Method, device and equipment for predicting sand production rate of coal seam

Abstract

The embodiment of the invention relates to the technical field of deep coal bed gas development, in particular to a coal bed sand production rate prediction method, device and equipment, and the method comprises the steps: calculating a first sand production coefficient and a second sand production coefficient of each production working condition parameter according to historical distribution data of a plurality of production working condition parameters in a plurality of well groups of a coal bed; the first sand production coefficient represents the influence degree of the production working condition parameter on the coal seam sand production amount; the second sand production coefficient represents the influence degree of the production working condition parameter on the sand production amount of each well group; determining a plurality of master control production working condition parameters of each well group according to the first sand production coefficient and the second sand production coefficient; and according to the current distribution data of the main control production working condition parameters of each well group, the sand production amount of the well group is predicted.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of deep coalbed methane development technology, specifically to a method, apparatus, and equipment for predicting coalbed sand production. Background Technology

[0002] In deep coalbed methane extraction, target coal seams are often located in loosely cemented sandstone formations with low and uneven cement content. The formation structure is easily disrupted under mining stress, causing skeletal sand particles to detach and transform into movable sand, which is carried to the wellbore and even the surface by the produced fluid flow. Furthermore, excessive production pressure differentials, unreasonable production rate control, frequent well workovers, and the natural decay of formation pressure all exacerbate sand production. Sand production not only causes severe erosion of equipment, accelerating its damage and shortening its service life, but also easily leads to operational failures such as pump sticking, affecting normal well production and even causing shutdowns, significantly reducing extraction efficiency and increasing operating costs.

[0003] However, existing sand production prediction technologies are mostly limited to qualitative judgments, typically relying on single parameters such as production pressure differential to predict whether a gas well will produce sand, and cannot achieve accurate quantitative prediction of sand production. This technological bottleneck makes it difficult to formulate effective sand control measures in advance during the extraction process, especially for deep coalbed methane wells, where there is currently a lack of reliable quantitative sand production prediction methods, which restricts the improvement of safe and efficient extraction levels. Summary of the Invention

[0004] The purpose of the embodiments in this specification is to provide a method, apparatus, and equipment for predicting coal seam sand production, so as to overcome the problem that existing methods cannot accurately and quantitatively predict coal seam sand production.

[0005] To solve the above-mentioned technical problems, the specific technical solutions of the embodiments in this specification are as follows:

[0006] On the one hand, the embodiments of this specification provide a method for predicting the amount of sand produced in a coal seam, including:

[0007] Based on the historical distribution data of multiple production condition parameters in multiple well groups of coal seams, the first sand production coefficient and the second sand production coefficient of each production condition parameter are calculated; the first sand production coefficient represents the degree of influence of the production condition parameter on the sand production of the coal seam; the second sand production coefficient represents the degree of influence of the production condition parameter on the sand production of each well group.

[0008] Based on the first sand production coefficient and the second sand production coefficient, multiple key production condition parameters for each well group are determined.

[0009] Based on the current distribution data of multiple key production parameters for each well group, the sand production of that well group is predicted.

[0010] Furthermore, the calculation of the first sand output coefficient for each production condition parameter includes:

[0011] Obtain historical sand production data for multiple well groups in the coal seam corresponding to the historical distribution data;

[0012] Determine the nonlinear converter corresponding to each production condition parameter;

[0013] Based on the nonlinear transformer, the historical distribution data of the production condition parameter is nonlinearly transformed to obtain the transformed distribution data of the production condition parameter.

[0014] Based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam, the first sand production coefficient of each production condition parameter is calculated.

[0015] Further, the step of calculating the first sand production coefficient for each production condition parameter based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam includes:

[0016] Based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam, the first sand production coefficient of each production condition parameter is calculated using the following formula:

[0017] ;

[0018] In the formula, For the first The first sand output coefficient of each production condition parameter; For a specific moment; The time lag is a preset set of time lags. ; It is a nonlinear converter. It belongs to the pre-defined set of nonlinear converters; This represents the total number of coal seam well groups. for Time of the first The production condition parameter is at the... Historical distribution data of each well group; for Time of the first Historical sand production data for each well group.

[0019] Furthermore, the calculation of the second sand output coefficient for each production condition parameter includes:

[0020] Based on the historical distribution data of each production condition parameter across multiple well groups, determine the information measurement weight of each production condition parameter in each well group.

[0021] Based on the historical distribution data of multiple production operating parameters in multiple well groups, the collinearity coefficient of each production operating parameter is determined. The collinearity coefficient represents the linear correlation between the production operating parameter and the remaining production operating parameters among the multiple production operating parameters.

[0022] Based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter, the second sand production coefficient of each production condition parameter at each well group is calculated; the information measurement weight is positively correlated with the second sand production coefficient; the collinearity coefficient is negatively correlated with the second sand production coefficient.

[0023] Further, the step of calculating the second sand production coefficient for each production condition parameter at each well group based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter includes:

[0024] Based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter, the second sand production coefficient of each production condition parameter at each well group is calculated using the following formula:

[0025] ;

[0026] In the formula, For the first The production condition parameter is at the... The second sand production coefficient at each well group; For a specific moment; The time lag is a preset set of time lags. ; for Time of the first The production condition parameter is at the... Information measurement weights for each well group; is the collinearity coefficient of the i-th production condition parameter; This is the preset threshold for the collinearity coefficient.

[0027] Furthermore, the determination of multiple key production condition parameters for each well group based on the first and second sand production coefficients includes:

[0028] By coupling the first sand production coefficient with the second sand production coefficient, a third sand production coefficient for each production condition parameter at each well group is obtained;

[0029] Based on the third sand output coefficient, multiple main control production condition parameters are selected from the multiple production condition parameters.

[0030] Further, the coupling of the first sand production coefficient and the second sand production coefficient to obtain the third sand production coefficient for each production condition parameter at each well group includes:

[0031] The third sand production coefficient for each production condition parameter at each well group is obtained by coupling the first sand production coefficient with the second sand production coefficient using the following formula:

[0032] ;

[0033] In the formula, For the first The production condition parameter is at the... The third sand production coefficient at each well group; For the coal seam development time; This is the preset scaling factor; This is a preset balance coefficient greater than 0; For the first The first sand output coefficient of each production condition parameter; For the first The production condition parameter is at the... The second sand production coefficient at each well group.

[0034] Furthermore, predicting the sand production of each well group based on the current distribution data of multiple key production parameters for each well group includes:

[0035] Based on the historical distribution data of multiple key production condition parameters of each well group and the historical sand production data of the well group, a multivariate regression prediction model for the well group is constructed.

[0036] The current distribution data of multiple key production parameters of the well group are input into the multivariate regression prediction model of the well group to obtain the sand production prediction data of the well group.

[0037] On another front, embodiments of this specification provide a device for predicting the amount of sand produced in a coal seam, comprising:

[0038] The calculation module is used to calculate the first sand production coefficient and the second sand production coefficient for each production condition parameter based on the historical distribution data of multiple production condition parameters in multiple well groups of the coal seam. The first sand production coefficient represents the degree of influence of the production condition parameter on the sand production of the coal seam. The second sand production coefficient represents the degree of influence of the production condition parameter on the sand production of each well group.

