A method for assessing the risk of rock burst in coal mines

Abstract

The present invention relates to the field of coal and rock mining technology, and in particular, provides a method for evaluating the rock burst hazard of coal mines. The method includes constructing an index system for evaluating the rock burst hazard of the amount of coal and rock ejected, and obtaining a training data set and a test data set through data regularization; optimizing and analyzing the constructed index system using a fuzzy neighborhood decision rough set model and the training data set to obtain a simplified subset; inputting the obtained simplified subset into a multi-classifier model to obtain a classification result, and performing a classification accuracy evaluation on the classification result, completing the weighting of the classifier based on the classification accuracy evaluation, and obtaining the classifier weight and the trained multi-classifier model; inputting the test data set into the trained multi-classifier model for testing to obtain a test result, and weighting and fusing the test results to obtain a hazard evaluation level, and the method realizes the accurate division of the rock burst hazard level of coal mines.

Description

Technical Field

[0001] The invention relates to the technical field of coal and rock mining, and in particular to a method for evaluating the danger of rock burst in coal mines. Background Art

[0002] Rock burst is a destructive dynamic disaster caused by elastic deformation of the mining face in coal and rock mining. Rock burst will cause great damage to coal and rock mining, causing the operation to stop at the least, and posing a threat to mining personnel at the worst. Therefore, mining engineering attaches great importance to the accurate evaluation and prediction of rock burst, and then takes effective response strategies to prevent catastrophic accidents.

[0003] At present, domestic and foreign scholars mainly use theoretical analysis and calculation, on-site monitoring and judgment, and regional early evaluation methods to evaluate and predict the risk of rock burst. The theoretical analysis and calculation method represented by the three-band load method uses quantitative calculation to realize rock burst risk analysis, but it is subject to the complex influence of mining pressure factors, resulting in the method not being able to give an accurate calculation formula, and it has strong subjectivity and the results are not credible. On-site monitoring and judgment analyzes and judges the signal data obtained from on-site monitoring of coal and rock mines, but it is subject to the influence of monitoring signal noise, discreteness, insufficient coverage, and difficulty in determining thresholds, resulting in the method not being ideal in practical applications. The regional early evaluation method is one of the commonly used methods at present. It realizes the classification and prediction of rock burst risk levels by comprehensively analyzing the multi-parameter factors affecting rock burst, and fusion of multi-source information. In this method, commonly used algorithm models include probabilistic neural network model, deep neural network model, hierarchical analysis method, etc.

[0004] The following deficiencies still need to be improved in the current mainstream rock burst hazard assessment methods. First, there are too many evaluation factors selected, and there is redundancy, which fails to accurately capture the most valuable influencing factors for hazard analysis; second, few methods directly incorporate the amount of coal and rock ejection into the indicator system of hazard analysis; third, the discriminant model used to generate decision-making conclusions is too single, and fails to give full play to the advantages of multi-model and multi-source information fusion. Summary of the invention

[0005] In view of this, the present invention provides a method for evaluating the risk of rock burst in coal mines, so as to achieve accurate classification of the risk level of rock burst in coal mines.

[0006] In a first aspect, the present invention provides a method for assessing the risk of rock burst in a coal mine, the method comprising:

[0007] Step 1: construct an index system for evaluating the rock burst hazard of coal and rock eruption, and obtain training data sets and test data sets through data regularization;

[0008] Step 2: Optimize and analyze the constructed index system using the fuzzy neighborhood decision rough set model and the training data set to obtain a simplified subset;

[0009] Step 3: Input the obtained reduced subset into the multi-classifier model to obtain the classification result, and perform classification accuracy evaluation on the classification result. According to the classification accuracy evaluation, the classifier is weighted to obtain the classifier weight and the trained multi-classifier model;

[0010] Step 4: Input the test data set into the trained multi-classifier model for testing, obtain the test results, and weightedly fuse the test results to obtain the hazard assessment level.

[0011] Optionally, the index system for evaluating the rock burst hazard of coal and rock eruption volume includes:

[0012] A total of 11 influencing factors, including coal thickness a1, inclination a2, burial depth a3, structural conditions a4, inclination change a5, coal thickness change a6, gas concentration a7, roof management a8, pressure relief a9, coal cannon sound a10 and coal and rock eruption a11, are used as evaluation index elements for rock burst.

