Rock burst intensity grade prediction method and system based on comprehensive weighting and set pair analysis

Abstract

The invention relates to the technical field of underground engineering, in particular to a rockburst intensity grade prediction method and system based on comprehensive weighting and set pair analysis, and the method comprises the steps: selecting prediction indexes, building a rockburst intensity grade prediction index system, and dividing the rockburst intensity grade into four grades based on the prediction index system, the method comprises the steps of collecting and preprocessing rockburst engineering case data, calculating a set pair connection degree function of prediction indexes by using a set pair analysis theory, determining subjective weights of rockburst intensity grade prediction indexes based on an improved analytic hierarchy process, determining objective weights of the rockburst intensity grade prediction indexes based on a variable coefficient method, introducing a distance function, and determining comprehensive weights of the indexes. And determining the relation degree of the rockburst intensity grade comprehensive set pair based on the index comprehensive weight, and determining the rockburst intensity grade by integrating the maximum membership criterion and the confidence criterion. The rockburst intensity grade prediction method has the beneficial effect of improving the accuracy and reliability of rockburst intensity grade prediction.

Description

Technical Field

[0001] The present invention relates to the field of underground engineering technology, and more specifically to a rockburst intensity level prediction method and system based on comprehensive weighting and set pair analysis. Background Art

[0002] Rockburst is a common dynamic geological disaster in the excavation of mines, tunnels and various underground space projects. It is often manifested as rock peeling, rock fragmentation in the empty area and explosive ejection. The occurrence of rockburst is usually related to the sudden release of elastic strain energy accumulated in rock excavation, other external disturbances or high tectonic stress rock mass. The sudden rupture and fall of rocks seriously endanger the safety of workers and equipment on site. In recent years, as my country's underground space projects have begun to develop deeper, rockburst disasters have occurred frequently during on-site construction, and it is very urgent to effectively and accurately predict rockbursts.

[0003] Many scholars at home and abroad have conducted extensive research on rockburst prediction from the perspectives of empirical criteria, mathematical theoretical models, machine learning, and intelligent algorithms. The empirical criteria are single-factor rockburst prediction methods, such as the Russenes criterion, Tao Zhenyu criterion, Singh rockburst proneness index criterion, energy criterion, and Turchaninov criterion, all of which use the established rockburst criteria to predict the rockburst level. Multi-factor prediction methods such as decision trees, cloud models, efficacy coefficient methods, gradient boosting trees, neural networks, ideal point methods, and extension theory have achieved certain results in the accuracy of rockburst prediction, but there are still some shortcomings. For example, traditional mathematical methods such as efficacy coefficient methods, decision trees, and Bayesian models are seriously affected by human subjectivity in determining indicator weights and qualitative factors, and it is difficult to objectively predict the rockburst intensity level. Deep learning algorithms such as SVM models, PCA-PNN models, and PCA-SOFM neural network models require a large amount of rockburst data for training, which will increase the complexity of the model and extend the calculation time. The key to predicting multi-factor rockburst problems using uncertainty theory is to reasonably determine the index weights and the reliability of rockburst grade judgment. The decision or evaluation results of a single weighting method such as subjective weights are highly arbitrary, while the objective weight method based on mathematical theory does not involve the intention of the decision maker. Therefore, it is very easy to cause a certain difference between the prediction results and the actual situation.

[0004] Similar prior art includes a Chinese patent application with publication number CN112699553A, which discloses a rockburst intensity level intelligent prediction system method, belonging to the field of underground engineering and geotechnical engineering technology. The method of the invention includes selecting rockburst intensity level prediction indicators; constructing a rockburst intensity level prediction indicator system; determining specific classification standards for rockburst intensity levels; collecting typical rockburst engineering case data at home and abroad; determining the subjective and objective weight values ​​of rockburst intensity level prediction indicators through the Delphi method and the improved entropy weight method; using the preference coefficient method to determine the optimal combination weight value of the rockburst intensity level prediction indicators; using the multidimensional normal cloud model theory and python programming means to determine the comprehensive membership of the rockburst intensity level; determining the rockburst intensity level according to the maximum membership principle; and independently developing a rockburst intensity level prediction system based on python software.

[0005] There is also a Chinese patent application with publication number CN115130375A, which provides a rockburst intensity prediction method, belonging to the technical field of geotechnical engineering and underground excavation engineering, the method comprising: determining evaluation indicators for rockburst intensity prediction and rockburst intensity classification; establishing a rockburst intensity prediction sample data set according to the determined rockburst intensity prediction evaluation indicators and rockburst intensity classification; expanding the minority classes in the data set; based on the expanded data set, using cross-validation and African vulture optimization algorithm to optimize the hyperparameters in the XGBoost integrated algorithm, and establishing an AVOV-XGBoost rockburst intensity prediction model; inputting the rockburst sample to be predicted into the established AVOV-XGBoost rockburst intensity prediction model to obtain the grade prediction value of the rockburst intensity.

[0006] However, when constructing the models, the above two methods for predicting rockburst intensity levels have failed to fully consider the problems that may be encountered in the process of collecting actual rockburst engineering data. Due to the limitations of sensor performance and complex and changeable environmental factors, the collected data may often contain errors or noise. If these factors are not properly handled, they will have an adverse effect on the accuracy of rockburst intensity level prediction, which may cause the prediction results to deviate from the actual situation. Therefore, the present invention provides a rockburst intensity level prediction method and system based on comprehensive weighting and set pair analysis to improve the accuracy and reliability of rockburst intensity level prediction. Summary of the invention

[0007] In order to better solve the above problems, the present invention provides a rockburst intensity prediction method based on comprehensive weighting and set pair analysis, and is implemented through the following steps:

[0008] Step S1: Select rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index Wet and rock mass integrity factor K v as a prediction indicator, and establishing the rockburst intensity level prediction indicator system based on the prediction indicator;

[0009] Step S2: Based on the prediction index system, the rockburst intensity level is divided into four levels: level I no rockburst, level II weak rockburst, level III moderate rockburst and level IV strong rockburst, so as to determine the rockburst intensity level classification standard;

[0010] Step S3: Collect the first rock burst engineering data from the historical rock burst engineering case database, and obtain the data within a preset time period from the first rock burst engineering data as the target data, and calculate the rock mass stress coefficient σ in the target data. θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock mass stress coefficient σ that changes with time c / σ t and the elastic energy index W et , and determine the target correction technology, use the target correction technology to correct the first rock burst engineering data, and obtain the corrected second rock burst engineering data, wherein the rock burst engineering data includes the first rock burst engineering data and the second rock burst engineering data, and also includes the rock mass stress coefficient σ θ / σ c , the rock brittleness coefficient σ c / σ t , the elastic energy index W et and the rock mass integrity factor K v ;

[0011] Step S4: Based on the second rockburst engineering data, using set pair analysis theory, calculate the rockburst intensity level prediction index x l The set pair connection function of ;

[0012] Step S5: determining the subjective weight of the rockburst intensity level prediction index based on the improved analytic hierarchy process;

[0013] Step S6: determining the objective weight of the rockburst intensity level prediction index based on the coefficient of variation method;

[0014] Step S7: Determine the comprehensive weight of the indicator by introducing the distance function;

[0015] Step S8: determining the comprehensive set pair connection degree of rockburst intensity levels based on the comprehensive weight of the indicators;

[0016] Step S9: Based on the rockburst intensity level comprehensive set pair connection degree, the maximum membership criterion and the confidence criterion are combined to determine the rockburst intensity level.

