Intelligent analysis method and analysis equipment for post-pressure pumping stop data of shale gas well

Abstract

The invention discloses an intelligent analysis method and analysis equipment for post-pressure pumping stopping data of a shale gas well, and relates to the technical field of oil and gas development. In the method, after analysis equipment automatically extracts and integrates fracturing data from a multi-source heterogeneous document, wellhead pressure is converted into bottom hole pressure, and bottom hole pressure data is filtered based on optimal parameters obtained through automatic optimization. Then, the analysis device generates a pressure double logarithmic curve graph, automatically divides flow feature segments based on the graph, and calculates an interpretation result. Finally, the analysis equipment can also establish a fracturing parameter optimization model based on a historical interpretation result, and an optimization design scheme is provided for a subsequent well section. According to the method, an oil field engineer can complete the whole process operation from data processing to intelligent decision making in a single integrated platform, the process of repeatedly importing and exporting data and debugging parameters through multiple kinds of software for fracturing effect evaluation is simplified, and then the efficiency of fracturing effect evaluation is improved.

Description

Technical Field

[0001] This application relates to the field of oil and gas development technology, and in particular to an intelligent analysis method and analysis equipment for post-pressure pump shutdown data of shale gas wells. Background Technology

[0002] With the deepening of oil and gas resource development, shale gas development has become an important part of my country's energy sector. Shale gas is characterized by ultra-low permeability, requiring formation modification through hydraulic fracturing technology for economical development. The pressure data obtained during fracturing operations contains rich information about the formation and fractures, serving as a crucial basis for evaluating the effectiveness of fracturing.

[0003] Currently, post-fracturing pump shutdown data from shale gas wells requires processing and analysis using various specialized software programs (such as Office, Matlab, and Saphir). Specifically, firstly, effective data from the pump shutdown stage is obtained from fracturing operation files using Office. Then, the data is imported into Matlab, where the pressure data is filtered by repeatedly adjusting the filtering algorithm code. Next, the filtered data is exported and imported into Saphir to generate and initially process the double logarithmic pressure curve. Subsequently, the time, pressure drop, and pressure conductivity data processed by Saphir are exported again from Saphir and imported into another interpretation algorithm software developed based on Matlab. Different flow characteristic stages are defined on the double logarithmic pressure curve. Finally, by repeatedly setting the data and interpretation algorithm, fracturing effect evaluation data such as fracture half-length and conductivity at different flow characteristic stages are obtained.

[0004] Existing fracturing effect analysis methods require the use of different processing software to process data and data transfer between different software. When one software processes data incorrectly, other software will process incorrect data. When it is necessary to trace the source of the incorrect data, it is necessary to check whether the data processed by each software is wrong one by one, thereby reducing the analysis efficiency of shale gas well fracturing effect assessment. Summary of the Invention

[0005] This application provides an intelligent analysis method and equipment for post-fracturing pump shutdown data of shale gas wells. By intelligently processing and analyzing a large amount of post-fracturing pump shutdown data of shale gas wells, the analysis efficiency of shale gas well fracturing effect assessment can be improved.

[0006] Firstly, an intelligent analysis method for shale gas well post-fracturing pump shutdown data is provided, comprising: an analysis device extracting well inclination data, perforation parameter data, and fracturing second point data for each fracturing stage from batch-loaded shale gas well test geological design documents and fracturing construction data documents based on a preset data dictionary, wherein the preset data dictionary defines multiple representations of the same data under different document formats; the analysis device determining that the fracturing second point data for each fracturing stage is wellhead pressure data located between the pump shutdown start time and the pump shutdown end time; based on the well inclination data and the perforation parameter data, the analysis device converting the wellhead pressure data into bottom hole pressure data; and based on the filtered data... The analysis equipment generates a double logarithmic pressure curve representing the pressure drop and pressure drop derivative changes over time based on the pressure drop derivative curve in the bottom hole pressure data. The analysis equipment divides the flow characteristic segments based on the slope of the pressure drop derivative curve in the double logarithmic pressure curve. These flow characteristic segments include inter-fracture flow segments, fracture-net filtration linear flow segments, fracture-net filtration bilinear flow segments, and fracture-net closure segments. Based on preset geological engineering data and fracturing segment data, the analysis equipment calls the interpretation algorithm corresponding to each fracturing segment of the flow characteristic segments to calculate interpretation result data corresponding one-to-one with each fracturing segment. The analysis equipment summarizes the interpretation result data for the entire well section of the shale gas well to generate a fracturing effect report.

[0007] By adopting the above technical solution, the steps of extracting post-fracturing pump shutdown data, pressure conversion, curve generation, and interpretation calculation of shale gas wells are integrated into a single analysis device. This simplifies the operation process of importing and exporting data between multiple independent software programs, reduces errors that may be caused by data transmission, and facilitates the traceability of the data processing process. In this way, the analytical efficiency and data consistency of shale gas well fracturing effect evaluation are improved.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the analysis device extracts well inclination data, perforation parameter data, and fracturing time point data for each fracturing stage from batch-loaded shale gas well test geological design documents and fracturing construction data documents based on a preset data dictionary. Specifically, this includes: the analysis device batch-converting batch-loaded shale gas well test geological design documents and fracturing construction data documents of different formats into intermediate documents of a preset format; based on the preset data dictionary, named entity recognition, and keyword matching technology, the analysis device matches and extracts well inclination data, perforation parameter data, and fracturing time point data from the intermediate documents of the preset format; the analysis device establishes a correlation relationship between the extracted data according to a three-layer structure of well-fracturing stage-fracturing stage data, and stores it in a relational database in a structured form.

[0009] By adopting the above technical solution, documents of different formats are batch converted into a unified intermediate format and stored in a structured and associated manner, which simplifies the operation of sorting out multi-source heterogeneous data, transforms scattered and unstructured information into standardized data that can be directly accessed, provides a unified data foundation for subsequent fracturing effect analysis, and improves the analytical efficiency of shale gas well fracturing effect evaluation.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, the analysis device determines the wellhead pressure data of each fracturing segment's fracturing second-point data located between the pump shutdown start time and the pump shutdown end time. Specifically, this includes: the analysis device filtering out displacement data and wellhead pressure data from the fracturing second-point data of each fracturing segment, the displacement data and the wellhead pressure data being arranged in chronological order; the analysis device determining the moment when the displacement value in the displacement data changes from being greater than a preset displacement threshold for the first time to being less than or equal to the preset displacement threshold for a continuously preset duration as the pump shutdown start time; from the pump shutdown start time, after a preset number of wellhead pressure data points, the analysis device determining the moment when the pressure difference between adjacent wellhead pressure data points first exceeds or equals a preset pressure change threshold as the pump shutdown end time; the analysis device determining the wellhead pressure data of each fracturing segment's fracturing second-point data located between the pump shutdown start time and the pump shutdown end time.