[0013] Optionally, the fuzzy neighborhood decision rough set model in step 2 includes:

[0014] Decision system: It is a binary group ,in For the domain, is the conditional attribute set, is the decision attribute, and ;

[0015] Fuzzy similarity relation: given domain , if the relationship The following conditions are met:

[0016] 1) Reflexivity: ;

[0017] 2) Symmetry: ;

[0018] Then it is called for The fuzzy similarity relationship defined in , the expression is:

[0019] ;

[0020] Where: For continuous data, It is discrete data;

[0021] Fuzzy Neighborhood Equivalence Class: Given a Domain , for any , its fuzzy neighborhood equivalence class, its expression is:

[0022] ;

[0023] Where: is the fuzzy neighborhood radius, and ;

[0024] Fuzzy Neighborhood Conditional Probability: Given a Decision System , , is a collection of objects with the same nominal decision class, in the condition attribute set Next, for any object , its fuzzy neighborhood equivalence class is , the fuzzy neighborhood conditional probability is expressed as:

[0025] ;

[0026] Among them, in the equivalence class Under the assumption that , if and only if hour, ;

[0027] In the framework of decision rough set theory, the loss function matrix based on data relationship dependency is given, and its expression is:

[0028] ;

[0029] Where i= P , B or N ; j =1 or c ;

[0030] is the fuzzy neighborhood conditional probability; is the compensation coefficient, satisfying For Marked as correspond The sample relative to The importance of is expressed as:

[0031] ;

[0032] For Marked as correspond The sample relative to The importance of is expressed as:

[0033] ;

[0034] For non-empty equivalence classes ,exist Under the assumption that , if and only if hour, ;

[0035] The loss function matrix based on data relationship dependence, and ,make , and get the probability threshold parameter The new expression is:

[0036] ;

[0037] ;

[0038] Therefore, the updated decision rule is expressed as:

[0039] P )like ,but ;

[0040] B )like ,but ;

[0041] N )like ,but ;

[0042] The two probability threshold parameters are expressed as and ,in , ;parameter After determination, we can get the distribution information of the original data and ; The only hyperparameters to be determined are unknown parameters and ; The rough set model of fuzzy neighborhood decision making is established as follows:

[0043] ;

[0044] ;

[0045] .

[0046] Optionally, obtaining the reduced subset includes:

[0047] Based on the reduction principle of minimizing decision risk in the fuzzy neighborhood decision rough set model, the expression of decision risk is:

[0048] ;

[0049] Where: ; For global decision-making risks;

[0050] Relative importance: Given a decision system , then for the conditional attribute The relative importance of is expressed as:

[0051] ;

[0052] Attribute reduction criterion: Given a decision system ,like The following two conditions are met, and the expression is:

[0053] 1) ;

[0054] 2) , both have ;

[0055] but For comparison The reduced subset with low risk is combined with the optimization algorithm to obtain the optimal reduced subset.

[0056] Optionally, step 4 includes:

[0057] The prediction results of each classifier are weighted and summed up into one result; the weight is determined by the training accuracy conversion coefficient of the classifier on the data set, and its expression is:

[0058] ;

[0059] in, For the i The result weight of the classifier is For its training classification accuracy, n is the number of classifiers;

[0060] The final decision result is a weighted voting result, that is, each classifier gives a judgment result for each category. , then the judgment of the multi-classifier for a category is The output result is the category with the largest discrimination result, so as to obtain the hazard assessment level.

[0061] In a second aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the method for assessing the risk of coal mine rock burst in the first aspect or any possible implementation of the first aspect.

[0062] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: one or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions, which, when executed by the device, enable the device to execute the method for assessing the hazard of coal mine rock burst in the first aspect or any possible implementation of the first aspect.

[0063] In the technical solution provided by the present invention, the method includes constructing an index system for evaluating the rock burst hazard of the amount of coal and rock ejected, and obtaining a training data set and a test data set through data regularization; optimizing and analyzing the constructed index system using a fuzzy neighborhood decision rough set model and the training data set to obtain a simplified subset; inputting the obtained simplified subset into a multi-classifier model to obtain a classification result, and performing a classification accuracy evaluation on the classification result, completing the weighting of the classifier based on the classification accuracy evaluation, and obtaining the classifier weight and the trained multi-classifier model; inputting the test data set into the trained multi-classifier model for testing to obtain a test result, and weightedly fusing the test results to obtain a hazard evaluation level. The method realizes the accurate classification of the rock burst hazard level of coal mines. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0065] Figure 1 A flow chart of a method for assessing the risk of rock burst in a coal mine provided by an embodiment of the present invention;

[0066] Figure 2 A histogram of attribute reduction statistics provided by an embodiment of the present invention;

[0067] Figure 3 A histogram showing the comparison results of the classifier accuracy before and after simplification provided by an embodiment of the present invention;

[0068] Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present invention.

