Coal rock gas yield scale prediction method based on data driving

Abstract

The invention relates to the technical field of coal rock gas exploration and development, in particular to a coal rock gas yield scale prediction method based on data driving, which comprises the following steps: building a basic data platform; preprocessing the basic data in the basic data platform to obtain a data matrix; constructing a standard curve based on the data matrix; performing fitting and extrapolation on the basis of the standard curve to form a full-life-cycle standard gas production curve; obtaining a multi-dimensional prediction result based on the full-life-cycle standard gas production curve; the multi-dimensional prediction result is obtained by carrying out yield superposition according to a time dimension and a space dimension. According to the method, data driving is taken as a core, a data closed loop is realized by relying on a digital intelligent platform, and dynamic prediction of the coal rock gas yield scale is realized through four key steps of standardization processing, typical curve construction, differential prediction and multi-dimensional superposition.

Description

Technical Field

[0001] This invention relates to the field of coal and shale gas exploration and development technology, and in particular to a data-driven method for predicting the scale of coal and shale gas production. Background Technology

[0002] Coal shale gas is an unconventional natural gas found in the pore-fracture system of coal seams. Its main component is methane, and it has both energy and environmental value. As a clean fossil energy source, its development can alleviate the contradiction between natural gas supply and demand, while reducing coal mine gas accidents and greenhouse gas emissions. It is an important supplement to my country's energy structure transformation.

[0003] Globally, my country possesses abundant coal-rock gas resources, with proven geological reserves exceeding 36 trillion cubic meters, primarily distributed in the Qinshui Basin and the eastern edge of the Ordos Basin. In recent years, my country has achieved breakthroughs in coal-rock gas development, becoming a key component of increased domestic natural gas production. However, coal-rock gas development faces dual constraints: the unique characteristics of the reservoirs and limitations in prediction technology, and it differs significantly from shale gas development. ① Complex reservoir and production characteristics: Coal and rock reservoirs are highly heterogeneous, with low porosity and low permeability. Furthermore, coal and rock gas wells exhibit a production pattern of "large fluctuations in production in the early stage and long periods of production stabilization and reduction in the middle and later stages." Production data is significantly affected by drainage and production systems and changes in reservoir pressure, with far more interfering factors than shale gas (shale gas is mainly characterized by "high initial production, rapid decline, and long period of stable production," without a long-term drainage phase). ② Existing prediction methods have prominent limitations: Current coal gas production prediction mainly relies on traditional reservoir engineering methods (such as analytical model method, Arps production decline method) and empirical analogy method, which have three major problems: (1) Lack of real-time performance: Traditional methods require manual processing of static parameters (such as reservoir thickness, permeability) and dynamic data (daily gas production, pressure), and the data update cycle is as long as 3 to 6 months, which cannot meet the needs of "dynamic changes in daily production data" of coal gas wells. Although existing shale gas methods focus on real-time performance, they are not designed for data processing in the coal gas drainage stage; (2) Poor adaptability: Coal gas development involves drilling, completion of drilling, and other aspects. There are many types of wells, such as drilling wells, fracturing wells, and wells waiting to be put into production. The "well construction-fracturing-production" cycle of each well is very different. The traditional analogy method is difficult to dynamically adjust the prediction model. The shale gas prediction method does not consider the additional cycle of coal gas "waiting for compression, waiting for production"; (3) Insufficient accuracy: The traditional method does not specifically deal with the problem of "data interference and abnormal values ​​(such as zero production caused by shutting in the well and repairing the well)" in the single-phase liquid discharge stage of coal gas. It also does not consider the impact of the length of the horizontal section of fracturing and the discharge system on the production. The prediction error generally exceeds 20%, which cannot meet the accuracy requirements of production capacity planning. The error control logic of the shale gas prediction method needs to be adapted to the characteristics of coal gas data.

[0004] Furthermore, existing technologies cannot meet the needs of multi-dimensional prediction. Existing methods mostly focus on single-well production prediction and lack the production superposition logic at the "well area-block-reservoir" level, making it difficult to support the calculation of the production scale of the entire region and annual deployment decisions. Although this is consistent with the "annual deployment prediction" requirement of shale gas, coal-rock gas requires more detailed well type cycle superposition.

