Horizontal well production and profile prediction method and system integrating well logging and numerical reservoir simulation
Through the method of fusion logging and reservoir numerical simulation, the random forest algorithm is used to predict the liquid production profile intensity of horizontal wellbores, which solves the problems of high-cost and high-precision measurement in the existing technology, and achieves a lower-cost accurate prediction effect.
Patent Information
- Application Number
- CN202411048930.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-11-26
AI Technical Summary
In the prior art, although the measurement method of the horizontal well production profile is accurate, it is costly and it is difficult to achieve accurate prediction at low cost.
A method of fusion logging and reservoir numerical simulation is used to select and pretreat the logging curve data, and combine a single-well geological model to construct a classification model based on random forests to predict the fluid production profile intensity type of horizontal wellbore wellbore.
It effectively reduces costs, improves the prediction accuracy of the liquid production profile, and improves the prediction effect of the machine learning model through feature selection and single-well numerical simulation model.
Smart Images

Figure CN119026028A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of horizontal wellbore fluid production profiles, and in particular to a method and system for predicting horizontal well production profiles that integrates well logging and reservoir numerical simulation. Background Art
[0002] Horizontal wells usually provide higher production capacity than vertical wells due to their large contact area with the reservoir, so they have become one of the mainstream drilling methods for oil field development. However, it is precisely because of the large contact area between horizontal wells and the reservoir that the bottom hole flow will be more complicated, and the uneven fluid production profile caused by reservoir heterogeneity may be more significant. The horizontal well production profile plays a vital role in determining the flow direction of reservoir fluid, clarifying the distribution of remaining oil underground and the development of dominant seepage channels. At present, the main methods for measuring the production profile include PLT (Production Logging Tool) testing, distributed fiber optic logging, etc. These methods are highly accurate, but the corresponding cost is high.
[0003] In view of this, this application is hereby filed. Summary of the invention
[0004] The purpose of the present invention is to provide a method and system for predicting the production profile of a horizontal well that integrates well logging and reservoir numerical simulation, so as to predict the production profile of a horizontal well bore through well logging curve data. Based on the well logging curve, the invention introduces the calculation results of the production profile of a single-well reservoir numerical model as the input parameter of the machine learning model, which effectively improves the prediction effect of the production profile. The invention selects features from the logging curve, integrates feature engineering of reservoir numerical simulation, and optimizes the hyperparameters of the machine learning model, and establishes a process for predicting the intensity type of the production profile of a horizontal well bore based on the well logging curve based on the random forest algorithm. The present invention effectively utilizes the actual measured data of the production profile on site and the more common logging curve data, and provides a method for predicting the production profile at a relatively low cost that is suitable for the field.
[0005] The present invention is achieved through the following technical solutions:
[0006] In a first aspect, the present invention provides a method for predicting horizontal well production profile by integrating well logging and reservoir numerical simulation, the method comprising:
[0007] Acquire measured logging data, which includes several logging curves and their fluid production profile curves; pre-process the logging data to obtain a fluid production intensity label at a certain position of the wellbore in the pre-processed fluid production profile curve;
[0008] The well logging curves are sorted and screened by using the sorting importance method to obtain the screened well logging curves;
[0009] For any well, a single well geological model is constructed according to the relevant parameters of the corresponding logging curve; and based on the single well geological model, the production of fixed liquid production at the wellhead is simulated to obtain the theoretical liquid production of the single well;
[0010] The theoretical liquid production of a single well and its corresponding screened logging curve are taken as input features, and the liquid production intensity label is taken as the target value to construct a classification model based on random forest. The classification model based on random forest is used for classification to obtain the predicted value of liquid production profile.
[0011] Furthermore, the well logging curve refers to the well logging curve of each horizontal production section of the oil wells in the oil field;
[0012] The fluid production profile curve refers to the fluid production profile curve obtained through PLT testing or optical fiber logging.
