluation feature parameters; ij is the standardized evaluation index value.

[0122] ③ Calculate the information entropy of the evaluation index;

[0123]

[0124] In the formula, E j is the information entropy; n is the number of data contained in each dessert evaluation feature parameter; P ij is the weight of the jth evaluation index in the i-th group of dessert evaluation feature parameters; if P ij If is 0, then define

[0125] ④ Calculate the weight of evaluation indicators based on information entropy;

[0126]

[0127] Where W j is the weight of evaluation index j; E j is the information entropy; m is the number of characteristic parameters for sweet spot evaluation, such as reservoir physical parameters such as porosity, permeability k, saturation Sg or So, and geomechanical parameters such as Young's modulus, Poisson's ratio, horizontal principal stress, brittleness, and horizontal stress difference coefficient.

[0128] According to the main control factors of dessert identification and the corresponding weights, the comprehensive coefficient of dessert evaluation is calculated, including:

[0129]

[0130] Where: δ——comprehensive coefficient of dessert evaluation;

[0131] G i ——The main control factors of dessert identification are normalized and dimensionless;

[0132] W i ——The weight of the main controlling factors for dessert identification.

[0133] The above steps obtain the weight coefficient by establishing an entropy weight method model of the dependent variable (gas production) and multiple independent variables (main controlling factors for sweet spot evaluation).

[0134] Step 5: Based on the main control factors of the dessert identification selected above and the corresponding weights calculated, the comprehensive coefficient of dessert evaluation is calculated comprehensively.

[0135] The above steps three to five take into account that there is a certain correlation between the characteristic parameters of the sweet spot evaluation and the gas production, and different parameters have different influences on the post-fracturing production. Therefore, combined with the post-fracturing production data, the grey correlation analysis method is used to select the factors with high correlation as the main control factors for evaluation, and the entropy weight method is used to determine the evaluation weights. The sweet spot evaluation standard is comprehensively determined to establish a comprehensive evaluation standard for block fracturing sweet spots.

[0136] Step 6. Through the above calculations, the logging data of different sub-layers in the target block reservoir are directly matched with the sweet spot comprehensive evaluation coefficient and the gas production data of the production wells, and then an intelligent sweet spot identification model training database is established.

[0137] In the above steps, the logging data may include: natural gamma, resistivity, sonic time difference, mud content, density, porosity, well diameter, etc.

[0138] Step 7. Establish a neural network model, and input part of the data (80% of the data) in the basic database of the intelligent fracturing sweet spot identification model training as the training set data into the neural network model for training, and continuously optimize and adjust the key parameters such as the activation function and the number of neurons in the neural network model until the loss function is minimized, and complete the optimization of the neural network model, that is, the establishment of a deep learning model, and use the remaining part of the data in the basic database of the intelligent fracturing sweet spot identification model training (the remaining 20% ​​of the data) as the test set data to test the deep learning model, verify that the accuracy of the sweet spot comprehensive evaluation coefficient and the gas production data of the production wells reaches the preset effect (more than 85%), and realize the verification of the optimized neural network model prediction results, that is, the deep learning model prediction results, to form an intelligent fracturing sweet spot identification model, or an intelligent fracturing comprehensive sweet spot model.

[0139] In the above steps, before establishing the neural network model, the deviation standard method can be used to normalize the logging data to eliminate the influence of different dimensions on data analysis and make the data comparable. After completing the normalization of the logging data, the neural network model can be established.

[0140] The normalization method is as follows:

[0141]

[0142] In the formula, x max 、x minare the maximum and minimum values ​​in the sequence respectively.

[0143] On this basis, considering that more logging curve categories and a large amount of data can improve the accuracy of the recognition model, there must be a certain degree of correlation between some logging curve categories, which causes data redundancy and reduces the model running speed and data processing efficiency. It is necessary to perform correlation analysis and principal component dimensionality reduction to select the main categories with poor mutual correlation as feature parameters for subsequent model training.