[0005] In summary, there is an urgent need for a technical solution that adapts to the characteristics of coal and gas reservoirs and production, possesses real-time and rolling capabilities, and supports multi-dimensional prediction, in order to overcome the bottlenecks of traditional methods.

[0006] Therefore, existing technologies still need improvement. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention proposes a data-driven method for predicting the scale of coal and shale gas production, thereby resolving the technical issues existing in the current method for predicting the scale of coal and shale gas production.

[0008] To address the aforementioned technical problems, some embodiments of the present invention disclose a data-driven method for predicting the scale of coal and rock gas production, including: Step 1: Establish a basic data platform; Step 2: Preprocess the basic data in the basic data platform to obtain a data matrix; Step 3: Construct a standard curve based on the data matrix; Step 4: Fit and extrapolate based on the standard curve to form a full life cycle standard gas production curve; Step 5: Obtain multi-dimensional prediction results based on the aforementioned full life cycle standard gas production curve; The multi-dimensional prediction results are obtained by superimposing outputs according to the time and spatial dimensions.

[0009] In some embodiments, the multi-dimensional prediction results include an annual production scale table, a well type production contribution ratio chart, and a production trend chart for the next five years.

[0010] In some embodiments, steps one through four are iteratively updated based on updates to the underlying data.

[0011] In some embodiments, updates to the basic data include: the latest daily production data of wells already in production; attribute change data from wells to drilling completion, from drilling completion to fracturing, and from fracturing to production; and parameters for newly deployed wells and / or planned wells.

[0012] In some embodiments, step one, building a basic data platform includes: real-time collection and storage of basic engineering parameters and daily production data of the smallest prediction unit, the wells that have been put into production.

[0013] In some embodiments, preprocessing the basic data in the basic data platform includes: Based on the daily gas production data of the wells already in production, interfering data is removed to obtain the filtered daily gas production data of a single well; The screened daily gas production data of single wells were standardized according to the length of the horizontal section of the fracturing to eliminate the impact of the difference in the length of the horizontal section on the production, and standardized data were obtained. Standardize the data of all wells that have been put into production and align them according to their effective production time to construct a data matrix.

[0014] In some embodiments, fitting and extrapolating based on the standard curve to form a full life cycle standard gas production curve includes: For each effective production time, calculate the arithmetic mean and median mean of the daily gas production per unit fracturing section length for all effective samples at time t; The median mean was then corrected using the arithmetic mean to obtain the standard gas production curve; The standard gas production curve is numerically fitted and extrapolated to the entire life cycle of coal and rock gas wells to obtain the full life cycle standard gas production curve.

[0015] In some embodiments, the numerical fitting employs a double exponential decreasing function, with the following formula:

[0016] in, For the fitted Standard daily gas production per unit length of fracturing section at any given time, expressed in m³ / (d·m)). These are the fitting coefficients; The effective production time after fitting is expressed in days (d). Using the "coefficient of determination" "and "adjusted coefficient of determination" "As an accuracy indicator, the control condition is..." and Ensure that the fitting error is less than 10%.

[0017] In some embodiments, the multi-dimensional prediction results obtained based on the full life cycle standard gas production curve include: predicting and superimposing the production of wells that have been put into production, wells that have been drilled, wells that have been completed, fractured wells, deployed wells, and planned wells.

[0018] In some embodiments, the prediction model formula for the fractured well is: ; in, This indicates the actual length (m) of the fractured section in the fractured well. For the fitted Standard daily gas production per unit length of fracturing section at any given time. To predict the daily gas production of a fractured well at time t, The average fracturing utilization rate of the block is dimensionless.

