Coal seam roof pre-grouting optimization method based on data driving

Abstract

The invention discloses a coal seam roof pre-grouting optimization method based on data driving. The coal seam roof pre-grouting optimization method comprises the following steps: S1, fine exploration and original information extraction of a geologic body in a research area; s2, construction information extraction; s3, performing grouting project classification by adopting a DBSCAN clustering algorithm; s4, establishing a quantitative relation model based on a classification result; the method is suitable for the technical field of coal water prevention and control, multi-source geological exploration data and dynamic grouting engineering parameters are fused, a geological-engineering feature coupling model is established, a nonlinear association rule of a geological structure and grouting response is quantitatively analyzed based on a machine learning algorithm, and the method has the advantages of being high in practicability and high in reliability. And a multi-objective collaborative optimization system is developed, and intelligent matching of grouting parameters, accurate prediction of the construction effect and dynamic optimization of an engineering scheme are achieved. Finally, the technical purposes of improving the historical data utilization rate, enhancing the grouting and water plugging success rate and reducing the engineering cost are achieved, and scientific decision support is provided for coal mine water disaster prevention and control.

Description

Technical Field

[0001] The present invention belongs to the technical field of coal mine water control, and particularly relates to an optimization method for pre-grouting of coal seam roof based on data-driven. Background Technique

[0002] In recent years, as the large-scale exploitation stage of Jurassic coalfields in western China has been entered, the prevention and control of water hazards in coal seam roofs has become a key technical problem restricting the safe and efficient production of mines. As the core technology for blocking the hydraulic connection of overlying aquifers, the pre-grouting in the plane before mining exposes the following technical bottlenecks in engineering practice:

[0003] Problem of data value attenuation: Limited by the construction period of several months, the existing technology lacks the full-process integration of multi-source heterogeneous data generated by grouting projects (including geological exploration data, grouting process parameters and real-time monitoring data), resulting in insufficient utilization of historical engineering data and forming the "data island" effect.

[0004] Problem of fuzzy mechanism understanding: There is a complex non-linear coupling relationship between geological structure parameters (such as fracture development degree, rock mass permeability coefficient) and grouting response parameters (such as leakage characteristics, grouting volume, pressure characteristics and slurry characteristics), and traditional empirical formulas cannot establish a quantitative mapping model.

[0005] Problem of lack of decision-making intelligence: The scheme design relies on the discrete knowledge transfer of the engineer experience library, and lacks a data-driven intelligent decision support system, resulting in significant subjective deviations in the optimization of grouting parameters (slurry ratio, grouting pressure threshold, borehole layout density).

[0006] Problem of inaccurate effect prediction: The existing prediction model does not consider the spatial variability of formation parameters, resulting in a large prediction error of the water blocking rate and seriously affecting the accuracy of engineering risk assessment.

[0007] Based on this, there is an urgent need for an optimization method for pre-grouting of coal seam roof based on data-driven to break through the technical bottlenecks of traditional empirical decision-making models, and finally achieve the technical goals of improving the utilization rate of historical data, enhancing the success rate of grouting water blocking and reducing engineering costs, so as to provide scientific decision-making support for coal mine water hazard prevention and control. Summary of the Invention

[0008] The purpose of the present invention is to overcome the defects of the existing technology and provide an optimization method for pre-grouting of coal seam roof based on data-driven.

[0009] To achieve the above purpose, the present invention adopts the following technical solutions:

[0010] An optimization method for pre-grouting of coal seam roof based on data-driven includes the following steps:

[0011] S1 Fine exploration of geological bodies in the study area and extraction of original information;

[0012] S2 Construction information extraction;

[0013] S3 Use the DBSCAN clustering algorithm to classify grouting projects;

[0014] S4 Establish a quantitative relationship model based on the classification results;

[0015] S5 Parameter optimization and database update during construction.

[0016] Preferably, in the step S1, it specifically includes:

[0017] Based on existing geophysical exploration, drilling, geological, and coal mine production data, complete the fine exploration of geological bodies in the study area, and construct a fine geological model of the study area; extract the original geological occurrence information at each grouting point location, including relevant parameters such as the distance from the grouting point to the coal seam roof, lithology, fracture development degree, and water abundance.