[0069] Description of the drawings: 21 - electronic device; 211 - processor; 212 - memory; 213 - computer program. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0071] It should be clear that the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "said" and "the" used in the embodiments of the present invention are also intended to include plural forms, unless the context clearly indicates other meanings.

[0073] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0074] The word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0075] Rock burst is a very harmful dynamic disaster in coal and rock mining. Accurately assessing the danger level of rock burst is a prerequisite for curbing its occurrence as soon as possible. The precise selection of indicator elements and the rational use of evaluation methods have always been a difficult problem that has troubled relevant researchers and engineering applications.

[0076] Aiming at the data conditions and decision-making requirements for rock burst hazard level evaluation, the present invention adopts a fuzzy neighborhood particle characterization form equivalent to human semantic expression, constructs a fuzzy neighborhood decision rough set, and realizes subset screening of rock burst influencing factors. Furthermore, by imitating the human group decision-making strategy, multiple classifiers are used to realize the hazard level voting decision in the form of weighted allocation. Comparative experiments and engineering application experiments show that the method proposed in the present invention has more advantages than other methods and can effectively guide the practical application of the method in engineering.

[0077] Figure 1 A flow chart of a method for assessing the risk of rock burst in a coal mine provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the method includes:

[0078] Step 1: Construct an index system for evaluating the rock burst hazard of coal and rock eruption, and obtain training and test data sets through data regularization.

[0079] In the embodiment of the present invention, the index system for evaluating the rock burst hazard of coal and rock ejection volume includes:

[0080] A total of 11 influencing factors, including coal thickness a1, inclination a2, burial depth a3, structural conditions a4, inclination change a5, coal thickness change a6, gas concentration a7, roof management a8, pressure relief a9, coal cannon sound a10 and coal and rock eruption a11, are used as evaluation index elements for rock burst. Among them, the coal and rock eruption a11 index is not included in the index elements in the research of many scholars. However, according to actual research analysis and engineering investigation, it is found that this index has a certain implicit relationship with rock burst and is also one of the main factors for evaluating rock burst in engineering. Therefore, the present invention incorporates it into the evaluation index system.

[0081] Among the above indicators, structural condition a4, inclination change a5, coal thickness change a6, roof management a8, pressure relief a9, and coal cannon sound a10 are qualitative indicators of expert evaluation. Through the existing technology, the transformation of qualitative indicators into quantitative indicators can be realized and constructed into discrete data.

[0082] In the experimental verification part based on the public data set, the data set collected in the prior art was selected for verification, with a total of 35 samples. First, the data set was normalized to eliminate the data annihilation problem caused by different dimensions. Subsequently, the data set was sent to the fuzzy neighborhood decision rough set model for attribute simplification calculation. Finally, the attribute subset obtained by simplification was sent to the multi-classifier to implement the classification ability test of the classifier and obtain the weights of different classifiers. Among them, the multiple classifiers selected by the present invention include: ID1 decision tree (C4.5), ID2 reduced error pruning tree (REPtree), ID3 naive Bayes classifier (NB), ID4 bagging classifier (Bagging).

[0083] 80% of the samples (i.e., 28 samples) were randomly selected as the training set, and the attributes were reduced by adjusting the two parameters of the fuzzy neighborhood decision rough set model. The statistical results of 191 reduced subsets were obtained, as shown in Figure 2. Figure 2 As shown. In the fuzzy neighborhood decision rough set model, The value changes from 0.05 to 0.95, with a step size of 0.05. The value varies from 0.1 to 0.9 with a step size of 0.1.