[0017] As a preferred technical solution of the present invention, in step S3, the rock mass stress coefficient σ in the target data is θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock mass stress coefficient σ that changes with time c / σ t and the elastic energy index W et The abnormal value of the rockburst is detected and the corrected second rockburst engineering data is obtained, which also includes:

[0018] Step S31: The rock mass stress coefficient σ in the target data θ / σ c and the elastic energy index W et Purification is performed to generate a normal time series of rock mass stress coefficients σ θ / σ c and elastic energy index W et , wherein the normal time series refers to a series of values ​​arranged in chronological order, and the rock mass stress coefficient σ θ / σ c and the elastic energy index W et They are usually collected at equal or unequal time intervals;

[0019] Step S32: Obtain the first rock mass stress coefficient σ in the target data θ / σ c When the time series is greater than the first threshold or less than the second threshold, the X data sets corresponding to the time series are statistically analyzed for the rock mass stress coefficient σ in the historical rock burst engineering data. θ / σ c , and randomly select X second outliers from the probability distribution of the first outliers, insert the X second outliers into the corresponding X data sets, and obtain the second rock mass stress coefficient σ θ / σ c , where X represents a positive integer greater than or equal to 1 and less than Sum, Sum represents the number of the first rock mass stress coefficients in the target data, and the first abnormal value refers to the rock mass stress coefficient σ in the historical rock burst engineering data θ / σ c The measured value in the equation is higher than the normal maximum value or lower than the normal minimum value; similarly, the second elastic energy index W is obtained. et ;

[0020] Step S33: Using a variety of data correction techniques, the second rock mass stress coefficient σ θ / σ c and the second elastic energy index Wet Perform correction to obtain multiple sets of correction results, based on the rock mass stress coefficient σ of the normal time series θ / σ c and the elastic energy index W et The rock mass stress coefficient σ is plotted against multiple sets of calibration results. θ / σ c and the elastic energy index W et The change curve over time determines the rock mass stress coefficient σ that is closest to the normal time series θ / σ c and elastic energy index W et The correction technology corresponding to the correction result of the change curve is the target correction technology;

[0021] Step S34: using the target correction technology to correct the first rockburst engineering data to obtain second rockburst engineering data.

[0022] As a preferred technical solution of the present invention, in step S31, the rock mass stress coefficient σ in the target data is θ / σ c and the elastic energy index W et Purification is performed to generate a normal time series of rock stress coefficient σ θ / σ c Number and elastic energy index W et ,include:

[0023] Step S311: acquiring multiple data records with completely identical data records from the first rockburst engineering data, and deleting duplicate data from the data records to acquire non-duplicate first rockburst engineering data;

[0024] Step S312: Obtain the rock mass stress coefficient σ from the first non-repeated rock burst engineering data θ / σ c , and the rock mass stress coefficient σ θ / σ c Convert to the data format of the same data type and the same unit;

[0025] Step S313: Retrieve the rock mass stress coefficient σ θ / σ c , and use the non-null value that is closest to the null value and earlier in time to supplement the null value;

[0026] Step S314: Determine the rock mass stress coefficient σ θ / σ c range, retrieve the rock mass stress coefficient σ θ / σ cThe outliers that deviate from the range are replaced by the median to generate the rock mass stress coefficient σ after replacing the outliers. θ / σ c The rock mass stress coefficient σ as a normal time series θ / σ c , wherein the median refers to the value located in the middle position after N values ​​are taken consecutively from the data adjacent to the abnormal value and arranged in order of size, and N represents a positive integer greater than or equal to 1;

[0027] Step S315: Repeat steps S312 to S314 to generate the elastic energy index W of the normal time series et .

[0028] As a preferred technical solution of the present invention, in step S4, the rockburst intensity level prediction index x is calculated. l The set pair connection function is as follows:

[0029] μ l =a l +b l,1 p1+b l,2 p2+…+b l,k-2 p k-2 +c l q

[0030] Among them, a l +b l,1 +b l,2 +…+b l,k-2 +c l =1, l represents the number of rockburst intensity prediction indicators l=1, 2, ..., m, m represents the number of indicators, m is less than or equal to k, k represents the rockburst intensity level, a l Represents the prediction index x l The probability of belonging to the level 1 standard, b l,1 , b l,2 ,……,b l,k-2 Respectively represent the prediction index x l The probability of belonging to the 1st, 2nd, 3rd, …, k-1th level standard, c l Represents the predicted index x l The probability of belonging to the k-level standard, p1, p2, ..., p k-2 ∈[-1, 1] represents the difference coefficient, q represents the opposition coefficient, q=-1.

[0031] As a preferred technical solution of the present invention, the step S5 further includes: introducing an exponential function to optimize the transfer matrix, improving the hierarchical analysis method for calculating the subjective weight of the prediction index, and the calculation steps are as follows:

[0032] Step S51: Determine the comparison matrix: Combined with expert opinions, the comparison matrix A of the prediction index is established using the 3-scale method:

[0033]

[0034] Among them, a ij represents the importance of factor i relative to factor j in the comparison matrix, i, j = 1, 2, 3, ..., n, n represents the total number of factors, and satisfies a ij =0,a ij +a ji =0,a ij The meanings of different values ​​are:

[0035]

[0036] Step S52: Determine the optimal transfer matrix: Use a weighted method to optimize the comparison matrix to obtain the optimal transfer matrix R:

[0037]

[0038] in,

[0039] Step S53: Determine the comprehensive judgment matrix: introduce an exponential function to calculate the optimal transfer matrix to obtain the optimized comprehensive judgment matrix D:

[0040]

[0041] Among them, d in =exp(r ij );

[0042] Step S54: Calculation of subjective weight: The subjective weight is calculated using the square root method. The calculation process is as follows:

[0043]

[0044] Among them, ω Ai * represents the subjective weight vector, ω Ai Represents the subjective weight of the prediction indicator.

[0045] As a preferred technical solution of the present invention, the step S6 determines the objective weight of the prediction index of the rockburst intensity level based on the coefficient of variation method, and the calculation formula is as follows:

[0046]

[0047] Among them, V irepresents the coefficient of variation of the objective indicator, Si represents the standard deviation, represents the average value, ω Ei Represents the objective weight of the prediction indicator.

[0048] As a preferred technical solution of the present invention, step S7 further includes:

[0049] By introducing the distance function, assigning different preference coefficients of objective weights of the coefficient of variation method and subjective weights of the improved analytic hierarchy process, the optimal comprehensive weights of the rockburst intensity prediction index are determined, and the calculation formula is as follows:

[0050]

[0051] Wherein, L represents the distance function between the coefficient of variation method and the improved analytic hierarchy process, ω represents the comprehensive weight, and y and z represent the preference coefficients of the improved analytic hierarchy process and the coefficient of variation method, respectively.

[0052] As a preferred technical solution of the present invention, step S8 further includes:

[0053] The process of determining the connection degree of the comprehensive set of rockburst intensity level prediction is as follows:

[0054] According to the rockburst intensity level set in step S4, the connection function μ l Construct the prediction indicator connection matrix T:

[0055]

[0056] Among them, t a1 Represents the prediction index connection function μ l Constant term, t l,1 represents the coefficient of p1 in the connection function, t l,2 represents the coefficient of p2 in the connection function, t cl represents the coefficient of q;

[0057] The optimized comprehensive weight in the above step S7 is combined with the set pair connection degree to determine the comprehensive connection degree μ for predicting the rockburst intensity level. A-B , and its calculation formula is expressed as formula 1:

[0058]

[0059] Wherein, u represents the comprehensive weight vector of the prediction index, [ω1, ω2, ..., ω l ], ω1 represents the comprehensive weight of the prediction index;

[0060] The formula 1 is simplified as:

[0061] μ A-B =f1+f2p1+f3p2+f4q

[0062] Among them, f1 represents the possibility that the rockburst intensity level belongs to level I without rockburst; f2 represents the possibility that it belongs to level II weak rockburst; f3 represents the possibility that it belongs to level III moderate rockburst; and f4 represents the possibility that it belongs to level IV strong rockburst.