[0039] The determination module is used to determine multiple main control production condition parameters for each well group based on the first sand production coefficient and the second sand production coefficient.

[0040] The prediction module is used to predict the sand production of each well group based on the current distribution data of multiple main control production condition parameters of each well group.

[0041] In another aspect, a computer device is provided, including a memory for storing computer programs and a processor for executing the computer programs to implement the aforementioned method for predicting coal seam sand production.

[0042] Furthermore, embodiments of this specification also provide a computer program product, which, when run by the processor of a computer device, executes instructions for any of the methods described above.

[0043] As can be seen from the technical solutions provided in the embodiments of this specification above, these embodiments can calculate the first and second sand production coefficients for each production condition parameter based on historical distribution data of multiple production condition parameters across multiple well groups in the coal seam. The first sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of the coal seam; the second sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of each well group. Based on the first and second sand production coefficients, multiple main control production condition parameters for each well group are determined. Based on the current distribution data of the multiple main control production condition parameters for each well group, the sand production of that well group is predicted. By calculating the first and second sand production coefficients respectively, a comprehensive factor evaluation system from the overall coal seam to individual well groups is constructed, overcoming the shortcomings of traditional methods that only consider global patterns while ignoring local characteristics. This system can simultaneously capture the general patterns and special performances of parameters, ensuring that the selection of main control factors conforms to the overall trend while adapting to the differences in local geological and development conditions. Furthermore, the selected main control production condition parameters have clear physical meaning and regional adaptability, enabling the subsequent prediction model to fully consider the uniqueness of each well group. This well-specific forecasting approach significantly improves the accuracy and reliability of sand production prediction, providing precise data support for developing differentiated sand control measures and effectively avoiding the problem of insufficient adaptability of the traditional one-size-fits-all method in complex coalbed methane reservoirs. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the accompanying drawings used in the description of the embodiments or prior art will be briefly introduced below.

[0045] Figure 1 This is a flowchart of a method for predicting coal seam sand production provided in the embodiments of this specification;

[0046] Figure 2 This is a flowchart illustrating the overall logic of a method for predicting coal seam sand production provided in the embodiments of this specification.

[0047] Figure 3This is a schematic diagram illustrating the relative error in the sand production prediction of coalbed methane well A provided in the embodiments of this specification;

[0048] Figure 4 This is a schematic diagram of the structural composition of a coal seam sand production prediction device provided in the embodiments of this specification;

[0049] Figure 5 This is a schematic diagram of the structural composition of the computer device provided in the embodiments of this specification. Detailed Implementation

[0050] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0051] It should be noted that the terms "first," "second," etc., used in this specification, claims, and the foregoing drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0052] In some embodiments, the coal seam production parameters may include at least: production days, wellhead pressure, fluid production, gas production, desanding time, total desanding volume, well depth, fluid intensity, displacement fluid volume, sand addition volume, average sand ratio, and hydrochloric acid volume.

[0053] These production parameters not only cover basic production records but also include key operational parameters such as well completion, fracturing, and production control. Specifically, production days can be used to characterize the continuity and cumulative production effect of gas well production; desanding time can reflect the frequency and cycle of desanding operations; wellhead pressure can serve as a key indicator directly reflecting the wellbore flow state and formation energy; fluid production and gas production can be used to assess the production intensity of reservoir fluids and gas well productivity, respectively, and are the basis for analyzing the fluid's sand-carrying capacity; total desanding volume can be a direct quantitative measure of the sand production scale; well depth is related to formation pressure and temperature environment; fluid intensity, displacement fluid volume, sand addition volume, average sand ratio, and hydrochloric acid volume—these parameters collectively describe the scale and intensity of reservoir stimulation, directly affecting the conductivity and rock mechanical stability of the near-wellbore zone.

[0054] This parameter system obtains information from multiple sources, including production dynamics, downhole conditions, and engineering interventions, aiming to build an input space with complete information and clear physical meaning for subsequent accurate quantitative prediction of sand volume.

[0055] Figure 1 This is a flowchart of a method for predicting coal seam sand production provided in the embodiments of this specification. Figure 2 This is a flowchart illustrating the overall logic of a method for predicting coal seam sand production provided in the embodiments of this specification. In specific implementation, it includes the following steps:

[0056] S10: Based on the historical distribution data of multiple production condition parameters in multiple well groups of the coal seam, calculate the first sand production coefficient and the second sand production coefficient for each production condition parameter; the first sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of the coal seam; the second sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of each well group.

[0057] Based on multi-dimensional historical production data of each well group in the coal seam, the first and second sand production coefficients corresponding to each production condition parameter are calculated, thereby constructing a parameter impact assessment system from global to local perspectives. The first sand production coefficient is used to quantify the overall impact of this parameter on the total sand production at the overall coal seam scale; the second sand production coefficient is used to characterize the local impact of this parameter on sand production behavior in each specific well group, thus enabling differentiated identification of sand production control factors under different geological and development conditions.

[0058] In some embodiments, the calculation of the first sand production coefficient for each production condition parameter in step S10 above may specifically include: obtaining historical sand production data of multiple well groups in the coal seam corresponding to the historical distribution data; determining the nonlinear transformer corresponding to each production condition parameter; performing a nonlinear transformation on the historical distribution data of the production condition parameter according to the nonlinear transformer to obtain the transformed distribution data of the production condition parameter; and calculating the first sand production coefficient for each production condition parameter based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of the multiple well groups in the coal seam.

[0059] Historical distribution data for each coalbed methane well group can include the actual values ​​of multiple production condition parameters recorded for each well group during a specific production stage. These parameters include, but are not limited to, production pressure, fluid production, gas production, and wellhead pressure, constituting a multidimensional time-series dataset reflecting the dynamic behavior of the well group. To establish a quantitative correlation between operating parameters and sand production behavior, the actual sand production data of each well group within the corresponding time period can be acquired synchronously as a benchmark for subsequent modeling and analysis.

[0060] To further capture the nonlinear dependence between parameters and sand production rate, a specially designed nonlinear transformer can be configured for each production condition parameter. This transformer can be based on the physical background of the parameter itself and its possible role in the sand production mechanism, such as critical pressure effects, flow rate saturation characteristics, and decay trends. Structured nonlinear mapping can be applied to the raw data. For example, a piecewise function can be introduced for the production pressure parameter to highlight its critical behavior, while logarithmic or power transformations can be applied to the liquid production rate to enhance its nonlinear characteristics of sand-carrying capacity. Through such transformations, the raw data is converted into transformed distribution data with greater physical meaning and discriminative power.

[0061] Based on this, by calculating the similarity index between the data sequence after each parameter transformation and the sand production sequence of the well group, the intensity of the parameter's influence on sand production can be quantitatively assessed. To further unify the dimensions and facilitate comparison, the obtained similarity index can be normalized to finally generate the first sand production coefficient for this parameter. By introducing a nonlinear mapping mechanism, the limitations of traditional linear correlation analysis in describing complex mining responses are effectively overcome, significantly improving the ability to identify the main controlling factors of sand production and the interpretability of the model.

