A coal wall stability evaluation method and system for a large mining height working face containing dirt bands

Abstract

This application provides a method and system for evaluating the stability of coal face in a high-mining face with interbedded rock, belonging to the field of coal mining and geotechnical engineering technology. The method involves: acquiring basic parameters of the coal face; calculating local and overall stability indices and numerical simulation indices, and classifying them into key indicators and general indicators; determining whether the key indicators meet critical thresholds; if any do not meet the threshold, a high-risk assessment is directly applied; otherwise, a random forest regression model is used to dynamically weight and normalize the general indicators, and a weighted sum is obtained to obtain a comprehensive evaluation score, which is then used to classify high, medium, and low risk levels; collecting on-site monitoring data to determine a reference score, comparing it with the comprehensive evaluation score, and updating the training sample set and retraining the model when the error exceeds the threshold. This invention combines the judgment of the safety baseline of key indicators with the dynamic weighting of general indicators, and introduces a feedback iteration mechanism to achieve adaptive, rapid, and accurate evaluation of coal face stability under complex geological conditions.

Description

Technical Field

[0001] This application belongs to the field of coal mining and geotechnical engineering technology, specifically relating to a method and system for evaluating the stability of coal face in a high-mining face with interbedded rock. Background Technology

[0002] As coal resource development continues to extend deeper, some mining areas have now entered the deep mining stage with depths exceeding 600 meters. Among these, coal seams with more complex mining conditions, such as those with steep dips (35° to 65°), high extraction heights (5 to 8.5 meters), and containing 1 to 3 layers of interbedded rock, account for an increasingly larger proportion and have become a key mining focus both now and in the future. However, the stability of the coal face is a significant challenge in the mining of such complex coal seams. Interbedded rock coal seams are typical heterogeneous composite materials; the weak interbedded rock surfaces greatly disrupt the integrity and continuity of the coal body. Under the influence of gravity, mining-induced stress, and steep dips, they are highly susceptible to instability phenomena such as slippage, crushing, and shear failure, severely hindering safe and efficient production.

[0003] Traditional coal seam stability evaluation methods are mainly applicable to gently dipping, simple coal seams without interbedded rock. For example, existing technologies often employ a single safety factor method or a method based on simple numerical simulation to predict the depth of rock spalling. These methods struggle to accurately characterize the discontinuous deformation and local instability mechanisms caused by interbedded rock interfaces, and cannot adapt to complex working conditions such as large dip angles, high mining heights, and changes in the occurrence state of interbedded rock. They generally suffer from high false alarm rates, poor model adaptability, and slow evaluation update cycles. In particular, existing evaluation systems are mostly statically weighted, failing to consider geological and working condition changes during actual mining, leading to significant discrepancies between evaluation results and actual engineering conditions.

[0004] Therefore, there is an urgent need to construct a new coal seam stability evaluation system that can accurately reflect the instability mechanism of coal seams with large dip angles and high mining heights containing interbedded gangue, and can be adaptive and rapidly updated, so as to provide reliable technical support for safe and efficient coal mining. Summary of the Invention

[0005] In a first aspect, embodiments of this application provide a method for evaluating the stability of a coal face with interbedded rock and high mining height, comprising the following steps: S1. Obtain the basic parameters of the coal face, calculate the local stability index and overall stability index of the coal face based on the basic parameters, and use numerical simulation tools to establish a working face model to obtain numerical simulation indexes under multiple working conditions. The basic parameters include the physical and mechanical parameters, occurrence characteristics, geostress data and hydrological conditions of the coal seam and interbedded rock. S2. Local stability indicators, overall stability indicators, and numerical simulation indicators are divided into key indicators and general indicators; S3. Determine whether the key indicators meet the preset critical thresholds; If any key indicator fails to meet its corresponding critical threshold, the coal wall stability is directly determined to be high risk. If all key indicators meet the corresponding critical thresholds, proceed to step S4. S4. A pre-trained random forest regression model is used to dynamically assign weights to general indicators to obtain the dynamic weights of each general indicator. At the same time, the general indicators are normalized to obtain normalized indicators. Then, the dynamic weights of each general indicator and the corresponding normalized indicator values ​​are weighted and summed to obtain the comprehensive evaluation score of coal wall stability. S5. Based on the comprehensive evaluation score, the coal wall stability is classified into high-risk, medium-risk, or low-risk levels; S6. Collect on-site monitoring data, determine the reference score for coal wall stability based on the on-site monitoring data, and compare it with the calculated comprehensive evaluation score for coal wall stability. When the absolute value of the difference between the two exceeds the preset error threshold, add the on-site monitoring data to the training sample set, retrain the random forest regression model, and update the dynamic weights of general indicators.