[0063] As a preferred technical solution of the present invention, in step S9, determining the rockburst intensity level further includes:

[0064] In the comprehensive connection degree μ A-B Belongs to the above-mentioned level III moderate rock burst or f h When it is greater than or equal to the third threshold, the maximum membership criterion is used to determine the rockburst intensity level;

[0065] When f h When λ is less than the third threshold and the difference between the memberships of the various levels is small, the maximum membership criterion is used to easily misjudge the rockburst level, and then the confidence criterion is used to judge the rockburst level by introducing the confidence λ;

[0066] The corresponding membership degree at any level is: f h =f1+f2+…+f h ≥λ, the predicted rockburst level is h, where h is a positive integer between 1 and 4, and the confidence level λ∈[0.5,0.7].

[0067] The present invention also provides a rockburst intensity prediction system based on comprehensive weighting and set pair analysis as described above, comprising the following modules:

[0068] Construction unit for selecting rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index W et and rock mass integrity factor K v as a prediction indicator, and establishing the rockburst intensity level prediction indicator system based on the prediction indicator;

[0069] A grading unit, used to classify the rockburst intensity level into four levels: level I no rockburst, level II weak rockburst, level III moderate rockburst and level IV strong rockburst, so as to determine the rockburst intensity level grading standard;

[0070] A preprocessing unit is used to collect first rock burst engineering data from a historical rock burst engineering case database, and obtain data within a preset time period from the first rock burst engineering data as target data, and perform a preprocessing on the rock mass stress coefficient σ in the target data. θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock mass stress coefficient σ that changes with time c / σ t and the elastic energy index W et , and determine the target correction technology, use the target correction technology to correct the first rock burst engineering data, and obtain the corrected second rock burst engineering data, wherein the rock burst engineering data includes the first rock burst engineering data and the second rock burst engineering data, and also includes the rock mass stress coefficient σ θ / σ c , the rock brittleness coefficient σ c / σ t , the elastic energy index W et and the rock mass integrity factor K v ;

[0071] The calculation unit is configured as follows: using set pair analysis theory, calculate the rockburst intensity level prediction index x l The set pair connection function; based on the improved analytic hierarchy process, the subjective weight of the prediction index of the rockburst intensity level is determined; based on the coefficient of variation method, the objective weight of the prediction index of the rockburst intensity level is determined; by introducing the distance function, the comprehensive weight of the index is determined; based on the comprehensive weight of the index, the comprehensive set pair connection degree of the rockburst intensity level is determined;

[0072] The prediction unit is used to determine the rockburst intensity level based on the rockburst intensity level comprehensive set pair connection degree, comprehensive maximum membership criterion and confidence criterion.

[0073] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0074] The present invention selects prediction indicators and establishes a rockburst intensity prediction index system to more accurately and objectively predict the rockburst intensity level and reduce the possibility of false alarms and missed alarms; the rockburst intensity level is divided into four levels based on the prediction index system to determine the rockburst intensity level classification standard; the first rockburst engineering data is corrected by target correction technology to improve the accuracy of the collected rockburst engineering case data.

[0075] The present invention also calculates the rockburst intensity level prediction index x by using set pair analysis theory. lThe set pair connection function is used to construct the comparison matrix of the hierarchical analysis method using the three-scaling method, and the exponential function is introduced to optimize the transfer matrix, which avoids the consistency test, simplifies the calculation process, and improves the hierarchical analysis method for calculating the subjective weight of the index; based on the improved hierarchical analysis method, the subjective weight of the rockburst intensity level prediction index is determined; based on the coefficient of variation method, the objective weight of the rockburst intensity level prediction index is determined; by introducing the distance function, the comprehensive weight of the index is determined, the rationality of obtaining the preference coefficient of different weighting methods is improved, and the reliability of the comprehensive weighting is improved. The comprehensive set pair connection degree of the rockburst intensity level is determined based on the comprehensive weight of the index; based on the comprehensive set pair connection degree of the rockburst intensity level, the maximum membership criterion and the confidence criterion are combined to determine the rockburst intensity level, so that the rockburst intensity level prediction method has good accuracy and reliability. Through the coordination between the above steps, the rockburst intensity level prediction method can improve the accuracy and reliability of the rockburst intensity level prediction, and provide a new way to predict the rockburst intensity level in the field of deep rock underground excavation. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0077] Figure 1 It is a flowchart of the steps of the rockburst intensity level prediction method based on comprehensive weighting and set pair analysis in the present invention;

[0078] Figure 2 It is a composition structure diagram of the rockburst intensity level prediction system based on comprehensive weighting and set pair analysis in the present invention. DETAILED DESCRIPTION

[0079] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0080] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0081] Rockburst is a common dynamic geological disaster in the excavation of mines, tunnels and various underground space projects. It is often manifested as rock peeling, rock fragmentation in the empty area and explosive ejection. The occurrence of rockburst is usually related to the sudden release of elastic strain energy accumulated in rock excavation, other external disturbances or high tectonic stress rock mass. The sudden rupture and fall of rocks seriously endanger the safety of workers and equipment on site. In recent years, as my country's underground space projects have begun to develop deeper, rockburst disasters have occurred frequently during on-site construction, and it is very urgent to effectively and accurately predict rockbursts.

[0082] However, due to the limitations of sensor performance and complex and changeable environmental factors during the acquisition process of actual rockburst engineering data, the collected data may often contain errors or noise. If these factors are not properly handled, they will have an adverse effect on the accuracy of rockburst intensity prediction and may cause the prediction results to deviate from the actual situation.

[0083] In view of the above technical problems, the present invention proposes Figure 1 The rockburst intensity prediction method based on comprehensive weighting and set pair analysis is implemented through the following steps:

[0084] Step S1: Select rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index W et and rock mass integrity factor K v As a prediction indicator, a rockburst intensity level prediction index system is established based on the prediction indicators.

[0085] Specifically, in order to solve the problem of predicting rockburst disasters, prediction indicators are selected as evaluation criteria for the rockburst intensity level prediction index system. The selection of rockburst prediction indicators is an important and complex task. Establishing a rockburst intensity level prediction index system can more accurately and objectively predict the rockburst intensity level and reduce the possibility of false alarms and missed reports. The selected rockburst prediction indicators should be able to fully reflect the characteristics of rockbursts, and can reflect the internal and external conditions for rockbursts. They must also meet conditions that are relatively easy to obtain, that is, indicators can be obtained through indoor tests or field tests. Therefore, the present invention selects the following factors as indicators for evaluating the occurrence of rockbursts: rock mass stress coefficient σ θ / σ c It can describe the stress characteristics of surrounding rock, σ θ is the maximum tangential stress of the chamber and the surrounding rock stress, σ c is the uniaxial compressive strength. The rock properties can be expressed by the rock brittleness index σ c / σ t To describe, σ t is the uniaxial tensile strength of rock. Elastic energy index W etReflects the energy characteristics of rocks, W et It refers to the ratio of the elastic strain energy accumulated by the rock sample before reaching the peak strength to the unloading loss energy. The larger the value, the more energy is released during the failure process, and the integrity of the rock mass K v Reflects the degree of development of fissures and joints.

[0086] Step S2: Based on the prediction index system, the rockburst intensity level is divided into four levels: level I no rockburst, level II weak rockburst, level III moderate rockburst and level IV strong rockburst, so as to determine the rockburst intensity level classification standard.

[0087] Specifically, referring to the research results of Wang Hanyuan in "Fuzzy Mathematical Comprehensive Evaluation Method for Rockburst Prediction" and Zhou Keping in "Study on Rockburst Intensity Grading Prediction Based on Entropy Weight-Normal Cloud Model", the classification standards for rockburst intensity prediction indicators are sorted out, as shown in Table 1.

[0088] Table 1 Rockburst prediction indicators and classification standards

[0089]

[0090] Step S3: Collect the first rock burst engineering data from the historical rock burst engineering case database, and obtain the data within a preset time period from the first rock burst engineering data as the target data, and calculate the rock mass stress coefficient σ in the target data. θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock stress coefficient σ that changes with time c / σ t and elastic energy index W et The abnormal value of the rock burst engineering data is determined, and the target correction technology is used to correct the first rock burst engineering data, and the corrected second rock burst engineering data is obtained, wherein the rock burst engineering data includes the first rock burst engineering data and the second rock burst engineering data, and also includes the rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index W et and rock mass integrity factor K v .

