Tight gas reservoir fracturing sweet spot prediction method and device, electronic equipment and medium

Abstract

The invention relates to a tight gas reservoir fracturing dessert prediction method and device, electronic equipment and a medium, and belongs to the technical field of hydraulic fracturing of petroleum and natural gas engineering.The method comprises the steps that historical values of reservoir parameters and dessert evaluation parameters are obtained; calculating a difference matrix of each reservoir parameter and the dessert evaluation parameter based on historical values of the reservoir parameters and the dessert evaluation parameter, calculating a correlation degree of each reservoir parameter and the dessert evaluation parameter based on the difference matrix, and selecting a target parameter of which the correlation degree is greater than a threshold value from the reservoir parameters; training the initial lightweight gradient elevator model based on the target parameters and the dessert evaluation parameters to obtain a completely trained lightweight gradient elevator model; and predicting the dessert based on the lightweight gradient elevator model which is completely trained. According to the method, the target parameters are screened from the aspect of trend matching, the non-linear relation between the target parameters is processed through the lightweight gradient elevator model, and high-precision sweet spot prediction is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of hydraulic fracturing in oil and natural gas engineering, and in particular to a method, device, electronic equipment and medium for predicting fracturing sweet spots in tight gas reservoirs. Background Art

[0002] Tight sandstone gas reservoirs are widespread and rich in reserves. However, development is challenging and less than ideal due to formation heterogeneity and poor engineering-geological compatibility. Currently, staged horizontal well fracturing is widely used for these reservoirs. This requires the prediction of fracturing sweet spots in tight gas reservoirs. Reservoir sweet spots serve as a key basis for well placement, fracturing staging, and perforation location optimization. These sweet spots include both geological sweet spots (high-quality resource areas) and engineering sweet spots (areas prone to complex fracture formation).

[0003] In the existing technology, traditional sweet spot prediction methods mainly include single factor analysis methods, which rely on single geological or engineering indicators.

[0004] However, there are complex nonlinear relationships between reservoir parameters (such as the interaction between permeability and brittleness index). Traditional single-factor analysis methods are difficult to deeply explore the intrinsic connections between multi-parameter coupling effects. For example, the coupling effect of permeability and brittleness index has poor prediction accuracy, which can easily lead to problems such as blind well and layer selection and high fracturing costs. Summary of the Invention

[0005] In view of this, it is necessary to provide a method, device, electronic equipment and medium for predicting fracturing sweet spots in tight gas reservoirs to solve the technical problem of poor prediction accuracy in the prior art.

[0006] In order to solve the above problems, in a first aspect, the present invention provides a method for predicting fracturing sweet spots in tight gas reservoirs, comprising: Obtain historical values ​​of reservoir parameters and sweet spot evaluation parameters; Calculating a difference matrix between each reservoir parameter and the sweet spot evaluation parameter based on historical values ​​of the reservoir parameter and the sweet spot evaluation parameter, calculating a correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix, and selecting a target parameter having the correlation greater than a threshold value from the reservoir parameters; Training an initial lightweight gradient boosting machine model based on historical values ​​of the target parameter and the sweet spot evaluation parameter to obtain a fully trained lightweight gradient boosting machine model; Obtain a target parameter's value to be measured, input the target parameter's value to be measured into the fully trained lightweight gradient boosting machine model, obtain a predicted value of a sweet spot evaluation parameter, and predict a sweet spot based on the predicted value.

[0007] In a possible implementation, after obtaining historical values ​​of reservoir parameters and sweet spot evaluation parameters, the method further includes: Performing forward normalization processing on historical values ​​of the forward parameters among the historical values ​​of the reservoir parameters and the sweet spot evaluation parameters, and performing reverse normalization processing on historical values ​​of the reverse parameters; After obtaining the value of the target parameter to be measured, the method further includes: The forward normalization processing is performed on the forward parameter to be measured values ​​of the target parameter to be measured, and the reverse normalization processing is performed on the reverse parameter to be measured values.

[0008] In one possible implementation, the reservoir parameters include geological parameters and engineering parameters. The geological parameters include porosity, permeability, gas saturation, mud content and total hydrocarbon average value. The engineering parameters include brittleness index, fracture toughness, Poisson's ratio, Young's modulus and compressive strength.

[0009] In a possible implementation, the sweet spot evaluation parameter is an open flow parameter.