[0013] Furthermore, the logging data is preprocessed to obtain a label of the fluid production intensity at a certain position of the wellbore in the preprocessed fluid production profile curve, including:
[0014] S11, denoted as oil field N well The N of each horizontal liquid production section of the oil wells log The well logging curve is L i,j The production profile curve obtained by PLT testing or optical fiber logging is P i , where i is the oil well number, i=1,2,…,N well , j is the logging curve type number;
[0015] S12, based on the linear interpolation of the borehole depth, the arbitrary logging curve L i,j , Liquid production profile curve P i Unified into a vector of preset elements;
[0016] S13, for any logging curve L i,j Perform normalization processing to obtain the normalized logging curve And for any liquid production profile curve P i Perform dimensionless processing to obtain the liquid production profile curve after dimensionless processing
[0017] S14, for any i and k, determine The dimensionless liquid production P at a certain position k in the wellbore i,k If the value is 0, obtain the label B of whether liquid is produced at this position i,k , if B i,k =1, it means there is liquid production; if B i,k =0, it means no produced fluid;
[0018] S15, for any i and k, according to The dimensionless liquid production P at a certain position k in the wellbore i,k The size of the liquid is classified to obtain the liquid production intensity label C at that position i,k .
[0019] Furthermore, the ranking importance method is used to sort and screen the well logging curves to obtain the screened well logging curves, including:
[0020] S21, based on the logging curve L at a certain position k in the wellbore i,j,k And liquid label B i,k Train a logistic regression classifier and record the baseline accuracy on the validation set;
[0021] S22, among the multiple logging curves, select any one logging curve and randomly shuffle it, and keep the other logging curves unchanged; train a logistic regression classifier based on the shuffled data set, and record the first accuracy on the same validation set as step S11, and calculate the difference between the first accuracy and the baseline accuracy. If the difference is larger, it means that the importance of the logging curve is higher;
[0022] S23, repeating step S22 until all logging curves are traversed, removing the logging curves whose difference is less than a preset value (ie removing the logging curves with low importance), and obtaining the screened logging curves.
[0023] Furthermore, the well logging curves are sorted and screened by using the sorting importance method to obtain the screened well logging curves, which also includes:
[0024] S24, retaining the porosity and permeability parameters in the well logging curves obtained and screened in step S23;
[0025] S25, the number of retained logging curve features is N l ′ og , the jth logging curve of any i-th well is
[0026] Furthermore, for any well, a single well geological model is constructed according to the relevant parameters of the well logging curve; and based on the single well geological model, the production of a fixed liquid production at the wellhead is simulated to obtain the theoretical liquid production of the single well, including:
[0027] S31, for any i-th well, construct a single well geological model through the relevant parameters of the well logging curve; the relevant parameters of the well logging curve include permeability, porosity, water saturation, wellbore trajectory, fluid PVT data, and horizontal section oil layer top and bottom structure data;
[0028] S32, based on the single well geological model, simulate the production of fixed liquid production at the wellhead, obtain the liquid production profile along the horizontal section of the wellbore, and interpolate the liquid production profile along the horizontal section of the wellbore to obtain the liquid production P at any i-th well k position i ′ ,k ;
[0029] S33, for any liquid production rate P i ′ ,k Perform dimensionless processing to obtain the theoretical liquid production of a single well after dimensionless processing
[0030] In a second aspect, the present invention further provides a horizontal well production profile prediction system integrating well logging and reservoir numerical simulation. The system uses the above-mentioned horizontal well production profile prediction method integrating well logging and reservoir numerical simulation. The system comprises:
[0031] An acquisition unit is used to acquire measured well logging data, the well logging data including a number of well logging curves and their fluid production profile curves;
[0032] A preprocessing unit, used to preprocess the logging data to obtain a label of the fluid production intensity at a certain position of the wellbore in the preprocessed fluid production profile curve;
[0033] A sorting and screening unit, used to sort and screen the well logging curves by using a sorting importance method to obtain screened well logging curves;
[0034] Theoretical value calculation unit is used to construct a single well geological model for any well according to the relevant parameters of the corresponding logging curve; and simulate the production of fixed liquid production at the wellhead based on the single well geological model to obtain the theoretical liquid production of the single well;
[0035] The prediction value calculation unit is used to take the theoretical liquid production of a single well and its corresponding screened logging curve as input features, take the liquid production intensity label as the target value, and build a classification model based on random forest; the classification model based on random forest is used for classification to obtain the prediction value of the liquid production profile.