[0019] By adopting the above technical solution, the present invention has at least the following beneficial effects: This invention provides a data-driven method for predicting coal and shale gas production scale. With data-driven principles at its core, it establishes a daily-granularity data platform to achieve real-time acquisition and dynamic updating of basic data (engineering parameters, production data), meeting the real-time requirements of prediction and overcoming the data lag shortcomings of traditional methods. It constructs a full-process model encompassing "data preprocessing - standard curve construction - well-type prediction - production overlay," specifically addressing the characteristics of coal and shale gas such as "single-phase drainage, outliers, and multiple well types," improving prediction accuracy (controlling error <10%), which is superior to existing methods. It supports multi-dimensional prediction from "well area (smallest prediction unit) - block - gas reservoir," adapting to the production capacity planning needs of different development stages, meeting the requirements of rolling forecasts and flexibility, and accommodating annual deployment decisions. It provides quantitative basis for calculating the workload of annual deployment wells and medium-to-long-term production planning for coal and shale gas, reducing decision-making costs and promoting improved development efficiency. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a data-driven method for predicting the scale of coal and rock gas production, as disclosed in some embodiments of the present invention. Detailed Implementation

[0022] The embodiments of this disclosure will be further described in detail below with reference to the accompanying drawings and examples. The detailed description of the embodiments and the accompanying drawings are used to illustrate the principles of this disclosure by way of example, but should not be used to limit the scope of this disclosure. This disclosure can be implemented in many different forms and is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

[0023] These embodiments are provided to make the disclosure thorough and complete, and to fully express the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specifically stated, the relative arrangement of components and steps, material composition, numerical expressions, and values ​​set forth in these embodiments should be interpreted as exemplary only and not as limiting.

[0024] Furthermore, words such as "include" or "contain" mean that the element preceding the word covers the element listed after the word, and do not exclude the possibility that it may also cover other elements.

[0025] It should also be noted that, in the description of this disclosure, unless otherwise expressly specified and limited, the specific meaning of each term in this disclosure can be understood by those skilled in the art as appropriate. All terms used in this disclosure have the same meaning as understood by those skilled in the art to which this disclosure pertains, unless otherwise specifically defined. It should also be understood that terms defined in general dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and should not be interpreted with an idealized or highly formalized meaning, unless expressly defined herein.

[0026] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, they should be considered part of the specification.

[0027] like Figure 1 As shown, some embodiments of the present invention disclose a data-driven method for predicting the scale of coal and shale gas production. Addressing the limitations of existing coal and shale gas production prediction methods, such as poor real-time performance, weak adaptability, low accuracy, and limited dimensionality, and considering the unique needs of coal and shale gas development, a multi-dimensional prediction method based on a digital intelligence platform covering the entire lifecycle has been developed, including the following steps: Step 1: Establish a basic data platform; Step 2: Preprocess the basic data in the basic data platform to obtain a data matrix; Step 3: Construct a standard curve based on the data matrix; Step 4: Fit and extrapolate based on the standard curve to form a full life cycle standard gas production curve; Step 5: Obtain multi-dimensional prediction results based on the aforementioned full life cycle standard gas production curve; The multi-dimensional prediction results are obtained by superimposing outputs according to the time and spatial dimensions.

[0028] This embodiment takes "data-driven" as its core, relying on a digital intelligence platform to achieve data closed-loop. Through four key steps—"standardized processing, typical curve construction, differentiated prediction, and multi-dimensional overlay"—it achieves dynamic prediction of coal and shale gas production scale. It realizes production overlay at the "well area-block-reservoir" level, supporting the calculation of production scale across the entire region and annual deployment decisions. By overlaying more granular well types and cycles, it achieves accurate prediction of coal and shale gas production, meeting the requirements of rolling forecasting and flexibility, and adapting to annual deployment decisions.

[0029] The data-driven method for predicting coal shale gas production scale disclosed in some embodiments of the present invention includes: Step 1: Building the basic data platform.

[0030] Data Acquisition: Real-time acquisition of basic engineering parameters (length of horizontal fracturing section) of the smallest prediction unit, already in production wells. Drilling cycle Fracturing cycle Waiting period Investment period (etc.) and daily production data (daily gas production) , sleeve pressure (Data such as drainage volume, etc.) are collected at the "daily level" to ensure data timeliness; Data storage: Historical and real-time data are stored using a distributed database (such as Hadoop), supporting data backtracking (up to 5 years) and batch retrieval; Data interface: Reserved interface for connection with oilfield ERP system and production monitoring system to achieve automatic data synchronization and avoid manual input errors.