[0010] In a possible implementation, calculating the correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix includes: Selecting a global minimum difference and a global maximum difference from the difference matrix; Calculating the correlation coefficient between each reservoir parameter and the sweet spot evaluation parameter at each moment based on the global minimum difference and the global maximum difference; The correlation coefficients at each moment are averaged to obtain the correlation degree between each reservoir parameter and the sweet spot evaluation parameter.

[0011] In one possible implementation, a grid search and 3-fold cross validation method are used in the initial lightweight gradient boosting machine model training process.

[0012] In one possible implementation, an early stopping mechanism is used during the initial lightweight gradient boosting machine model training process.

[0013] In a second aspect, the present invention further provides a device for predicting sweet spots in tight gas reservoir fracturing, comprising: A data acquisition unit, used for acquiring historical values ​​of reservoir parameters and sweet spot evaluation parameters; a parameter screening unit, configured to calculate a difference matrix between each reservoir parameter and the sweet spot evaluation parameter based on historical values ​​of the reservoir parameter and the sweet spot evaluation parameter, calculate a correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix, and select a target parameter having the correlation greater than a threshold value from the reservoir parameters; A model training unit, configured to train an initial lightweight gradient boosting machine model based on the historical values ​​of the target parameter and the sweet spot evaluation parameter to obtain a fully trained lightweight gradient boosting machine model; A model prediction unit is used to obtain a target parameter's value to be measured, input the target parameter's value to be measured into the fully trained lightweight gradient boosting machine model, obtain a predicted value of a sweet spot evaluation parameter, and predict the sweet spot based on the predicted value.

[0014] In a third aspect, the present invention further provides an electronic device comprising a memory and a processor; The memory is used to store programs; The processor is coupled to the memory and is configured to execute the program stored in the memory to implement the steps of the above-mentioned method for predicting the fracturing sweet spot of a tight gas reservoir.

[0015] In a fourth aspect, the present invention further provides a computer-readable storage medium having a program or instruction stored thereon, which, when executed by a processor, implements the steps of the above-mentioned method for predicting fracturing sweet spots in tight gas reservoirs.

[0016] The beneficial effects of the present invention are as follows: the method for predicting sweet spots in fracturing of tight gas reservoirs provided by the present invention, targeting the characteristics of strong heterogeneity and complex nonlinear interaction of parameters in tight gas reservoirs, first uses the grey correlation analysis method to screen out core target parameters with high matching degree with target trends, and explores the relationship between reservoir parameters and sweet spot evaluation parameters from the perspective of "trend matching", incorporates the logic of domain analysis, solves the problems of complex relationships between reservoir parameters and uneven data distribution, and then uses a lightweight gradient boosting machine model to process the nonlinear relationships and complex interaction patterns between high-dimensional and sparse target parameters, and captures the deep laws that are difficult to characterize by the grey correlation analysis method. The combination of the two not only retains the analysis logic of the traditional method, but also gives play to the modeling advantages of machine learning, effectively considers the coupling effects between multiple parameters, and realizes high-precision sweet spot prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic flow chart of an embodiment of a method for predicting sweet spots in tight gas reservoir fracturing provided by the present invention; Figure 2 For the present invention Figure 1 A schematic flow chart of an embodiment of step S102; Figure 3 The association ranking diagram provided by the present invention; Figure 4 A comparison chart of the model prediction results provided by the present invention; Figure 5 A schematic structural diagram of an embodiment of a tight gas reservoir fracturing sweet spot prediction device provided by the present invention; Figure 6 This is a schematic structural diagram of an embodiment of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0019] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in the present invention illustrate operations implemented according to some embodiments of the present invention. It should be understood that the operations of the flowcharts can be implemented out of sequence, and steps that have no logical contextual relationship can be reversed in order or implemented simultaneously. In addition, those skilled in the art, guided by the content of the present invention, can add one or more other operations to the flowcharts or remove one or more operations from the flowcharts. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0020] The terms "first" and "second" in the embodiments of the present invention are used only for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, technical features specified as "first" or "second" may explicitly or implicitly include at least one of these features. "And / or" describes the association between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone.

[0021] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0022] Before presenting the embodiments, the following terms are explained.

[0023] The present invention provides a method, device, electronic equipment and medium for predicting fracturing sweet spots in tight gas reservoirs, which are described below respectively.