[0018] Preferably, in the step S1, it specifically includes:

[0019] S11 Fine exploration work:

[0020] Comprehensively use existing geophysical exploration, drilling, and conduct in-depth analysis and integration of existing detailed geological data and coal mine production data to comprehensively and meticulously complete the fine exploration of geological bodies in the study area. Through the coordinated use of these means, accurately master the geological structures and stratigraphic distribution conditions at different positions in the study area, laying a foundation for the subsequent construction of a fine geological model;

[0021] S12 Construct a fine geological model:

[0022] With the help of the professional geological modeling software petrel, digitize the collected geophysical exploration, drilling, and various geological data, and construct a three-dimensional fine geological model that can accurately reflect the geological characteristics of the study area according to the actual spatial coordinates and geological attribute information; this model can intuitively present key information such as the undulation of the strata, the distribution of rocks, and the morphology of various geological structures;

[0023] S13 Extract the original geological occurrence information:

[0024] Based on the constructed fine geological model, accurately extract the original geological occurrence information at each grouting point location; specifically, it covers the following key parameters: the distance from the grouting point to the coal seam roof, lithology, fracture development degree, and water abundance.

[0025] Preferably, in the step S2, it specifically includes:

[0026] For all kinds of information generated and recorded during the actual construction of each grouting point, a comprehensive and detailed extraction work is carried out, specifically including the following important aspects of engineering parameters: leakage characteristics, grouting volume, pressure characteristics, and slurry characteristics.

[0027] Preferably, in step S3, it specifically includes:

[0028] Use the DBSCAN clustering algorithm to classify the grouting project, introduce the Mahalanobis distance to eliminate the influence of dimensions, optimize the clustering radius ε through the silhouette coefficient, and establish a feature library of grouting project types.

[0029] Preferably, in step S4, it specifically includes:

[0030] Based on the classification results, use LASSO regression for feature selection, determine the penalty coefficient λ through cross-validation, and establish an interpretable quantitative relationship model between the geological occurrence information of each grouting point and the grouting project.

[0031] Preferably, in step S4, it specifically includes:

[0032] S41 Use LASSO regression for feature selection:

[0033] Based on the classification results of the grouting project obtained by the DBSCAN clustering algorithm above, further use the LASSO regression method for feature selection;

[0034] S42 Determine the penalty coefficient λ through cross-validation:

[0035] Divide the data set into K subsets, sequentially select one subset as the validation set, and the remaining subsets as the training set. Train the model at different λ values and evaluate the model performance on the validation set. By repeating this process multiple times, find the penalty coefficient λ value that makes the model perform optimally in terms of average performance, ensure that the established model can accurately capture the quantitative relationship between the geological occurrence information of each grouting point and the grouting project without overfitting the data, and finally establish an interpretable quantitative relationship model.

[0036] Preferably, in step S5, it specifically includes:

[0037] During the construction process, solve the optimal grouting parameter combination for the geological occurrence information of each grouting point, update the database, and optimize the engineering plan based on data-driven.

[0038] Preferably, in step S5, it specifically includes:

[0039] S51 Solve the optimal grouting parameter combination:

[0040] During the actual construction process, for each specific grouting point, first, based on the geological occurrence information obtained from the pre-precise exploration, these parameters are input into the established quantitative relationship model. Through model calculation, the optimal grouting parameter combination corresponding to this grouting point is solved, including but not limited to the most appropriate grouting volume, grouting pressure, and specific values of slurry characteristics;

[0041] S52 database update:

[0042] As the construction progresses and new grouting point data is generated, the parameters in the actual construction process of each grouting point are timely fed back and updated to the database. At the same time, the established clustering model, quantitative relationship model, etc. are regularly re-evaluated and optimized using the new data, so that the entire data-driven engineering plan optimization system can continuously learn and adapt to the new situations and changes in the actual project, always maintaining its effectiveness and accuracy, and further improving the optimization level of subsequent engineering plans.