[0011] By adopting the above technical solution, the analysis equipment determines the start and end points of the pump shutdown section for each fracturing stage by setting objective thresholds for displacement and pressure change rate, replacing the subjective judgment method relying on manual observation. This ensures that the data intervals extracted for each analysis have a unified standard, helping to guarantee data consistency and thus improving the accuracy and comparability of subsequent shale gas well fracturing effect analysis results.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, the step of generating a double logarithmic pressure curve representing the change of pressure drop and pressure drop derivative over time based on the pressure drop and pressure drop derivative of the filtered bottom hole pressure data specifically includes: the analysis device filtering the bottom hole pressure data based on a preset combination of preset filtering parameters of a preset filtering algorithm; and the analysis device generating a double logarithmic pressure curve representing the change of pressure drop and pressure drop derivative over time based on the pressure drop and pressure drop derivative of the filtered bottom hole pressure data.

[0013] By adopting the above technical solution, the complex process of filtering, which requires writing and modifying code in software such as Matlab, is simplified into a standardized operation of calling a preset algorithm within a unified analysis device. This simplifies the existing technical process of programming and debugging in software such as Matlab, reduces the programming skills required of operators, decreases filtering failures caused by code errors or improper parameter settings, and improves the accuracy of shale gas well fracturing effect analysis results.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, after the step of generating a double logarithmic pressure curve representing the change of pressure drop and pressure drop derivative over time based on the pressure drop and pressure drop derivative of the filtered bottom hole pressure data, the method further includes: generating an original double logarithmic pressure curve representing the change of pressure drop and pressure drop derivative over time based on the pressure drop and pressure drop derivative of the unfiltered bottom hole pressure data; the analysis device determines the degree of fit between the original double logarithmic pressure curve and the original double logarithmic pressure curve, wherein the degree of fit refers to the original... The root mean square error between the initial pressure logarithmic curve and the pressure logarithmic curve is calculated. The analysis device determines the smoothness of the pressure logarithmic curve based on the second derivative of the pressure logarithmic curve. Using the combined evaluation index of the fit and the smoothness as the target, the analysis device automatically searches for and determines the optimal combination of filtering parameters using a hyperparameter tuning algorithm, which includes at least a grid search algorithm, a random search algorithm, and a Bayesian optimization algorithm. Based on the optimal combination of filtering parameters of the preset filtering algorithm, the analysis device performs filtering processing on the bottom hole pressure data.

[0015] By adopting the above technical solution, establishing quantitative evaluation indicators for fit and smoothness, and automatically searching for optimal filtering parameters using a hyperparameter tuning algorithm, the filtering scheme is determined in a more objective way based on the noise characteristics of the data, thereby improving the calculation accuracy and quality of the pressure derivative curve.

[0016] In some embodiments, in conjunction with the first aspect, the method further includes: based on the interpretation results data of each fracturing segment in the fracturing effect report, the analysis device uses a clustering analysis method to group fracturing segments with similar characteristics into one category; the analysis device uses the geological engineering parameters and fracturing segment parameters of each category of fracturing segments as feature variables, and the interpretation results data of each category of fracturing segments as target variables to construct a training sample set; based on the correlation analysis of the feature variables in the training sample set, the analysis device removes redundant feature variables and constructs a set of necessary feature variables; the analysis device uses a multiple regression analysis method to establish a mapping relationship between the set of necessary feature variables and the target variables; the analysis device uses a cross-validation method to verify and optimize the mapping relationship to construct a fracturing parameter optimization model; and the analysis device optimizes the construction parameters of subsequent well sections to be fractured based on the fracturing parameter optimization model.

[0017] By adopting the above technical solution, through a series of machine learning steps such as cluster analysis, feature screening and regression modeling, the isolated data points originally scattered in various fracturing effect reports are transformed into fracturing parameter optimization models that can reveal the intrinsic relationship between geological engineering conditions and fracturing effects. This enables the analysis equipment to provide quantitative optimization suggestions for the design of construction parameters for subsequent fracturing well sections.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, based on the interpretation results data of each fracturing segment in the fracturing effect report, the analysis device uses a clustering analysis method to group fracturing segments with similar characteristics into one category. Specifically, this includes: the analysis device standardizing the interpretation results data of each fracturing segment; the analysis device using principal component analysis to reduce the dimensionality of the standardized interpretation results data; and based on the dimensionality-reduced feature vectors, the analysis device using a density clustering algorithm to group fracturing segments with similar characteristics into one category.

[0019] By adopting the above technical solution, the interpretation results data are standardized and subjected to principal component analysis for dimensionality reduction before cluster analysis, reducing the interference of data dimensions and noise on the clustering results. This allows for a more objective identification of the data's inherent structure, classifying fracturing segments with similar characteristics, and providing higher-quality training data for constructing fracturing parameter optimization models, thereby improving the accuracy of the fracturing parameter optimization models.

[0020] In a second aspect, embodiments of this application provide an analysis device comprising: one or more processors and a memory; the memory is coupled to the one or more processors and is used to store computer program code, the computer program code including computer instructions, wherein the one or more processors invoke the computer instructions to cause the analysis device to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on an analysis device, cause the analysis device to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on an analysis device, cause the analysis device to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the analysis device provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. Because the analysis equipment integrates the steps of extracting post-fracturing pump shutdown data, pressure conversion, curve generation, and interpretation calculation of shale gas wells into a single analysis device, it simplifies the operation process of importing and exporting data between multiple independent software programs, reduces errors that may be caused by data transmission, and facilitates the traceability of the data processing process, thereby improving the analysis efficiency and data consistency of shale gas well fracturing effect evaluation.

[0025] 2. The complex process of filtering in software like Matlab, which previously required writing and modifying code, has been simplified to a standardized operation of calling a preset algorithm within a unified analysis device. This simplifies the existing programming and debugging process in software like Matlab, reduces the programming skills required of operators, and decreases filtering failures caused by code errors or improper parameter settings, thereby improving the accuracy of shale gas well fracturing effect analysis results.

[0026] 3. Because the analysis equipment uses a series of machine learning steps such as cluster analysis, feature selection and regression modeling, it transforms the isolated data points that were originally scattered in various fracturing effect reports into fracturing parameter optimization models that can reveal the intrinsic relationship between geological engineering conditions and fracturing effects. This enables the analysis equipment to provide quantitative optimization suggestions for the design of construction parameters for subsequent fracturing well sections. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating an intelligent analysis method for post-pressure pump shutdown data of shale gas wells, as described in an embodiment of this application.

[0028] Figure 2 This is another flowchart illustrating an intelligent analysis method for post-pressure pump shutdown data of shale gas wells, as described in this application.

[0029] Figure 3 This is a schematic diagram of the physical device structure of the analysis equipment in the embodiments of this application. Detailed Implementation

[0030] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0031] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0032] This application provides an intelligent analysis method and equipment for post-fracturing pump shutdown data of shale gas wells. By intelligently processing and analyzing a large amount of post-fracturing pump shutdown data of shale gas wells, the analysis efficiency of shale gas well fracturing effect assessment can be improved.