[0024] Figure 1 A schematic flow chart of an embodiment of the method for predicting sweet spots in tight gas reservoir fracturing provided by the present invention is shown in FIG. Figure 1As shown in Figure 2, the tight gas reservoir fracturing sweet spot prediction method includes: S101, obtaining historical values ​​of reservoir parameters and sweet spot evaluation parameters; In order to better determine the sweet spot for fracturing tight gas reservoirs, in some embodiments of the present invention, the reservoir parameters and sweet spot evaluation parameters that need to be considered in the study area are analyzed, screened, and determined. The reservoir parameters include geological parameters and engineering parameters. The geological parameters are used to reflect the key geological attributes of the reservoir's oil and gas enrichment capacity, including porosity, permeability, gas saturation, mud content, and total hydrocarbon average value. The engineering parameters are used to evaluate the reservoir's fracturing ability and development benefits, including brittleness index, fracture toughness, Poisson's ratio, Young's modulus, and compressive strength. In a specific embodiment, the absolute open flow (AOF) parameter is selected as the sweet spot evaluation parameter.

[0025] Taking into account that reservoir parameters do not distinguish between parameter directions, which may lead to dimensional confusion between mud content (negative) and permeability (positive), affecting the rationality of subsequent prediction models, in some embodiments of the present invention, the positive parameters and negative parameters in the reservoir parameters and sweet spot evaluation parameters are normalized separately to ensure that the larger the eigenvalue, the more favorable it is for fracturing. Specifically, after step S101, the method further includes: performing positive normalization processing on the historical values ​​of the positive parameters in the historical values ​​of the reservoir parameters and the sweet spot evaluation parameters, and performing reverse normalization processing on the historical values ​​of the reverse parameters.

[0026] It should be noted that the larger the positive parameter, that is, the parameter value, the more beneficial it is to fracturing, and the smaller the reverse parameter, that is, the parameter value, the more beneficial it is to fracturing. The forward normalization formula is (x-min) / (max-min), that is, the characteristic value is scaled to [0,1]. The larger the value, the better the characteristic (for example, for permeability POR, the larger the value, the better. After forward normalization, the larger the value, the higher the permeability). The reverse normalization formula is (max-x) / (max-min), that is, the characteristic value is scaled to [0,1]. The larger the value, the better the characteristic (for example, for shale content VSH, the smaller the value, the better. After reverse normalization, the larger the value, the lower the shale content). In a specific embodiment, the sweet spot evaluation parameter AOF is processed by forward normalization to ensure consistency with the characteristic scale.

[0027] S102, calculating a difference matrix between each reservoir parameter and the sweet spot evaluation parameter based on historical values ​​of the reservoir parameter and the sweet spot evaluation parameter, calculating a correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix, and selecting a target parameter with a correlation greater than a threshold value from the reservoir parameters; In some embodiments of the present invention, Figure 2 As shown, in step S102, the correlation between each reservoir parameter and the sweet spot evaluation parameter is calculated based on the difference matrix, including: S201, selecting a global minimum difference and a global maximum difference from the difference matrix; It should be noted that the difference matrix is ​​obtained by comparing the normalized reservoir parameter matrix with the column vector of the sweet spot evaluation parameters.

[0028] S202, calculating the correlation coefficient between each reservoir parameter and the sweet spot evaluation parameter at each moment based on the global minimum difference and the global maximum difference; It should be noted that the calculation formula of the correlation coefficient is:

[0029]

[0030] Where, Indicates the Reservoir parameters and Dessert evaluation parameters The correlation coefficient of the moment, Indicates the Reservoir parameters The historical value of the moment, Indicates the Dessert evaluation parameters The historical value of the moment, represents the global minimum difference, represents the global maximum difference, represents the resolution factor (set to 0.5 in a specific embodiment)

[0031] S203. Average the correlation coefficients at each moment to obtain the correlation degree between each reservoir parameter and the sweet spot evaluation parameter.

[0032] It should be noted that the calculation formula for correlation is:

[0033] Where, Indicates the The correlation between reservoir parameters and sweet spot evaluation parameters, Indicates the The number of historical values ​​of a reservoir parameter.

[0034] It should be noted that the grey relational analysis method is mostly used in traditional statistical analysis. In the present invention, taking into account the advantages of the grey relational analysis (GRA) method such as nonlinear adaptability, small sample robustness, and parameter adjustability, it is suitable for dealing with complex relationships between reservoir parameters and uneven data distribution problems. The grey relational analysis algorithm is used to screen the target parameters.