[0043] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0044] In the present invention, by integrating multi-source geological exploration data and dynamic grouting engineering parameters, a geological-engineering feature coupling model is established, breaking through the technical bottleneck of the traditional empirical decision-making mode. Based on machine learning algorithms, the non-linear correlation law between geological structures and grouting responses is quantitatively analyzed, and a multi-objective collaborative optimization system is developed to realize intelligent matching of grouting parameters, accurate prediction of construction effects, and dynamic optimization of engineering plans. Finally, the technical goals of improving the utilization rate of historical data, enhancing the success rate of grouting water blocking, and reducing engineering costs are achieved, providing scientific decision-making support for the prevention and control of coal mine water disasters. Brief Description of the Drawings

[0045] Figure 1 is a flowchart of an optimization method for pre-grouting of coal seam roof based on data driving according to the present invention. Detailed Description of the Specific Embodiments

[0046] The following details the specific embodiments of the present invention.

[0047] The "range" disclosed in the present invention is defined in the form of a lower limit and an upper limit. A given range is defined by selecting a lower limit and an upper limit, and the selected lower limit and upper limit define the boundaries of a particular range. The ranges defined in this way can include or exclude the end values, and can be combined arbitrarily, that is, any lower limit can be combined with any upper limit to form a range. For example, if a range of 10 to 50 is listed for a specific parameter, ranges of 10 to 40 and 20 to 50 are also contemplated. In addition, if the minimum range values 1 and 2 are listed, and if the maximum range values 3, 4, and 5 are listed, then the following ranges are all contemplated: 1 to 3, 1 to 4, 1 to 5, 2 to 3, 2 to 4, and 2 to 5. In this application, unless otherwise specified, the numerical range "a to b" represents an abbreviated representation of any real number combination between a and b, where a and b are both real numbers. For example, the numerical range "0 to 5" means that all real numbers between "0 to 5" are fully listed herein, and "0 to 5" is only an abbreviated representation of these numerical combinations.

[0048] If there is no special instruction, all embodiments and optional embodiments of this application can be combined with each other to form a new technical solution.

[0049] If there is no special instruction, all technical features and optional technical features of this application can be combined with each other to form a new technical solution.

[0050] If there is no special instruction, all steps of this application can be carried out in sequence or randomly, preferably in sequence. For example, the method includes steps (a) and (b), which means that the method can include steps (a) and (b) carried out in sequence, or can also include steps (b) and (a) carried out in sequence. For example, it is mentioned that the method may further include step (c), which means that step (c) can be added to the method in any order. For example, the method can include steps (a), (b), and (c), or can also include steps (a), (c), and (b), or can also include steps (c), (a), and (b), etc.

[0051] If there is no special instruction, the "including" and "containing" mentioned in this application mean open-ended or can also be closed-ended. For example, the "including" and "containing" can mean that other components not listed can also be included or contained, or can only include or contain the listed components.

[0052] If there is no special instruction, the reaction is carried out under normal temperature and normal pressure conditions.

[0053] If there is no special instruction, all parts or percentages are by weight or weight percentage.

[0054] In the present invention, all substances used are known substances, which can be purchased or synthesized by known methods.

[0055] In the present invention, all devices or equipment used are conventional devices or equipment known in the field, and they can all be purchased.

[0056] The following further illustrates the specific implementation manners of a data-driven optimization method for pre-grouting of coal seam roof in the present invention in conjunction with embodiments. The data-driven optimization method for pre-grouting of coal seam roof in the present invention is not limited to the descriptions of the following embodiments.

[0057] Embodiment 1:

[0058] A data-driven optimization method for pre-grouting of coal seam roof, as Figure 1 shown, includes the following steps:

[0059] 1. Based on existing geophysical exploration, drilling, geological, and coal mine production data, complete the fine exploration work of geological bodies in the study area and construct a fine geological model of the study area. Extract the original geological occurrence information at each grouting point, including relevant parameters such as the distance from the grouting point to the coal seam roof, lithology, fracture development degree, and water abundance.

[0060] 2. According to the construction information of each grouting point, extract relevant engineering parameters such as leakage characteristics, grouting volume, pressure characteristics, and slurry characteristics.

[0061] 3. Use the DBSCAN clustering algorithm to classify grouting projects, introduce the Mahalanobis distance to eliminate the influence of dimensions, optimize the clustering radius ε through the silhouette coefficient, and establish a feature library of grouting project types.