[0091] Specifically, due to the limitations of sensor performance and complex and changeable environmental factors during the collection process of actual rock burst engineering case data, the collected data may often contain errors or noise. Since the accuracy of rock burst engineering data directly affects the accuracy of rock burst intensity level prediction, it is necessary to first use effective data correction technology to correct the errors caused by sensor performance limitations and environmental factors to improve the accuracy of the collected rock burst engineering case data. The first rock burst engineering data is the data directly extracted from the actual rock burst engineering case, and the second rock burst engineering data is the rock burst engineering data after preprocessing. The rock burst engineering data used below will all be the second rock burst engineering data. The data within the preset time period is obtained from the first rock burst engineering data as the target data because there are many first rock burst engineering data collected. If all are processed, the execution efficiency will be very low and it will take more time. Therefore, a part of the target data is selected through the preset time period to verify which correction method is better for processing or correcting rock burst engineering data, with higher accuracy, and which correction method is used to process the entire first rock burst engineering data. The specific method will be described in detail below.

[0092] At the same time, in order to verify the reliability and accuracy of the rockburst intensity prediction method proposed in the present invention, referring to the research results of Yin Xin in "Rockburst intensity classification prediction model based on combined weighting and attribute interval identification theory" and Zhou Keping in "Study on rockburst intensity classification prediction based on entropy weight-normal cloud model", 17 groups of typical rockburst engineering cases at home and abroad were selected, as shown in Table 2.

[0093] Table 2 Typical rock burst cases at home and abroad

[0094]

[0095]

[0096] Step S4: Based on the second rockburst engineering data, use the set pair analysis theory to calculate the rockburst intensity level prediction index x l The set pair connection function of .

[0097] Specifically, assuming that the rockburst intensity prediction index x l The set is A l , l = 1, 2, ..., m, set B h represents the prediction level corresponding to the prediction index, h = 1, 2, ..., k. Based on the set pair analysis theory, the rockburst intensity level prediction index x l The set pair connection function is the above expression.

[0098] According to the four rockburst intensity grade standards divided in step S2 above, the prediction index x is calculated. lThe three-element set connection function μ composed of the rockburst intensity level dividing points S1, S2, S3 (as shown in Table 3) l The rockburst intensity prediction index σ selected in step S1 is c / σ t The larger the better, σ θ / σ c , W et , K v The smaller the better. For the prediction indicators, the larger the better and the smaller the prediction indicators, the better, please see the detailed description below.

[0099] Table 3 Rockburst intensity level standards and standard thresholds

[0100]

[0101]

[0102] Step S5: Based on the improved analytic hierarchy process, the subjective weights of rockburst intensity level prediction indicators are determined.

[0103] Specifically, referring to the research of Yin Xin in "Rockburst intensity classification prediction model based on combined weighting and attribute interval identification theory" and Zhou Keping in "Study on rockburst intensity classification prediction based on entropy weight-normal cloud model", combined with the opinions of relevant experts, the rock mass stress characteristics, i.e., σ θ / σ c The most important factor for rock burst is the rock mass lithology condition σ c / σ t , the development of rock mass joints and fissures is more important than its energy characteristics. Therefore, the subjective weights of the rockburst intensity level prediction indicators are obtained according to the subjective weight calculation steps. The specific steps for determining the subjective weights of the rockburst intensity level prediction indicators will be described in detail below.

[0104] Step S6: Determine the objective weight of the rockburst intensity level prediction index based on the coefficient of variation method.

[0105] Specifically, the prediction index data of the rockburst engineering case collected in step S3 are used in combination with the coefficient of variation method to calculate the objective weight of the rockburst intensity level prediction index, and the specific content will be described in detail below.

[0106] Step S7: Determine the comprehensive weight of the indicators by introducing the distance function.

[0107] Specifically, in order to eliminate the influence of subjective and objective factors on the weight accuracy and comprehensively consider the reliability and accuracy of the prediction results, the present invention introduces a distance function to assign different preference coefficients to the subjective weights of the improved hierarchical analysis method and the objective weights of the coefficient of variation method, and determines the optimal comprehensive weights of the rockburst intensity level prediction indicators. The specific calculation will be described in detail below.

[0108] Step S8: Determine the comprehensive set-to-connectivity of rockburst intensity levels based on the comprehensive weights of the indicators.

[0109] Specifically, the process of determining the connection degree of the comprehensive set of rockburst intensity levels will be described below.

[0110] Step S9: Based on the rockburst intensity level comprehensive set pair connection degree, the maximum membership criterion and the confidence criterion are combined to determine the rockburst intensity level. The judgment method will be described in detail below.

[0111] Through the coordination of the above steps, the rockburst intensity level prediction method can improve the accuracy and reliability of rockburst intensity level prediction, and provide a new approach to the problem of rockburst intensity level prediction in the field of deep rock underground excavation.

[0112] Furthermore, in the above step S3, the rock mass stress coefficient σ in the target data is θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock stress coefficient σ that changes with time c / σ t and elastic energy index W et The abnormal value of the rockburst is detected and the corrected second rockburst engineering data is obtained, which also includes:

[0113] Step S31: The rock mass stress coefficient σ in the target data θ / σ c and elastic energy index W et Purification is performed to generate a normal time series of rock mass stress coefficients σ θ / σ c and elastic energy index W et , where the normal time series refers to a series of values ​​arranged in chronological order, and the rock mass stress coefficient σ θ / σ c and elastic energy index W et Data are usually collected at equal or unequal time intervals.

[0114] Specifically, the rock mass stress coefficient σ in the target data is selected θ / σ c and elastic energy index W etPurification is performed to improve data quality and reduce redundant and repeated data. Because the rock mass stress coefficient usually reflects the stress state of a certain point in the rock mass, it is the fundamental force that causes surrounding rock damage during underground engineering excavation. It is monitored over time or mining progress. It belongs to time series data and can reflect the changes in stress state over time. The elastic energy index reflects the situation in which the elastic strain energy accumulated in the rock after underground engineering excavation exceeds the energy required for rock damage. It is also monitored over time or mining progress. It also belongs to time series data and can reflect the changes in elastic strain energy in the rock over time. These characteristics make the rock mass stress coefficient σ θ / σ c and elastic energy index W et During purification, special consideration needs to be given to its temporal continuity and correlation, while the rock brittleness coefficient and rock mass integrity coefficient do not belong to time series data in the present invention. Of course, this method is also used to purify other time series data, which is not limited in the present invention.

[0115] Step S32: Obtain the first rock mass stress coefficient σ in the target data θ / σ c When it is greater than the first threshold or less than the second threshold, the X data sets corresponding to the time series are used to statistically analyze the rock mass stress coefficient σ in the historical rock burst engineering data. θ / σ c In the above example, X second outliers are randomly selected from the probability distribution of the first outliers, and the X second outliers are respectively inserted into the corresponding X data sets to obtain the second rock mass stress coefficient σ θ / σ c , where X represents a positive integer greater than or equal to 1 and less than Sum, Sum represents the number of the first rock mass stress coefficient in the target data, and the first outlier refers to the rock mass stress coefficient σ in the historical rock burst engineering data θ / σ c The measured value in the equation is higher than the normal maximum value or lower than the normal minimum value; similarly, the second elastic energy index W is obtained. et ;

[0116] Specifically, since the rock mass stress coefficient reflects the change of stress state over time, the change of rock mass stress coefficient in the time series is large or small according to the actual construction process. For several data sets with large changes, the above steps are further processed to increase the rock mass stress coefficient σ θ / σ cThe gap between them is used to verify the correction technology below. For example, the rock stress coefficients are 0.65, 0.45, 0.8, 0.66, 0.38, 0.42, 0.53, 0.24, 0.42, 0.39 and 0.255, which obviously do not conform to the rock stress coefficients corresponding to the time series. They may be abnormal values ​​caused by the environment during the acquisition process. The first threshold can be set to 0.7 and the second threshold can be set to 0.3. Then the above data is divided into 3 data sets. The first data set is 0.65, 0.45 and 0.8, the second data set is 0.66, 0.38, 0.42, 0.53 and 0.24, and the third data set is 0.42, 0.39 and 0.255. The first abnormal values ​​obtained are: 1.5, 0.9, 0.05 and 0.12, etc. Then select any 3 abnormal values ​​from the first abnormal values ​​and insert them into the corresponding 3 data sets respectively. The normal maximum value and the normal minimum value refer to the safe and appropriate range of rock stress coefficients.