[0062] By configuring a nonlinear converter for each parameter, the calculation of the first sand output coefficient no longer depends on the linear assumption. It can effectively identify non-monotonic influence mechanisms such as the critical point of production pressure and the saturation range of liquid output, and significantly improve the physical rationality and prediction accuracy of the screening of the main control factors of sand output.

[0063] In some embodiments, the calculation of the second sand production coefficient for each production condition parameter in step S10 above may specifically include: determining the information measurement weight of each production condition parameter in each well group based on the historical distribution data of each production condition parameter in multiple well groups; determining the collinearity coefficient of each production condition parameter based on the historical distribution data of multiple production condition parameters in multiple well groups, wherein the collinearity coefficient characterizes the linear correlation between the production condition parameter and the remaining production condition parameters among the multiple production condition parameters; calculating the second sand production coefficient of each production condition parameter in each well group based on the information measurement weight of each production condition parameter in each well group and the collinearity coefficient of each production condition parameter; wherein the information measurement weight is positively correlated with the second sand production coefficient; and the collinearity coefficient is negatively correlated with the second sand production coefficient.

[0064] Based on the historical data distribution characteristics of various production parameters at multiple time points within different well groups, their corresponding information measurement weights can be calculated. These weights quantify the dispersion of parameter values ​​within a specific well group's historical time period and their contribution to sand production prediction. Information measurement weights can be measured using information theory indices such as Shannon entropy and distribution entropy. Lower entropy values ​​indicate a more concentrated data distribution and lower uncertainty for the parameter within the well group, resulting in a greater amount of effective information and thus a more significant role in local sand production mechanisms.

[0065] In some embodiments, the historical distribution data of various production condition parameters within different well groups can be standardized to eliminate the influence of dimensions. Then, the data is substituted into the following formula to calculate the information entropy, which measures the uncertainty of the factor; the lower the entropy value, the greater the factor weight.

[0066] ;

[0067] In the formula, This refers to the historical distribution data of the j-th generated operating condition parameter at time t in the i-th well group. The standardized value eliminates the influence of dimensions and has a range of (0,1). The information entropy of the j-th production condition parameter in the i-th well group measures the uncertainty of its value at different times. Through information entropy calculation, an objective assessment of the importance of each production condition parameter is achieved. This method is based on the inherent distribution characteristics of the parameter data, rather than relying on manual experience-based assignment, effectively avoiding the influence of subjective bias on factor selection. The magnitude of the information entropy value directly reflects the stability of the parameter's value at different time points. The lower the entropy value, the more regular the parameter's change over time, and the lower the uncertainty. This characteristic enables it to effectively screen out those control factors that have a continuous and stable impact on sand production, and eliminate those interfering parameters that fluctuate drastically and lack regularity, thereby improving the reliability of the prediction model. The weight calculation method based on information entropy can provide refined input parameters for subsequent prediction models. These parameters are not only highly important but also have good stability, significantly improving the accuracy and robustness of the prediction model.

[0068] To further avoid the interference of overlapping information among multiple parameters in identifying the controlling factors, a collinearity analysis mechanism can be introduced. By using methods such as variance inflation factor or eigenvalue decomposition based on the correlation coefficient matrix, the linear dependence strength between each parameter and the others can be accurately assessed, thereby identifying redundant information in the parameter set. A higher collinearity coefficient indicates that the information provided by that parameter can be more easily replaced by other parameters, and its value in independently explaining sand phenomena is lower.

[0069] Based on the separately obtained information measure weights and collinearity coefficients, these two are fused together with a negative correlation to construct a second sand production coefficient. Specifically, the higher the information measure weight and the lower the collinearity coefficient, the larger the second sand production coefficient. This integration mechanism ensures that parameters with high discriminative and independent characteristics in local well groups are preferentially identified as key control factors, thereby improving the adaptability and accuracy of the sand production prediction model under heterogeneous coal seam conditions and providing a reliable basis for differentiated management of well groups.

[0070] The second sand production coefficient integrates information measurement weights and collinearity corrections, which not only reflects the local sensitivity of parameters in specific well groups, but also suppresses the evaluation bias caused by multi-parameter collinearity. This supports the construction of a sand production prediction model for well groups that adapts to geological heterogeneity, and can provide a reliable basis for differentiated sand control decisions.

[0071] Overall, the first sand production coefficient identifies key control variables affecting sand production from a global perspective, while the second sand production coefficient reveals the spatial differentiation of this variable across different well groups at a local scale. The combination of these two approaches constructs a global-local two-layer assessment framework, effectively overcoming the shortcomings of traditional methods in complex, heterogeneous coalbed methane reservoirs, such as poor adaptability and weak explanatory power. This provides a more comprehensive and reliable data foundation for accurate quantitative prediction of overall coalbed methane production.

[0072] S20: Determine multiple key production parameters for each well group based on the first and second sand production coefficients.

[0073] In some embodiments, step S20 may specifically include: coupling the first sand production coefficient and the second sand production coefficient to obtain a third sand production coefficient for each production condition parameter at each well group; and selecting a plurality of master control production condition parameters from the plurality of production condition parameters based on the third sand production coefficient.

[0074] The first and second sand production coefficients for the same production condition parameters in the same well group can be coupled for calculation to generate a composite index—the third sand production coefficient—that comprehensively reflects both its global influence and local specificity. This coupling process can employ weighted fusion, geometric averaging, or rule-based comprehensive evaluation methods to ensure that both the overall importance of the parameter and its adaptability in a specific well group are taken into account.

[0075] After obtaining the third sand production coefficient of each parameter across all well groups, the main control parameters can be selected accordingly. Specifically, for each well group, all its operating parameters can be sorted from highest to lowest according to the third sand production coefficient, and the top K parameters can be selected as the main control production operating parameters for that well group. Alternatively, a threshold can be set, and parameters with a third sand production coefficient higher than this threshold can be identified as main control parameters. Through this process, a customized set of main control parameters can be determined for each well group.

[0076] By coupling the first sand production coefficient, which reflects the global impact of the parameter, with the second sand production coefficient, which characterizes the specificity of the well group, the third sand production coefficient effectively balances the general regularity and local anomalies of the parameter. This avoids the drawbacks of relying solely on global coefficients, which may ignore key local factors, or relying solely on local coefficients, which may amplify random errors. This ensures that the selected master control parameters are consistent with the overall mechanism understanding and closely reflect the actual production dynamics of the well group.

[0077] Furthermore, the screening mechanism based on the third sand production coefficient ensures that the final selected parameter set not only has significant influence but also possesses a certain degree of independence and representativeness. This helps improve the interpretability of the subsequently constructed sand production prediction model and provides clear and reliable input for formulating targeted sand control measures, facilitating refined production management with a tailored approach for each well.

[0078] S30: Based on the current distribution data of multiple main control production parameters of each well group, predict the sand production of that well group.