[0062] 4. Based on the classification results, use LASSO regression for feature selection, determine the penalty coefficient λ through cross-validation, and establish an interpretable quantitative relationship model between the geological occurrence information of each type of grouting point and the grouting project.

[0063] 5. During the construction process, solve the optimal grouting parameter combination for the geological occurrence information of each grouting point, update the database, and optimize the engineering plan based on data driving.

[0064] Embodiment 2:

[0065] A data-driven optimization method for pre-grouting of coal seam roof, as Figure 1 shown, includes the following steps:

[0066] (I) Fine exploration of geological bodies in the study area and extraction of original information:

[0067] Fine exploration work:

[0068] Comprehensively utilize existing geophysical exploration methods (such as combining various advanced geophysical exploration methods like seismic exploration and transient electromagnetic method), drilling operations (conduct drilling operations according to a scientific and reasonable borehole layout plan to obtain accurate geological samples and data at different depths and positions), and conduct in-depth analysis and integration of existing detailed geological data and coal mine production data to comprehensively and meticulously complete the fine exploration work of geological bodies in the study area. Through the coordinated application of these means, accurately grasp the geological structures, stratigraphic distributions, etc. at different positions in the study area as much as possible, laying a solid foundation for the subsequent construction of a fine geological model.

[0069] Construct a fine geological model:

[0070] With the help of the professional geological modeling software petrel, digitally process the geophysical exploration, drilling, and various geological data collected, and construct a three-dimensional fine geological model that can accurately reflect the geological characteristics of the study area according to the actual spatial coordinates and geological attribute information. This model can visually present key information such as the undulation of the strata, the distribution of rocks, and the morphology of various geological structures.

[0071] Extract the original geological occurrence information:

[0072] Based on the constructed fine geological model, accurately extract the original geological occurrence information at each grouting point location. Specifically, it includes the following key parameters:

[0073] Distance from the grouting point to the coal seam roof: Through the spatial positioning of each geological interface in the model and the accurate identification of the coal seam roof, accurately measure the vertical distance value from the grouting point to the coal seam roof, providing a basis for the subsequent analysis of the basic spatial position relationship.

[0074] Lithology: Based on the analysis of the core samples obtained from drilling and the geological attribute identification of the rocks at the corresponding positions in the model, determine the specific types of rocks (such as sandstone, shale, limestone, etc.) at the location of the grouting point. Different lithologies have important influences on the diffusion and penetration characteristics of the slurry during the grouting process.

[0075] Degree of fracture development: Combining the characteristics of fractures reflected in geophysical exploration data (such as seismic wave reflection, electromagnetic signal anomalies, etc.) and the actual observation records of fractures during drilling, use professional fracture analysis methods (such as fracture rate calculation, fracture strike and dip statistics, etc.) to quantitatively evaluate the degree of fracture development at the grouting point location, which plays a key role in the flow path and diffusion range of the slurry in the rock formation.

[0076] Water abundance: Refer to the distribution description of aquifers in geological reports, relevant data such as pumping tests, and the simulation of water-bearing areas in the model to comprehensively judge the water abundance status at the grouting point location, which is directly related to the difficulty of achieving the goals of water blocking and water reduction in the grouting project.

[0077] (2) Construction information extraction:

[0078] For all kinds of information generated and recorded during the actual construction of each grouting point, a comprehensive and detailed extraction work is carried out, specifically including the following important aspects of engineering parameters:

[0079] Leakage characteristics: Analyze in detail the leakage situation of the grout during different stages and at different pressures in the grouting process in the construction records, such as the change trend of the leakage volume, the key positions and time nodes where leakage occurs, etc. By sorting out these leakage characteristics, it is possible to understand the permeability of the rock formation at the grouting point and the possible weak channels, etc.

[0080] Grout volume: Accurately count the total volume of the grout injected from the start to the end of the grouting process for each grouting point, and at the same time pay attention to the change of the grout volume over time, which is one of the important indicators for measuring the grouting effect and judging the filling situation of the rock formation.