[0006] Secondly, embodiments of this application also provide a coal wall stability evaluation system for a high-mining working face containing interbedded rock, comprising: The basic parameter acquisition and index calculation module is used to acquire the basic parameters of the coal face, calculate the local stability index and overall stability index of the coal face based on the basic parameters, and use numerical simulation tools to establish a working face model to obtain numerical simulation indexes under multiple working conditions. The basic parameters include the physical and mechanical parameters, occurrence characteristics, geostress data and hydrological conditions of the coal seam and interbedded rock. The indicator classification module is used to classify local stability indicators, overall stability indicators, and numerical simulation indicators into key indicators and general indicators. The key indicator determination module is used to determine whether the key indicators meet the preset critical thresholds. If any key indicator does not meet its corresponding critical threshold, the coal wall stability is directly determined to be high risk. If all key indicators meet their corresponding critical thresholds, the dynamic weight evaluation module is triggered. The dynamic weight evaluation module is used to dynamically assign weights to general indicators using a pre-trained random forest regression model to obtain the dynamic weights of each general indicator. At the same time, the general indicators are normalized to obtain normalized indicators. Then, the dynamic weights of each general indicator are weighted and summed with the corresponding normalized indicator values ​​to obtain the comprehensive evaluation score of coal wall stability. The risk level classification module is used to classify coal wall stability into high-risk, medium-risk, or low-risk levels based on the comprehensive evaluation score. The feedback correction module is used to collect on-site monitoring data, determine the reference score for coal wall stability based on the on-site monitoring data, and compare it with the calculated comprehensive evaluation score for coal wall stability. When the absolute value of the difference between the two exceeds the preset error threshold, the on-site monitoring data is added to the training sample set, the random forest regression model is retrained, and the dynamic weights of general indicators are updated.

[0007] As can be seen from the above technical solutions, this application has the following advantages: The coal face stability evaluation method and system for high-extraction working faces with interbedded rock provided in this application employs a strategy that combines key indicators with general indicators in the training sample set. This ensures that a high-risk alarm can be triggered immediately when extreme dangerous indicators appear, while also allowing for refined risk level classification under the premise of basic safety compliance. By introducing a random forest regression model, the dynamic weights of each general indicator can be automatically calculated according to different geological conditions, resulting in evaluation results that closely reflect the actual complex mining environment. Through self-learning, the system can compare field monitoring data with model predictions. Once the error exceeds the limit, the training sample set is automatically updated and the model is retrained to complete closed-loop optimization, thereby solving the problem of decreased accuracy of static models over time. Attached Figure Description

[0008] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating the method for evaluating the stability of the coal face in a high-mining face with interbedded rock, as described in this invention.

[0010] Figure 2 This is a schematic diagram of the coal wall stability evaluation system for a high-extraction working face with interbedded rock, as described in this invention. Detailed Implementation

[0011] The various embodiments of this disclosure will be described more fully in the following detailed description of the specific steps of the coal wall stability evaluation method for high-mining faces with interbedded rock. This disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of this disclosure to the specific embodiments disclosed herein, but rather this disclosure should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of this disclosure.

[0012] This embodiment provides a method for evaluating the stability of coal walls in longwall faces with interbedded rock and high mining height. By combining the judgment of the safety baseline of key indicators with the dynamic weighting of general indicators by random forest, and introducing an iterative mechanism for feedback of field monitoring data, it can adapt to changes in complex geological conditions and achieve rapid, accurate, and low-miss rate evaluation of the stability of coal walls in longwall faces with interbedded rock and high mining height, thereby improving the reliability of early warning.