[0033] The following describes an intelligent analysis method for shale gas well post-pressure pump shutdown data from an embodiment of this application: Please see Figure 1 This is a flowchart illustrating an intelligent analysis method for post-pressure pump shutdown data of shale gas wells in an embodiment of this application.

[0034] S101. The analysis equipment extracts well inclination data, perforation parameter data, and fracturing time point data of each fracturing section from the batch-loaded shale gas well test geological design documents and fracturing construction data documents based on a preset data dictionary.

[0035] The analysis equipment refers to the hardware entity that executes the intelligent analysis method for post-compression shutdown data of a shale gas well, as described in this application embodiment. The pre-built data dictionary is a pre-constructed knowledge base storing various synonymous or near-synonymous expressions of key parameters; for example, mapping the parameter Young's modulus to various textual expressions such as elastic modulus and rock modulus that may appear in different documents. The oil testing geological design document is an unstructured Word or PDF document that records design information such as wellbore structure, reservoir geological characteristics, and perforation scheme. The fracturing construction data document is a structured or semi-structured file (such as CSV, LAS, or a text file in a specific format) containing real-time monitoring data, recording second-level data on pressure, displacement, sand concentration, etc., changing over time during the construction process. Well inclination data represents the wellbore trajectory's position information in three-dimensional space, typically including depth, inclination angle, and azimuth. Perforation parameter data refers to the specific engineering parameters for establishing an oil and gas passage at the reservoir location, such as perforation interval, perforation depth, perforation density, phase angle, and perforation diameter. Fracturing second-point data refers to a continuous data stream recorded second by second during the fracturing operation.

[0036] Specifically, the analysis equipment first uses the open-source and cross-platform LibreOffice suite to batch convert raw documents in various formats (such as .doc, .docx, .pdf, .csv, .txt) into a unified, easily parsable intermediate format (such as XML or JSON). Then, for the intermediate-formatted oil testing geological design documents, the analysis equipment uses a pre-defined data dictionary, combined with Named Entity Recognition (NER) and keyword matching technologies, to accurately locate and extract geological engineering parameters such as Young's modulus and Poisson's ratio, as well as perforation parameters, from a large amount of descriptive text. For fracturing operation data documents, the analysis equipment directly extracts complete fracturing point data such as pressure and displacement according to pre-defined column definitions or format specifications. Finally, the analysis equipment associates all the data extracted from different documents with the well number as the highest index, following a three-layer logical structure of well-fracturing section-fracturing section parameters, and stores it in a structured relational database.

[0037] In some embodiments, parameters can be extracted from batch-loaded shale gas well testing geological design documents and fracturing construction data documents in various ways: Optionally, a template matching and rule extraction-based approach: The analysis device pre-establishes parsing templates for common document formats (such as fracturing design documents with specific layouts) of different oilfields or service companies. The templates define the location rules for key parameters (such as mid-depth) near page numbers, paragraphs, table rows and columns, or specific keywords. When documents are loaded in batches, the analysis device first performs layout analysis, automatically identifying and matching the most suitable parsing template. The analysis device then directly extracts parameter values ​​from precise locations in the document according to the rules defined in the template, achieving fast and accurate extraction from fixed-format documents. Optionally, parameters can also be extracted from batch-loaded shale gas well testing geological design documents and fracturing construction data documents adaptively based on machine learning: The analysis device uses a document parameter recognition model (such as a BERT-based sequence labeling model) trained with a large number of labeled documents (i.e., samples where parameters and their locations in the documents are manually specified). For a newly loaded document, the analysis device first inputs its text content into the model, which then outputs parameter labels (such as B-Young's modulus, I-Young's modulus) that each word or text fragment may belong to. Next, based on the label sequence output by the model, the analysis device automatically extracts the complete parameter names and corresponding values. This method has better generalization ability and robustness for unseen, non-standardized document formats.

[0038] It is understandable that other methods can be used to extract parameters from the batch-loaded shale gas well test geological design documents and fracturing construction data documents. For example, the two methods mentioned above can be combined, with template matching given priority, and machine learning models can be called to process documents that fail to match. No limitation is made here.

[0039] S102, The analysis equipment determines the wellhead pressure data of each fracturing section, which is located between the start time of pump shutdown and the end time of pump shutdown.

[0040] The pump shutdown start time refers to the precise moment at the end of a fracturing operation when the pump displacement first drops from a significantly positive value to near zero and remains low thereafter. The pump shutdown end time refers to the moment after pump shutdown when the fracture network begins to close significantly due to pressure release within the wellbore and fractures, resulting in a clear inflection point on the pressure monitoring curve. Data after this inflection point is not very meaningful for analyzing the flow characteristics before fracture closure. Wellhead pressure data is a sequence of pressure values ​​measured by pressure sensors located at the surface wellhead, recorded in the fracturing second-point data.

[0041] Specifically, the analysis equipment first retrieves the chronologically ordered displacement and wellhead pressure data for the specified fracturing section from the database. To determine the start time of pump shutdown, the analysis equipment searches in reverse chronological order, starting from the end of the displacement data sequence. When it finds that the displacement value first exceeds a preset displacement threshold (e.g., 0.1 m³ / s), it continues searching. 3 When the wellhead pressure data changes from a state where it is below a certain threshold (e.g., 0.5 MPa / s, used to filter out equipment noise or small fluctuations) to a state where it remains below or equal to that threshold for a subsequent preset duration (e.g., 30 consecutive seconds), the moment of this transition is precisely recorded as the pump shutdown start time. Next, to determine the pump shutdown end time, starting from the determined pump shutdown start time, the analysis equipment first ignores a preset number of wellhead pressure data points (to skip early unstable phases such as wellbore storage effects), and then begins calculating the pressure difference between adjacent pressure data points point by point. When this pressure difference first exceeds or equals a preset pressure change threshold (e.g., 0.5 MPa / s, marking a sudden change in the slope of the pressure curve, i.e., the start of fracture closure), the analysis equipment determines that moment as the pump shutdown end time. Finally, the analysis equipment saves all wellhead pressure data between these two time points as a single, valid analytical data segment.

[0042] In some embodiments, the precise identification of the start and end times of pump shutdown can be achieved in several ways: Optionally, an identification method based on sliding window statistics: For determining the start time of pump shutdown, the analysis device uses a fixed-size time window (e.g., 60 seconds) to slide from back to front on the displacement data, calculating the average and standard deviation of the displacement values ​​within the window. When the average value within the window is first lower than a preset displacement threshold, and the standard deviation is also lower than a very small fluctuation threshold, the right boundary of the window is determined as the start time of pump shutdown. For determining the end time of pump shutdown, the analysis device applies a sliding window to the wellhead pressure data after pump shutdown, calculating the slope of the linear regression of the pressure data within the window. When the rate of change of the slope of adjacent windows first exceeds a preset inflection point judgment threshold, that position is determined as the end time of pump shutdown. Optionally, the precise identification of the start and end times of pump shutdown can also be achieved based on signal processing and pattern recognition: The analysis device first applies a step signal detection algorithm (e.g., a one-dimensional form of Canny edge detection) to the displacement data, directly locating the most significant step point from high to low displacement, and using it as the start time of pump shutdown. The analysis equipment performs wavelet transform on the wellhead pressure data after pump shutdown, analyzing the wavelet coefficients at different scales. The inflection point of the pressure curve caused by fracture closure will appear as a significant modulus maxima on the wavelet coefficients at a specific scale. The analysis equipment determines the pump shutdown end time by detecting the location of this modulus maxima point.