[0036] Furthermore, the well logging curve refers to the well logging curve of each horizontal production section of the oil wells in the oil field;
[0037] The fluid production profile curve refers to the fluid production profile curve obtained through PLT testing or optical fiber logging.
[0038] Furthermore, the theoretical value calculation unit includes:
[0039] The single well geological model construction subunit is used to construct a single well geological model for any i-th well through the relevant parameters of the well logging curve; the relevant parameters of the well logging curve include permeability, porosity, water saturation, wellbore trajectory, fluid PVT data, and horizontal section oil layer top and bottom structure data;
[0040] Theoretical liquid production calculation subunit is used to simulate the production of fixed liquid production at the wellhead based on the single well geological model, obtain the liquid production profile along the horizontal section of the wellbore, and interpolate the liquid production profile along the horizontal section of the wellbore to obtain the liquid production P at any i-th well k position i ′ ,k ;
[0041] Theoretical liquid production processing subunit is used to process any liquid production P i ′ ,k Perform dimensionless processing to obtain the theoretical liquid production of a single well after dimensionless processing
[0042] In a third aspect, the present invention further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned horizontal well production profile prediction method integrating well logging and reservoir numerical simulation.
[0043] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0044] The present invention integrates well logging and reservoir numerical simulation to predict the horizontal well production profile method and system. The present invention realizes the prediction of the strength type of the horizontal wellbore production profile using the well logging curve, which reduces the cost. The well logging curve is selected by ranking importance method, and the well logging curve that is invalid or interferes with the prediction of the production profile is excluded; a single well numerical simulation model is also constructed, through which the theoretical production profile of the horizontal wellbore is calculated, and it is input into the classification model based on random forest as a feature together with the well logging curve, which effectively improves the prediction accuracy of the machine learning model. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:
[0046] Figure 1 This is a flow chart of a method for predicting production profile of a horizontal well by integrating well logging and reservoir numerical simulation according to the present invention;
[0047] Figure 2 This is a schematic diagram of the DTS logging interpretation of the production profile results of a certain well of the present invention;
[0048] Figure 3A schematic diagram of a well logging curve for a horizontal section of a well of the present invention;
[0049] Figure 4 This is a heat map of feature correlation coefficients before feature selection in the present invention;
[0050] Figure 5 This is a feature distribution diagram after feature selection of the present invention;
[0051] Figure 6 It is a schematic diagram of a single well geological model of the present invention;
[0052] Figure 7 It is a schematic diagram of the numerical simulation calculation results of a single well of the present invention;
[0053] Figure 8 Schematic diagram of classification accuracy of the model of the present invention; a. The theoretical liquid production profile characteristics are not introduced; b. The theoretical liquid production profile characteristics are introduced;
[0054] Fig. 9 It is a schematic diagram of the strength classification of the liquid production profile of the present invention;
[0055] Fig.10 This is a structural block diagram of the horizontal well production profile prediction system that integrates well logging and reservoir numerical simulation in the present invention. DETAILED DESCRIPTION
[0056] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0057] At present, the main methods for measuring production profiles include PLT (Production Logging Tool) testing, distributed fiber optic logging, etc. These methods have high accuracy but are correspondingly more expensive.
[0058] For oil fields that have accumulated a large amount of production profile data, machine learning technology provides a new way to predict production profiles based on well logging curves. Well logging curves are standard operating data for oil fields and are easy to obtain. By constructing a mapping relationship between well logging curves and production profiles, it can not only serve as a supplement to traditional measurement technology, but also help to deeply understand the relationship between the static geological characteristics reflected by the well logging curves and production dynamics, thereby optimizing subsequent drilling locations and oil and gas field development strategies.