[0084] pass Figure 2 The results show that the attribute a4 structural condition is the most significant factor affecting the danger of rock burst. This is because the structural condition of coal and rock mines directly affects the production safety of coal and rock mines, and design defects are also the root cause of production problems. The second is the sound of coal cannons of attribute a10, the inclination of attribute a2, the pressure relief of attribute a9, the burial depth of attribute a3, the eruption of coal and rock of attribute a11, the gas concentration of attribute a7, and the change of inclination of attribute a5. The factors with relatively small influence are attribute a1 coal thickness, attribute a6 coal thickness change, and attribute a8 roof management. The above results are consistent with the results obtained by the main cause analysis in the prior art, and are also in line with the actual analysis of rock burst danger in coal and rock mines. Therefore, in the attribute simplification, the factors with less influence, namely a1, a6, and a8, can be deleted, and the remaining attributes can be retained, namely A={a2, a3, a4, a5, a7, a9, a10, a11}. Therefore, in the subsequent evaluation of rock burst hazard level in the coal and rock mining industry, only 8 indicators in A need to be collected to achieve a more accurate hazard level evaluation, thereby reducing the cost of the indicator collection process.

[0085] Step 2: Use the fuzzy neighborhood decision rough set model and training data set to optimize and analyze the constructed indicator system to obtain a simplified subset.

[0086] In the embodiment of the present invention, a simplified subset is obtained to enhance the interpretability of the algorithm model for risk assessment.

[0087] The fuzzy neighborhood decision-theoretic rough set model (FNDTRS) is a rough set model that is determined by introducing the fuzzy neighborhood particle metric domain division on the basis of the decision rough set model.

[0088] In the embodiment of the present invention, the fuzzy neighborhood decision rough set model in step 2 includes:

[0089] Decision system: It is a binary group ,in For the domain, is the conditional attribute set, is the decision attribute, and ;

[0090] The fuzzy neighborhood decision rough set model is the same as the classical rough set and decision rough set models, and the research object is the decision system. The basic input data of the rock burst hazard assessment studied in this invention is a typical decision system, including the condition attribute set composed of the influencing factors affecting the occurrence of rock burst, the decision attribute composed of the hazard level, and the domain composed of the collected indicator data.

[0091] For decision-making systems such as rock burst hazard assessment, the relevant data consists of discrete data (such as structural conditions, inclination changes, etc.) and continuous data (such as coal thickness, burial depth, etc.), which makes the classical rough set model that can only process discrete data not directly applicable, and some discretization methods will destroy the spatial structure of continuous data. Therefore, scholars usually use measurement methods such as fuzzy particles and neighborhood particles to describe the relationship between spatial data.

[0092] Fuzzy similarity relation: given domain If the relationship The following conditions are met:

[0093] 1) Reflexivity: ;

[0094] 2) Symmetry: ;

[0095] Then it is called for The fuzzy similarity relationship defined in , the expression is:

[0096] ;

[0097] Where: For continuous data, It is discrete data;

[0098] Fuzzy Neighborhood Equivalence Class: Given a Domain , for any , its fuzzy neighborhood equivalence class, its expression is:

[0099] ;

[0100] Where: is the fuzzy neighborhood radius, and ;

[0101] Fuzzy Neighborhood Conditional Probability: Given a Decision System , , is a collection of objects with the same nominal decision class, in the condition attribute set Next, for any object , its fuzzy neighborhood equivalence class is , the fuzzy neighborhood conditional probability is expressed as:

[0102] ;

[0103] Among them, in the equivalence class Under the assumption that , if and only if hour, ;

[0104] In the framework of decision rough set theory, the loss function matrix based on data relationship dependence is given, and its expression is:

[0105] ;

[0106] Where i= P , B or N ; j =1 or c ;

[0107] is the fuzzy neighborhood conditional probability; is the compensation coefficient, satisfying For Marked as correspond The sample relative to The importance of is expressed as:

[0108] ;

[0109] For Marked as correspond The sample relative to The importance of is expressed as:

[0110] ;

[0111] For non-empty equivalence classes ,exist Under the assumption that , if and only if hour, ;

[0112] The loss function matrix based on data relationship dependence, and ,make , and get the probability threshold parameter The new expression is:

[0113] ;

[0114] ;

[0115] Therefore, the updated decision rule is expressed as:

[0116] P )like ,but ;

[0117] B )like ,but ;

[0118] N )like ,but ;

[0119] The two probability threshold parameters are expressed as and ,in , ;parameter After determination, we can get and ; The only hyperparameters to be determined are unknown parameters and ; The rough set model of fuzzy neighborhood decision making is established as follows:

[0120] ;

[0121] ;

[0122] .