[0081] Pressure characteristics: Collect the grouting pressure data at different times during the grouting process, and analyze the characteristics such as the peak value, average value of the pressure and the pressure change curve, etc. The grouting pressure is closely related to the bearing capacity of the rock formation, the degree of crack opening and closing, etc., and is crucial for reasonably controlling the grouting process.

[0082] Grout characteristics: Cover key attribute information such as the density, viscosity, and composition of the grout. Different grout characteristics will affect its flow performance, setting time in the rock formation, and the cementing effect with the rock formation, etc., and thus affect the final quality of the grouting project.

[0083] (3) Using the DBSCAN clustering algorithm for grouting project classification:

[0084] Introduce the Mahalanobis distance to eliminate the influence of dimension:

[0085] When using the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm to classify the data related to the grouting project, since the various geological occurrence information and engineering parameters extracted often have different dimensions (such as the distance unit is meters, the pressure unit is megapascals, etc.), these dimension differences will have a deviation impact on the clustering result. To solve this problem, the Mahalanobis distance is introduced to replace the traditional Euclidean distance for measuring the distance between data points. The Mahalanobis distance can take into account the correlation between variables and the variance differences of different variables, and by standardizing the data, effectively eliminate the influence of dimension on clustering, making the clustering result more accurately reflect the internal distribution law of the data.

[0086] Optimizing the clustering radius ε through the silhouette coefficient:

[0087] The clustering radius ε is a key parameter in the DBSCAN algorithm, and its value directly determines the density of clustering and the rationality of the final clustering result. To determine the optimal value of the clustering radius ε, the silhouette coefficient is used as an evaluation index. The silhouette coefficient comprehensively measures the closeness of a data point to other points within the cluster it belongs to and the separation from points in adjacent clusters. By continuously adjusting the value of the clustering radius ε within a certain range and calculating the corresponding silhouette coefficients, the value of ε that maximizes the overall silhouette coefficient is found, thereby optimizing the clustering result, making the various grouting engineering data after clustering highly similar internally and significantly different between different classes, and then establishing a scientific and reasonable grouting engineering type feature library. This feature library can clearly present the differences and commonalities in geological occurrence and construction characteristics of different types of grouting projects.

[0088] (4) Establishing a quantitative relationship model based on the classification results:

[0089] Using LASSO regression for feature selection:

[0090] Based on the grouting engineering classification results obtained by the DBSCAN clustering algorithm above, the LASSO (Least Absolute Shrinkage and Selection Operator) regression method is further used for feature selection. Among many geological occurrence information and engineering parameters, there are some variables that may have relatively little impact on the grouting engineering results or have multicollinearity problems. LASSO regression can shrink the coefficients corresponding to some unimportant variables to zero by imposing an absolute value penalty term on the regression coefficients, thereby screening out the characteristic variables that play a key role in establishing the relationship between the geological occurrence information of grouting points and grouting projects, simplifying the model structure, and improving the interpretability and prediction accuracy of the model.

[0091] Determining the penalty coefficient λ through cross-validation:

[0092] The penalty coefficient λ is an important hyperparameter in LASSO regression. Its value determines the degree of penalty on the regression coefficients, thereby affecting the result of feature selection and the performance of the model. To determine the appropriate value of the penalty coefficient λ, the cross-validation (such as the commonly used K-fold cross-validation method) technique is applied. The dataset is divided into K subsets. One subset is sequentially selected as the validation set, and the remaining subsets are used as the training set. The model is trained with different values of λ and the performance of the model is evaluated on the validation set (such as evaluation metrics like mean squared error, coefficient of determination, etc.). By repeating this process multiple times, the value of the penalty coefficient λ that makes the model perform optimally in terms of average performance is found, ensuring that the established model can accurately capture the quantitative relationship between the geological occurrence information of various grouting points and the grouting project without overfitting the data. Finally, a quantitative relationship model with strong interpretability is established. This model can clearly reveal the degree of mutual influence and variation law among the parameters of each grouting project under different geological conditions, providing a reliable theoretical basis for the optimization of subsequent engineering plans.