[0059] The present invention designs a method and system for predicting the production profile of a horizontal well that integrates well logging and reservoir numerical simulation, and realizes the prediction of the production profile of a horizontal well bore through well logging curve data. On the basis of the well logging curve, the invention introduces the calculation results of the production profile of a single-well reservoir numerical model as the input parameter of the machine learning model, which effectively improves the prediction effect of the production profile. The invention selects the features of the logging curve, integrates the feature engineering of the reservoir numerical simulation, and optimizes the hyperparameters of the machine learning model, and establishes a process for predicting the intensity type of the production profile of a horizontal well bore based on the well logging curve based on the random forest algorithm. The present invention effectively utilizes the actual measured data of the production profile on site and the more common logging curve data, and provides a method for predicting the production profile at a lower cost that is suitable for the field.
[0060] The present invention realizes the prediction of the strength type of the horizontal wellbore fluid production profile using the logging curve. By combining the logging curve and parameters such as fluid properties, a single well numerical simulation model is constructed, and the calculated horizontal wellbore fluid production profile is used as the input feature of the classification model based on random forest, which effectively improves the prediction accuracy of the machine learning model.
[0061] Example 1
[0062] like Figure 1 As shown, the horizontal well production profile prediction method of the present invention integrates well logging and reservoir numerical simulation, and the method comprises:
[0063] S1: Acquire measured logging data, which includes several logging curves and their fluid production profile curves; pre-process the logging data to obtain a fluid production intensity label at a certain position of the wellbore in the pre-processed fluid production profile curve;
[0064] S2: sorting and screening the well logging curves using the sorting importance method to obtain screened well logging curves;
[0065] S3: For any well, a single well geological model is constructed according to relevant parameters of the corresponding well logging curve; and based on the single well geological model, the production of a fixed liquid production at the wellhead is simulated to obtain the theoretical liquid production of the single well;
[0066] S4: The theoretical liquid production of a single well and its corresponding screened logging curve are used as input features, and the liquid production intensity label is used as the target value to construct a classification model based on random forest; the classification model based on random forest is used for classification to obtain the predicted value of the liquid production profile.
[0067] As a further implementation, the well logging curve refers to the well logging curve of each horizontal production section of the oil wells in the oil field;
[0068] The fluid production profile curve refers to the fluid production profile curve obtained through PLT testing or optical fiber logging.
[0069] As a further implementation, the logging data is preprocessed to obtain the fluid production intensity label of a certain position of the wellbore in the preprocessed fluid production profile curve, including:
[0070] S11, denoted as oil field N well The N of each horizontal liquid production section of the oil wells log The well logging curve is L i,j The production profile curve obtained by PLT testing or optical fiber logging is P i , where i is the oil well number, i=1,2,…,N well , j is the logging curve type number;
[0071] S12, based on the linear interpolation of the borehole depth, the arbitrary logging curve L i,j , Liquid production profile curve P i Unified into a vector of preset elements;
[0072] S13, for any logging curve L i,j Perform normalization processing to obtain the normalized logging curve And for any liquid production profile curve P i Perform dimensionless processing to obtain the liquid production profile curve after dimensionless processing
[0073] S14, for any i and k, determine The dimensionless liquid production P at a certain position k in the wellbore i,k If the value is 0, obtain the label B of whether liquid is produced at this position i,k , if B i,k =1, it means there is liquid production; if B i,k =0, it means no produced fluid;
[0074] S15, for any i and k, according to The dimensionless liquid production P at a certain position k in the wellbore i,k The size of the liquid is classified to obtain the liquid production intensity label C at that position i,k .
[0075] As a further implementation, the ranking importance method is used to sort and screen the well logging curves to obtain the screened well logging curves, including:
[0076] S21, based on the logging curve L at a certain position k in the wellbore i,j,k And liquid label B i,k Train a logistic regression classifier and record the baseline accuracy on the validation set;
[0077] S22, among the multiple logging curves, select any one logging curve and randomly shuffle it, and keep the other logging curves unchanged; train a logistic regression classifier based on the shuffled data set, and record the first accuracy on the same validation set as step S11, and calculate the difference between the first accuracy and the baseline accuracy. If the difference is larger, it means that the importance of the logging curve is higher;
[0078] S23, repeating step S22 until all logging curves are traversed, removing the logging curves whose difference is less than a preset value (ie removing the logging curves with low importance), and obtaining the screened logging curves.