[0043] It is understandable that other methods can be used to automatically extract the pump stop data segment, such as cross-validating the results of multiple algorithms to improve the accuracy of identification; this is not limited here.

[0044] S103. Based on the well inclination data and the perforation parameter data, the analysis device converts the wellhead pressure data into bottom hole pressure data.

[0045] Among them, bottom hole pressure data refers to the pressure sequence obtained through calculation that can truly reflect the pressure at the inlet of the reservoir fracturing fracture. It is the direct input for all subsequent flow characteristic analyses. Compared with wellhead pressure, bottom hole pressure eliminates the influence of hydrostatic pressure and frictional pressure drop in the wellbore, accurately restoring the easily disturbed pressure readings on the surface to the pressure underground that can represent the true response of the reservoir, providing accurate input for subsequent scientific seepage mechanics analysis.

[0046] Specifically, the analysis equipment first retrieves the wellbore inclination data and perforation parameter data corresponding to the fracturing section extracted in step S101 from the database. Using the wellbore inclination data (including depth, inclination angle, and azimuth sequence), the analysis equipment calculates the precise three-dimensional coordinates of the perforation location in the fracturing section, thereby obtaining its true vertical depth, i.e., the mid-depth. Simultaneously, the analysis equipment obtains the density parameter of the fracturing fluid inside the wellbore (this parameter can also be extracted from the well test geological design document). Subsequently, for each pressure point in the wellhead pressure data sequence extracted in step S102, the analysis equipment performs a pressure conversion formula calculation. The core of this formula is: Bottomhole pressure = Wellhead pressure + Fracturing fluid hydrostatic pressure - Wellbore friction. During the pressure drop phase after pump shutdown, since the fluid is essentially still, wellbore friction can be ignored. Therefore, the formula simplifies to: Bottomhole pressure = Wellhead pressure + Fracturing fluid density × Gravitational acceleration × Mid-depth. By performing point-by-point calculations on the entire wellhead pressure data segment, the analysis equipment generates a completely new, one-to-one corresponding bottom hole pressure data sequence.

[0047] In some embodiments, wellhead pressure data can be converted to bottomhole pressure data in several ways: Optionally, a fine conversion method based on segmented calculation: The analysis equipment divides the wellbore path from the wellhead to the perforation location into several micro-segments based on the measurement points of the well inclination data. The analysis equipment calculates the vertical height increment of each micro-segment and considers possible fluid density or temperature changes within the wellbore (if there is multiple measurement data), calculating the hydrostatic pressure generated by each micro-segment. By accumulating the hydrostatic pressures of all micro-segments, the total hydrostatic pressure value is obtained, and then added to the wellhead pressure to obtain a more accurate bottomhole pressure. This method is more accurate when the wellbore structure is complex or the fluid properties vary greatly. Optionally, a complex conversion method considering the influence of multiphase fluids: In some cases (such as the initial stage of fracturing fluid flowback), there may be gas-liquid two-phase flow within the wellbore. The analysis equipment first determines the phase state of the fluid within the wellbore based on the design documents or flowback data. If the flow is determined to be multiphase, the analysis equipment will invoke a pre-set multiphase flow wellbore pressure calculation model. This model requires input parameters such as gas-liquid ratio and fluid properties. The analysis equipment uses wellhead pressure, temperature, and fluid composition as inputs, and calculates the bottom hole pressure by solving complex momentum and energy conservation equations. This method is suitable for special operating conditions where there is rapid gas backflow after pump shutdown.

[0048] Understandably, other methods can be used to convert wellhead pressure data into bottomhole pressure data. For example, a simplified average vertical depth method can be used under standard operating conditions, while a more complex model can be started in high-requirement analyses. The analysis equipment can automatically switch according to the data completeness and user selection, which is not limited here.

[0049] S104. Based on the pressure drop and pressure drop derivative of the bottom hole pressure data after filtering, the analysis device generates a double logarithmic pressure curve characterizing the change of pressure drop and pressure drop derivative over time.

[0050] Filtering refers to applying digital signal processing algorithms to remove random, high-frequency noise from bottomhole pressure data caused by sensors, circuits, or environmental factors, while preserving the true trend and key characteristics of pressure drop. Pressure drop refers to the difference between the bottomhole pressure and a reference pressure (usually the initial pressure at the moment of pump shutdown). The pressure drop derivative is the derivative of the pressure drop with respect to a function of time; in well test analysis, it is usually the derivative of the pressure drop with respect to the natural logarithm of time. It is highly sensitive to small changes in the pressure curve and can amplify different flow characteristics. A double logarithmic pressure curve is a professional diagnostic chart; both its horizontal and vertical axes use logarithmic scales, and it plots two curves: pressure drop versus time and pressure drop derivative versus time.

[0051] Specifically, the analysis device first filters the bottom hole pressure data sequence generated in step S103. In this embodiment, the analysis device calls a preset filtering algorithm (e.g., wavelet transform) and uses preset filtering parameters to complete this operation. The preset filtering parameters are obtained based on industry experience or a large amount of historical data statistics, aiming to provide a reasonable basic filtering effect for most operating conditions. Filtering is crucial because derivative operations are extremely sensitive to noise in the data. Without filtering, the calculated derivative curve will be full of spikes, making subsequent slope analysis impossible. After obtaining a smooth bottom hole pressure curve, the analysis device calculates the corresponding pressure drop and pressure drop derivative. Finally, the analysis device creates a double logarithmic coordinate system and plots the calculated pressure drop and pressure drop derivative sequence, along with the corresponding time series, on this coordinate system, thereby generating a double logarithmic pressure curve graph that engineers can interpret.

[0052] In some embodiments, filtering and the generation of double logarithmic curves can be achieved in several ways: Optionally, a standardized processing method is used: the analysis device has a built-in Standard Operating Procedure (SOP) that specifies a unique filtering algorithm (such as a Savitzky-Golay filter) and a fixed set of parameters (e.g., a window length of 21 data points and a polynomial order of 2). When this step is executed, the analysis device does not provide any options and directly calls the algorithm and fixed parameters to process the bottom hole pressure data, ensuring high comparability of results obtained from different wells and by different analysts. After processing, the analysis device automatically calculates the pressure drop and its derivative, and generates a standard-format double logarithmic pressure curve. Optionally, an interactive processing method with manual parameter adjustment is also available: the analysis device provides a user interface with a drop-down menu for selecting a filtering algorithm (such as moving average or Gaussian filter) and several corresponding parameter adjustment sliders (such as window size and standard deviation). Users can select a filtering algorithm from the menu and then manually adjust its parameters by dragging the sliders. Whenever the user adjusts the parameters, the analysis device immediately recalculates the filter curve and pressure drop derivative curve, and refreshes the results in real time on the double logarithmic pressure graph on the screen. By observing the smoothness of the curve and its fit with the original trend, the user can manually find a satisfactory set of filter parameters based on experience, and then confirm the use of the results.