[0093] (V) Parameter Optimization and Database Update during Construction:

[0094] Solving the Optimal Grouting Parameter Combination:

[0095] During the actual construction process, for each specific grouting point, first, based on the geological occurrence information obtained from the pre-precise exploration (such as the distance from the coal seam roof, lithology, fracture development degree, water abundance, etc.), these parameters are input into the established quantitative relationship model. Through model calculation, the optimal grouting parameter combination corresponding to this grouting point is solved, including but not limited to the specific values of the most appropriate grouting volume, grouting pressure, slurry characteristics, and other parameters. This can ensure that during the construction process of each grouting point, the most optimized grouting plan can be implemented according to its unique geological conditions, improving the effect of the grouting project on preventing water disasters in the coal seam roof and the overall construction efficiency.

[0096] Database Update

[0097] As the construction progresses and new data of grouting points are generated, various parameters during the actual construction process of each grouting point (including geological occurrence information, construction information, and the final evaluation of grouting effect, etc.) are timely fed back and updated to the database. At the same time, the established clustering model, quantitative relationship model, etc. are regularly re-evaluated and optimized (such as re-conducting clustering analysis, re-determining the parameters of the regression model, etc.) using the new data, enabling the entire data-driven engineering plan optimization system to continuously learn and adapt to new situations and changes in the actual project, always maintaining its effectiveness and accuracy, and further improving the optimization level of subsequent engineering plans.

[0098] Economic Analysis:

[0099] Based on this method in terms of cost savings, by constructing a quantitative relationship model to solve the optimal grouting parameter combination, it can reduce the waste of grouting materials, lower the loss and energy consumption of construction equipment, and also avoid the cost of repeated construction. On the other hand, in terms of benefit improvement, it can improve the coal mining efficiency, increase the coal sales revenue, and extend the service life of the mine, bringing considerable potential benefits. Although there are investments in fine geological exploration, database establishment, and data analysis and modeling in the early stage, in the long run, in the several years after implementation, it is expected to bring considerable net economic benefits, with a good return on investment, overall demonstrating the feasibility and advantages of this solution in terms of economy, providing strong support for its popularization and application.

[0100] By adopting the above technical solution:

[0101] The present invention aims at the treatment of roof sandstone water in Jurassic coal resources in western China, and proposes a data-driven optimization method for pre-grouting of coal seam roofs. The present invention realizes the fine exploration of the geological body in the research area through the combination of a variety of advanced means, not only comprehensively obtains the geological occurrence information, but also deeply excavates various engineering parameters in the construction process, breaking the limitations of insufficient and incomplete data utilization in the past, deeply integrating the key data in both geology and construction, and providing a rich and high-quality data basis for subsequent accurate analysis and scheme optimization. A complete data-driven engineering scheme optimization closed-loop is constructed, from determining the initial optimal grouting parameter combination based on the fine geological model and data classification and relationship model before construction, to collecting new data in real time during construction and feeding back to update the database, and then continuously optimizing and adjusting the model using the updated data, so that the engineering scheme can be dynamically optimized continuously with the change of the actual construction situation, significantly improving the ability of the pre-grouting project of coal seam roofs to cope with complex geological conditions and water disaster prevention and control, effectively ensuring the safety and efficiency in the coal mining process, and having obvious advantages and innovation compared with the traditional relatively fixed engineering scheme lacking dynamic adjustment.

[0102] The above content is a further detailed description of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention belongs, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A data-driven optimization method for pre-grouting of coal seam roof, characterized in that It includes the following steps: S1 Fine exploration of geological bodies in the study area and extraction of original information; S2 Extraction of construction information; S3 Using the DBSCAN clustering algorithm for grouting project classification; S4 Establishing a quantitative relationship model based on the classification results; S5 Parameter optimization and database update during the construction process.

2. The data-driven optimized method for pre-grouting of coal seam roof according to claim 1, characterized in that In the step S1, it specifically includes: Based on the existing geophysical exploration, drilling, geological and coal mine production data, complete the fine exploration of geological bodies in the study area, and construct a fine geological model of the study area; extract the original geological occurrence information at each grouting point, including relevant parameters such as the distance from the grouting point to the coal seam roof, lithology, fracture development degree, and water abundance.