[0079] As a further implementation, the well logging curves are sorted and screened using a sorting importance method to obtain screened well logging curves, and the method also includes:
[0080] S24, retaining the porosity and permeability parameters in the well logging curves obtained and screened in step S23;
[0081] S25, the number of retained logging curve features is N l ′ og , the jth logging curve of any i-th well is
[0082] As a further implementation, for any well, a single well geological model is constructed according to the relevant parameters of the well logging curve; and based on the single well geological model, the production of a fixed liquid production at the wellhead is simulated to obtain the theoretical liquid production of the single well, including:
[0083] S31, for any i-th well, construct a single well geological model through the relevant parameters of the well logging curve; the relevant parameters of the well logging curve include permeability, porosity, water saturation, wellbore trajectory, fluid PVT data, and horizontal section oil layer top and bottom structure data;
[0084] S32, based on the single well geological model, simulate the production of fixed liquid production at the wellhead, obtain the liquid production profile along the horizontal section of the wellbore, and interpolate the liquid production profile along the horizontal section of the wellbore to obtain the liquid production P at any i-th well k position i ′ ,k ;
[0085] S33, for any liquid production rate P i ′ ,k Perform dimensionless processing to obtain the theoretical liquid production of a single well after dimensionless processing
[0086] As a further implementation, in step S4, the theoretical liquid production of a single well is And its corresponding selected logging curve As input features, the production intensity label C i,k As the target value, classification is performed through a random forest-based classification model.
[0087] Specifically, the training steps of the random forest-based classification model are:
[0088] (1) Divide the dataset into a training set and a validation set;
[0089] (2) Based on the grid search method, the depth, number of decision trees, and minimum number of node samples of the random forest-based classification model were adjusted;
[0090] (3) retain the hyperparameters of the classification model with the highest accuracy on the validation set;
[0091] (4) Using the optimal hyperparameters, use all the data in the training set and validation set as training data for training.
[0092] This embodiment is implemented by taking an open hole horizontal oil well in an oil field in the Middle East as an example, and the specific implementation is as follows:
[0093] First, data preprocessing and feature selection
[0094] A total of 14 oil wells in the oil field have been subjected to distributed temperature sensing (DTS) logging interpretation. The logging curves are as follows: Figure 2 As shown, the DTS liquid production profile is as follows Figure 3 shown. Figures 2 to 9 The corresponding Chinese meanings of the English words are as follows:
[0095] MD (Measured Depth): measured depth;
[0096] TVD (True Vertical Depth): True vertical depth;
[0097] Schematic: schematic diagram;
[0098] Formation: stratum;
[0099] DTS (Distributed Temperature Sensing) Traces: Distributed temperature sensing curves;
[0100] Q oil: oil flow;
[0101] CALS (Caliper Log): Caliper logging;
[0102] FC13Four-Arm Caliper Log: Four-arm caliper logging;
[0103] GR (Gamma Ray Log): Gamma ray logging;
[0104] MSFL (Micro Spherically Focused Log): Micro spherically focused resistivity logging;
[0105] PERM_CA (Permeability, Core Analysis): permeability (core analysis);
[0106] POR (Porosity): Porosity;
[0107] POTA (Potassium Log): Potassium logging;
[0108] SP (Spontaneous Potential Log): natural potential logging;
[0109] SW_AR (Water Saturation, Archie): water saturation (Archie formula);
[0110] THOR (Thorium Log): Thorium well logging;
[0111] URAN (Uranium Log): Uranium logging;
[0112] VSH (Volume of Shale): shale content;
[0113] Oil Rate: oil flow;
[0114] Liquid Rate: Liquid flow rate;
[0115] Watercut: moisture content;
[0116] Oil Productivity Index: Oil production index;
[0117] Effective R0: effective well diameter;
[0118] Effective Kh: effective permeability thickness product (Kh);
[0119] Drawdown: pressure drop;
[0120] Porosity: porosity;
[0121] Permeability: permeability;
[0122] Initial Saturation: initial saturation;
[0123] Iterations: number of iterations;
[0124] Test Accuracy: Test accuracy.