[0031] Step 2: Basic data preprocessing.

[0032] Given the unique characteristics of coal and rock gas production data, the following steps are used to process the data from individual wells already in production and construct a standardized data matrix: Data filtering: Based on daily gas production data, two types of interfering data were removed: Single-phase drainage stage data: In the early stage of production after the pressure reduction of coal gas wells, there is a brief period where drainage is the main process (no gas production). It is necessary to identify and remove the data of this stage based on the drainage curve. Abnormal shutdown data: Exclude data where daily gas production is zero or abnormally low (less than 10% of normal production) due to well shutdown, well repair, or equipment failure. Standardization processing: The daily gas production data of the screened single wells are standardized according to the length of the horizontal section of the fracturing process to eliminate the impact of differences in the length of the horizontal section on the production. The calculation formula is as follows:

[0033] in: For the first Effective production time of the wells already in production The corresponding daily gas production per unit length of fracturing section (unit: m³ / (d·m)); For the first Koujing Actual daily gas production at any given moment (unit: m³ / d). For the first Length of the horizontal section of the fractured well (unit: m); Data matrix construction: Standardize the data of all production wells and align them according to "effective production time" to construct a data matrix:

[0034] in: Effective production time (unit: days); The maximum effective production time of all wells that have been put into production (in days). for The number of valid samples at a given time (i.e., the number of wells with valid data at that time, in units of wells).

[0035] Step 3: Constructing the standard curve.

[0036] A standard gas production curve for the entire life cycle of a coal gas well is constructed through statistical calculation and fitting correction to adapt to the prediction needs of multiple well types. Standard curve statistical calculation: For each effective production time Calculate the two sets of statistical values: ① average : ① The arithmetic mean of daily gas production per unit fracturing section length of all valid samples at any given time, reflecting the overall average level; ② M50 value :Will The valid samples at any given time are sorted by value, and the arithmetic mean of the range from the 25th percentile (P25) to the 75th percentile (P75) is taken to effectively remove the interference of abnormally high values ​​(such as high-yield wells) and abnormally low values ​​(such as low-yield wells), which is suitable for the high dispersion of coal and rock gas data. Sample truncation processing: when ( As the lower limit of the effective sample size, it is recommended to take a value of 20 to 50 wells, which can be adjusted according to the size of the well area (different from the sample size requirement for shale gas). When the sample size is insufficient, statistical calculations should be stopped after that point to avoid errors caused by insufficient sample size. Equivalence Correction: Since the M50 value focuses on the characteristics of intermediate samples, it may deviate from the overall average level. Therefore, an equivalence correction coefficient is introduced. The M50 value is corrected to ensure that the standard curve reflects the true average capacity of the smallest forecast unit. The calculation formula is as follows:

[0037] in: It is a dimensionless coefficient, and its value is usually between 0.8 and 1.2. The closer it is to 1, the better the consistency between the M50 value and the average value. Final standard curve: Corrected daily gas production per unit fracturing section length is This generates a continuous and smooth standard curve, as shown in Table 1 below: Table 1

[0038] Step 4: Standard Curve Fitting and Extrapolation Numerical fitting is performed on the corrected standard curve data and extrapolated to the entire life cycle of coal and rock gas wells (usually 20 years) to ensure the complete production cycle of the predicted cover well.