[0079] In some embodiments, step S30 may specifically include: constructing a multivariate regression prediction model for the well group based on the historical distribution data of multiple main control production condition parameters of each well group at the well group and the historical sand production data of the well group; inputting the current distribution data of multiple main control production condition parameters of the well group into the multivariate regression prediction model of the well group to obtain the sand production prediction data of the well group.

[0080] For each well group, historical distribution data of its multiple key production parameters can be used as independent variables, while historical sand production data corresponding to that well group can be used as the dependent variable. A multivariate regression prediction model specific to that well group can be constructed using regression analysis. This model aims to establish a mathematical mapping relationship between key parameters and sand production. Its construction fully considers the interaction effects between parameters and possible nonlinear characteristics, and regularization and other methods can be introduced to verify the robustness and effectiveness of the model.

[0081] After the model is built and validated, the current distribution data of multiple key production parameters of the well group, collected in real time, can be input into its corresponding prediction model to calculate the predicted sand production of the well group over a certain period in the future. This process can realize the transformation from static historical analysis to dynamic real-time prediction, providing direct and quantitative basis for sand control decisions at the production site.

[0082] In some embodiments, the sand production multiple regression model matrix representation for each well group in a coalbed methane well field can be established as follows:

[0083] ;

[0084] In the formula, Let the sand production quality of the i-th group of gas wells be denoted as . These are the regression coefficients for the sand production model in gas wells; To convert nonlinear influencing factors into gas well sand production quality in the regression model. The linearization term; This is the residual.

[0085] In some embodiments, regression coefficients can be calculated, and a sand output quality prediction model can be determined:

[0086] ;

[0087] In the formula: This represents the predicted sand production quality of the i-th group of gas wells; This is an estimated value of the regression coefficient for sand output quality; The operating parameters are generated for the t-th master control of the i-th gas well. The above-mentioned multivariate regression-based sand production quality prediction model can maintain good prediction accuracy while having the advantages of high computational efficiency and convenient implementation. It only requires conventional production data to complete modeling and prediction, making it suitable for rapid field application and meeting real-time decision-making needs.

[0088] By constructing independent regression models for each well group based on its own master control parameters, the spatial heterogeneity of coalbed methane reservoirs is fully respected, overcoming the limitations of a one-size-fits-all prediction model. This personalized modeling approach significantly improves the fit of prediction results to the specific geological and production conditions of well groups, thereby enhancing the accuracy of sand production forecasts.

[0089] Because the input variables are only a few key control parameters selected through prior scientific screening, rather than all operating condition parameters, the structure of the multiple regression model is simplified. This not only reduces the risk of model overfitting and improves computational efficiency, but also enhances the model's physical transparency and decision support value.

[0090] In some embodiments, calculating the first sand production coefficient for each production condition parameter based on the similarity between the transformed distribution data of the plurality of production condition parameters and the historical sand production data of multiple well groups in the coal seam can specifically include: calculating the first sand production coefficient for each production condition parameter using the following formula based on the similarity between the transformed distribution data of the plurality of production condition parameters and the historical sand production data of multiple well groups in the coal seam:

[0091] ;

[0092] In the formula, For the first The first sand output coefficient of each production condition parameter; For a specific moment; The time lag is a preset set of time lags. ; It is a nonlinear converter. It belongs to the pre-defined set of nonlinear converters; This represents the total number of coal seam well groups. for Time of the first The production condition parameter is at the... Historical distribution data of each well group; for Time of the first Historical sand production data for each well group.

[0093] The above formula, by incorporating time-delay effect analysis, can accurately capture the delayed response mechanism of sand. Specifically, the above formula introduces a time-delay parameter. and in the preset time delay set This study systematically examined the time lag effect between changes in production parameters and the sand production response. Sand production in coalbed methane wells is a dynamic process; the impact of changes in production parameters on formation stability and the detection of sand particle migration at the wellhead both involve a certain time lag. Traditional methods ignore this time lag, potentially leading to underestimation or even misjudgment of the importance of certain parameters. The above formula, by introducing time lag effect analysis, can accurately capture the delayed response mechanism of sand production. Specifically, the above formula introduces a time lag parameter... and in the preset time delay set The study systematically examined the time delay effect between changes in production parameters and the sand output response. By automatically identifying the time difference corresponding to the most significant correlation between each parameter and the sand output, the study more realistically and accurately reflects the actual causal relationship mechanism, greatly improving the physical rationality and accuracy of factor identification.

[0094] Introducing a nonlinear transformation helps to reveal the complex laws governing nonlinear effects. Specifically, the above formula introduces a nonlinear transformer. and in the preset function set The search is performed on the original operating condition parameter data. Nonlinear mapping is employed. The relationship between coalbed methane sand production and many operating parameters is not a simple linear one, but may exhibit critical points, such as a sharp increase in sand production after the pressure exceeds a certain threshold, or a saturation effect where the sand-carrying capacity slows down after the production volume increases to a certain level. Traditional linear correlation coefficients cannot effectively capture these patterns. By automatically adapting to the mathematical expression that best reveals the underlying nonlinear physical laws, the implicit nonlinear correlations can be made explicit, thus overcoming the limitations of traditional linear analysis models and revealing the intrinsic mechanism of sand production more profoundly.

[0095] The above formula is passed through This dual optimization operation seeks the optimal combination of time delay and optimal nonlinear transformation that maximizes the correlation between the operating parameters and the sand output. It elevates the factor identification problem from a static, linear judgment to a dynamic, nonlinear pattern search problem. Through this global search strategy, it is possible to tailor the nonlinear representation and time delay scale most relevant to the sand output for each parameter. The final result... This is the correlation strength between the parameter and the sand output under the best observation perspective. This result is far more reliable and powerful than the correlation coefficient under a single perspective, ensuring that the first sand output coefficient can truly screen out the key controlling factors with deep physical connections.

[0096] Overall, a powerful and flexible correlation analysis framework was constructed using three major techniques: time delay analysis, nonlinear transformation, and dual optimization search. This significantly enhances the ability to detect nonlinear causal relationships between variables in complex dynamic systems, laying a solid theoretical foundation for accurately and quantitatively identifying the main controlling factors of sand production in deep coalbed methane.

[0097] In some embodiments, the calculation of the second sand production coefficient for each production condition parameter at each well group based on the information measure weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter may specifically include: calculating the second sand production coefficient for each production condition parameter at each well group using the following formula based on the information measure weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter:

[0098] ;

[0099] In the formula, For the first The production condition parameter is at the... The second sand production coefficient at each well group; For a specific moment; The time lag is a preset set of time lags. ; for Time of the first The production condition parameter is at the... Information measurement weights for each well group; is the collinearity coefficient of the i-th production condition parameter; This is the preset threshold for the collinearity coefficient.

[0100] The above formula can capture the time-delay effect of local influences through a dynamic time-matching mechanism. Specifically, the above formula uses... It can be within a preset time delay set Find an optimal time delay that maximizes the computational result. This is because the impact of production parameters on a specific well group may take time. For example, changes in wellhead pressure may not immediately cause sand production, but rather show an effect at the bottom of the well after a period of time. This mechanism can automatically match the time difference between parameter changes and the most significant sand production response in that well group, ensuring that the assessed parameter has the most realistic and intense local impact, thus more accurately characterizing the different dynamic response processes that may exist in different well groups due to differences in geological conditions.