[0117] Regarding the randomly selected X second abnormal values, inserting them into the corresponding X data sets respectively, further comprising: when the selected second abnormal value is greater than the first rock mass stress coefficient σ in the target data θ / σ c When any value is found in , the second largest outlier is inserted into the rock mass stress coefficient σ θ / σ c In the data set with the smallest average value, the second outlier selected is smaller than the first rock mass stress coefficient σ in the target data. θ / σ c When any value is found in , the smallest second abnormal value is inserted into the rock mass stress coefficient σ θ / σ c In order to further increase the rock mass stress coefficient σ θ / σ c The gap between them can better verify the correction technology described below.

[0118] Step S33: Use a variety of data correction techniques to correct the stress coefficient σ of the second rock mass. θ / σ c and the second elastic energy index W et Perform correction and obtain multiple sets of correction results based on the rock mass stress coefficient σ of the normal time series θ / σ c and elastic energy index W et Compared with multiple sets of calibration results, the rock stress coefficient σ is plotted respectively. θ / σ c and elastic energy index W et The curve of change over time determines the rock mass stress coefficient σ that is closest to the normal time series θ / σc and elastic energy index W et The correction technology corresponding to the correction result of the change curve is the target correction technology;

[0119] Step S34: Correct the first rockburst engineering data using a target correction technique to obtain second rockburst engineering data.

[0120] Specifically, a suitable correction technology is selected to reduce the impact of outliers and restore the authenticity of the data. For example, the Mann filter, LPF filter, Savitzky-Golay filter, moving average filter, etc. These correction technologies are all existing technologies and will not be described here. The method of each correction technology is independently applied to the rock stress coefficient σ after the outlier is inserted. θ / σ c and elastic energy index W et , obtain multiple sets of correction results, the correction results are closer to the true value, make the data smoother, and can reflect the change trend of the data, and determine the rock stress coefficient σ closest to the normal time series through the change trend θ / σ c and elastic energy index W et The correction technology corresponding to the correction result of the change curve is the target correction technology. The target correction technology is determined to determine a method suitable for correcting the first rockburst engineering data, and all the first rockburst engineering data are corrected with the determined correction technology.

[0121] Furthermore, in the above step S31, the rock mass stress coefficient σ in the target data is θ / σ c and elastic energy index W et Purification is performed to generate a normal time series of rock stress coefficient σ θ / σ c Number and elastic energy index W et ,include:

[0122] Step S311: acquiring multiple data records with completely identical data records from the first rockburst engineering data, and deleting duplicate data in the data records to acquire non-duplicate first rockburst engineering data;

[0123] Step S312: Obtain rock mass stress coefficient σ from the first non-repeated rock burst engineering data θ / σ c , and the rock mass stress coefficient σ θ / σ c Convert to the data format of the same data type and the same unit;

[0124] Step S313: Retrieve rock mass stress coefficient σ θ / σ cThe null value in is replaced by the non-null value closest to the null value and earlier in time;

[0125] Among them, the null value refers to the rock mass stress coefficient σ θ / σ c , such as values ​​represented by null, spaces, or NaN.

[0126] Step S314: Determine rock mass stress coefficient σ θ / σ c range, retrieve the rock mass stress coefficient σ θ / σ c The outliers that deviate from the range are replaced by the median to generate the rock mass stress coefficient σ after replacing the outliers. θ / σ c The rock mass stress coefficient σ as a normal time series θ / σ c , where the median refers to the value in the middle position after taking N consecutive values ​​from the data adjacent to the outlier and arranging them in order of size, and N represents a positive integer greater than or equal to 1;

[0127] Step S315: Repeat steps S312 to S314 to generate the elastic energy index W of the normal time series et .

[0128] Specifically, in the normally collected rock engineering data, there are inevitably duplications, null values, different numerical units or types, or there are many abnormal values. These situations will affect the quality of rock burst engineering data, and thus affect the accuracy of rock burst intensity level prediction. Therefore, the rock engineering data within the preset time period is used as the target data, and it is purified to provide a basis for selecting appropriate correction technology as reference data.

[0129] Furthermore, in the above step S4, the rockburst intensity level prediction index x is calculated. l The set pair connection function is as follows:

[0130] μ l =a l +b l,1 p1+b l,2 p2+…+b l,k-2 p k-2 +c l q

[0131] Among them, a l +b l,1 +b l,2 +…+b l,k-2 +c l=1, l represents the number of rockburst intensity prediction indicators l = 1, 2, ..., m, m represents the number of indicators, m is less than or equal to k, k represents the rockburst intensity level, a l Represents the prediction index x l The probability of belonging to the level 1 standard, b l,1 , b l,2 ,……,b l,k-2 Represent the prediction index x l The probability of belonging to the 1st, 2nd, 3rd, …, k-1th level standard, c l Represents the prediction index x l The probability of belonging to the k-level standard, p1, p2, ..., p k-2 ∈[-1, 1] represents the difference coefficient, q represents the opposition coefficient, q=-1.

[0132] Specifically, for the larger the prediction index, the better the model, μ l The expression is:

[0133]

[0134] In the formula: S1≥S2≥S3.

[0135] For the smaller the prediction index, the better the model, μ l The expression is:

[0136]

[0137] In the formula: S1≤S2≤S3.

[0138] Furthermore, the above step S5 also includes:

[0139] The exponential function is introduced to optimize the transfer matrix and improve the analytic hierarchy process for calculating the subjective weights of the prediction indicators. The calculation steps are as follows:

[0140] Step S51: Determine the comparison matrix: Combined with expert opinions, the comparison matrix A of the prediction index is established using the 3-scale method:

[0141]

[0142] Among them, a ij represents the importance of factor i relative to factor j in the comparison matrix, i, j = 1, 2, 3, ..., n, n represents the total number of factors, and satisfies a ij =0,a ij +a ji =0,a ij The meanings of different values ​​are:

[0143]

[0144] Step S52: Determine the optimal transfer matrix: Use the weighted method to optimize the comparison matrix to obtain the optimal transfer matrix R:

[0145]

[0146] in,

[0147] Step S53: Determine the comprehensive judgment matrix: introduce an exponential function to calculate the optimal transfer matrix to obtain the optimized comprehensive judgment matrix D:

[0148]

[0149] Among them, d in =exp(r ij );

[0150] Step S54: Calculate subjective weight: Calculate subjective weight using square root method. The calculation process is as follows:

[0151]

[0152] Among them, ω Ai * represents the subjective weight vector, ω Ai Represents the subjective weight of the prediction indicator.

[0153] Specifically, since the traditional hierarchical analysis method is greatly affected by the randomness of samples, a separate consistency test is required, and the calculation process is complicated. Therefore, the comparison matrix of the hierarchical analysis method is constructed using the three-scaling method, and the exponential function is introduced to optimize the transfer matrix, which avoids the consistency test and simplifies the calculation process. The hierarchical analysis method of the subjective weight of the calculation index is improved. The specific calculation steps are as mentioned above. Referring to the research of Yin Xin in "Rockburst Intensity Grading Prediction Model Based on Combined Weighting and Attribute Interval Recognition Theory" and Zhou Keping in "Research on Rockburst Intensity Grading Prediction Based on Entropy Weight-Normal Cloud Model", combined with the opinions of relevant experts, the rock mass stress characteristics, that is, σ θ / σ c The most important factor for rock burst is the rock mass lithology condition σ c / σ t , the development of rock mass joints and fissures is more important than its energy characteristics. Therefore, according to the subjective weight calculation steps, the subjective weights of each prediction index of rock burst intensity level are obtained as shown in Table 4.