[0035] It's also worth noting that GRA quantifies the "importance" of reservoir parameters by calculating correlation coefficients with sweet spot evaluation parameters (such as AOF), thereby selecting core target parameters that closely match the target trend. These features serve as inputs to subsequent prediction models, reducing redundant information and lowering model training complexity. This reduces computational costs, as subsequent prediction models no longer need to process a large number of irrelevant features, resulting in faster training and lower resource consumption. This efficiency improvement is particularly pronounced in scenarios with high feature dimensionality.

[0036] In a specific embodiment, the reservoir parameters are porosity, permeability, gas saturation, mud content, total hydrocarbon average, brittleness index, fracture toughness, Poisson's ratio, Young's modulus and compressive strength, and the sweet spot evaluation parameter is the open flow parameter. The correlation is sorted from high to low, and the correlation ranking diagram is as follows: Figure 3 As shown in the figure, the first six reservoir parameters (permeability, total hydrocarbon average, porosity, brittleness index, fracture toughness, and shale content) are selected as target parameters.

[0037] S103, training the initial lightweight gradient boosting machine model based on the target parameters and the sweet spot evaluation parameters to obtain a fully trained lightweight gradient boosting machine model; It should be noted that the Light Gradient Boosting Machine (LightGBM) model, while inheriting the GBDT framework, innovatively introduces three optimization techniques: a histogram-based feature discretization algorithm, which maps continuous features into discrete interval statistics, significantly reducing computational effort and improving splitting efficiency; gradient one-side sampling (GOSS), which reduces data size while ensuring model accuracy by retaining high-gradient samples and randomly sampling low-gradient samples; and a leaf-wise growth strategy with depth restriction, which prioritizes splitting leaf nodes with the largest gains and limits tree depth, effectively controlling overfitting risk while improving fitting capabilities. These improvements enable LightGBM to increase training speed by 2-20 times and reduce memory consumption by over 80% while maintaining the high accuracy advantages of GBDT. It can efficiently process large-scale data and balance classification and regression tasks, and is primarily used for purely data-driven modeling. In this invention, considering that the target parameters belong to high-dimensional sparse data, the lightweight gradient boosting machine model performs well in interpreting the feature importance of high-dimensional sparse data, and has better training speed, computational efficiency, interpretability and anti-overfitting ability. It is particularly suitable for complex geological-engineering parameter modeling scenarios such as sweet spot prediction of tight sandstone gas reservoirs. The lightweight gradient boosting machine model is selected for sweet spot prediction.

[0038] In order to better train the model, in some embodiments of the present invention, during training, a data set is constructed based on the historical values ​​of the target parameters and the dessert evaluation parameters, and the data set is divided into a training set (80%) and a test set (20%). A fixed random seed is set to ensure that the results are reproducible; the model is trained with the training set to simulate real application scenarios, and the model generalization ability is evaluated with the test set; during the training process, a regression task is specified (the target variable AOF is a continuous value), and the mean square error (MSE) is used as the evaluation indicator during the training process. The larger the mean square error value, the better the model performance. The random seed is fixed to ensure that the model training results are consistent; and grid search and 3-fold cross-validation are used to find the optimal parameters. During cross-validation, the training set is divided into 3 subsets, which are used as validation sets in turn. When the grid search traverses the combination, there are 3×3×3=27 parameter combinations to find the parameters that minimize the MSE of the validation set.

[0039] To prevent overfitting, in some embodiments of the present invention, an early stopping mechanism is used during model training. The early stopping mechanism will terminate the model training early if the validation set MSE does not improve within a preset number of iterations (such as 10 rounds) and automatically select the optimal number of iterations.

[0040] In a specific embodiment, after the model is trained with the training set, the model will use the validation set to predict the target variable AOF, and finally fit the predicted value with the actual value to reflect the model prediction accuracy and output the result. The comparison chart of the model prediction results is as follows: Figure 4 As shown, from Figure 4 It can be seen that the model uses MSE (mean square error) to reflect the overall deviation of the model. The final MSE value is 0.1962, indicating that the prediction error is small and the model accuracy is high. Model fit (R 2 ) value is 0.9150, R 2 The closer the value is to 1, the stronger the model's ability to explain the data. A value of 0.9150 indicates that the model can explain 91.5% of the data variation, indicating excellent fitting and good prediction model performance. The data range of 0-2 is densely populated and has higher prediction accuracy. This is because the target reservoir is highly heterogeneous and the reservoir parameters are unevenly distributed. Data samples within this range account for the vast majority of the original dataset, resulting in a sufficient sample size, allowing the model to learn better and achieve higher prediction accuracy. However, the data range of 3-5, due to the smaller amount of original data, results in insufficient model training accuracy, and the predicted values ​​deviate from the actual values.