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] Please see Figure 1 The diagram shows a flowchart of a method for evaluating the stability of a coal face with interbedded rock and high mining height, according to a specific embodiment. The method includes the following steps: S1. Obtain the basic parameters of the coal face, calculate the local stability index and overall stability index of the coal face based on the basic parameters, and use numerical simulation tools to establish a working face model to obtain numerical simulation indexes under multiple working conditions. The basic parameters include the physical and mechanical parameters, occurrence characteristics, geostress data and hydrological conditions of the coal seam and interbedded rock. It should be noted that the basic parameters collected in this step not only include traditional physical and mechanical parameters, but also incorporate hydrological conditions and geostress data to meet the needs of deep and complex mining conditions; by using numerical simulation tools to conduct multi-condition simulations, the stress distribution at different stages of the mining process can be predicted before actual mining, thereby providing early warning. S2. Local stability indicators, overall stability indicators, and numerical simulation indicators are divided into key indicators and general indicators; It should be noted that this step classifies and manages the indicators, listing the strength indicators that directly affect the bearing capacity of the coal face as key indicators, and listing the simulation indicators that reflect deformation characteristics as general indicators, thus achieving differentiated evaluation through stratified processing. S3. Determine whether the key indicators meet the preset critical thresholds; If any key indicator fails to meet its corresponding critical threshold, the coal wall stability is directly determined to be high risk. If all key indicators meet the corresponding critical thresholds, proceed to step S4. It should be noted that before performing complex comprehensive calculations, this step involves checking key safety factors and screening the coal wall's resistance to slippage or shear, thus improving the safety and reliability of the evaluation method. S4. A pre-trained random forest regression model is used to dynamically assign weights to general indicators to obtain the dynamic weights of each general indicator. At the same time, the general indicators are normalized to obtain normalized indicators. Then, the dynamic weights of each general indicator and the corresponding normalized indicator values ​​are weighted and summed to obtain the comprehensive evaluation score of coal wall stability. It should be noted that this step uses a random forest regression model to replace traditional human experience in determining the weights of each indicator. This model can automatically identify indicators that affect stability based on the current geological conditions and assign them high weights, so that the evaluation results can be dynamically adjusted as the working conditions change, making them more objective and accurate. S5. Based on the comprehensive evaluation score, the coal wall stability is classified into high-risk, medium-risk, or low-risk levels; It should be noted that this step transforms the abstract comprehensive evaluation score into specific levels of high risk, medium risk, and low risk, achieving a combination of qualitative and quantitative evaluation, which facilitates on-site engineering technicians to intuitively obtain the stability state of the coal face. S6. Collect on-site monitoring data, determine the reference score for coal wall stability based on the on-site monitoring data, and compare it with the calculated comprehensive evaluation score for coal wall stability. When the absolute value of the difference between the two exceeds the preset error threshold, add the on-site monitoring data to the training sample set, retrain the random forest regression model, and update the dynamic weights of general indicators. It should be noted that this step compares the actual on-site monitoring results with the model prediction results. If the model prediction is found to be inaccurate, new data is used to retrain the model, so that the model can continuously adapt to subtle changes in on-site geological conditions, maintain high prediction accuracy, and complete closed-loop optimization.

[0015] This embodiment uses a bottom-line determination of key indicators to impose rigid safety constraints, uses a random forest model to dynamically adjust the weights of general indicators according to working conditions, and iteratively updates the model based on field monitoring data. This solves the problem that existing technologies have fixed weights and cannot adapt to complex geological changes, and achieves accurate and adaptive evaluation of the stability of coal face in high-extraction working faces with interbedded rock.

[0016] Furthermore, as a refinement and extension of the specific implementation methods of the above embodiments, in order to fully illustrate the specific implementation process in this embodiment, another method for evaluating the stability of the coal wall in a high-mining face containing interbedded rock is provided. This method includes the following steps: S1. Obtain the basic parameters of the coal face, calculate the local stability index and overall stability index of the coal face based on the basic parameters, and use numerical simulation tools to establish a working face model to obtain numerical simulation indexes under multiple working conditions. The basic parameters include the physical and mechanical parameters, occurrence characteristics, geostress data and hydrological conditions of the coal seam and interbedded rock. The specific steps of step S1 are as follows: S11. The physical and mechanical parameters of the coal seam and interbedded rock are determined by indoor tests. The physical and mechanical parameters include compressive strength, tensile strength, shear strength, elastic modulus, Poisson's ratio, cohesion, internal friction angle, and cohesion, internal friction angle and bond strength at the interface between the coal seam and interbedded rock. S12. The occurrence characteristics of the coal seam and interbedded rock, as well as the geostress field, are determined through field testing methods, and the hydrological conditions of the working face are obtained based on existing exploration data; the occurrence characteristics include the thickness of the interbedded rock. Location of interbedded coal and coal seam dip angle The geostress field includes vertical stress. and horizontal stress ; S13. Calculate the local stability index of the coal wall based on the aforementioned basic parameters; the local stability index includes the anti-slip safety factor. Compression safety factor and shear safety factor ; Anti-slip safety factor The calculation formula is:

[0017] In the formula, Indicates the cohesion of the slip surface. Indicates the in-plane friction angle. This represents the normal stress acting on the slip surface. This represents the shear stress borne by the slip surface; Compression safety factor The calculation formula is:

[0018] In the formula, This indicates the uniaxial compressive strength of the coal body and interbedded rock. This indicates the actual pressure borne by the coal body and interbedded rock; Shear safety factor The calculation formula is:

[0019] In the formula, The uniformly distributed load acting on the sliding body. Indicates the height of the shear body. This indicates the thickness of the intercalated rock mass within the shear body. This indicates the angle between the shear plane and the coal face. Indicates the angle of coal seam upward mining; , These represent the cohesive forces of coal and intercalated gangue, respectively. , These represent the internal friction angles of coal and intercalated gangue, respectively. S14. Determine the overall stability index of the coal wall based on the aforementioned basic parameters; the overall stability index is the safety margin of the ultimate bearing capacity of the coal wall. Wherein, the safety margin The safety margin is determined by whether the above inequality holds. Safety margin when the critical threshold is met and the inequality does not hold The critical threshold is not met; Safety margin of ultimate bearing capacity of the coal face The bearing capacity of the coal face is determined using the following formula:

[0020] In the formula, The bearing capacity coefficient at the end of the coal sample. For the continuity of the coal seam at the working face, The coefficient of deformation due to pressure on the coal body. The coal wall strain softening coefficient is given. Where n is the height of the coal face, and n is the number of different rock strata above the coal seam. , , These are the unit weight, elastic modulus, and thickness of the j-th rock stratum above the coal seam, respectively, where j is the integer sequence number of the rock strata above the coal seam from bottom to top, with a value range of 1-n; When the value on the left side of the inequality exceeds the value on the right side, the ultimate bearing capacity safety margin It was determined that the critical threshold was not met; It should be noted that when the above inequality holds, it indicates that the actual load effect of the coal wall is not greater than its ultimate bearing capacity, and the coal wall has sufficient safety margin; conversely, when the inequality does not hold, the load effect of the coal wall exceeds its bearing limit, the safety margin is insufficient, and it is determined that the critical threshold is not met. S15. Use numerical simulation tools UDEC or FLAC3D to build a working face model, set model boundary conditions (e.g., bottom fixed, horizontal constraints on both sides, and equivalent overburden load applied to the top), arrange monitoring points and monitoring lines on the coal face and in the surrounding rock, perform step-by-step excavation calculations and solutions, and after the solution is completed, perform the following index extraction: Extract the horizontal displacement values ​​from each monitoring point on the coal face surface, and take the maximum value as the maximum horizontal displacement. ; Extract the state variables of all elements in the working face model, and calculate the percentage of the area of ​​elements that have entered the plastic state relative to the total area of ​​the coal face and surrounding rock. This percentage is used as the proportion of the plastic zone area. ; Extract the peak stress in the stress concentration region in front of the coal face and compare it with the original vertical stress. The ratio of the two values ​​is used as the stress concentration factor. ; S2. Local stability indicators, overall stability indicators, and numerical simulation indicators are divided into key indicators and general indicators; The specific steps of step S2 are as follows: S21. Increase the anti-slip safety factor Shear safety factor Compression safety factor and ultimate bearing capacity safety margin These are classified as key indicators for determining the safety baseline. S22. Maximum horizontal displacement , percentage of plastic zone area and stress concentration factor These are categorized into general indicators for comprehensive stability scoring; S3. Determine whether the key indicators meet the preset critical thresholds; If any key indicator fails to meet its corresponding critical threshold, the coal wall stability is directly determined to be high risk. If all key indicators meet the corresponding critical thresholds, proceed to step S4. The specific steps of step S3 are as follows: S31. Let the critical threshold of the anti-slip safety factor be... The critical threshold for the shear safety factor is The critical threshold for the compressive strength safety factor is The critical threshold for the ultimate bearing capacity safety margin is ; S32. Determining the anti-slip safety factor Shear safety factor Compression safety factor and ultimate bearing capacity safety margin Are they all greater than their respective critical thresholds? , , and ; like , , , If any one of the following conditions is met, the coal wall stability is directly determined to be high-risk, and the process ends. like , , , If both conditions are met, the key indicators are determined to meet the preset critical threshold, and the process proceeds to step S4. S4. A pre-trained random forest regression model is used to dynamically assign weights to general indicators to obtain the dynamic weights of each general indicator. At the same time, the general indicators are normalized to obtain normalized indicators. Then, the dynamic weights of each general indicator and the corresponding normalized indicator values ​​are weighted and summed to obtain the comprehensive evaluation score of coal wall stability. The specific steps of step S4 are as follows: S41. Determine the input feature vector of the random forest regression model, the input feature vector including sampling height. Coal seam dip angle Thickness of interbedded rock Location of interbedded coal Cohesion of coal body , friction angle in coal body Cohesion of intercalated coal , internal friction angle of the interbedded rock Vertical stress Horizontal stress Maximum horizontal displacement , percentage of plastic zone area and stress concentration factor ; S42. Construct a training sample set and train a random forest regression model, wherein each sample in the training sample set consists of an input feature vector and a corresponding coal wall stability state label; The specific steps of step S42 are as follows: S421. Collect data from historical working faces, working faces with similar geological conditions, and the target working face to construct a training sample set. Where m is the total number of samples in the training sample set. Let be the input feature vector of the i-th sample. Output target This is the coal wall stability state label corresponding to the i-th sample; S422. Establish a random forest regression model. The random forest consists of T regression trees. The training process for each regression tree is as follows: The Bootstrap self-sampling method is used to sample data from the training sample set. Random sampling with replacement Sample subsets of the same size , as training samples for the t-th regression tree; It should be noted that by sampling with replacement, a subset with the same number of samples as the original dataset is obtained, but which may contain duplicate samples. At each node split, a subset of features is randomly selected from all M input features. The size of the feature subset is k, where 1≤k≤M, and the search for the optimal splitting feature is only allowed within this feature subset; Based on sample subset and feature subset The t-th regression tree is trained using the regression tree algorithm; where M is the total number of input features, and here M=13, which is the number of input features in the input feature vector, including sampling height. Coal seam dip angle Thickness of interbedded rock Location of interbedded coal Cohesion of coal body , friction angle in coal body Cohesion of intercalated coal , internal friction angle of the interbedded rock Vertical stress Horizontal stress Maximum horizontal displacement , percentage of plastic zone area and stress concentration factor ; Specifically, for the t-th regression tree, based on the sample subset and feature subset The regression tree is trained according to the following recursive splitting process: S4221. Determine whether one of the following conditions is met: The number of samples contained in the current node is less than the preset minimum number of samples, or the mean square error of the current node is less than the preset threshold. If so, set the current node as a leaf node and use the mean of the target values ​​of the samples output within the current node as the predicted value; If not, proceed to step S4222; S4222. Traverse and locate feature subsets For each feature in the dataset, for each possible splitting threshold of the localized feature, the sample is divided into a left subset and a right subset. The weighted mean square error after splitting is calculated, and the feature and splitting threshold that maximize the reduction of the mean square error are selected as the optimal split. S4223. Divide the current node into a left child node and a right child node according to the optimal split; S4224. Recursively execute steps S4221 to S4224 for each child node until the stopping condition is met; S4225. The generated complete binary tree is the t-th regression tree. The t-th regression tree corresponds to the input feature vector. Predicted value Input feature vector The mean of the target values ​​of the samples falling into the leaf nodes; S423. The final output of the random forest regression model is the average of the predictions from each regression tree:

[0021] in, For the input feature vector, Let t be the prediction result of the t-th regression tree. This is the final predicted output of the random forest model, where T is the total number of regression trees; S424. Calculate the importance contribution of each general indicator to the output target to obtain the feature importance matrix. , in, The calculation method is as follows: In a random forest, for the i-th feature, the average reduction in mean squared error caused by splitting nodes across all regression trees is calculated, and then normalized to obtain the result. ,and ; S43. Extract the feature importance matrix of each general indicator using the trained random forest regression model. ,in, This represents the degree of contribution of the i-th general index to the prediction of the coal wall stability state label, and , where n is the total number of general indicators; S44. Introducing the Main Controlling Factor Correction Matrix ,in, is the working condition correction coefficient for the i-th general index; S45. Calculate the dynamic weights of each general indicator according to the following formula. :

[0022] Where n represents the number of general indicators; S46. Normalize the general indicators of the current working face to obtain normalized indicator values. ; The specific steps of step S46 are as follows: For positive indicators where larger values ​​generally indicate higher stability, the following formula is used for normalization: ; For negative indicators where a larger value generally signifies higher risk, the following formula is used for normalization: ; Obtain the normalized index value ; in, Let be the measured value of the i-th general index. and Let be the minimum and maximum values ​​of the i-th indicator in the training sample set, respectively. is the normalized value of the i-th general index, and its value ranges from 0 to 1; S47. Calculate the comprehensive evaluation score S for coal wall stability according to the following formula: ; S5. Based on the comprehensive evaluation score, the coal wall stability is classified into high-risk, medium-risk, or low-risk levels; The specific steps of step S5 are as follows: Preset low-risk threshold lower limit and high-risk threshold upper limit ,and ; When the comprehensive evaluation score At that time, the coal wall stability was determined to be high risk; When the comprehensive evaluation score At that time, the stability of the coal wall was determined to be of medium risk; When the comprehensive evaluation score At that time, the coal wall stability was determined to be low risk; S6. Collect on-site monitoring data, determine the reference score for coal wall stability based on the on-site monitoring data, and compare it with the calculated comprehensive evaluation score for coal wall stability. When the absolute value of the difference between the two exceeds the preset error threshold, add the on-site monitoring data to the training sample set, retrain the random forest regression model, and update the dynamic weights of general indicators. The specific steps of step S6 are as follows: S61. Collect on-site monitoring data, including coal wall horizontal displacement, coal wall spalling depth, support working resistance, roof pressure, number of microseismic events, and surrounding rock deformation rate; S62. Based on the coal wall spalling depth in the field monitoring data, determine the reference score for coal wall stability according to the preset mapping rule between spalling depth and risk score. ; The mapping rule is as follows: When the maximum depth of the board hour, ; When the maximum depth of the board hour, ; When the maximum depth of the board hour, ; S63. Calculate the comprehensive evaluation score. As a model prediction score Calculate the prediction error And preset error threshold ; when At the same time, keep the current random forest regression model unchanged; when At that time, the field monitoring data collected this time is combined with the corresponding basic parameters to form a new training sample, which is added to the training sample set to retrain the random forest regression model and update the dynamic weights of general indicators.