[0154] Table 4 Results of subjective weight calculation using improved analytic hierarchy process

[0155] Predictive indicators <![CDATA[σ c / s t ]]> <![CDATA[σ θ / s c ]]> <![CDATA[W et ]]> <![CDATA[K v ]]> Subjective weight value 0.2760 0.4551 0.1015 0.1674

[0156] Furthermore, in the above step S6, based on the coefficient of variation method, the objective weight of the rockburst intensity level prediction index is determined, and the calculation formula is as follows:

[0157]

[0158] Among them, V i represents the coefficient of variation of the objective indicator, Si represents the standard deviation, represents the average value, ω Ei Represents the objective weight of the prediction indicator.

[0159] Specifically, the prediction index data of typical rockburst engineering cases collected in step S3 are used, and the objective weights of the prediction indexes of the rockburst intensity level are calculated in combination with the coefficient of variation method as shown in Table 5:

[0160] Table 5 Objective weight calculation results using the coefficient of variation method

[0161] Predictive indicators <![CDATA[σ c / s t ]]> <![CDATA[σ θ / s c ]]> <![CDATA[W et ]]> <![CDATA[K v ]]> Objective weight value 0.2943 0.2797 0.2685 0.1575

[0162] Furthermore, the above step S7 also includes:

[0163] By introducing the distance function, assigning different preference coefficients of objective weights of the coefficient of variation method and subjective weights of the improved analytic hierarchy process, the optimal comprehensive weights of the rockburst intensity prediction index are determined. The calculation formula is as follows:

[0164]

[0165] Among them, L represents the distance function between the coefficient of variation method and the improved hierarchical analysis method, ω represents the comprehensive weight, y and z represent the preference coefficients of the improved hierarchical analysis method and the coefficient of variation method, respectively.

[0166] Specifically, the single weight determination method can only obtain the indicator weight from a subjective or objective perspective, which can easily lead to deviations in the results. The comprehensive weighting rule is to assign different preference coefficients to subjective weights and objective weights, and combine the two to determine the comprehensive method of indicator weights, so as to obtain more reasonable indicator weights. Therefore, the comprehensive weights of rockburst intensity level prediction indicators are determined by introducing distance function, comprehensive coefficient of variation method and improved hierarchical analysis method.

[0167] In order to eliminate the influence of subjective and objective factors on the weight accuracy and comprehensively consider the reliability and accuracy of the prediction results, the present invention introduces a distance function to assign different preference coefficients to the subjective weight of the improved hierarchical analysis method and the objective weight of the coefficient of variation method, and determines the optimal comprehensive weight of the rockburst intensity level prediction index. The calculation formula is as above.

[0168] The calculation results of preference coefficients by improved analytic hierarchy process and coefficient of variation method are 0.586 and 0.414 respectively.

[0169] The comprehensive weights of various prediction indicators for rockburst intensity levels are shown in Table 6.

[0170] Table 6 Comprehensive weight calculation results

[0171] Predictive indicators <![CDATA[o c / the t ]]> <![CDATA[o θ / the c ]]> <![CDATA[W et ]]> <![CDATA[K v ]]> Comprehensive weight value 0.2836 0.3825 0.1706 0.1633

[0172] Furthermore, the above step S8 also includes:

[0173] The process of determining the connection degree of the comprehensive set of rockburst intensity prediction is as follows:

[0174] According to the rockburst intensity level set in step S4, the connection function μ l Construct the prediction indicator connection matrix T:

[0175]

[0176] Among them, t a1 Represents the prediction index connection function μ l Constant term, t l,1 represents the coefficient of p1 in the contact function, t l,2 represents the coefficient of p2 in the contact function, t cl represents the coefficient of q;

[0177] In the above step S7, the optimized comprehensive weight is combined with the set pair connection degree to determine the comprehensive connection degree μ for rockburst intensity level prediction. A-B , and its calculation formula is expressed as formula 1:

[0178]

[0179] Among them, u represents the comprehensive weight vector of prediction indicators, [ω1, ω2, ..., ω l ], ω1 represents the comprehensive weight of the prediction index;

[0180] Formula 1 can be simplified as:

[0181] μ A-B =f1+f2p1+f3p2+f4q

[0182] Among them, f1 indicates the possibility of no rockburst belonging to level I of rockburst intensity; f2 indicates the possibility of weak rockburst belonging to level II; f3 indicates the possibility of moderate rockburst belonging to level III; and f4 indicates the possibility of strong rockburst belonging to level IV.

[0183] By introducing the distance function, the rationality of obtaining the preference coefficients of different weighting methods is improved, and the reliability of comprehensive weighting is enhanced.

[0184] The comprehensive correlation of rockburst intensity levels of typical rockburst engineering cases collected in step S3 is shown in Table 7.

[0185] Table 7 Calculation results of comprehensive connection degree for typical rock burst engineering cases

[0186]

[0187] Furthermore, in the above step S9, determining the rockburst intensity level also includes:

[0188] In the comprehensive connection degree μ A-B Belongs to level III moderate rock burst or f h When it is greater than or equal to the third threshold, the maximum membership criterion is used to determine the rockburst intensity level;

[0189] When f h When it is less than the third threshold and the difference between the memberships of each level is small, the maximum membership criterion is prone to misjudgment of the rockburst level. Then, by introducing the confidence λ, the confidence criterion is used to judge the rockburst level.

[0190] The corresponding membership degree at any level is: f h =f1+f2+…+f h When ≥λ, the predicted rockburst level is h, where h is a positive integer between 1 and 4, and the confidence level λ∈[0.5,0.7].

[0191] Specifically, the comprehensive connection degree μ A-B The intermediate level membership, i.e. f h ≥0.5 (h=1, 2, 3, 4), the maximum membership criterion is used to determine the rockburst intensity level. The third threshold or λ is 0.5, when f h When <0.5 and the difference between the memberships of each level is small, the use of the maximum membership criterion is prone to misjudgment of the rockburst level, thereby reducing the reliability of the prediction model. Therefore, by introducing the confidence λ, the confidence criterion is used to judge the rockburst level, making the prediction result more accurate and reliable. When there is a certain level corresponding to the membership: f h =f1+f2+…+f h When ≥λ, the predicted rockburst level is h. Generally, the confidence level λ∈[0.5,0.7].

[0192] The prediction results of rockburst intensity levels for 17 groups of typical rockburst cases collected in step S3 are shown in Table 8.

[0193] Table 8 Prediction results of rockburst intensity levels for typical rockburst engineering cases

[0194]

[0195]

[0196] It can be seen from Table 8 that the actual rockburst level of the third group of samples is moderate rockburst, which is predicted to be slight to moderate rockburst using this model. The prediction results of the other 16 groups are completely consistent with the actual level. The prediction accuracy of the rockburst intensity level of the present invention is 16 / 17=94.1%.

[0197] The rockburst intensity prediction method proposed in the present invention is applied to the 750m horizontal middle section of Jiaojia Gold Mine, and the prediction results are shown in Table 9. Through on-site investigation and Peng Xiao's geological disaster investigation of Jiaojia Gold Mine in "Characteristics and mechanism of rockburst at five deep gold mines in Jiaodong Peninsula of China", it is found that during the construction process, rock blocks will fall off the top of the tunnel, accompanied by rockburst phenomena such as slight rock tearing sounds, and the rockburst intensity prediction results of the present invention are consistent with the actual situation.