[0039] Fitting function selection: Considering the characteristics of coal-rock gas—"small initial fluctuations, gradual decline in the middle stage, and stable production in the long term"—the double exponential decreasing function in document 04 is referenced, and the double exponential decreasing function is selected, as shown in the following formula:

[0040] in: For the fitted Standard daily gas production per unit length of fracturing section at any given time (unit: m³ / (d·m)). These are the fitting coefficients (obtained using the least squares method). The effective production time after fitting (unit: days); Fitting accuracy control: using the coefficient of determination "and "adjusted coefficient of determination" "As an accuracy indicator, the control condition is..." and Ensure that the fitting error is less than 10%; Coefficient of determination: ,in For actual calibration data, To fit the predicted values, This is the average of the actual data; Adjust the coefficient of determination: ,in For the number of data points, This represents the number of independent variables (including constant terms) in the fitted function. Curve extrapolation: When the fitting accuracy meets the requirements, the curve is extrapolated to... d (approximately 25 years, the average lifespan of a coal-rock gas well) forms a standard gas production curve for the entire lifespan. .

[0041] Step 5: Production Prediction by Well Type For the six categories of wells in coal and shale gas development—namely, "wells in production, wells under drilling, wells completed, fractured wells, deployment wells, and planned wells"—differentiated prediction models are designed, and a new "cycle overlay" logic specific to coal and shale gas is added, as shown in Table 2. Table 2

[0042] Step Six: Production Scale Stacking and Result Output Production is overlaid using both time (annual / quarterly) and spatial (well area / block / gas reservoir) dimensions to output multi-dimensional prediction results: The time dimension overlay includes: Production wells: Based on the block's average production time rate (usually 0.9 to 0.95), the predicted production is allocated to the calendar year (e.g., remaining effective production time for the year = total number of days in the year × production time rate - number of days already produced). Other types of wells: The production time is determined based on "drilling start / drilling completion / fracturing time + corresponding cycle", the predicted production is allocated to the corresponding year, and a new "waiting for fracturing, waiting for commissioning cycle" specifically for coal and rock gas is added.

[0043] Spatial dimension overlay includes: Well area production: The annual projected production of all wells in the well area is superimposed.

[0044] Block output: The output of all wells within the block.

[0045] Gas reservoir production: The production of all blocks within the gas reservoir is superimposed.

[0046] Output results: Generates an annual production scale table, a well-type production contribution ratio chart, and a production trend chart for the next five years. It supports export in Excel and PDF formats and can be directly used for capacity planning decisions.

[0047] Based on the above embodiments, in order to further improve the accuracy of the prediction results, it also includes real-time iterative updates based on the updates of the basic data.

[0048] To meet the requirements of rolling updates and real-time performance, the data platform automatically updates the following data daily, triggering prediction model iterations. The iteration cycle can be set to "daily (real-time update)" or "monthly (deep correction)" to ensure that prediction results are dynamically optimized as the development process progresses.

[0049] Latest daily production data (daily gas production, pressure) for wells that have been put into production.

[0050] Attribute change data for wells from drilling start to drilling completion, from drilling completion to fracturing, and from fracturing to production.

[0051] Parameters for newly deployed / planned wells (such as design horizontal section length and drilling start time).

[0052] In summary, the data-driven coal and shale gas production scale prediction method disclosed in this invention achieves real-time acquisition and dynamic updating of basic data (engineering parameters, production data) by building a daily granular data intelligence platform, meeting the real-time requirements of prediction and overcoming the data lag defects of traditional methods. By constructing a full-process model of "data preprocessing - standard curve construction - well type prediction - production superposition", it specifically addresses the characteristics of coal and shale gas such as "single-phase liquid discharge, outliers, and multiple well types", effectively improving prediction accuracy (controlling error <10%). This invention supports multi-dimensional prediction of "well area (smallest prediction unit) - block - gas reservoir", adapting to the production capacity planning needs of different development stages, meeting the rolling and flexible requirements of prediction, and adapting to annual deployment decisions. It provides quantitative basis for the annual deployment well workload calculation and medium- and long-term production planning of coal and shale gas, reduces decision-making costs, and promotes development efficiency improvement.

[0053] The embodiments of this disclosure have now been described in detail. To avoid obscuring the concept of this disclosure, some details known in the art have not been described. Those skilled in the art can fully understand how to implement the technical solutions disclosed herein based on the above description.