3. The data-driven optimized method for pre-grouting of coal seam roof according to claim 1, wherein In the step S1, it specifically includes: S11 Fine exploration work: Comprehensively use the existing geophysical exploration, drilling, and conduct in-depth analysis and integration of the existing detailed geological data and coal mine production data to comprehensively and meticulously complete the fine exploration of geological bodies in the study area. Through the coordinated use of these means, accurately master the geological structures and stratigraphic distribution in different positions in the study area as much as possible, laying a foundation for the subsequent construction of a fine geological model; S12 Constructing a fine geological model: With the help of the professional geological modeling software petrel, digitally process the collected geophysical exploration, drilling and various geological data, and construct a three-dimensional fine geological model that can accurately reflect the geological characteristics of the study area according to the actual spatial coordinates and geological attribute information; this model can intuitively present key information such as the undulation of the strata, the distribution of rocks, and the morphology of various geological structures; S13 Extracting the original geological occurrence information: Based on the constructed fine geological model, accurately extract the original geological occurrence information at each grouting point; specifically covering the following key parameters: the distance from the grouting point to the coal seam roof, lithology, fracture development degree and water abundance.

4. The data-driven optimized method for pre-grouting of coal seam roof according to claim 1, characterized in that In the step S2, it specifically includes: For all kinds of information generated and recorded during the actual construction of each grouting point, conduct a comprehensive and meticulous extraction work, specifically including the following important engineering parameters: leakage characteristics, grouting volume, pressure characteristics and slurry characteristics.

5. A data-driven optimization method for pre-grouting of coal seam roof as claimed in claim 1, characterized in that, In the step S3, it specifically includes: Using the DBSCAN clustering algorithm for grouting project classification, introducing the Mahalanobis distance to eliminate the influence of dimension, optimizing the clustering radius ε through the silhouette coefficient, and establishing a characteristic library of grouting project types.

6. The data-driven optimized method for pre-grouting of coal seam roof as claimed in claim 1, wherein In the step S4, it specifically includes: Based on the classification results, use LASSO regression for feature selection, determine the penalty coefficient λ through cross-validation, and establish an interpretable quantitative relationship model between the geological occurrence information of each grouting point and the grouting project.

7. The data-driven optimized method for pre-grouting of coal seam roof according to claim 1, characterized in that, In the step S4, it specifically includes: S41 Using LASSO regression for feature selection: Based on the above grouting project classification results obtained by the DBSCAN clustering algorithm, further use the LASSO regression method for feature selection; S42 Determining the penalty coefficient λ through cross-validation: The dataset is divided into K subsets. One subset is sequentially selected as the validation set, and the remaining subsets are used as the training set. The model is trained under different λ values and the model performance is evaluated on the validation set. By repeating this process multiple times, the penalty coefficient λ value that makes the model perform optimally in terms of average performance is found, ensuring that the established model can accurately capture the quantitative relationship between the geological occurrence information of various grouting points and the grouting project without overfitting the data. Finally, a quantitative relationship model with strong interpretability is established.

8. A data-driven optimization method for pre-grouting of coal seam roof as claimed in claim 1, characterized in that In step S5, it specifically includes: During the construction process, the optimal grouting parameter combination is solved for the geological occurrence information of each grouting point, and the database is updated for data-driven engineering plan optimization.

9. The data-driven optimized method for pre-grouting of coal seam roof according to claim 1, characterized in that, In step S5, it specifically includes: S51 Solve the optimal grouting parameter combination: During the actual construction process, for each specific grouting point, first, according to the geological occurrence information obtained by pre-precise exploration, these parameters are input into the established quantitative relationship model, and the optimal grouting parameter combination corresponding to this grouting point is solved through model calculation, including but not limited to the most appropriate grouting volume, grouting pressure, and specific values of slurry characteristics; S52 Database update: As the construction progresses and new grouting point data is generated, the parameters during the actual construction process of each grouting point are timely fed back and updated to the database. At the same time, the established clustering model, quantitative relationship model, etc. are regularly re-evaluated and optimized using the new data, enabling the entire data-driven engineering plan optimization system to continuously learn and adapt to new situations and changes in the actual project, always maintaining its effectiveness and accuracy, and further improving the optimization level of subsequent engineering plans.

Images