[0123] In the embodiment of the present invention, obtaining the reduced subset includes:

[0124] Based on the reduction principle of minimizing decision risk in the fuzzy neighborhood decision rough set model, the expression of decision risk is:

[0125] ;

[0126] Where: ; It is a global decision-making risk; in the decision-making process, All possible risks are fully taken into consideration, so it can be called a global decision-making risk.

[0127] Relative importance: Given a decision system , then for the conditional attribute The relative importance of is expressed as:

[0128] ;

[0129] Attribute reduction criterion: Given a decision system ,like The following two conditions are met, and the expression is:

[0130] 1) ;

[0131] 2) , both have ;

[0132] but For comparison The reduced subset with low risk is combined with the optimization algorithm to obtain the optimal reduced subset.

[0133] Therefore, a heuristic forward greedy algorithm can be described as a pseudo code, as shown in Table 1 below.

[0134] Table 1 Pseudocode of the heuristic forward greedy algorithm

[0135] ;

[0136] As shown in Table 1, the above algorithm is based on the idea of ​​heuristic greedy search. First, the algorithm tries to find the attribute subset with the lowest risk by using a while loop. Then, a reverse deletion operation is performed to refine the selected subset according to the attribute reduction criterion. Finally, the optimal reduced subset is obtained after two rounds of loop operations.

[0137] Step 3: Input the obtained simplified subset into the multi-classifier model to obtain the classification result, and perform classification accuracy evaluation on the classification result. Based on the classification accuracy evaluation, complete the classifier weighting to obtain the classifier weight and the trained multi-classifier model.

[0138] In the embodiment of the present invention, classification as one of the three major tasks of machine learning is a supervised learning process, the purpose of which is to train a classification model based on a manually annotated classification sample data set, which is often referred to as a classifier. The classifier can be used to predict new sample data and map the new sample data to a certain category in the existing categories. In the classification learning process, the input training sample data set must contain information such as features (Feature, also known as Attribute), labels (Label, also known as Class), and the specific sample data can be expressed as, which has the same specific meaning as in the decision system. The essence of classification learning is to find the mapping relationship between features and labels, that is, the classifier is essentially a mapping function from input variables (features) to output variables (labels). When a set of unknown sample data with features but no labels is input, the label of the unknown sample data can be predicted by the classifier.

[0139] For classification tasks, scholars have proposed a variety of strategies to achieve the fitting and prediction of mapping relationships, such as decision tree classifiers, naive Bayes classifiers, support vector machine classifiers, etc. Since each classifier has a different classification mechanism, the data applicable conditions of each classifier are different, and the classification results are different.

[0140] The reduced subsets were fed into multiple classifiers, and the classification results were 0.9429, 0.9143, 0.9238, and 0.9190, respectively. The weights representing the credibility of each classifier were: {0.2548, 0.2471, 0.2497, 0.2484}. However, before reduction, the reduced results of all data fed into the four classifiers were 0.9400, 0.8714, 0.8857, and 0.8810, respectively. The comparison of classification accuracy before and after reduction is shown in Figure 2. Figure 3 As shown, this proves that after data reduction, the classification accuracy of all classifiers has been improved.

[0141] Therefore, based on the calculation of the above process, the risk assessment results of all test data can be obtained, as shown in Table 2. In the evaluation results, the FDA method, CPM method, and DBA method were selected for comparative analysis. The evaluation accuracy analysis of the method was carried out using a 10-fold cross validation method, that is, the original sample was randomly divided into 10 parts, 9 of which were used as training sets and 1 as a test set. After 20 rounds of testing, the average results were shown in Table 2.

[0142] Table 2 Hazard assessment results

[0143] ;

[0144] The results in Table 2 show that the classification accuracy of the proposed method is the highest, and a more accurate assessment of rock burst risk status can be obtained, followed by the FDA method, while the CPM method has the lowest classification accuracy, which is only 81.82%.

[0145] Step 4: Input the test data set into the trained multi-classifier model for testing, obtain the test results, and weightedly fuse the test results to obtain the hazard assessment level.

[0146] In the embodiment of the present invention, step 4 includes:

[0147] The prediction results of each classifier are weighted and summed up into one result; the weight is determined by the training accuracy conversion coefficient of the classifier on the data set, and its expression is:

[0148] ;

[0149] in, For the i The result weight of the classifier is For its training classification accuracy, n is the number of classifiers;

[0150] The final decision result is a weighted voting result, that is, each classifier gives a judgment result for each category. , then the judgment of the multi-classifier for a category is The output result is the category with the largest discrimination result, so as to obtain the hazard assessment level.