[0198] Table 9 Prediction results of rockburst grade in the middle section of the 750m level of Jiaojia Gold Mine

[0199]

[0200] This rockburst intensity level prediction method has good accuracy and reliability, and can provide a new approach to predicting rockburst intensity levels in the field of deep rock underground excavation.

[0201] The present invention also provides a Figure 2 The rockburst intensity level prediction system based on comprehensive weighting and set pair analysis shown in the figure includes the following modules:

[0202] Construction unit for selecting rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index W et and rock mass integrity factor K v As a prediction index, a rockburst intensity level prediction index system is established based on the prediction index;

[0203] The classification unit is used to classify the rockburst intensity into four levels: level I no rockburst, level II weak rockburst, level III moderate rockburst and level IV strong rockburst, so as to determine the rockburst intensity classification standard;

[0204] The preprocessing unit is used to collect the first rock burst engineering data from the historical rock burst engineering case database, and obtain the data within a preset time period from the first rock burst engineering data as the target data, and calculate the rock mass stress coefficient σ in the target data. θ / σc and elastic energy index W et Preprocessing is performed to correct the rock stress coefficient σ that changes with time c / σ t and elastic energy index W et The abnormal value of the rock burst engineering data is determined, and the target correction technology is used to correct the first rock burst engineering data, and the corrected second rock burst engineering data is obtained, wherein the rock burst engineering data includes the first rock burst engineering data and the second rock burst engineering data, and also includes the rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index W et and rock mass integrity factor K v ;

[0205] The calculation unit is configured as follows: Using set pair analysis theory, calculate the rockburst intensity level prediction index x l The set pair connection function is used to determine the subjective weight of rockburst intensity prediction indicators based on the improved analytic hierarchy process. The objective weight of rockburst intensity prediction indicators is determined based on the coefficient of variation method. The comprehensive weight of indicators is determined by introducing the distance function. The comprehensive set pair connection degree of rockburst intensity is determined based on the comprehensive weight of indicators.

[0206] The prediction unit is used to determine the rockburst intensity level on the basis of the rockburst intensity level comprehensive set pair connection degree, comprehensive maximum membership criterion and confidence criterion.

[0207] In summary, prediction indicators are selected and a rockburst intensity prediction index system is established to predict the rockburst intensity level more accurately and objectively and reduce the possibility of false alarms and missed alarms. The rockburst intensity level is divided into four levels based on the prediction index system to determine the rockburst intensity level classification standard. The first rockburst engineering data is corrected by target correction technology to improve the accuracy of the collected rockburst engineering case data.

[0208] Based on the data of the second rockburst project, the set pair analysis theory is used to calculate the rockburst intensity level prediction index x lThe set pair connection function is used to construct the comparison matrix of the hierarchical analysis method using the three-scaling method, and the exponential function is introduced to optimize the transfer matrix, which avoids the consistency test, simplifies the calculation process, and improves the hierarchical analysis method for calculating the subjective weight of the index; based on the improved hierarchical analysis method, the subjective weight of the rockburst intensity level prediction index is determined; based on the coefficient of variation method, the objective weight of the rockburst intensity level prediction index is determined; by introducing the distance function, the comprehensive weight of the index is determined, the rationality of obtaining the preference coefficient of different weighting methods is improved, and the reliability of the comprehensive weighting is improved. The comprehensive set pair connection degree of the rockburst intensity level is determined based on the comprehensive weight of the index; based on the comprehensive set pair connection degree of the rockburst intensity level, the maximum membership criterion and the confidence criterion are combined to determine the rockburst intensity level, so that the rockburst intensity level prediction method has good accuracy and reliability. Through the coordination between the above steps, the rockburst intensity level prediction method can improve the accuracy and reliability of the rockburst intensity level prediction, and provide a new way to predict the rockburst intensity level in the field of deep rock underground excavation.

[0209] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0210] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The above-mentioned program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0211] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0212] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0213] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A rockburst intensity prediction method based on comprehensive weighting and set pair analysis, characterized in that: The method comprises the following steps: Step S1: Select rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index W et and rock mass integrity factor K v as a prediction indicator, and establishing the rockburst intensity level prediction indicator system based on the prediction indicator; Step S2: Based on the prediction index system, the rockburst intensity level is divided into four levels: level I no rockburst, level II weak rockburst, level III moderate rockburst and level IV strong rockburst; Step S3: Collect the first rock burst engineering data from the historical rock burst engineering case database, and obtain the data within a preset time period from the first rock burst engineering data as the target data, and calculate the rock mass stress coefficient σ in the target data. θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock mass stress coefficient σ that changes with time c / σ t and the elastic energy index W et , and determine the target correction technology, use the target correction technology to correct the first rock burst engineering data, and obtain the corrected second rock burst engineering data, wherein the rock burst engineering data includes the first rock burst engineering data and the second rock burst engineering data, and also includes the rock mass stress coefficient σ θ / σ c , the rock brittleness coefficient σ c / σ t , the elastic energy index W et and the rock mass integrity factor K v ; Step S4: Based on the second rockburst engineering data, using set pair analysis theory, calculate the rockburst intensity level prediction index x l The set pair connection function of ; Step S5: determining the subjective weight of the rockburst intensity level prediction index based on the improved analytic hierarchy process; Step S6: determining the objective weight of the rockburst intensity level prediction index based on the coefficient of variation method; Step S7: Determine the comprehensive weight of the indicator by introducing the distance function; Step S8: determining the comprehensive set pair connection degree of rockburst intensity levels based on the comprehensive weight of the indicators; Step S9: Based on the rockburst intensity level comprehensive set pair connection degree, the maximum membership criterion and the confidence criterion are combined to determine the rockburst intensity level.

2. The method according to claim 1, characterized in that In step S3, the rock mass stress coefficient σ in the target data is θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock mass stress coefficient σ that changes with time c / σ t and the elastic energy index W et The abnormal value of the rockburst is detected and the corrected second rockburst engineering data is obtained, which also includes: Step S31: The rock mass stress coefficient σ in the target data θ / σ c and the elastic energy index W et Purification is performed to generate a normal time series of rock mass stress coefficients σ θ / σ c and elastic energy index W et , wherein the normal time series refers to a series of values ​​arranged in chronological order, and the rock mass stress coefficient σ θ / σ c and the elastic energy index W et They are usually collected at equal or unequal time intervals; Step S32: Obtain the first rock mass stress coefficient σ in the target data θ / σ c When the time series is greater than the first threshold or less than the second threshold, the X data sets corresponding to the time series are statistically analyzed for the rock mass stress coefficient σ in the historical rock burst engineering data. θ / σ c , and randomly select X second outliers from the probability distribution of the first outliers, insert the X second outliers into the corresponding X data sets, and obtain the second rock mass stress coefficient σ θ / σ c , where X represents a positive integer greater than or equal to 1 and less than Sum, Sum represents the number of the first rock mass stress coefficients in the target data, and the first abnormal value refers to the rock mass stress coefficient σ in the historical rock burst engineering data θ / σ c The measured value in the equation is higher than the normal maximum value or lower than the normal minimum value; similarly, the second elastic energy index W is obtained. et ; Step S33: Using a variety of data correction techniques, the second rock mass stress coefficient σ θ / σ c and the second elastic energy index W et Perform correction to obtain multiple sets of correction results, based on the rock mass stress coefficient σ of the normal time series θ / σ c and the elastic energy index W et The rock mass stress coefficient σ is plotted against multiple sets of calibration results. θ / σ c and the elastic energy index W et The change curve over time determines the rock mass stress coefficient σ that is closest to the normal time series θ / σ c and elastic energy index W et The correction technology corresponding to the correction result of the change curve is the target correction technology; Step S34: using the target correction technology to correct the first rockburst engineering data to obtain second rockburst engineering data.