[0101] The above formula, by balancing local information with global independence, can highlight the key production parameters for each well group. Specifically, the numerator of the above formula is... .in, Characterizes the uncertainty or disorder of parameters in specific well group data (such as entropy). The smaller the value, the more concentrated and regular the data distribution of this parameter within the well group, and the higher its value in terms of local information. Therefore, It is an indicator that characterizes the clarity or importance of local information. This ensures that parameters with regular data distribution and that provide clear information receive a higher base score in the evaluation of the well group. It focuses on the performance of parameters in a specific well group, rather than their global average performance, thus enhancing its adaptability to geological heterogeneity.

[0102] Furthermore, the above formula can suppress the weights of redundant information through adaptive collinearity penalty. Specifically, the denominator of the above formula... It is an adaptive penalty term. Among them, It is the first The collinearity coefficients of the parameters measure the degree of information overlap between them and other parameters. It is a preset threshold, representing the upper limit of tolerable collinearity. When When the penalty value is 1, no penalty is applied. When the penalty is greater than 1, and increases with the percentage exceeding the limit, it reduces the final penalty. This approach effectively addresses the multicollinearity problem. If a parameter is locally important (large numerator) but its information can be linearly interpreted by other parameters (high collinearity), its importance is reduced by the denominator penalty. This forces the model to prioritize parameters that are both locally important and provide independent information as controlling factors, thus preventing the model from becoming unstable and difficult to interpret due to redundant input parameters, and improving the robustness of the final prediction equation.

[0103] Overall, the above formula is passed through The optimal time window is identified, and the ratio of local information clarity to information redundancy of the parameter within that window is calculated. This maximum value is the second sand production coefficient. For a parameter to achieve a high second sand production coefficient, three conditions must be met simultaneously: 1) it must have a significant and consistent impact on the well group; 2) this impact must be independent and cannot be replaced by other parameters; and 3) its impact must be most prominent at a specific time lag. This makes the second sand production coefficient an excellent indicator for accurately quantifying the unique contribution of production condition parameters in a specific well group.

[0104] In some embodiments, the coupling of the first sand production coefficient and the second sand production coefficient to obtain a third sand production coefficient for each production condition parameter at each well group may specifically include: coupling the first sand production coefficient and the second sand production coefficient using the following formula to obtain a third sand production coefficient for each production condition parameter at each well group:

[0105] ;

[0106] In the formula, For the first The production condition parameter is at the... The third sand production coefficient at each well group; For the coal seam development time; This is the preset scaling factor; This is a preset balance coefficient greater than 0; For the first The first sand output coefficient of each production condition parameter; For the first The production condition parameter is at the... The second sand production coefficient at each well group.

[0107] The above formula enables adaptive weight transfer from global experience to local data. Specifically, in the above formula, represents the first sand output coefficient. Second sand output coefficient Allocated according to development time Dynamically changing weights. Among them, The weight is The weight of ; is When development time When smaller, Close to 1, and The value is close to 0. This means that in the early stages of gas field development, production data from individual well groups is limited, and local patterns are not obvious. Therefore, it relies more on universally applicable global patterns learned from data from all well groups across the entire coal seam to screen for controlling factors, effectively overcoming the bottleneck of insufficient early data. However, in the later stages of development, with... Increase Decay to near 0, The weight decreased to ,and The weight then rises to At this point, with the accumulation of a large amount of local production data in the well group, the model tends to trust local patterns that reflect its unique geological and production characteristics. This precisely aligns with the objective process of development understanding moving from macro to micro, and from the general to the specific.

[0108] Furthermore, the balance coefficient in the above formula This determined the overall experience in the later stages of development. The base weights that can be retained. (Settings) A value greater than 0 ensures that even in the later stages of development, robust global insights validated across all gas wells can still be effective, preventing the model from being completely skewed by noise or short-term fluctuations in local data and enhancing the robustness of decision-making.

[0109] Meanwhile, the scaling factor in the above formula The speed at which weights are transferred from the global to the local level is controlled. For gas reservoirs with rapidly changing geological conditions and strong heterogeneity, a larger weight can be set. This allows for a rapid transition to a locally dependent model; for relatively homogeneous, stable-producing gas reservoirs, a smaller [model] can be set. This allows for broader application of experience over a longer period. Secondly, utilizing an exponential function further smooths and continuously changes the weights, avoiding potential abrupt changes in model output and unstable prediction results caused by hard weight switching at a specific point in time. It ensures that the set of controlling factors and the final sand production prediction results evolve gradually with the development process, which aligns with the continuous nature of gas field production and facilitates stable production management.

[0110] These two parameters provide a control lever, enabling the coupling strategy to be flexibly customized according to the specific geological conditions and development strategies of different coalbed methane reservoirs, thereby enhancing the universality and engineering applicability of the method.

[0111] Overall, the above formula, by introducing a time variable, intelligently balances the contributions of global prior knowledge and local real-time data at different development stages. This enables the selection of key control factors to overcome data sparsity problems in the early stages and accurately characterize local features in the later stages, thereby achieving adaptive and precise identification of key sand production control factors throughout the entire gas field lifecycle.

[0112] In some embodiments, determining the information measurement weight of each production condition parameter in each well group based on the historical distribution data of each production condition parameter across multiple well groups may specifically include: acquiring geological feature data for each well group, the geological feature data including at least coal body structure type, burial depth grade, and structural complexity index; calculating the basic information entropy weight of each production condition parameter based on the degree of dispersion of its distribution in the historical distribution data of the well group; determining the geological feature fit of the production condition parameter based on the similarity between the geological feature data of each well group and the typical geological feature data of the coal seam; calculating the dynamic sensitivity coefficient of each production condition parameter based on the coefficient of variation and trend coefficient of its historical distribution data in the well group; and weighting and fusing the basic information entropy weight with the geological feature fit and dynamic sensitivity coefficient to obtain the information measurement weight of each production condition parameter in each well group.

[0113] It is possible to obtain detailed geological feature data for each well group. This data includes at least the coal body structure type, such as primary structured coal, fractured coal, and granular coal; burial depth level, such as classification based on formation pressure gradient; and structural complexity index, which can be calculated by combining parameters such as fault density and formation dip angle. These geological features constitute the unique geological fingerprint of the well group, providing a basis for subsequent weighting.

[0114] Based on the dispersion of each production condition parameter in the historical distribution data of a specific well group, the basic information entropy weight can be calculated using the information entropy algorithm. This weight reflects the statistical information content of the parameter data. The similarity between the actual geological feature data of each well group and the database of typical geological features of coal seams can be compared, and a similarity function based on Euclidean distance can be used for quantitative evaluation to determine the matching degree between the well group and various standard geological models, thus obtaining the geological feature fit with the production condition parameter. Simultaneously, the dynamic sensitivity coefficient of each production condition parameter is calculated. Specifically, the coefficient of variation of the parameter in the historical data of the well group and the trend coefficient obtained through linear regression can be comprehensively considered to capture the dynamic change characteristics of the parameter. The coefficient of variation reflects the amplitude of data fluctuations, while the trend coefficient reflects the direction and rate of parameter change. The basic information entropy weight, geological feature fit, and dynamic sensitivity coefficient can be weighted and fused after normalization, with the weight coefficients adjusted according to specific coal seam characteristics, ultimately yielding an information measurement weight that comprehensively reflects the importance of the parameter in a specific well group.