[0151] The present invention is based on experimental verification analysis of Luwa Coal Mine:

[0152] Luwa Coal Mine is located in Zhanghuang Town, Yutai County, Jining City. The 3305 working face is currently mining at the -450m level, with a maximum mining depth of 390m, an average coal seam thickness of 2.2m, a coal seam inclination of 8°, and no folds with drastic changes in inclination.

[0153] In identifying the influencing factors of rock burst in Luwa Coal Mine, we only focus on the influencing factor A={a2, a3, a4, a5, a7, a9, a10, a11}. The relevant information is shown in Table 3.

[0154] Table 3 Statistics of factors affecting rock burst in coal mines

[0155] ;

[0156] Using the trained multi-classifier model, we can obtain the results shown in Table 4.

[0157] Table 4 Evaluation results of multiple classifiers for rock burst in coal mines

[0158] ;

[0159] Through the calculation in Table 4, the four classifiers all determined that the hazard level of the Luwa module was 1. After being fused with the weighted information, the weighted level was obtained. Then, through accumulation, it was found that the rock burst hazard level of the Luwa Coal Mine was 1, which was the same as the actual rock burst hazard level of the mine assessed by experts, proving that the method of the present invention is practical and effective.

[0160] In order to realize the assessment and prediction of rock burst hazard level in rock coal mines, the present invention proposes a method for evaluating the rock burst hazard of coal mines from the perspective of data mining and machine learning, based on the characteristics of rock burst hazard level evaluation data sets. The method is based on fuzzy neighborhood decision rough sets and multi-classifier fusion methods, which reduces the collection cost of influencing factors in the rock burst hazard level assessment and improves the classification and recognition ability of the classifier, thereby providing a more accurate hazard level assessment result. The comparative experimental results show that the method proposed in the present invention is more advantageous than other commonly used methods in terms of hazard level identification accuracy. In the hazard level evaluation verification based on Luwa Coal Mine, the present invention provides a set of definite solutions, which is convenient for the rapid promotion of the method of the present invention, and also verifies the effectiveness of the method of the present invention. In addition, the method proposed in the present invention can also be applied to other fields such as safety level evaluation and state determination.

[0161] Each step of the embodiment of the present invention may be performed by an electronic device, which includes but is not limited to a tablet computer, a portable PC, a desktop computer, etc.

[0162] In the technical solution provided by the present invention, the method includes constructing an index system for evaluating the rock burst hazard of the amount of coal and rock ejected, and obtaining a training data set and a test data set through data regularization; optimizing and analyzing the constructed index system using a fuzzy neighborhood decision rough set model and the training data set to obtain a simplified subset; inputting the obtained simplified subset into a multi-classifier model to obtain a classification result, and performing a classification accuracy evaluation on the classification result, completing the weighting of the classifier based on the classification accuracy evaluation, and obtaining the classifier weight and the trained multi-classifier model; inputting the test data set into the trained multi-classifier model for testing to obtain a test result, and weightedly fusing the test results to obtain a hazard evaluation level. The method realizes the accurate classification of the rock burst hazard level of coal mines.

[0163] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the electronic device where the computer-readable storage medium is located is controlled to execute the above-mentioned embodiment of the method for assessing the risk of rock burst in coal mines.

[0164] Figure 4A schematic diagram of an electronic device provided by an embodiment of the present invention, such as Figure 4 As shown, the electronic device 21 includes: a processor 211, a memory 212, and a computer program 213 stored in the memory 212 and executable on the processor 211. When the computer program 213 is executed by the processor 211, the method for assessing the danger of rock burst in coal mines in the embodiment is implemented. To avoid repetition, they are not described one by one here.

[0165] The electronic device 21 includes, but is not limited to, a processor 211 and a memory 212. Those skilled in the art will appreciate that Figure 4 It is only an example of the electronic device 21 and does not constitute a limitation of the electronic device 21. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0166] The processor 211 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0167] The memory 212 may be an internal storage unit of the electronic device 21, such as a hard disk or memory of the electronic device 21. The memory 212 may also be an external storage device of the electronic device 21, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card (FlashCard), etc. equipped on the electronic device 21. Further, the memory 212 may also include both an internal storage unit of the electronic device 21 and an external storage device. The memory 212 is used to store computer programs and other programs and data required by network devices. The memory 212 may also be used to temporarily store data that has been output or is to be output.