3. The method according to claim 2, characterized in that In step S31, the rock mass stress coefficient σ in the target data is θ / σ c and the elastic energy index W et Purification is performed to generate a normal time series of rock stress coefficient σ θ / σ c Number and elastic energy index W et ,include: Step S311: acquiring multiple data records with completely identical data records from the first rockburst engineering data, and deleting duplicate data from the data records to acquire non-duplicate first rockburst engineering data; Step S312: Obtain the rock mass stress coefficient σ from the first non-repeated rock burst engineering data θ / σ c , and the rock mass stress coefficient σ θ / σ c Convert to the data format of the same data type and the same unit; Step S313: Retrieve the rock mass stress coefficient σ θ / σ c , and use the non-null value that is closest to the null value and earlier in time to supplement the null value; Step S314: Determine the rock mass stress coefficient σ θ / σ c range, retrieve the rock mass stress coefficient σ θ / σ c The outliers that deviate from the range are replaced by the median to generate the rock mass stress coefficient σ after replacing the outliers. θ / σ c The rock mass stress coefficient σ as a normal time series θ / σ c , wherein the median refers to the value located in the middle position after N values ​​are taken consecutively from the data adjacent to the abnormal value and arranged in order of size, and N represents a positive integer greater than or equal to 1; Step S315: Repeat steps S312 to S314 to generate the elastic energy index W of the normal time series et .

4. The method according to claim 1, characterized in that: In step S4, the rockburst intensity level prediction index x is calculated. l The set pair connection function is as follows: μ l =a l +b l,1 p1+b l,2 p2+…+b l,k-2 p k-2 +c l q Among them, a l +b l,1 +b l,2 +…+b l,k-2 +c l =1, l represents the number of rockburst intensity prediction indicators l=1, 2, ..., m, m represents the number of indicators, m is less than or equal to k, k represents the rockburst intensity level, a l Represents the prediction index x l The probability of belonging to the level 1 standard, b l,1 、b l,2 ,……,b l,k-2 Respectively represent the prediction index x l The probability of belonging to the 1st, 2nd, 3rd, …, k-1th level standard, c l Represents the predicted index x l The probability of belonging to the k-level standard, p1, p2, ..., p k-2 ∈[-1, 1] represents the difference coefficient, q represents the opposition coefficient, q=-1.

5. The method according to claim 1, characterized in that The step S5 further includes: introducing an exponential function to optimize the transfer matrix and improving the analytic hierarchy process for calculating the subjective weight of the prediction index, and the calculation steps are as follows: Step S51: Determine the comparison matrix: Combined with expert opinions, the comparison matrix A of the prediction index is established using the 3-scale method: Among them, a ij represents the importance of factor i relative to factor j in the comparison matrix, i, j = 1, 2, 3, ..., n, n represents the total number of factors, and satisfies a ij =0,a ij +a ji =0,a ij The meanings of different values ​​are: Step S52: Determine the optimal transfer matrix: Use a weighted method to optimize the comparison matrix to obtain the optimal transfer matrix R: in, Step S53: Determine the comprehensive judgment matrix: introduce an exponential function to calculate the optimal transfer matrix to obtain the optimized comprehensive judgment matrix D: Among them, d in =exp(r ij ); Step S54: Calculation of subjective weight: The subjective weight is calculated using the square root method. The calculation process is as follows: Among them, ω Ai * represents the subjective weight vector, ω Ai Represents the subjective weight of the prediction indicator.

6. The method according to claim 1, characterized in that In step S6, based on the coefficient of variation method, the objective weight of the prediction index of the rockburst intensity level is determined, and the calculation formula is as follows: Among them, V i represents the coefficient of variation of the objective indicator, Si represents the standard deviation, represents the average value, ω Ei Represents the objective weight of the prediction indicator.

7. The method according to claim 1, characterized in that The step S7 further comprises: By introducing the distance function, assigning different preference coefficients of objective weights of the coefficient of variation method and subjective weights of the improved analytic hierarchy process, the optimal comprehensive weights of the rockburst intensity prediction index are determined, and the calculation formula is as follows: Wherein, L represents the distance function between the coefficient of variation method and the improved analytic hierarchy process, ω represents the comprehensive weight, and y and z represent the preference coefficients of the improved analytic hierarchy process and the coefficient of variation method, respectively.

8. The method according to claim 7, characterized in that The step S8 further comprises: The process of determining the connection degree of the comprehensive set of rockburst intensity level prediction is as follows: According to the rockburst intensity level set in step S4, the connection function μ l Construct the prediction indicator connection matrix T: Among them, t a1 Represents the prediction index connection function μ l Constant term, t l,1 represents the coefficient of p1 in the connection function, t l,2 represents the coefficient of p2 in the connection function, t cl represents the coefficient of q; The optimized comprehensive weight in the above step S7 is combined with the set pair connection degree to determine the comprehensive connection degree μ for predicting the rockburst intensity level. A-B , and its calculation formula is expressed as formula 1: Wherein, u represents the comprehensive weight vector of the prediction index, [ω1, ω2, ..., ω l ], ω1 represents the comprehensive weight of the prediction index; The formula 1 is simplified as: μ A-B =f1+f2p1+f3p2+f4q Among them, f1 represents the possibility that the rockburst intensity level belongs to level I without rockburst; f2 represents the possibility that it belongs to level II weak rockburst; f3 represents the possibility that it belongs to level III moderate rockburst; and f4 represents the possibility that it belongs to level IV strong rockburst.

9. The method according to claim 8, characterized in that In step S9, determining the rockburst intensity level further includes: In the comprehensive connection degree μ A-B Belongs to the above-mentioned level III moderate rock burst or f h When it is greater than or equal to the third threshold, the maximum membership criterion is used to determine the rockburst intensity level; When f h When the confidence level is less than the third threshold and the difference between the membership levels of the various levels is small, the confidence level λ is introduced to determine the rockburst level using the confidence level criterion; The corresponding membership degree at any level is: f h =f1+f2+…+f h ≥λ, the predicted rockburst level is h, where h is a positive integer between 1 and 4, and the confidence level λ∈[0.5,0.7].

10. A rockburst intensity prediction system based on comprehensive weighting and set pair analysis, used to implement the method according to any one of claims 1 to 9, characterized in that: The system includes the following modules: Construction unit for selecting rock mass stress coefficient σ θ / σ c , rock brittleness coefficient σ c / σ t , elastic energy index W et and rock mass integrity factor K v as a prediction indicator, and establishing the rockburst intensity level prediction indicator system based on the prediction indicator; A grading unit, used to classify the rockburst intensity level into four levels: level I no rockburst, level II weak rockburst, level III moderate rockburst and level IV strong rockburst, so as to determine the rockburst intensity level grading standard; A preprocessing unit is used to collect first rock burst engineering data from a historical rock burst engineering case database, and obtain data within a preset time period from the first rock burst engineering data as target data, and perform a preprocessing on the rock mass stress coefficient σ in the target data. θ / σ c and elastic energy index W et Preprocessing is performed to correct the rock mass stress coefficient σ that changes with time c / σ t and the elastic energy index W et , and determine the target correction technology, use the target correction technology to correct the first rock burst engineering data, and obtain the corrected second rock burst engineering data, wherein the rock burst engineering data includes the first rock burst engineering data and the second rock burst engineering data, and also includes the rock mass stress coefficient σ θ / σ c , the rock brittleness coefficient σ c / σ t , the elastic energy index W et and the rock mass integrity factor K v ; The calculation unit is configured as follows: using set pair analysis theory, calculate the rockburst intensity level prediction index x l The set pair connection function; based on the improved analytic hierarchy process, the subjective weight of the prediction index of the rockburst intensity level is determined; based on the coefficient of variation method, the objective weight of the prediction index of the rockburst intensity level is determined; by introducing the distance function, the comprehensive weight of the index is determined; based on the comprehensive weight of the index, the comprehensive set pair connection degree of the rockburst intensity level is determined; The prediction unit is used to determine the rockburst intensity level based on the rockburst intensity level comprehensive set pair connection degree, comprehensive maximum membership criterion and confidence criterion.

Images