[0041] S104: Obtain the target parameter's value to be measured, input the target parameter's value to be measured into a fully trained lightweight gradient boosting machine model, obtain a predicted value of the sweet spot evaluation parameter, and predict the sweet spot based on the predicted value.

[0042] It should be noted that, correspondingly, the measured values ​​of the target parameters also need to be subjected to forward and reverse normalization processing to ensure that the larger the characteristic value, the more favorable it is for fracturing. Specifically, after obtaining the measured values ​​of the target parameters, the method also includes: performing forward normalization processing on the measured values ​​of the forward parameters among the measured values ​​of the target parameters, and performing reverse normalization processing on the measured values ​​of the reverse parameters.

[0043] In a specific embodiment, the open flow parameter AOF is selected as the sweet spot evaluation parameter. The AOF reflects the maximum theoretical production capacity of the reservoir under the current development conditions. The larger the better. The AOF threshold can be set based on the balance between development costs and benefits. The sweet spot range is determined by judging whether the open flow parameter is greater than the AOF threshold.

[0044] It should also be noted that the lightweight gradient boosting machine model can evaluate a feature's contribution to the model by calculating the number of splits (Gini coefficient gain) or the number of samples covered (frequency) in a tree node. However, feature importance is only calculated after training is complete and redundant features cannot be removed before training, resulting in a waste of computational resources. Furthermore, feature importance is based on post-training gradient information, not directly on data correlation. Lightweight gradient boosting models typically use recursive feature elimination (RFE) or SHAP for feature selection, but these methods rely on data training and may be dataset-specific. Furthermore, while EFB can reduce computational complexity when processing high-dimensional data, irrelevant features still increase split complexity, impacting training speed and memory consumption. Noisy features in the dataset can also lead to model overfitting. Especially in small sample sizes, the lightweight gradient boosting machine may misrepresent noise patterns. Therefore, the present invention introduces the grey correlation analysis algorithm to construct the GRA-LightGBM model. In terms of correlation analysis, GRA measures the importance of variables by the relative change trend between data, is not affected by the unit and distribution of the data, and screens the main controlling factors based on the correlation of the data itself, making up for the deficiency that the feature importance of the lightweight gradient boosting machine model must be obtained after the training is completed. Using the grey correlation analysis algorithm for preliminary feature screening can also avoid the lightweight gradient boosting machine model from directly using all features for training, reduce redundant features, and improve the model training efficiency and generalization ability. The main controlling factors screened out by the grey correlation analysis algorithm are combined with the lightweight gradient boosting machine model for scoring, which can better verify the actual contribution of the screened main controlling factors to the prediction from the "model perspective". The present invention proposes the "GRA-LightGBM" joint framework for the first time, and customizes the input features for the lightweight gradient boosting machine model through the grey correlation analysis algorithm to form a "screening-modeling" closed loop, which is a deep fusion of geological engineering logic and machine learning.

[0045] Compared with the existing technology, the present invention targets the characteristics of tight gas reservoirs with strong heterogeneity and complex nonlinear parameter interactions. It first uses the grey correlation analysis method to screen out core target parameters with high matching degree with the target trend, and then explores the relationship between reservoir parameters and sweet spot evaluation parameters from the perspective of "trend matching". It incorporates the logic of domain analysis and solves the problems of complex relationships between reservoir parameters and uneven data distribution. It then uses a lightweight gradient boosting machine model to process the nonlinear relationships and complex interaction patterns between high-dimensional and sparse target parameters, and captures the deep laws that are difficult to characterize using the grey correlation analysis method. The combination of the two retains the analysis logic of the traditional method and gives play to the modeling advantages of machine learning, effectively considering the coupling effects between multiple parameters and achieving high-precision sweet spot prediction.