[0125] According to step S1, the well logging curve and the fluid production profile obtained by DTS interpretation are processed. Figure 4 As shown in the figure, there are 20 well logging curves, and the collinearity between the features is not strong. Some of the well logging curves have abnormal data, which are removed. The production profile is divided into 4 categories: 1, 2, 3, and 4. The larger the value, the higher the production intensity.
[0126] According to step S2, five well logging curves are retained by screening through the permutation importance method, such as Figure 5 shown.
[0127] Second, calculate the theoretical profile through the single well numerical simulation model (i.e. single well geological model)
[0128] Execute step S3. Specifically, the geological model of each well is established based on the logging porosity, permeability, PVT, water saturation and other parameters, such as Figure 6 Based on the black oil model, according to the actual production system, given the wellhead liquid production, the theoretical liquid production profile is obtained, as shown in Figure 7 shown.
[0129] Third, the classification model based on random forest
[0130] Execute step S4. Set the parameter search range: depth 3 to 10, number of decision trees 16 to 128, minimum number of node samples 4 to 32. Finally, the optimal parameters are 5, 45, and 7 respectively.
[0131] Under the optimal parameters, the model accuracy comparison between removing the theoretical liquid production profile and retaining the theoretical liquid production profile is shown in the figure below: Figure 8 shown.
[0132] The intensity classification of the liquid production profile obtained using the optimal model is as follows: Fig. 9 shown.
[0133] The present invention realizes the prediction of the intensity type of the horizontal wellbore fluid production profile by using the well logging curve, thereby reducing the cost. The well logging curve is feature selected by using the ranking importance method, and the well logging curves that are invalid or interfere with the prediction of the fluid production profile are excluded. A single well numerical simulation model is also constructed, through which the theoretical fluid production profile of the horizontal wellbore is calculated, and it is input into the classification model based on random forest as a feature together with the well logging curve, which effectively improves the prediction accuracy of the machine learning model.
[0134] Example 2
[0135] like Fig.10 As shown, the difference between this embodiment and embodiment 1 is that this embodiment provides a horizontal well production profile prediction system integrating well logging and reservoir numerical simulation, which is a horizontal well production profile prediction method integrating well logging and reservoir numerical simulation using embodiment 1, and the system corresponds one-to-one to the horizontal well production profile prediction method integrating well logging and reservoir numerical simulation in embodiment 1; the system includes:
[0136] An acquisition unit is used to acquire measured well logging data, the well logging data including a number of well logging curves and their fluid production profile curves;
[0137] A preprocessing unit, used to preprocess the logging data to obtain a label of the fluid production intensity at a certain position of the wellbore in the preprocessed fluid production profile curve;
[0138] A sorting and screening unit, used to sort and screen the well logging curves by using a sorting importance method to obtain screened well logging curves;
[0139] Theoretical value calculation unit is used to construct a single well geological model for any well according to the relevant parameters of the corresponding logging curve; and simulate the production of fixed liquid production at the wellhead based on the single well geological model to obtain the theoretical liquid production of the single well;
[0140] The prediction value calculation unit is used to take the theoretical liquid production of a single well and its corresponding screened logging curve as input features, take the liquid production intensity label as the target value, and build a classification model based on random forest; the classification model based on random forest is used for classification to obtain the prediction value of the liquid production profile.
[0141] As a further implementation, the well logging curve refers to the well logging curve of each horizontal production section of the oil wells in the oil field;
[0142] The fluid production profile curve refers to the fluid production profile curve obtained through PLT testing or optical fiber logging.