[0115] By introducing geological feature adaptation, the weight calculation fully considers the geological heterogeneity of different well groups. The introduction of a dynamic sensitivity coefficient allows the weight calculation to move beyond static historical data distribution and capture real-time parameter trends. When a parameter exhibits abnormal fluctuations or significant trend changes, its weight is adjusted accordingly, providing a more timely and sensitive indicator for sand production early warning. The weighted fusion mechanism of these three factors effectively avoids biases that may arise from a single indicator. The basic information entropy weight ensures statistical reliability, the geological feature adaptation guarantees geological rationality, and the dynamic sensitivity coefficient provides real-time sensitivity; the synergistic effect of these three factors significantly improves the accuracy and stability of the weight calculation.

[0116] In some embodiments, the above-mentioned selection of multiple master production condition parameters from the multiple production condition parameters based on the third sand production coefficient may specifically include: determining a sand production coefficient threshold for each well group based on the statistical distribution of multiple third sand production coefficients corresponding to multiple production condition parameters of each well group; selecting multiple production condition parameters of the well group according to the sand production threshold to obtain multiple candidate production condition parameters; performing collinear clustering analysis on the multiple candidate production condition parameters to obtain multiple information clusters, each information cluster including multiple candidate production condition parameters; and taking the candidate production condition parameter with the highest third sand production coefficient in each information cluster as the master production condition parameter.

[0117] For each specific well group, the statistical distribution characteristics of the third sand production coefficient corresponding to all its production operating parameters can be analyzed to determine an adaptive sand production coefficient threshold. This threshold can dynamically reflect the significance level of parameters within the well group. Based on this threshold, all operating parameters within the well group can be initially screened, retaining parameters not lower than the threshold as candidate production operating parameters, forming a preliminary set of important parameters. Collinearity clustering analysis is performed on the screened candidate parameters. By calculating the correlation matrix between parameters, highly correlated parameters are grouped into the same information cluster, thereby identifying redundant information among parameters. Within each information cluster, the parameter with the highest third sand production coefficient is selected as the representative of that cluster and determined as the main control production operating parameter, ensuring that each selected parameter has the highest independent representativeness.

[0118] By determining dynamic thresholds based on the distribution of the third sand production coefficient of each well group, this method overcomes the adaptability differences of fixed threshold or fixed-quantity screening methods across different well groups. This allows the screening criteria to fit the unique geological conditions and production dynamics of each well group, effectively distinguishing parameters that truly have a significant impact within the context of that well group, thus improving the accuracy and specificity of the screening. By selecting only the most representative parameters from each information cluster, multicollinearity is eliminated to the greatest extent while retaining core information. This provides independent and complementary input variables for the subsequently constructed prediction model, significantly enhancing the mathematical stability of the model and the reliability of the prediction results. The final set of key control parameters not only has high statistical significance but also low redundancy. This allows for a clearer interpretation of the key driving factors affecting the sand production behavior of the well group, providing a direct and clear basis for making precise decisions such as optimizing production systems and implementing targeted sand control measures, thereby improving the practicality and efficiency of the entire sand production prediction and management process.

[0119] As can be seen from the coal seam sand production prediction method provided in the embodiments of this specification above, the embodiments of this specification can calculate the first sand production coefficient and the second sand production coefficient for each production condition parameter based on the historical distribution data of multiple production condition parameters in multiple well groups of the coal seam. The first sand production coefficient characterizes the degree of influence of the production condition parameter on the coal seam sand production; the second sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of each well group. Based on the first and second sand production coefficients, multiple main control production condition parameters for each well group are determined. Based on the current distribution data of the multiple main control production condition parameters for each well group, the sand production of the well group is predicted. By calculating the first and second sand production coefficients respectively, a comprehensive factor evaluation system from the overall coal seam to individual well groups is constructed, overcoming the shortcomings of traditional methods that only consider global laws and ignore local characteristics. It can simultaneously capture the general laws and special performances of parameters, ensuring that the selection of main control factors conforms to the overall trend and adapts to the differences in local geological and development conditions. Furthermore, the selected main control production condition parameters have clear physical meaning and regional adaptability, enabling the subsequent prediction model to fully consider the uniqueness of each well group. This well-specific forecasting approach significantly improves the accuracy and reliability of sand production prediction, providing precise data support for developing differentiated sand control measures and effectively avoiding the problem of insufficient adaptability of the traditional one-size-fits-all method in complex coalbed methane reservoirs.

[0120] The following is a specific embodiment of this specification:

[0121] Production conditions and operational data of a coalbed methane well site were obtained. The factors affecting the sand production quality of coalbed methane were production pressure, liquid production, gas production, and liquid consumption intensity. When analyzing the main controlling factors of sand production quality for each well group in the coal seam, the first and second sand production coefficients were calculated, and then the third sand production coefficient was obtained by combining them. The relevant calculation results for a certain well group are shown in Table 1.

[0122] Table 1

[0123]

[0124] Based on the above multi-dimensional data analysis and comprehensive consideration, it can be clearly observed that production pressure and fluid production are the main control parameters for the sand production quality of this well group.

[0125] Based on the range of changes in production pressure and fluid production during production operations, a multiple regression model for sand production quality in gas wells can be established as follows:

[0126] ;

[0127] In the formula: The sand output quality of the well group is expressed in g / d. Well production pressure, MPa; The daily fluid production of the well group is expressed in m³ / d. It is a dimensionless number and is related to factors such as coal seam characteristics and fracturing technology. A dimensionless number, representing the pressure coefficient; is a dimensionless number, representing the liquid production coefficient.

[0128] Based on the calculated regression coefficients and multiple regression method, the main control production parameters for multiple well groups, namely production pressure and fluid production rate, were fitted together. The fitting results are as follows:

[0129] Low-pressure, low-production well group (P≤3.0MPa, Q) W ≤5m³ / d):

[0130] ;

[0131] Low-pressure, high-yield fluid well group (P≤3.0MPa, QW>5m³ / d):

[0132] ;

[0133] High-pressure, low-production fluid well group (P > 3.0 MPa, QW ≤ 5 m³ / d):

[0134] ;

[0135] High-pressure, high-yield fluid well group (P > 3.0 MPa, QW > 5 m³ / d):

[0136] ;

[0137] Reference Figure 3 The sand production factors of the four well groups were analyzed:

[0138] Low-pressure, low-production well groups: Pressure differential-sensitive failure. Under low-pressure, low-production conditions, the production pressure differential becomes the core controlling factor. When the pressure differential exceeds the critical value (approximately 2.8 MPa), a significant shear failure zone will form near the wellbore, inducing sudden sheet-like spalling.