[0046] In order to better implement a method for predicting sweet spots in fracturing of a tight gas reservoir in an embodiment of the present invention, based on a method for predicting sweet spots in fracturing of a tight gas reservoir, correspondingly, as follows Figure 5 As shown, an embodiment of the present invention further provides a tight gas reservoir fracturing sweet spot prediction device 500, comprising: A data acquisition unit 501 is used to acquire historical values ​​of reservoir parameters and sweet spot evaluation parameters; a parameter screening unit 502 for calculating a difference matrix between each reservoir parameter and the sweet spot evaluation parameter based on historical values ​​of the reservoir parameter and the sweet spot evaluation parameter, calculating a correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix, and selecting a target parameter from the reservoir parameters whose correlation is greater than a threshold; A model training unit 503 is used to train the initial lightweight gradient boosting machine model based on the target parameters and the sweet spot evaluation parameters to obtain a fully trained lightweight gradient boosting machine model; The model prediction unit 504 is used to obtain the target parameter's value to be measured, input the target parameter's value to be measured into a fully trained lightweight gradient boosting machine model, obtain the predicted value of the dessert evaluation parameter, and predict the dessert based on the predicted value.

[0047] The tight gas reservoir fracturing sweet spot prediction device 500 provided in the above embodiment can implement the technical solution described in the above embodiment of the tight gas reservoir fracturing sweet spot prediction method. The specific implementation principles of the above units can refer to the corresponding contents in the above embodiment of the tight gas reservoir fracturing sweet spot prediction method, which will not be repeated here.

[0048] like Figure 6 As shown, the present invention also provides an electronic device 600. The electronic device 600 includes a processor 601, a memory 602 and a display 603. Figure 6 Only some of the components of the electronic device 600 are shown, but it should be understood that implementation of all of the shown components is not required, and more or fewer components may be implemented instead.

[0049] In some embodiments, the memory 602 may be an internal storage unit of the electronic device 600, such as a hard disk or memory of the electronic device 600. In other embodiments, the memory 602 may also be an external storage device of the electronic device 600, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 600.

[0050] Furthermore, the memory 602 may include both an internal storage unit of the electronic device 600 and an external storage device. The memory 602 is used to store application software installed in the electronic device 600 and various data.

[0051] In some embodiments, the processor 601 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program codes or process data stored in the memory 602 , such as the method for predicting sweet spots in tight gas reservoir fracturing according to the present invention.

[0052] In some embodiments, the display 603 can be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 603 is used to display information on the electronic device 600 and to display a visual user interface. Components 601-603 of the electronic device 600 communicate with each other via a system bus.

[0053] In some embodiments of the present invention, when the processor 601 executes the tight gas reservoir fracturing sweet spot prediction program in the memory 602, the following steps may be implemented: Obtain historical values ​​of reservoir parameters and sweet spot evaluation parameters; Calculating a difference matrix between each reservoir parameter and the sweet spot evaluation parameter based on historical values ​​of the reservoir parameter and the sweet spot evaluation parameter, calculating a correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix, and selecting a target parameter with a correlation greater than a threshold value from the reservoir parameters; The initial lightweight gradient boosting machine model is trained based on the target parameters and the sweet spot evaluation parameters to obtain a fully trained lightweight gradient boosting machine model; Obtain the target parameter's measured value, input the target parameter's measured value into a fully trained lightweight gradient boosting machine model, obtain the predicted value of the sweet spot evaluation parameter, and predict the sweet spot based on the predicted value.

[0054] It should be understood that, when the processor 601 executes the tight gas reservoir fracturing sweet spot prediction program in the memory 602 , in addition to the above functions, it can also implement other functions. For details, please refer to the description of the corresponding method embodiment above.

[0055] Furthermore, the embodiment of the present invention does not specifically limit the type of the electronic device 600 mentioned. The electronic device 600 may be a portable electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, or the like. Exemplary embodiments of portable electronic devices include but are not limited to portable electronic devices equipped with IOS, Android, Microsoft, or other operating systems. The above-mentioned portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 600 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0056] Accordingly, an embodiment of the present invention also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions of the tight gas reservoir fracturing sweet spot prediction method provided by the above-mentioned method embodiments can be implemented.

[0057] Those skilled in the art will appreciate that all or part of the process steps of the above-described embodiments can be implemented by instructing related hardware through a computer program, and the above-described program can be stored in a computer-readable storage medium, such as a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0058] The above is a detailed introduction to a method for predicting sweet spots in tight gas reservoir fracturing provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as limiting the present invention.