[0023] It should be understood that the sequence number of each step in the above embodiments 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 the present invention.

[0024] like Figure 2 As shown, the following are embodiments of the coal wall stability evaluation system for longwall faces with interbedded rock and high mining height provided by this disclosure. This system and the coal wall stability evaluation methods for longwall faces with interbedded rock and high mining height in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the coal wall stability evaluation system for longwall faces with interbedded rock and high mining height, please refer to the embodiments of the coal wall stability evaluation methods for longwall faces with interbedded rock and high mining height described above.

[0025] The system includes: The basic parameter acquisition and index calculation module is used to acquire the basic parameters of the coal face, calculate the local stability index and overall stability index of the coal face based on the basic parameters, and use numerical simulation tools to establish a working face model to obtain numerical simulation indexes under multiple working conditions. The basic parameters include the physical and mechanical parameters, occurrence characteristics, geostress data and hydrological conditions of the coal seam and interbedded rock. The indicator classification module is used to classify local stability indicators, overall stability indicators, and numerical simulation indicators into key indicators and general indicators. The key indicator determination module is used to determine whether the key indicators meet the preset critical thresholds. If any key indicator does not meet its corresponding critical threshold, the coal wall stability is directly determined to be high risk. If all key indicators meet their corresponding critical thresholds, the dynamic weight evaluation module is triggered. The dynamic weight evaluation module is used to dynamically assign weights to general indicators using a pre-trained random forest regression model to obtain the dynamic weights of each general indicator. At the same time, the general indicators are normalized to obtain normalized indicators. Then, the dynamic weights of each general indicator are weighted and summed with the corresponding normalized indicator values ​​to obtain the comprehensive evaluation score of coal wall stability. The risk level classification module is used to classify coal wall stability into high-risk, medium-risk, or low-risk levels based on the comprehensive evaluation score. The feedback correction module is used to collect on-site monitoring data, determine the reference score for coal wall stability based on the on-site monitoring data, and compare it with the calculated comprehensive evaluation score for coal wall stability. When the absolute value of the difference between the two exceeds the preset error threshold, the on-site monitoring data is added to the training sample set, the random forest regression model is retrained, and the dynamic weights of general indicators are updated.

[0026] This embodiment overcomes the shortcomings of fixed evaluation weights in traditional methods by integrating the basic parameter acquisition and index calculation module, index division module, key index judgment module, dynamic weight evaluation module, risk level division module, and feedback correction module. This improves the accuracy, reliability, and adaptability of coal wall stability evaluation in high-extraction working faces.

[0027] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for evaluating the stability of coal face in a high-mining face with interbedded rock, characterized in that, Includes the following steps: S1. Obtain the basic parameters of the coal face, calculate the local stability index and overall stability index of the coal face based on the basic parameters, and use numerical simulation tools to establish a working face model to obtain numerical simulation indexes under multiple working conditions. The basic parameters include the physical and mechanical parameters, occurrence characteristics, geostress data and hydrological conditions of the coal seam and interbedded rock. S2. Local stability indicators, overall stability indicators, and numerical simulation indicators are divided into key indicators and general indicators; S3. Determine whether the key indicators meet the preset critical thresholds; If any key indicator fails to meet its corresponding critical threshold, the coal wall stability is directly determined to be high risk. If all key indicators meet the corresponding critical thresholds, proceed to step S4. S4. A pre-trained random forest regression model is used to dynamically assign weights to general indicators to obtain the dynamic weights of each general indicator. At the same time, the general indicators are normalized to obtain normalized indicators. Then, the dynamic weights of each general indicator and the corresponding normalized indicator values ​​are weighted and summed to obtain the comprehensive evaluation score of coal wall stability. S5. Based on the comprehensive evaluation score, the coal wall stability is classified into high-risk, medium-risk, or low-risk levels; S6. Collect on-site monitoring data, determine the reference score for coal wall stability based on the on-site monitoring data, and compare it with the calculated comprehensive evaluation score for coal wall stability. When the absolute value of the difference between the two exceeds the preset error threshold, add the on-site monitoring data to the training sample set, retrain the random forest regression model, and update the dynamic weights of general indicators.

2. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 1, characterized in that, The specific steps of step S1 are as follows: S11. The physical and mechanical parameters of the coal seam and interbedded rock are determined by indoor tests. The physical and mechanical parameters include compressive strength, tensile strength, shear strength, elastic modulus, Poisson's ratio, cohesion, internal friction angle, and cohesion, internal friction angle and bond strength at the interface between the coal seam and interbedded rock. S12. The occurrence characteristics of the coal seam and interbedded rock, as well as the geostress field, are determined through field testing methods, and the hydrological conditions of the working face are obtained based on existing exploration data; the occurrence characteristics include the thickness of the interbedded rock. Location of interbedded coal and coal seam dip angle The geostress field includes vertical stress. and horizontal stress ; S13. Calculate the local stability index of the coal wall based on the aforementioned basic parameters; the local stability index includes the anti-slip safety factor. Compression safety factor and shear safety factor ; Shear safety factor The calculation formula is: In the formula, The uniformly distributed load acting on the sliding body. Indicates the height of the shear body. This indicates the thickness of the intercalated rock mass within the shear body. This indicates the angle between the shear plane and the coal face. Indicates the angle of coal seam upward mining; , These represent the cohesive forces of coal and intercalated gangue, respectively. , These represent the internal friction angles of coal and intercalated gangue, respectively. S14. Determine the overall stability index of the coal wall based on the aforementioned basic parameters; the overall stability index is the safety margin of the ultimate bearing capacity of the coal wall. ; Safety margin of ultimate bearing capacity of the coal face The bearing capacity of the coal face is determined using the following formula: In the formula, The bearing capacity coefficient at the end of the coal sample. For the continuity of the coal seam at the working face, The coefficient of deformation due to pressure on the coal body. The coal wall strain softening coefficient is given. Where n is the height of the coal face, and n is the number of different rock strata above the coal seam. , , These are the unit weight, elastic modulus, and thickness of the j-th rock stratum above the coal seam, respectively, where j is the integer sequence number of the rock strata above the coal seam from bottom to top, with a value range of 1-n; When the value on the left side of the inequality exceeds the value on the right side, the ultimate bearing capacity safety margin It was determined that the critical threshold was not met; S15. Use numerical simulation tools UDEC or FLAC3D to build a working face model, set model boundary conditions, arrange monitoring points and monitoring lines on the coal face and in the surrounding rock, perform step-by-step excavation calculations and solutions, and after the solution is completed, extract the following indicators: Extract the horizontal displacement values ​​from each monitoring point on the coal face surface, and take the maximum value as the maximum horizontal displacement. ; Extract the state variables of all elements in the working face model, and calculate the percentage of the area of ​​elements that have entered the plastic state relative to the total area of ​​the coal face and surrounding rock. This percentage is used as the proportion of the plastic zone area. ; Extract the peak stress in the stress concentration region in front of the coal face and compare it with the original vertical stress. The ratio of the two values ​​is used as the stress concentration factor. .

3. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 1, characterized in that, The specific steps of step S2 are as follows: S21. Increase the anti-slip safety factor Shear safety factor Compression safety factor and ultimate bearing capacity safety margin These are classified as key indicators for determining the safety baseline. S22. Maximum horizontal displacement , percentage of plastic zone area and stress concentration factor These are general indicators used for comprehensive stability scoring.

4. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 3, characterized in that, The specific steps of step S3 are as follows: S31. Let the critical threshold of the anti-slip safety factor be... The critical threshold for the shear safety factor is The critical threshold for the compressive strength safety factor is The critical threshold for the ultimate bearing capacity safety margin is ; S32. Determining the anti-slip safety factor Shear safety factor Compression safety factor and ultimate bearing capacity safety margin Are they all greater than their respective critical thresholds? , , and ; like , , , If any one of the following conditions is met, the coal wall stability is directly determined to be high-risk, and the process ends. like , , , If both conditions are met, the key indicators are determined to meet the preset critical threshold, and the process proceeds to step S4.

5. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 1, characterized in that, The specific steps of step S4 are as follows: S41. Determine the input feature vector of the random forest regression model, the input feature vector including sampling height. Coal seam dip angle Thickness of interbedded rock Location of interbedded coal Cohesion of coal body , friction angle in coal body Cohesion of intercalated coal , internal friction angle of the interbedded rock Vertical stress Horizontal stress Maximum horizontal displacement , percentage of plastic zone area and stress concentration factor ; S42. Construct a training sample set and train a random forest regression model, wherein each sample in the training sample set consists of an input feature vector and a corresponding coal wall stability state label; S43. Extract the feature importance matrix of each general indicator using the trained random forest regression model. ,in, This represents the degree of contribution of the i-th general index to the prediction of the coal wall stability state label, and , where n is the total number of general indicators; S44. Introducing the Main Controlling Factor Correction Matrix ,in, is the working condition correction coefficient for the i-th general index; S45. Calculate the dynamic weights of each general indicator according to the following formula. : Where n represents the number of general indicators; S46. Normalize the general indicators of the current working face to obtain normalized indicator values. ; S47. Calculate the comprehensive evaluation score S for coal wall stability according to the following formula: 。 6. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 5, characterized in that, The specific steps of step S42 are as follows: S421. Collect data from historical working faces, working faces with similar geological conditions, and the target working face to construct a training sample set. Where m is the total number of samples in the training sample set. Given the input feature vector of the i-th sample, the output target... This is the coal wall stability state label corresponding to the i-th sample; S422. Establish a random forest regression model. The random forest consists of T regression trees. The training process for each regression tree is as follows: The Bootstrap self-sampling method is used to sample data from the training sample set. Random sampling with replacement Sample subsets of the same size , as training samples for the t-th regression tree; At each node split, a subset of features is randomly selected from all M input features. The size of the feature subset is k, where 1≤k≤M, and the search for the optimal splitting feature is only allowed within this feature subset; Based on sample subset and feature subset Train the t-th regression tree according to the regression tree algorithm; where M is the total number of input features; S423. The final output of the random forest regression model is the average of the predictions from each regression tree: in, For the input feature vector, Let t be the prediction result of the t-th regression tree. This is the final predicted output of the random forest model, where T is the total number of regression trees; S424. Calculate the importance contribution of each general indicator to the output target to obtain the feature importance matrix. , in, The calculation method is as follows: In a random forest, for the i-th feature, the average reduction in mean squared error caused by splitting nodes across all regression trees is calculated, and then normalized to obtain the result. ,and .

7. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 5, characterized in that, The specific steps of step S46 are as follows: For positive indicators where larger values ​​generally indicate higher stability, the following formula is used for normalization: ； For negative indicators where a larger value generally signifies higher risk, the following formula is used for normalization: ； Obtain the normalized index value ; in, Let be the measured value of the i-th general index. and Let be the minimum and maximum values ​​of the i-th indicator in the training sample set, respectively. is the normalized value of the i-th general index, ranging from 0 to 1.

8. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 1, characterized in that, The specific steps of step S5 are as follows: Preset low-risk threshold lower limit and high-risk threshold upper limit ,and ; When the comprehensive evaluation score At that time, the coal wall stability was determined to be high risk; When the comprehensive evaluation score At that time, the stability of the coal wall was determined to be of medium risk; When the comprehensive evaluation score At that time, the stability of the coal wall was determined to be low risk.

9. The method for evaluating the stability of the coal face in a high-mining face with interbedded rock as described in claim 1, characterized in that, The specific steps of step S6 are as follows: S61. Collect on-site monitoring data, including coal wall horizontal displacement, coal wall spalling depth, support working resistance, roof pressure, number of microseismic events, and surrounding rock deformation rate; S62. Based on the coal wall spalling depth in the field monitoring data, determine the reference score for coal wall stability according to the preset mapping rule between spalling depth and risk score. ; S63. Calculate the comprehensive evaluation score. As a model prediction score Calculate the prediction error And preset error threshold ; when At the same time, keep the current random forest regression model unchanged; when At that time, the field monitoring data collected this time is combined with the corresponding basic parameters to form a new training sample, which is added to the training sample set to retrain the random forest regression model and update the dynamic weights of general indicators.

10. A coal wall stability evaluation system for a high-extraction working face containing interbedded rock, characterized in that, include: The basic parameter acquisition and index calculation module is used to acquire the basic parameters of the coal face, calculate the local stability index and overall stability index of the coal face based on the basic parameters, and use numerical simulation tools to establish a working face model to obtain numerical simulation indexes under multiple working conditions. The basic parameters include the physical and mechanical parameters, occurrence characteristics, geostress data and hydrological conditions of the coal seam and interbedded rock. The indicator classification module is used to classify local stability indicators, overall stability indicators, and numerical simulation indicators into key indicators and general indicators. The key indicator determination module is used to determine whether the key indicators meet the preset critical thresholds. If any key indicator does not meet its corresponding critical threshold, the coal wall stability is directly determined to be high risk. If all key indicators meet their corresponding critical thresholds, the dynamic weight evaluation module is triggered. The dynamic weight evaluation module is used to dynamically assign weights to general indicators using a pre-trained random forest regression model to obtain the dynamic weights of each general indicator. At the same time, the general indicators are normalized to obtain normalized indicators. Then, the dynamic weights of each general indicator are weighted and summed with the corresponding normalized indicator values ​​to obtain the comprehensive evaluation score of coal wall stability. The risk level classification module is used to classify coal wall stability into high-risk, medium-risk, or low-risk levels based on the comprehensive evaluation score. The feedback correction module is used to collect on-site monitoring data, determine the reference score for coal wall stability based on the on-site monitoring data, and compare it with the calculated comprehensive evaluation score for coal wall stability. When the absolute value of the difference between the two exceeds the preset error threshold, the on-site monitoring data is added to the training sample set, the random forest regression model is retrained, and the dynamic weights of general indicators are updated.

Images