[0143] As a further implementation, the theoretical value calculation unit includes:
[0144] The single well geological model construction subunit is used to construct a single well geological model for any i-th well through the relevant parameters of the well logging curve; the relevant parameters of the well logging curve include permeability, porosity, water saturation, wellbore trajectory, fluid PVT data, and horizontal section oil layer top and bottom structure data;
[0145] Theoretical liquid production calculation subunit is used to simulate the production of fixed liquid production at the wellhead based on the single well geological model, obtain the liquid production profile along the horizontal section of the wellbore, and interpolate the liquid production profile along the horizontal section of the wellbore to obtain the liquid production P at any i-th well k position i ′ ,k ;
[0146] Theoretical liquid production processing subunit is used to process any liquid production P i ′ ,k Perform dimensionless processing to obtain the theoretical liquid production of a single well after dimensionless processing
[0147] Among them, the execution process of each unit can be executed according to the process steps of the horizontal well production profile prediction method integrating well logging and reservoir numerical simulation in Example 1, and will not be repeated one by one in this embodiment.
[0148] At the same time, the present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the horizontal well production profile prediction method integrating well logging and reservoir numerical simulation is implemented.
[0149] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0150] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0151] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0152] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0153] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A horizontal well production profile prediction method integrating well logging and reservoir numerical simulation, characterized in that: The method includes: Acquire measured logging data, the logging data including a plurality of logging curves and a fluid production profile curve thereof; preprocess the logging data to obtain a fluid production intensity label of a certain position of the wellbore in the preprocessed fluid production profile curve; sorting and screening the well logging curves by using a sorting importance method to obtain screened well logging curves; For any well, a single well geological model is constructed according to relevant parameters of the corresponding well logging curve; and based on the single well geological model, the production of a fixed liquid production at the wellhead is simulated to obtain the theoretical liquid production of the single well; The theoretical liquid production of a single well and its corresponding screened logging curve are taken as input features, and the liquid production intensity label is taken as the target value to construct a classification model based on random forest. The classification model based on random forest is used for classification to obtain the predicted value of liquid production profile.
2. The method for predicting horizontal well production profile by integrating well logging and reservoir numerical simulation according to claim 1, characterized in that: The well logging curve refers to the well logging curve of each horizontal liquid production section of the oil wells in the oil field; The fluid production profile curve refers to the fluid production profile curve obtained through PLT testing or optical fiber logging.
3. The method for predicting horizontal well production profile by integrating well logging and reservoir numerical simulation according to claim 1, characterized in that: Preprocessing the logging data to obtain a label of the fluid production intensity at a certain position of the wellbore in the preprocessed fluid production profile curve includes: S11, denoted as oil field N well The N of each horizontal liquid production section of the oil wells log The well logging curve is L i,j The fluid production profile curve obtained by PLT testing or optical fiber logging is P i , where i is the oil well number, i=1,2,…,N well , j is the logging curve type number; S12, based on the linear interpolation of the borehole depth, the arbitrary logging curve L i,j , Liquid production profile curve P i Unified into a vector of preset elements; S13, for any logging curve L i,j Perform normalization processing to obtain the normalized logging curve And for any liquid production profile curve P i Perform dimensionless processing to obtain the liquid production profile curve after dimensionless processing S14, for any i and k, determine The dimensionless liquid production P at a certain position k in the wellbore i,k If the value is 0, obtain the liquid production label B indicating whether liquid is produced at this position i,k , if B i,k =1, it means there is liquid production; if B i,k =0, it means no produced fluid; S15, for any i and k, according to The dimensionless liquid production P at a certain position k in the wellbore i,k The size of the liquid is classified to obtain the liquid production intensity label C at that position i,k .
4. The method for predicting horizontal well production profile by integrating well logging and reservoir numerical simulation according to claim 1, characterized in that: The well logging curves are sorted and screened using a sorting importance method to obtain screened well logging curves, including: S21, based on the logging curve L at a certain position k in the wellbore i,j,k And liquid label B i,k Train a logistic regression classifier and record the baseline accuracy on the validation set; S22, among the multiple logging curves, select any one logging curve and randomly shuffle it, and keep the other logging curves unchanged; train a logistic regression classifier based on the shuffled data set, and record the first accuracy on the same validation set as step S11, and calculate the difference between the first accuracy and the baseline accuracy. If the difference is larger, it means that the importance of the logging curve is higher; S23, repeating step S22 until all logging curves are traversed, removing the logging curves whose difference is less than a preset value, and obtaining the screened logging curves.