[0168] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0169] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for assessing the risk of rock burst in a coal mine, characterized in that: The method comprises: Step 1: construct an index system for evaluating the rock burst hazard of coal and rock eruption, and obtain training data sets and test data sets through data regularization; Step 2: Use the fuzzy neighborhood decision rough set model FNDTRS and the training data set to optimize and analyze the constructed indicator system to obtain a simplified subset; Step 3: Input the obtained reduced subset into the multi-classifier model to obtain the classification result, and perform classification accuracy evaluation on the classification result. According to the classification accuracy evaluation, the classifier is weighted to obtain the classifier weight and the trained multi-classifier model; Step 4: Input the test data set into the trained multi-classifier model for testing to obtain the test results, and perform weighted fusion of the test results to obtain the risk assessment level; The fuzzy neighborhood decision rough set model in step 2 includes: Decision system: It is a binary group ,in For the domain, is the conditional attribute set, is the decision attribute, and ; Fuzzy similarity relation: given domain If the relationship The following conditions are met: 1) Reflexivity: ; 2) Symmetry: ; Then it is called for The fuzzy similarity relationship defined in , the expression is: ； Where: For continuous data, It is discrete data; Fuzzy Neighborhood Equivalence Class: Given a Domain , for any , its fuzzy neighborhood equivalence class, its expression is: ； Where: is the fuzzy neighborhood radius, and ; Fuzzy Neighborhood Conditional Probability: Given a Decision System , , is a collection of objects with the same nominal decision class, in the condition attribute set Next, for any object , its fuzzy neighborhood equivalence class is , the fuzzy neighborhood conditional probability is expressed as: ； Among them, in the equivalence class Under the assumption that , if and only if hour, ; In the framework of decision rough set theory, the loss function matrix based on data relationship dependence is given, and its expression is: ； Where i= P , B or N ; j =1 or c ; is the fuzzy neighborhood conditional probability; is the compensation coefficient, satisfying For Marked as correspond The sample relative to The importance of is expressed as: ； For Marked as correspond The sample relative to The importance of is expressed as: ； For non-empty equivalence classes ,exist Under the assumption that , if and only if hour, ; The loss function matrix based on data relationship dependence, and ,make , and get the probability threshold parameter The new expression is: ； ； Therefore, the updated decision rule is expressed as: P )like ,but ; B )like ,but ; N )like ,but ; The two probability threshold parameters are expressed as and ,in , ;parameter After determination, we can get the distribution information of the original data and ; The only hyperparameters to be determined are unknown parameters and ; The rough set model of fuzzy neighborhood decision making is established as follows: ； ； ； The obtaining of the reduced subset comprises: Based on the reduction principle of minimizing decision risk in the fuzzy neighborhood decision rough set model, the expression of decision risk is: ； Where: ; For global decision-making risks; Relative importance: Given a decision system , then for the conditional attribute The relative importance of is expressed as: ； Attribute reduction criterion: Given a decision system ,like The following two conditions are met, and the expression is: 1） ； 2) , both have ; but For comparison The reduced subset with low risk is combined with the optimization algorithm to obtain the optimal reduced subset.

2. The method according to claim 1, characterized in that The index system for evaluating the rock burst hazard of coal and rock eruption includes: A total of 11 influencing factors, including coal thickness a1, inclination a2, burial depth a3, structural conditions a4, inclination change a5, coal thickness change a6, gas concentration a7, roof management a8, pressure relief a9, coal cannon sound a10 and coal and rock eruption a11, are used as evaluation index elements for rock burst.

3. The method according to claim 1, characterized in that The step 4 comprises: The prediction results of each classifier are weighted and summed up into one result; the weight is determined by the training accuracy conversion coefficient of the classifier on the data set, and its expression is: ； in, For the i The result weight of the classifier is For its training classification accuracy, n is the number of classifiers; The final decision result is a weighted voting result, that is, each classifier gives a judgment result for each category. , then the judgment of the multi-classifier for a category is The output result is the category with the largest discrimination result, so as to obtain the hazard assessment level.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for assessing the risk of rock burst in coal mines as described in any one of claims 1 to 3.

5. An electronic device, characterized in that: include: one or more processors; Memory; And one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions, which, when executed by the device, enable the device to execute the method for assessing the risk of rock burst in coal mines as described in any one of claims 1 to 3.

Images