[0053] It is understandable that other methods can be used to complete the filtering process and generate the double logarithmic curve, such as using a median filter, which is particularly effective for processing data containing spike impulse noise. This is not a limitation here.

[0054] S105. The analysis device divides the flow characteristic segment based on the slope of the pressure drop derivative curve in the double logarithmic pressure curve.

[0055] The slope of the pressure drop derivative curve refers to the geometric slope of the curve on a certain segment in a logarithmic coordinate system. Specific values ​​of this slope (such as 0.25, 0.5, 1) are the theoretical basis for identifying specific seepage physical processes. The flow characteristic segment refers to a dominant flow pattern or stage that can be described by theoretical models during the pressure drop process in the fracture network and reservoir matrix.

[0056] Specifically, the analysis device initiates a curve feature recognition module. The core algorithm of this module calculates the local slope at each point on the pressure drop derivative curve. The analysis device moves along the pressure drop derivative curve using a sliding window, calculating the linear regression slope of the data points within the window. Subsequently, the analysis device identifies continuous curve segments with slope values ​​within a preset tolerance range (e.g., slopes between 0.45 and 0.55 are considered 0.5) as potential flow characteristic segments. For example, when the analysis device detects a curve segment with a slope consistently stable around 0.5, it labels it as a slotted mesh filter bilinear flow segment. The analysis device identifies all segments conforming to the theoretical slope characteristics from left to right and automatically marks the start and end boundaries of these segments on the graph, thus completing the automated flow stage division of the entire pressure drop process.

[0057] In some embodiments, automatic segmentation and optimization of flow characteristic segments can be achieved in several ways: Optionally, a segmentation method based on decision trees or rule engines: The analysis device has a built-in expert rule base that associates curve morphological features (such as slope, shape, and relative position) with flow stages to form a series of IF-THEN rules. For example, an early rising segment with a slope of 1 in the IF derivative curve is identified as a wellbore storage segment, and a U-shaped valley after a segment with a slope of 0.5 is identified as a boundary response. The analysis device scans the derivative curves chronologically, applies the rules in the rule base one by one, cuts and labels the curves, and completes the initial automatic segmentation. Users can see the segmentation results on the graph and can manually fine-tune them. The analysis device records the user's adjustments for future optimization of the rule base. Optionally, another segmentation method provides interactive iterative optimization: After completing the initial automatic segmentation, the analysis device presents the start and end boundary lines of each flow characteristic segment to the user in a draggable form on the pressure double logarithmic curve graph. If a user is dissatisfied with the division of a segment (for example, believing that the end point of a linear flow segment should be moved slightly to the right), they can directly drag the corresponding boundary line to the new position using the mouse. The analysis device responds immediately upon releasing the drag: it automatically recalculates the derivative slope of the adjusted flow characteristic segment (and adjacent segments) and updates the slope value on the graph in real time. Simultaneously, the input parameters required by subsequent interpretation algorithms associated with that segment (such as the duration of the segment) are also updated accordingly, and new interpretation results are immediately recalculated.

[0058] It is understandable that other methods can be used to divide the flow feature segments, such as using deep learning-based image segmentation technology, taking the double logarithmic curve as an image input, and having the model directly output the masks of different flow regions. This is not a limitation here.

[0059] S106. Based on preset geological engineering data and fracturing section data, the analysis device calls the interpretation algorithm corresponding to each fracturing section of the flow characteristic section to calculate the interpretation result data corresponding one-to-one with each fracturing section of the flow characteristic section.

[0060] The preset geological engineering data refers to the parameters extracted in step S101 that describe the physical properties of the reservoir and fluids, such as reservoir porosity, permeability, fluid viscosity, compressibility coefficient, and Young's modulus. The fracturing section data refers to parameters directly related to the currently analyzed fracturing section, such as the total amount of fracturing fluid injected, the total amount of proppant added, and the operational displacement. The interpretation algorithm refers to a series of mathematical models based on seepage mechanics theory, used to invert reservoir and fracture parameters from pressure dynamic data. Each flow characteristic section has its corresponding specialized interpretation algorithm; for example, the linear flow section corresponds to a linear flow model used to calculate the product of fracture half-length and conductivity. The interpretation result data refers to a series of key parameters calculated by the interpretation algorithm that can quantitatively evaluate the fracturing effect and describe reservoir characteristics, such as fracture half-length, fracture conductivity, fracture spacing, overall permeability, and fracture complexity index.

[0061] Specifically, the analysis device iterates through each flow characteristic segment defined in step S105. For a given flow segment (e.g., a fractured mesh filtration linear flow segment), the analysis device automatically calls its corresponding interpretation algorithm (i.e., a linear flow interpretation model) from its built-in algorithm library. Then, the analysis device automatically collects and organizes all the input parameters required by the algorithm from the database. This includes: parameters extracted from the flow segment itself (e.g., duration, intercept of the pressure curve in that segment), geological engineering parameters extracted from step S101 (e.g., porosity, viscosity), and fracturing segment parameters (e.g., injection volume). All parameters are automatically filled into the interpretation algorithm, and the analysis device performs calculations to obtain one or more interpretation result data (e.g., fracture half-length multiplied by fracture area). The analysis device repeats this process for all defined flow segments, ultimately obtaining a series of interpretation result data corresponding one-to-one with each flow segment.

[0062] S107. The analysis equipment summarizes the interpretation results data of the entire shale gas well section to generate a fracturing effect report.

[0063] The entire well section refers to the sum of all sections in a shale gas horizontal well that have undergone fracturing. The fracturing effect report is a comprehensive, visual summary document that organizes, compares, and presents the scattered interpretation results data from all fracturing sections of a well. It aims to provide oilfield engineers with a macroscopic and intuitive view of fracturing effectiveness evaluation and to provide a basis for subsequent production decisions and engineering optimization.

[0064] Specifically, the analysis equipment first queries and extracts interpretation results data for all calculated fracturing sections under a specified shale gas well from the database. Then, it summarizes the key interpretation results data for each fracturing section (such as fracture half-length, conductivity, and fracture complexity) and generates a series of comparative charts. For example, it generates a bar chart with the fracturing section number on the x-axis and fracture half-length on the y-axis, making the scale of fracture stimulation in each section immediately clear. Alternatively, it generates a scatter plot with fracture complexity on the x-axis and overall permeability on the y-axis to analyze the production differences under different fracture morphologies. Finally, the analysis equipment integrates these automatically generated charts, along with key raw data and parameter tables, into a pre-formatted report file (such as PDF or Word format), automatically filling in metadata such as well name and analysis date, generating a complete fracturing effect report that can be directly submitted or archived with a single click.