5. The method for predicting horizontal well production profile by integrating well logging and reservoir numerical simulation according to claim 4, characterized in that: The well logging curves are sorted and screened by using a sorting importance method to obtain screened well logging curves, and the method further includes: S24, retaining the porosity and permeability parameters in the well logging curves obtained and screened in step S23; S25, the number of retained logging curve features is N l ′ og , the jth logging curve of any i-th well is 6. The method for predicting horizontal well production profile by integrating well logging and reservoir numerical simulation according to claim 1, characterized in that: For any well, construct a single well geological model according to the relevant parameters of the logging curve; Based on the single well geological model, the production of fixed liquid production at the wellhead is simulated to obtain the theoretical liquid production of the single well, including: S31, for any i-th well, construct a single well geological model through relevant parameters of the well logging curve; the relevant parameters of the well logging curve include permeability, porosity, water saturation, wellbore trajectory, fluid PVT data, and horizontal section oil layer top and bottom structure data; S32, based on the single well geological model, simulate the production of fixed liquid production at the wellhead, obtain the liquid production profile along the horizontal section of the wellbore, and interpolate the liquid production profile along the horizontal section of the wellbore to obtain the liquid production P at any i-th well k position i ′ ,k ; S33, for any liquid production rate P i ′ ,k Perform dimensionless processing to obtain the theoretical liquid production of a single well after dimensionless processing 7. A horizontal well production profile prediction system integrating well logging and reservoir numerical simulation, characterized in that: The system includes: An acquisition unit, used for acquiring measured well logging data, wherein the well logging data includes a plurality of well logging curves and a fluid production profile curve thereof; A preprocessing unit, used to preprocess the logging data to obtain a fluid production intensity label at a certain position of the wellbore in the preprocessed fluid production profile curve; A sorting and screening unit, used to sort and screen the well logging curves by using a sorting importance method to obtain screened well logging curves; The theoretical value calculation unit is used to construct a single well geological model for any well according to the relevant parameters of the corresponding well logging curve; and simulate the production of a fixed liquid production at the wellhead based on the single well geological model to obtain the theoretical liquid production of the single well; The prediction value calculation unit is used to take the theoretical liquid production of a single well and its corresponding screened logging curve as input features, take the liquid production intensity label as the target value, and build a classification model based on random forest; the classification model based on random forest is used for classification to obtain the prediction value of the liquid production profile.
8. The horizontal well production profile prediction system integrating well logging and reservoir numerical simulation according to claim 7 is characterized in that: The well logging curve refers to the well logging curve of each horizontal liquid production section of the oil wells in the oil field; The fluid production profile curve refers to the fluid production profile curve obtained through PLT testing or optical fiber logging.
9. The horizontal well production profile prediction system integrating well logging and reservoir numerical simulation according to claim 7, characterized in that: The theoretical value calculation unit comprises: The single well geological model construction subunit is used to construct a single well geological model for any i-th well through the relevant parameters of the well logging curve; the relevant parameters of the well logging curve include permeability, porosity, water saturation, wellbore trajectory, fluid PVT data, and horizontal section oil layer top and bottom structure data; The theoretical liquid production calculation subunit is used to simulate the production of fixed liquid production at the wellhead based on the single well geological model, obtain the liquid production profile along the horizontal section of the wellbore, and interpolate the liquid production profile along the horizontal section of the wellbore to obtain the liquid production P at any i-th well k position i ′ ,k ; Theoretical liquid production processing subunit is used to process any liquid production P i ′ ,k Perform dimensionless processing to obtain the theoretical liquid production of a single well after dimensionless processing 10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for predicting horizontal well production profile by integrating well logging and reservoir numerical simulation as described in any one of claims 1 to 6 is implemented.
Citation Information
Cited By
Multi-well lithofacies identification method and device and medium
CN120742447A