[0139] Low-pressure, high-yield fluid well groups: fluid dragging is dominant. The low confining pressure weakens the capillary binding force of formation particles, while the high fluid production significantly enhances the fluid scouring effect.

[0140] High-pressure, low-production fluid well groups: Rock stability control. The high-pressure environment enhances the confining pressure, suppressing large-scale damage, but local weak cementation zones are still prone to instability.

[0141] High-pressure, high-yield fluid well groups: Dynamic equilibrium effect. High pressure conditions provide confining pressure support, promoting sand grain reorganization to form a stable sand arch. As the fluid production increases, the two-phase flow enhances the sand-carrying capacity. However, a sudden increase in pressure differential can disrupt the sand arch equilibrium and increase sand production.

[0142] Based on the above-described method for predicting coal seam sand production, this specification also provides embodiments of a coal seam sand production prediction device. For example... Figure 4 As shown, the coal seam sand production prediction device 400 may specifically include the following modules:

[0143] Calculation module 401 is used to calculate a first sand production coefficient and a second sand production coefficient for each production condition parameter based on historical distribution data of multiple production condition parameters in multiple well groups of coal seams; the first sand production coefficient represents the degree of influence of the production condition parameter on the amount of sand produced in the coal seam; the second sand production coefficient represents the degree of influence of the production condition parameter on the amount of sand produced in each well group.

[0144] The determination module 402 is used to determine multiple main control production condition parameters for each well group based on the first sand production coefficient and the second sand production coefficient.

[0145] The prediction module 403 is used to predict the sand production of each well group based on the current distribution data of multiple main control production condition parameters of each well group.

[0146] In some embodiments, the calculation module 401 described above can be specifically used for:

[0147] Obtain historical sand production data for multiple well groups in the coal seam corresponding to the historical distribution data;

[0148] Determine the nonlinear converter corresponding to each production condition parameter;

[0149] Based on the nonlinear transformer, the historical distribution data of the production condition parameter is nonlinearly transformed to obtain the transformed distribution data of the production condition parameter.

[0150] Based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam, the first sand production coefficient of each production condition parameter is calculated.

[0151] In some embodiments, the calculation module 401 described above can also be used for:

[0152] Based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam, the first sand production coefficient of each production condition parameter is calculated using the following formula:

[0153] ;

[0154] In the formula, For the first The first sand output coefficient of each production condition parameter; For a specific moment; The time lag is a preset set of time lags. ; It is a nonlinear converter. It belongs to the pre-defined set of nonlinear converters; This represents the total number of coal seam well groups. for Time of the first The production condition parameter is at the... Historical distribution data of each well group; for Time of the first Historical sand production data for each well group.

[0155] In some embodiments, the calculation module 401 described above can also be used for:

[0156] Based on the historical distribution data of each production condition parameter across multiple well groups, determine the information measurement weight of each production condition parameter in each well group.

[0157] Based on the historical distribution data of multiple production operating parameters in multiple well groups, the collinearity coefficient of each production operating parameter is determined. The collinearity coefficient represents the linear correlation between the production operating parameter and the remaining production operating parameters among the multiple production operating parameters.

[0158] Based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter, the second sand production coefficient of each production condition parameter at each well group is calculated; the information measurement weight is positively correlated with the second sand production coefficient; the collinearity coefficient is negatively correlated with the second sand production coefficient.

[0159] In some embodiments, the calculation module 401 described above can also be used for:

[0160] Based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter, the second sand production coefficient of each production condition parameter at each well group is calculated using the following formula:

[0161] ;

[0162] In the formula, For the first The production condition parameter is at the... The second sand production coefficient at each well group; For a specific moment; The time lag is a preset set of time lags. ; for Time of the first The production condition parameter is at the... Information measurement weights for each well group; is the collinearity coefficient of the i-th production condition parameter; This is the preset threshold for the collinearity coefficient.

[0163] In some embodiments, the determining module 402 described above can be specifically used for:

[0164] By coupling the first sand production coefficient with the second sand production coefficient, a third sand production coefficient for each production condition parameter at each well group is obtained;

[0165] Based on the third sand output coefficient, multiple main control production condition parameters are selected from the multiple production condition parameters.

[0166] In some embodiments, the determining module 402 may also be used for:

[0167] The third sand production coefficient for each production condition parameter at each well group is obtained by coupling the first sand production coefficient with the second sand production coefficient using the following formula:

[0168] ;

[0169] In the formula, For the first The production condition parameter is at the... The third sand production coefficient at each well group; For the coal seam development time; This is the preset scaling factor; This is a preset balance coefficient greater than 0; For the first The first sand output coefficient of each production condition parameter; For the first The production condition parameter is at the... The second sand production coefficient at each well group.

[0170] In some embodiments, the prediction module 403 described above can be specifically used for:

[0171] Based on the historical distribution data of multiple key production condition parameters of each well group and the historical sand production data of the well group, a multivariate regression prediction model for the well group is constructed.

[0172] The current distribution data of multiple key production parameters of the well group are input into the multivariate regression prediction model of the well group to obtain the sand production prediction data of the well group.

[0173] As can be seen from the coal seam sand production prediction device provided in the embodiments of this specification above, the embodiments of this specification can calculate the first sand production coefficient and the second sand production coefficient for each production condition parameter based on the historical distribution data of multiple production condition parameters in multiple well groups of the coal seam. The first sand production coefficient characterizes the degree of influence of the production condition parameter on the coal seam sand production; the second sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of each well group. Based on the first and second sand production coefficients, multiple main control production condition parameters for each well group are determined; based on the current distribution data of the multiple main control production condition parameters for each well group, the sand production of the well group is predicted. By calculating the first and second sand production coefficients respectively, a comprehensive factor evaluation system from the overall coal seam to individual well groups is constructed, overcoming the shortcomings of traditional methods that only consider global laws and ignore local characteristics. It can simultaneously capture the general laws and special performances of parameters, ensuring that the selection of main control factors conforms to the overall trend and adapts to the differences in local geological and development conditions. Furthermore, the selected main control production condition parameters have clear physical meaning and regional adaptability, enabling the subsequent prediction model to fully consider the uniqueness of each well group. This well-specific forecasting approach significantly improves the accuracy and reliability of sand production prediction, providing precise data support for developing differentiated sand control measures and effectively avoiding the problem of insufficient adaptability of the traditional one-size-fits-all method in complex coalbed methane reservoirs.

[0174] This specification also provides a computer device for predicting coal seam sand production, including a processor and a memory for storing processor-executable instructions. Specifically, the processor can perform the following tasks according to the instructions: calculating a first sand production coefficient and a second sand production coefficient for each production condition parameter based on historical distribution data of multiple production condition parameters across multiple well groups in the coal seam; the first sand production coefficient characterizes the degree of influence of the production condition parameter on the coal seam sand production; the second sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of each well group; determining multiple main control production condition parameters for each well group based on the first and second sand production coefficients; and predicting the sand production of the well group based on the current distribution data of the multiple main control production condition parameters for each well group.