[0065] In some embodiments, fracturing effect reports can be generated in several ways: Optionally, a customizable interactive report (dashboard) method: Instead of generating a static PDF report, the analysis device generates an interactive dashboard on a web interface. The dashboard displays multiple interconnected charts. Users can select parameters to compare via drop-down menus, filter specific ranges of fracturing segments using sliders, and click on a data point in a chart to highlight related data in other charts. Users can freely combine and arrange charts according to their analysis needs, and save the current view layout as a personal template or export it as a static report.

[0066] It is understandable that other methods can be used to generate fracturing effect reports, and this is not limited to these methods.

[0067] In the above embodiments, the analysis equipment improves the efficiency of shale gas well fracturing effect assessment by intelligently processing and analyzing a large amount of post-fracturing pump shutdown data. Through preset filtering parameters and analytical models, the consistency and comparability of the analysis results are enhanced. However, when processing data from different well conditions, its analytical accuracy is limited by the universality of the preset parameters.

[0068] Please refer to the following: Figure 2This is another flowchart illustrating an intelligent analysis method for shale gas well post-pressure pump shutdown data in an embodiment of this application.

[0069] S201. The analysis equipment extracts well inclination data, perforation parameter data, and fracturing time point data of each fracturing section from the batch-loaded shale gas well test geological design documents and fracturing construction data documents based on a preset data dictionary.

[0070] S202, The analysis equipment determines the wellhead pressure data of each fracturing section, which is located between the start time of pump shutdown and the end time of pump shutdown.

[0071] S203. Based on the well inclination data and the perforation parameter data, the analysis device converts the wellhead pressure data into bottom hole pressure data.

[0072] Step S201 is similar to step S101, step S202 is similar to step S102, and step S203 is similar to step S103, so they will not be described again here.

[0073] S204. Based on the optimal combination of filtering parameters of the preset filtering algorithm, the analysis device performs filtering processing on the bottom hole pressure data.

[0074] Specifically, the analysis device first generates a list of preset candidate combinations of filtering parameters based on the user-selected filtering algorithm (such as wavelet transform), containing a variety of possible parameter settings. This list can be a grid covering a wide range (grid search) or a randomly generated set (random search). Next, the analysis device iterates through this candidate combination list. For each candidate combination in the list, the analysis device performs an internal trial-and-error evaluation. The trial-and-error involves using the current candidate combination to perform a temporary trial filtering on the original bottom-hole pressure data, generating a temporary filtered curve. The evaluation involves the analysis device calculating the root mean square error between this temporary curve and the original bottom-hole pressure curve to obtain its fit score. Then, it calculates the second derivative of the temporary curve to obtain its smoothness score. These two scores are combined into a single combined evaluation index. After iterating through all candidate combinations, the analysis device compares the combined evaluation indices of all candidate combinations and determines the candidate combination with the highest score as the optimal filtering parameter combination. Finally, the analysis equipment calls the preset filtering algorithm and uses the newly selected optimal filtering parameter combination to perform final filtering on the raw bottom hole pressure data generated in step S203, thereby producing a smooth curve that has been scientifically proven to be the best for subsequent steps.

[0075] In some embodiments, the optimal combination of filter parameters can be determined in several ways: Optionally, a coarse-to-fine progressive search approach: The analysis device first performs a complete optimization loop on a large-scale, low-density sparse parameter grid to quickly locate a potential region containing the optimal solution. Then, within this potential region, the analysis device automatically generates a smaller but denser fine parameter grid. The analysis device performs the optimization loop again on the new fine grid, thereby determining the final optimal combination of filter parameters with higher accuracy. This approach balances search efficiency and solution accuracy. Optionally, a Bayesian optimization-based intelligent search approach: Instead of blindly traversing all candidate combinations, the analysis device first randomly tries several parameter combinations and evaluates their effects. Based on the performance of these initial points, the analysis device constructs a probabilistic surrogate model (such as a Gaussian process) to predict the potential effects of other untried parameter combinations. The analysis device uses an acquisition function to intelligently decide which parameter combination to try next. This decision balances exploration (trying areas with high uncertainty to avoid missing the global optimum) and exploitation (digging deeper into areas that are known to perform well), thereby finding the optimal solution more efficiently with fewer attempts.

[0076] It is understandable that other methods can be used to determine the optimal combination of filter parameters, such as using genetic algorithms or particle swarm optimization algorithms to search for the optimal combination of parameters. These methods are particularly suitable for dealing with complex optimization problems that are high-dimensional and non-convex, and are not limited here.

[0077] S205. Based on the pressure drop and pressure drop derivative of the filtered bottom hole pressure data, the analysis device generates a double logarithmic pressure curve characterizing the change of pressure drop and pressure drop derivative over time.

[0078] This step, performed after obtaining the clean, final-filtered bottomhole pressure data sequence in step S204, is a crucial step connecting the raw data with professional interpretation, aiming to transform one-dimensional time-series data into visual charts rich in diagnostic information. Specifically, the analysis device receives the smoothed bottomhole pressure data sequence from step S204. First, the analysis device uses the first point of the sequence as the initial pressure and calculates the pressure drop sequence for all subsequent points. Then, the analysis device uses a numerical differentiation algorithm (such as the Bourdet derivative algorithm) to calculate the pressure drop derivative sequence. This algorithm calculates the slope by performing linear regression on the data points within a logarithmic time window, effectively suppressing the interference of residual noise on the derivative calculation. After obtaining the two new time series, pressure drop and pressure drop derivative, the analysis equipment creates a double logarithmic coordinate system canvas, with time as the horizontal axis and pressure drop and pressure drop derivative as the two vertical axes. The two curves are then plotted on the canvas. Typically, the pressure drop derivative curve is represented by a more prominent symbol (such as a circle), while the pressure drop curve is represented by a solid line. The final result is a clear double logarithmic pressure curve that can be interpreted by experts.

[0079] S206. The analysis device divides the flow characteristic segment based on the slope of the pressure drop derivative curve in the double logarithmic pressure curve.

[0080] S207. Based on preset geological engineering data and fracturing section data, the analysis device calls the interpretation algorithm corresponding to each fracturing section of the flow characteristic section to calculate the interpretation result data corresponding one-to-one with each fracturing section of the flow characteristic section.

[0081] S208. The analysis equipment summarizes the interpretation results data of the entire shale gas well section to generate a fracturing effect report.

[0082] Step S206 is similar to step S105, step S207 is similar to step S106, and step S208 is similar to step S107, so they will not be described again here.

[0083] S209. Based on the interpretation results data of each fracturing segment in the fracturing effect report, the analysis equipment establishes a fracturing parameter optimization model.