[0054] While specific embodiments of this disclosure have been described in detail by way of examples, those skilled in the art should understand that the examples are for illustrative purposes only and not intended to limit the scope of this disclosure. Those skilled in the art should understand that modifications can be made to the above embodiments or equivalent substitutions can be made to some technical features without departing from the scope and spirit of this disclosure. In particular, as long as there is no structural conflict, the technical features mentioned in the various embodiments can be combined in any manner.

Claims

1. A data-driven method for predicting the scale of coal and rock gas production, characterized in that, include: Step 1: Establish a basic data platform; Step 2: Preprocess the basic data in the basic data platform to obtain a data matrix; Step 3: Construct a standard curve based on the data matrix; Step 4: Fit and extrapolate based on the standard curve to form a full life cycle standard gas production curve; Step 5: Obtain multi-dimensional prediction results based on the aforementioned full life cycle standard gas production curve; The multi-dimensional prediction results are obtained by superimposing outputs according to the time and spatial dimensions.

2. The data-driven method for predicting coal and rock gas production scale according to claim 1, characterized in that, The multi-dimensional prediction results include an annual production scale table, a well type production contribution ratio chart, and a production trend chart for the next five years.

3. The data-driven method for predicting coal and rock gas production scale according to claim 1, characterized in that, The updates based on the basic data are used to iteratively update steps one through four.

4. The data-driven method for predicting coal and rock gas production scale according to claim 3, characterized in that, The updates to the basic data include: the latest daily production data of wells already in production; attribute change data from wells that have started drilling to wells that have been completed, from wells that have been completed to fractured wells, and from fractured wells to wells that have been put into production; and parameters for newly deployed wells and / or planned wells.

5. The data-driven method for predicting coal and rock gas production scale according to claim 1, characterized in that, Step one, building the basic data platform includes: real-time collection and storage of basic engineering parameters and daily production data of the smallest prediction unit, the wells that have been put into production.

6. The data-driven method for predicting coal and rock gas production scale according to claim 1, characterized in that, Preprocessing the basic data in the aforementioned basic data platform includes: Based on the daily gas production data of the wells already in production, interfering data is removed to obtain the filtered daily gas production data of a single well; The screened daily gas production data of single wells were standardized according to the length of the horizontal section of the fracturing to eliminate the impact of the difference in the length of the horizontal section on the production, and standardized data were obtained. Standardize the data of all wells that have been put into production and align them according to their effective production time to construct a data matrix.

7. The data-driven method for predicting coal and rock gas production scale according to claim 1, characterized in that, Based on the standard curve, fitting and extrapolation are performed to form a full life cycle standard gas production curve, including: For each effective production time, calculate the arithmetic mean and median mean of the daily gas production per unit fracturing section length for all effective samples at time t; The median mean was then corrected using the arithmetic mean to obtain the standard gas production curve; The standard gas production curve is numerically fitted and extrapolated to the entire life cycle of coal and rock gas wells to obtain the full life cycle standard gas production curve.

8. The data-driven method for predicting coal and rock gas production scale according to claim 7, characterized in that, The numerical fitting uses a double exponential decreasing function, as shown in the formula: in, For the fitted Standard daily gas production per unit length of fracturing section at any given time, expressed in m³ / (d·m); These are the fitting coefficients; The effective production time after fitting is expressed in days (d). Using the "coefficient of determination" "and" Adjusted coefficient of determination "As an accuracy indicator, the control condition is..." and Ensure that the fitting error is less than 10%.

9. The data-driven method for predicting coal and rock gas production scale according to claim 1, characterized in that, Based on the standard gas production curve of the entire life cycle, multi-dimensional prediction results are obtained, including: predicting the production of wells that have been put into production, wells that have been drilled, wells that have been completed, fractured wells, deployed wells, and planned wells respectively, and superimposing them.

10. The data-driven method for predicting coal and rock gas production scale according to claim 9, characterized in that, The prediction model formula for the fractured well is as follows: ； in, This indicates the actual length (m) of the fractured section in the fractured well. For the fitted Standard daily gas production per unit length of fracturing section at any given time. To predict the daily gas production of a fractured well at time t, The average fracturing utilization rate of the block is dimensionless.

Images