[0175] To execute the above instructions more accurately, please refer to... Figure 5 As shown in the embodiments of this specification, another specific computer device 500 is also provided, wherein the computer device 500 includes a network communication port 501, a processor 502 and a memory 503, and the above structures are connected by internal cables so that the various structures can perform specific data interaction.

[0176] The processor 502 can be specifically used to: calculate a first sand production coefficient and a second sand production coefficient for each production condition parameter based on historical distribution data of multiple production condition parameters in multiple well groups of coal seams; the first sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of the coal seam; the second sand production coefficient characterizes the degree of influence of the production condition parameter on the sand production of each well group; determine multiple main control production condition parameters for each well group based on the first sand production coefficient and the second sand production coefficient; and predict the sand production of the well group based on the current distribution data of the multiple main control production condition parameters for each well group.

[0177] The memory 503 can be used to store the corresponding instruction program.

[0178] In this embodiment, the network communication port 501 can be a virtual port bound to different communication protocols, thereby enabling the sending or receiving of different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0179] In this embodiment, the processor 502 can be implemented in any suitable manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc. This specification is not limiting.

[0180] In this embodiment, the memory 503 includes volatile memory and non-volatile memory. The memory 503 can include multiple layers. In digital systems, anything that can store binary data can be a memory; in integrated circuits, a circuit with storage function but no physical form is also called a memory, such as RAM, FIFO, etc.; in a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0181] This specification also provides a computer program product, including at least one instruction or at least one program segment, wherein the at least one instruction or the at least one program segment is loaded and executed by a processor to achieve the following: Figure 1 The method shown.

[0182] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0183] It should also be understood that, in the embodiments of this specification, the terms and / or are merely descriptions of the relationships between related objects, indicating that three relationships may exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this specification generally indicates that the preceding and following related objects have an "or" relationship.

[0184] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0185] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0186] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0187] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational tasks to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The task is a function specified in one or more boxes.

[0188] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting the amount of sand produced in a coal seam, characterized in that, include: Based on the historical distribution data of multiple production condition parameters in multiple well groups of coal seams, calculate the first sand production coefficient and the second sand production coefficient for each production condition parameter. The first sand production coefficient characterizes the degree of influence of the production condition parameter on the amount of sand produced from the coal seam; The second sand production coefficient characterizes the degree of influence of this production condition parameter on the sand production of each well group; Based on the first sand production coefficient and the second sand production coefficient, multiple key production condition parameters for each well group are determined. Based on the current distribution data of multiple key production parameters for each well group, predict the sand production of that well group.

2. The method according to claim 1, characterized in that, The calculation of the first sand output coefficient for each production condition parameter includes: Obtain historical sand production data for multiple well groups in the coal seam corresponding to the historical distribution data; Determine the nonlinear converter corresponding to each production condition parameter; Based on the nonlinear transformer, the historical distribution data of the production condition parameter is nonlinearly transformed to obtain the transformed distribution data of the production condition parameter. Based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam, the first sand production coefficient of each production condition parameter is calculated.

3. The method according to claim 2, characterized in that, The step of calculating the first sand production coefficient for each production condition parameter based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam includes: Based on the similarity between the transformed distribution data of the multiple production condition parameters and the historical sand production data of multiple well groups in the coal seam, the first sand production coefficient of each production condition parameter is calculated using the following formula: ； In the formula, For the first The first sand output coefficient of each production condition parameter; For a specific moment; The time lag is a preset set of time lags. ; It is a nonlinear converter. It belongs to the pre-defined set of nonlinear converters; This represents the total number of coal seam well groups. for Time of the first The production condition parameter is at the... Historical distribution data of each well group; for Time of the first Historical sand production data for each well group.

4. The method according to claim 1, characterized in that, The calculation of the second sand output coefficient for each production condition parameter includes: Based on the historical distribution data of each production condition parameter across multiple well groups, determine the information measurement weight of each production condition parameter in each well group. Based on the historical distribution data of multiple production operating parameters in multiple well groups, the collinearity coefficient of each production operating parameter is determined. The collinearity coefficient represents the linear correlation between the production operating parameter and the remaining production operating parameters among the multiple production operating parameters. Based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter, the second sand production coefficient of each production condition parameter at each well group is calculated; the information measurement weight is positively correlated with the second sand production coefficient; the collinearity coefficient is negatively correlated with the second sand production coefficient.

5. The method according to claim 4, characterized in that, The calculation of the second sand production coefficient for each production condition parameter at each well group, based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter, includes: Based on the information measurement weight of each production condition parameter at each well group and the collinearity coefficient of each production condition parameter, the second sand production coefficient of each production condition parameter at each well group is calculated using the following formula: ； In the formula, For the first The production condition parameter is at the... The second sand production coefficient at each well group; For a specific moment; The time lag is a preset set of time lags. ; for Time of the first The production condition parameter is at the... Information measurement weights for each well group; is the collinearity coefficient of the i-th production condition parameter; This is the preset threshold for the collinearity coefficient.

6. The method according to claim 1, characterized in that, The determination of multiple key production condition parameters for each well group based on the first and second sand production coefficients includes: By coupling the first sand production coefficient with the second sand production coefficient, a third sand production coefficient for each production condition parameter at each well group is obtained; Based on the third sand output coefficient, multiple main control production condition parameters are selected from the multiple production condition parameters.

7. The method according to claim 6, characterized in that, The coupling of the first sand production coefficient and the second sand production coefficient yields the third sand production coefficient for each production condition parameter at each well group, including: The third sand production coefficient for each production condition parameter at each well group is obtained by coupling the first sand production coefficient with the second sand production coefficient using the following formula: ； In the formula, For the first The production condition parameter is at the... The third sand production coefficient at each well group; For the coal seam development time; This is the preset scaling factor; This is a preset balance coefficient greater than 0; For the first The first sand output coefficient of each production condition parameter; For the first The production condition parameter is at the... The second sand production coefficient at each well group.

8. The method according to claim 1, characterized in that, The step of predicting the sand production of a well group based on the current distribution data of multiple key production parameters for each well group includes: Based on the historical distribution data of multiple key production condition parameters of each well group and the historical sand production data of the well group, a multivariate regression prediction model for the well group is constructed. The current distribution data of multiple key production parameters of the well group are input into the multivariate regression prediction model of the well group to obtain the sand production prediction data of the well group.

9. A device for predicting the amount of sand produced in a coal seam, characterized in that, include: The calculation module is used to calculate the first and second sand production coefficients for each production condition parameter based on the historical distribution data of multiple production condition parameters in multiple well groups of coal seams. The first sand production coefficient characterizes the degree of influence of the production condition parameter on the amount of sand produced from the coal seam; The second sand production coefficient characterizes the degree of influence of this production condition parameter on the sand production of each well group; The determination module is used to determine multiple main control production condition parameters for each well group based on the first sand production coefficient and the second sand production coefficient. The prediction module is used to predict the sand production of each well group based on the current distribution data of multiple main control production condition parameters of each well group.

10. A computer device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the method of any one of claims 1-8.

Images