[0084] The interpretation results data refer to the numerical values ​​calculated in step S207 that can quantitatively characterize the fracturing effect, such as fracture half-length and fracture conductivity. In this step, these are used as target variables that the machine learning model needs to predict. The fracturing parameter optimization model is a mathematical model established through machine learning algorithms that can describe the complex nonlinear relationship between input and output. Its inputs are the geological and engineering parameters that affect the fracturing effect, and its output is the quality of the fracturing effect. This model is used to guide future fracturing designs.

[0085] Specifically, the analysis equipment first integrates the data from all fracturing segments involved in the S208 report to construct a training sample set. In this sample set, each row represents a fracturing segment and contains three types of data: geological engineering parameters (such as Young's modulus, Poisson's ratio, and in-situ stress), fracturing construction parameters (such as displacement, sand-to-fluid ratio, and fluid viscosity), and interpretation result data (such as fracture half-length and conductivity). The analysis equipment first uses cluster analysis methods (such as K-means or density clustering) to automatically group these fracturing segments according to their interpretation results, for example, into high-yield and high-efficiency groups, low-efficiency groups, etc. Then, the analysis equipment uses geological and construction parameters as feature variables and interpretation result data or their respective groups as target variables. To improve the model's performance and interpretability, the analysis equipment performs feature engineering, for example, by calculating the correlation between each feature variable and the target variable (such as the Pearson correlation coefficient) or using methods such as recursive feature elimination (RFE) to remove redundant or irrelevant features, constructing a concise and necessary set of feature variables. Finally, the analysis equipment uses one or more regression analysis methods (such as multiple linear regression, gradient boosting tree, or neural network) to establish a mapping relationship between the feature variable set and the target variable, and optimizes the hyperparameters of the model through cross-validation to obtain a validated and robust fracturing parameter optimization model.

[0086] In some embodiments, fracturing parameter optimization models can be constructed in several ways: Optionally, an ensemble learning-based model construction approach: Instead of training a single regression model, the analysis device simultaneously trains multiple different types of models, such as a linear model, a tree-based model (e.g., random forest), and a neural network model. The analysis device integrates the predictions of these models using a voting or weighted averaging strategy to form a final ensemble model. This approach is generally more robust and accurate than any single model because it combines the advantages of different models, reducing the risk of inappropriate model selection. Optionally, another model construction approach considers parameter interaction effects: When performing feature engineering, the analysis device not only considers the influence of individual parameters but also automatically generates combined features representing the interaction effects between parameters. For example, multiplying displacement and fluid viscosity generates a new hydraulic power feature. The analysis device employs machine learning models capable of automatically capturing complex relationships between features, such as Gradient Boosting Decision Tree (GBDT), which can discover and utilize these high-order interaction relationships during the model construction process. Models built in this way can reveal the physical mechanisms behind fracturing effects more deeply. For example, they can show that higher displacement or viscosity is not always better, but rather that a specific combination of the two is needed to achieve the optimal effect.

[0087] It is understandable that other methods can be used to construct fracturing parameter optimization models, such as using causal inference models to try to separate causality from correlation in order to provide more reliable optimization suggestions, which are not limited here.

[0088] S210. The analysis equipment optimizes the construction parameters of the subsequent well section to be fractured based on the fracturing parameter optimization model.

[0089] In this context, "subsequent fracturing sections" refers to new wells or sections planned for fracturing but not yet constructed within the same block or under similar geological conditions. Construction parameters refer to engineering variables that can be controlled by engineers during fracturing operations, such as pump injection rate, fracturing fluid type and volume, and proppant type and concentration.

[0090] Specifically, when optimizing the design of a subsequent fracturing well section, engineers first input the known, uncontrollable geological and engineering parameters of the well section (such as in-situ stress and Young's modulus obtained through well logging) into the analysis equipment. Then, engineers set an optimization objective, such as maximizing fracture conductivity, and can set constraints on some construction parameters (such as the discharge rate not exceeding the pump's maximum capacity, and the total fluid volume not exceeding the budget). After receiving this information, the analysis equipment uses the uncontrollable geological parameters as fixed inputs to the model, and then, within the user-defined constraints, uses optimization algorithms (such as genetic algorithms or particle swarm optimization) to perform a search for controllable construction parameters. For each set of candidate construction parameters during the search process, the equipment inputs it along with the fixed geological parameters into the fracturing parameter optimization model established in step S209 to obtain a predicted value for the fracturing effect. The optimization algorithm iterates continuously until it finds the set of construction parameters that maximizes the model's predicted optimization objective. Finally, the analysis equipment presents this optimal set of construction parameters as a design recommendation to the engineers.

[0091] In some embodiments, the construction parameters for subsequent fracturing well sections can be optimized in several ways: Optionally, a multi-objective optimization design approach is provided: the analysis device allows engineers to simultaneously set multiple, even potentially conflicting, optimization objectives, such as maximizing fracture conductivity while minimizing the total cost of fracturing fluid. The analysis device employs a multi-objective optimization algorithm (such as NSGA-II), which no longer seeks a single optimal point, but rather a set of solutions known as the Pareto front. Ultimately, the analysis device presents engineers with a Pareto front curve, where each point represents a different trade-off (e.g., one point offers the best performance but also the highest cost, while another offers slightly worse performance but significantly lower cost). Engineers can then select the solution that best suits the current operating conditions from this set of optimal solutions based on actual needs. Optionally, a robust optimization design approach that considers uncertainty is also available: the analysis device recognizes that the input geological parameters themselves contain measurement errors and uncertainties. When inputting geological parameters, engineers are allowed to input a range or a probability distribution (such as the mean and standard deviation), rather than just a fixed value. During the optimization process, the analysis device employs stochastic simulation methods (such as Monte Carlo simulation). For each set of candidate construction parameters, the analysis equipment performs thousands of calculations. Each calculation randomly samples from the probability distribution of geological parameters to obtain a probability distribution of fracturing effect, rather than just a predicted value. Ultimately, the optimization goal of the analysis equipment is no longer to maximize the expected value of the predicted value, but rather, for example, to ensure that the fracture half-length is not lower than a certain threshold at a 90% confidence level, thereby providing a more robust and lower-risk design scheme in the face of uncertainty.

[0092] It is understandable that other methods can be used to optimize the design of construction parameters for subsequent fracturing well sections, which are not limited here.

[0093] This application introduces an automatic hyperparameter optimization mechanism based on dual indices of fit and smoothness, improving the objectivity and adaptability of data processing. Furthermore, this application constructs a fracturing parameter optimization model that reveals the quantitative relationship between engineering parameters and fracturing effects by performing cluster analysis and feature extraction on historical fracturing effect data. Based on this model, it provides a data-driven optimization design scheme for the well section to be fractured. This application not only improves the accuracy of fracturing effect evaluation but also utilizes historical analysis results to provide data decision support for the optimization design of subsequent operations.

[0094] The above describes an intelligent analysis method for shale gas well post-pressure pump shutdown data in the embodiments of this application. The following describes an exemplary analysis device 300 provided in the embodiments of this application.