[0144] Usually the correlation coefficient method or regression analysis method can be used, as follows:

[0145]

[0146]

[0147] In the formula, r xy is the sample correlation coefficient, s xy is the sample covariance, s x is the sample standard deviation of x, s y is the sample standard deviation of y.

[0148] Step 8. During the drilling construction process, the logging data while drilling is fed back in real time and uploaded to the intelligent fracturing sweet spot identification model as input parameters. Through the calculation and analysis of the model, the corresponding sweet spot comprehensive evaluation coefficient and production well gas production data are output in real time. The sweet spot area of ​​the current drilled layer is judged based on the output sweet spot comprehensive evaluation coefficient and production well gas production data, guiding the adjustment of the wellbore trajectory, accurately hitting the target area, ensuring the reservoir drilling rate, and realizing intelligent comprehensive fracturing sweet spot identification while drilling.

[0149] In the above steps, when the corresponding sweet spot comprehensive evaluation coefficient and production well gas production data are output in real time, the oil and water layer identification of the actual drilling layer can also be output in real time to realize the identification and adjustment of the target layer during the drilling process.

[0150] The above steps 1 to 8 realize the determination of the comprehensive sweet spot of fracturing.

[0151] During the drilling process, the reservoir physical property characteristic parameters and geomechanical parameters output by the intelligent fracturing sweet spot identification model can be transmitted to the target block's surrounding geological model in real time to realize the input of basic parameters for fracturing model construction, so as to continuously update the target block's surrounding geological model, which is conducive to updating reservoir understanding, guiding subsequent well drilling and fracturing scheme design, and improving the accuracy of block production analysis and prediction.

[0152] After the drilling construction is completed, the intelligent fracturing sweet spot identification model can directly output the geomechanical parameter profile along the wellbore trajectory (including Young's modulus, Poisson's ratio, brittleness index, horizontal principal stress, etc.) and the integrated geological and engineering sweet spot profile, so as to efficiently, quickly and intuitively guide the next step of fracturing layer selection.

[0153] In summary, an intelligent comprehensive fracturing sweet spot identification method while drilling in the present application can deeply mine multi-dimensional data such as well logging, reservoir physical properties, geomechanics, and post-fracturing production, organically combine engineering sweet spots with geological sweet spots, and use machine learning methods. Taking expert analysis results as a reference, it can comprehensively realize the real-time identification of "sweet spots" in the drilling process, solve the problems of excessive influence of human factors and experience and insufficient reliability in the process of fracturing section and layer selection, reduce the impact of decision-making results caused by human factors on the final fracturing effect, improve fracturing design and construction efficiency, reduce costs, and achieve the purpose of increasing production in a more targeted manner.

[0154] The same or similar numbers in the drawings of this embodiment correspond to the same or similar parts; in the description of this application, it should be understood that if the terms "upper", "lower", "left", "right", etc. indicate an orientation or position relationship based on the orientation or position relationship shown in the drawings, it is only for the convenience of describing this application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as a limitation on this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0155] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An intelligent method for identifying sweet spots during drilling comprehensive fracturing, characterized in that: The following steps are involved: Collect reservoir characteristics data of target block reservoirs and gas production data of production wells; According to the reservoir characteristic data, the characteristic parameters for evaluating reservoir physical quality and the characteristic parameters for evaluating completion quality are screened to form the sweet spot evaluation characteristic parameters, and the corresponding relationship is established with the corresponding production well gas production data to form a basic database; Obtain the grey correlation between the sweet spot evaluation characteristic parameters and the gas production data of the production wells in the basic database, and determine the main control factors for sweet spot identification according to the grey correlation; Analyze the main control factors of dessert identification and determine the weights of the main control factors of dessert identification; According to the main control factors of dessert identification and the corresponding weights, the comprehensive coefficient of dessert evaluation is calculated; The logging data of different layers in the target block reservoir are matched with the comprehensive evaluation coefficient of the sweet spot and the gas production data of the production wells to establish an intelligent sweet spot identification model training database; Establish a neural network model, input part of the data in the basic database of the intelligent fracturing sweet spot identification model training as a training set into the neural network model for training, complete the establishment of the deep learning model, and use the remaining part of the data in the basic database of the intelligent fracturing sweet spot identification model training as a test set to test the deep learning model, verify that the accuracy of the sweet spot comprehensive evaluation coefficient and the gas production data of the production wells reaches the preset effect, and form an intelligent fracturing sweet spot identification model; During the drilling construction process, the logging while drilling data is fed back in real time and uploaded to the intelligent fracturing sweet spot identification model as input parameters. The corresponding sweet spot comprehensive evaluation coefficient and production well gas production data are output in real time. Based on the output sweet spot comprehensive evaluation coefficient and production well gas production data, intelligent comprehensive fracturing sweet spot identification while drilling is realized.

2. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: During the drilling construction process, the reservoir physical property characteristic parameters and geomechanical parameters output by the intelligent fracturing sweet spot identification model are transmitted to the target block's surrounding geological model in real time to update the target block's surrounding geological model.

3. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: After the drilling construction is completed, the intelligent fracturing sweet spot identification model outputs the geomechanical parameter profile along the wellbore trajectory and the integrated geological and engineering sweet spot profile to guide the next step of fracturing layer selection.

4. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: Obtain the grey correlation between the sweet spot evaluation characteristic parameters and the gas production data of the production wells in the basic database, including: ① Establish evaluation matrix X i and the reference matrix X0, corresponding to the sweet spot evaluation characteristic parameters and the gas production of the production wells; X0=[X0(1),X2(2)…,X m (n)] Where: m is the number of characteristic parameters for sweet spot evaluation, pieces; n——The number of data contained in each dessert evaluation feature parameter, pieces; X i ——The value of the sweet spot identification characteristic parameter, pcs; X0——gas production of production well, cubic meter / day; ②Use mean standardization to normalize the matrix [X0, X1, X2…, X n ] T Make it dimensionless; ③Calculate the correlation coefficient ξ i ; Where: i ——correlation coefficient, dimensionless; ρ——resolution coefficient, ρ=0.5; k——the number of rows in the evaluation matrix, rows; i——the number of columns of the evaluation matrix, columns; ④Calculate the grey correlation degree; Where: R i ——Grey correlation.

5. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: Analyze the main control factors of dessert identification and determine the weights of the main control factors of dessert identification, including: ① Use the range method to identify the main control factor X of the sweet spot as an evaluation index ij Standardization When X ij When it is a positive indicator: When X ij When it is a negative indicator: Where, X ij Y is the value of the jth evaluation index in the i-th group of dessert evaluation feature parameters; ij is the standardized evaluation index value; max(X ij ) and min(X ij ) represent X ij The maximum and minimum values ​​of ; ② Calculate the weight of evaluation indicators Where P ij Y is the weight of the jth evaluation index in the i-th group of dessert evaluation feature parameters; ij is the standardized evaluation index value; ③Calculate the information entropy of evaluation indicators In the formula, E j is the information entropy; n is the number of data contained in each dessert evaluation feature parameter; P ij is the weight of the jth evaluation index in the i-th group of dessert evaluation feature parameters; if P ij If is 0, then define ④ Calculate the weight of evaluation indicators based on information entropy Where W j is the weight of evaluation index j; E j is the information entropy; m is the number of sweet spot evaluation feature parameters.

6. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: According to the main control factors of dessert identification and the corresponding weights, the comprehensive coefficient of dessert evaluation is calculated, including: Where: δ——comprehensive coefficient of dessert evaluation; G i ——The main control factors of dessert identification are normalized and dimensionless; W i ——The weight of the main controlling factors for dessert identification.

7. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: Before building the neural network model, the logging data was normalized.

8. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: Well logging data include: natural gamma, resistivity, sonic time difference, mud content, density, porosity, and well diameter.

9. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: Characteristic parameters for evaluating reservoir physical quality include permeability, porosity, and saturation.

10. The intelligent while-drilling comprehensive fracturing sweet spot identification method according to claim 1, characterized in that: Characteristic parameters for evaluating completion quality include Young's modulus, Poisson's ratio, horizontal principal stress, brittleness index, and horizontal stress difference coefficient.

Images