[0059] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for predicting sweet spots in tight gas reservoir fracturing, characterized in that: include: Obtain historical values ​​of reservoir parameters and sweet spot evaluation parameters; Calculating a difference matrix between each reservoir parameter and the sweet spot evaluation parameter based on historical values ​​of the reservoir parameter and the sweet spot evaluation parameter, calculating a correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix, and selecting a target parameter having the correlation greater than a threshold value from the reservoir parameters; Training an initial lightweight gradient boosting machine model based on historical values ​​of the target parameter and the sweet spot evaluation parameter to obtain a fully trained lightweight gradient boosting machine model; Obtain a target parameter's value to be measured, input the target parameter's value to be measured into the fully trained lightweight gradient boosting machine model, obtain a predicted value of a sweet spot evaluation parameter, and predict a sweet spot based on the predicted value.

2. The method for predicting sweet spots in tight gas reservoir fracturing according to claim 1, characterized in that: After obtaining the historical values ​​of the reservoir parameters and the sweet spot evaluation parameters, the method further includes: Performing forward normalization processing on historical values ​​of the forward parameters among the historical values ​​of the reservoir parameters and the sweet spot evaluation parameters, and performing reverse normalization processing on historical values ​​of the reverse parameters; After obtaining the value of the target parameter to be measured, the method further includes: The forward normalization processing is performed on the forward parameter to be measured values ​​of the target parameter to be measured, and the reverse normalization processing is performed on the reverse parameter to be measured values.

3. The method for predicting sweet spots in tight gas reservoir fracturing according to claim 1, characterized in that: The reservoir parameters include geological parameters and engineering parameters. The geological parameters include porosity, permeability, gas saturation, mud content and total hydrocarbon average value. The engineering parameters include brittleness index, fracture toughness, Poisson's ratio, Young's modulus and compressive strength.

4. The method for predicting sweet spots in tight gas reservoir fracturing according to claim 1, characterized in that: The sweet spot evaluation parameter is the unobstructed flow parameter.

5. The method for predicting sweet spots in tight gas reservoir fracturing according to claim 1, characterized in that: Calculating the correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix includes: Selecting a global minimum difference and a global maximum difference from the difference matrix; Calculating the correlation coefficient between each reservoir parameter and the sweet spot evaluation parameter at each moment based on the global minimum difference and the global maximum difference; The correlation coefficients at each moment are averaged to obtain the correlation degree between each reservoir parameter and the sweet spot evaluation parameter.

6. The method for predicting sweet spots in tight gas reservoir fracturing according to claim 1, characterized in that: The initial lightweight gradient boosting machine model training process adopts grid search and 3-fold cross validation method.

7. The method for predicting sweet spots in tight gas reservoir fracturing according to claim 1, characterized in that: The early stopping mechanism is used during the training of the initial lightweight gradient boosting machine model.

8. A device for predicting sweet spots in tight gas reservoir fracturing, characterized in that: include: A data acquisition unit, used for acquiring historical values ​​of reservoir parameters and sweet spot evaluation parameters; a parameter screening unit, configured to calculate a difference matrix between each reservoir parameter and the sweet spot evaluation parameter based on historical values ​​of the reservoir parameter and the sweet spot evaluation parameter, calculate a correlation between each reservoir parameter and the sweet spot evaluation parameter based on the difference matrix, and select a target parameter having the correlation greater than a threshold value from the reservoir parameters; A model training unit, configured to train an initial lightweight gradient boosting machine model based on the historical values ​​of the target parameter and the sweet spot evaluation parameter to obtain a fully trained lightweight gradient boosting machine model; A model prediction unit is used to obtain a target parameter value to be measured, input the target parameter value to be measured into the trained lightweight gradient boosting machine model, obtain a predicted value of a sweet spot evaluation parameter, and predict the sweet spot based on the predicted value.

9. An electronic device, characterized in that: including memory and processor; The memory is used to store programs; The processor is coupled to the memory and is configured to execute the program stored in the memory to implement the steps of the method for predicting fracturing sweet spots in tight gas reservoirs as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The readable storage medium stores a program or instruction, and when the program or instruction is executed by a processor, the steps of the method for predicting the fracturing sweet spot of a tight gas reservoir according to any one of claims 1 to 7 are implemented.

Images