[0095] Figure 3This is a schematic diagram of an exemplary hardware structure of the analysis device 300 provided in an embodiment of this application. In some embodiments, the analysis device 300 is a computer device, which includes a processor, a memory, and a network interface connected via an analysis device bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operation analysis device, computer programs, and a database. The internal memory provides an environment for the operation analysis device and computer programs in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements an intelligent analysis method for shale gas well post-pressure pump shutdown data according to an embodiment of this application.

[0096] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0097] In some embodiments of this application, a computer-readable storage medium is also provided, including instructions that, when executed on the analysis device 300, cause the analysis device 300 to perform an intelligent analysis method for shale gas well post-pressure pump shutdown data according to an embodiment of this application.

[0098] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0099] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0100] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0101] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. An intelligent analysis method for shale gas well post-pressure pump shutdown data, applied to analysis equipment, characterized in that, The method includes: The analysis equipment extracts well inclination data, perforation parameter data, and fracturing second point data of each fracturing section from batch-loaded shale gas well test geological design documents and fracturing construction data documents based on a preset data dictionary. The preset data dictionary defines multiple representations of the same data under different document formats. The analysis equipment determines the wellhead pressure data of each fracturing stage as located between the start and end times of pump shutdown. Based on the well inclination data and the perforation parameter data, the analysis device converts the wellhead pressure data into bottom hole pressure data; Based on the pressure drop and pressure drop derivative of the filtered bottom hole pressure data, the analysis equipment generates a double logarithmic curve of pressure, representing the change of pressure drop and pressure drop derivative over time. The analytical device divides the flow characteristic segments based on the slope of the pressure drop derivative curve in the pressure double logarithmic curve diagram. The flow characteristic segments include the inter-slit crossflow segment, the slit mesh filtration linear flow segment, the slit mesh filtration bilinear flow segment, and the slit mesh closure segment. Based on preset geological engineering data and fracturing section data, the analysis device calls the interpretation algorithm corresponding to each fracturing section of the flow characteristic section to calculate the interpretation result data that corresponds one-to-one with each fracturing section of the flow characteristic section. The analysis equipment summarizes the interpretation results data of the entire shale gas well section to generate a fracturing effect report.

2. The method according to claim 1, characterized in that, The analysis equipment extracts well inclination data, perforation parameter data, and fracturing time point data for each fracturing stage from batch-loaded shale gas well testing geological design documents and fracturing construction data documents based on a preset data dictionary. Specifically, this includes: The analysis equipment will batch convert the different formats of shale gas well test geological design documents and fracturing construction data documents into intermediate documents of a preset format. Based on a preset data dictionary, named entity recognition, and keyword matching technology, the analysis device matches and extracts well deviation data, perforation parameter data, and fracturing second point data from the intermediate document in the preset format. The analysis equipment establishes a correlation between the extracted data according to the three-layer structure of well-fractured section-fractured section data, and stores it in a relational database in a structured form.

3. The method according to claim 1, characterized in that, The analysis equipment determines the wellhead pressure data for each fracturing stage, which is located between the pump shutdown start time and the pump shutdown end time. Specifically, this includes: The analysis device filters out the displacement data and wellhead pressure data from the fracturing second point data of each fracturing section, and the displacement data and wellhead pressure data are arranged in chronological order. The analysis device determines the moment when the displacement value in the displacement data changes from being greater than a preset displacement threshold for the first time to being less than or equal to the preset displacement threshold for a continuous preset duration as the pump stop start time point. Starting from the pump shutdown start time, after a preset number of wellhead pressure data points, the analysis device determines the moment when the pressure difference between adjacent wellhead pressure data points first exceeds or equals a preset pressure change threshold as the pump shutdown end time. The analysis equipment determines the wellhead pressure data for each fracturing stage as located between the pump shutdown start time and the pump shutdown end time.

4. The method according to claim 1, characterized in that, Based on the filtered bottomhole pressure data, the analysis equipment generates a double logarithmic pressure curve representing the change of pressure drop and pressure drop derivative over time, specifically including: Based on a preset filter algorithm and a preset combination of filter parameters, the analysis device filters the bottom hole pressure data. Based on the pressure drop and pressure drop derivative of the filtered bottom hole pressure data, the analysis equipment generates a double logarithmic pressure curve representing the change of pressure drop and pressure drop derivative over time.

5. The method according to claim 4, characterized in that, After the step of generating a double logarithmic pressure curve characterizing the pressure drop and pressure drop derivative over time based on the filtered bottom hole pressure data, the method further includes: Based on the pressure drop and pressure drop derivative of the bottom hole pressure data before filtering, the analysis equipment generates a raw double logarithmic pressure curve characterizing the change of pressure drop and pressure drop derivative over time. The analytical device determines the degree of fit between the original log-log pressure curve and the log-log pressure curve, wherein the degree of fit refers to the root mean square error value between the original log-log pressure curve and the log-log pressure curve. The analysis device determines the smoothness of the pressure double logarithmic curve based on the second derivative value of the pressure double logarithmic curve. Taking the combined evaluation index of fit and smoothness as the target, the analysis device uses a hyperparameter tuning algorithm to automatically search and determine the optimal combination of filter parameters. The hyperparameter tuning algorithm includes at least a grid search algorithm, a random search algorithm, and a Bayesian optimization algorithm. Based on the optimal combination of filtering parameters according to the preset filtering algorithm, the analysis device filters the bottom hole pressure data.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: Based on the interpretation results of each fracturing segment in the fracturing effect report, the analysis equipment uses cluster analysis to group fracturing segments with similar characteristics into one category. The analysis device uses the geological and engineering parameters and fracture section parameters of each type of fracture section as feature variables, and uses the interpretation results data of each type of fracture section as target variables to construct a training sample set. Based on the correlation analysis of the feature variables in the training sample set, the analysis device removes redundant feature variables and constructs a set of necessary feature variables. The analytical device uses a multiple regression analysis method to establish a mapping relationship between the set of necessary feature variables and the target variable; The analysis equipment verifies and optimizes the mapping relationship using cross-validation to construct an optimized fracturing parameter model; The analysis equipment optimizes the construction parameters of subsequent well sections to be fractured based on the fracturing parameter optimization model.

7. The method according to claim 6, characterized in that, Based on the interpretation results data of each fracturing segment in the fracturing effect report, the analysis equipment uses cluster analysis to group fracturing segments with similar characteristics into one category, specifically including: The analysis equipment standardizes the interpretation results data of each fracturing section; The analytical device uses principal component analysis to reduce the dimensionality of the standardized interpretation results data. Based on the dimensionality-reduced feature vectors, the analysis device uses a density clustering algorithm to group fracturing segments with similar characteristics into one class.

8. An analytical device, characterized in that, The analysis device includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the analysis device to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on the analysis device, the analysis device performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the analysis device, the analysis device performs the method as described in any one of claims 1-7.

Images