A method, medium, and system for risk assessment in steeply inclined coal mines

Abstract

This invention provides a method, medium, and system for risk assessment in steeply inclined coal mines, belonging to the technical field of risk assessment in this field. The method includes: first, standardizing multimodal data and organizing it into a multidimensional matrix; then, performing time series analysis and spatial correlation analysis on the matrix to obtain a risk analysis matrix reflecting the correlation between different data types and potential risk factors; next, using singular value decomposition (SVD) to decompose the risk analysis matrix into a stable matrix and a risk matrix, thereby extracting the main risk feature vectors; finally, based on these main risk indicator vectors, calculating the roof instability risk index, gas exceedance risk index, and flood risk index respectively, and synthesizing the overall risk assessment result. This addresses the technical problem of existing risk assessment methods lacking in-depth mining and detailed analysis of key risk factors, making it difficult to derive targeted risk control measures.

Description

Technical Field

[0001] This invention belongs to the technical field of risk assessment for steeply inclined coal mines, and specifically relates to a method, medium, and system for risk assessment of steeply inclined coal mines. Background Technology

[0002] Steeply inclined coal seams refer to coal seams with a dip angle of more than 45 degrees. These coal seams have harsh mining conditions and many safety risks, such as roof instability, gas exceeding limits, and water disasters, which bring huge challenges to mine production and operation.

[0003] To address the safety concerns of steeply inclined coal mines, the industry has developed several risk assessment and control methods. A common approach involves collecting various monitoring data, such as geological data, mining parameters, and real-time monitoring data, and then analyzing and processing this data to identify key factors affecting safe production. For example, some studies have used time series analysis and grey relational analysis to model monitoring data such as gas concentration and roof displacement, exploring the interrelationships between these key factors. Other studies have attempted to establish coal mine safety early warning models based on BP neural networks or fuzzy comprehensive evaluation, comprehensively considering the impact of multiple factors on coal mine safety. These existing methods have, to some extent, improved the ability to recognize and warn of risks in steeply inclined coal mines.

[0004] However, existing risk assessment methods still suffer from a lack of in-depth exploration and detailed analysis of key risk factors, making it difficult to derive targeted risk control measures. Summary of the Invention

[0005] In view of this, the present invention provides a method, medium and system for risk assessment in steeply inclined coal mines, which can solve the technical problem that existing risk assessment methods still lack in-depth exploration and detailed analysis of key risk factors, making it difficult to derive targeted risk control measures.

[0006] This invention is implemented as follows:

[0007] The first aspect of the present invention provides a method for risk assessment in steeply inclined coal mines, comprising the following steps:

[0008] S10. Collect multimodal data of steeply inclined coal mines, including geological data, mining parameters, and monitoring data; wherein, the geological data includes stratigraphic structure and fault distribution; the mining parameters include mining methods and mining depth; and the monitoring data includes gas concentration, roof displacement, and microseismic signals.

[0009] S20. Standardize and organize the multimodal data into a multidimensional matrix of multimodal data, wherein the first dimension of the multidimensional matrix represents different data types, the second dimension represents different time points, and the third dimension represents spatial location;

[0010] S30. Perform time series analysis and spatial correlation analysis on the multidimensional matrix to obtain a two-dimensional risk analysis matrix. The elements of the risk analysis matrix represent the correlation between different data types and potential risk factors.

[0011] S40. The risk analysis matrix is ​​decomposed into a stable matrix and a risk matrix using the singular value decomposition method, wherein the stable matrix reflects the stability characteristics of the steeply inclined coal mine, and the risk matrix contains the potential risk information of the steeply inclined coal mine.

[0012] S50. Perform a detailed analysis of the risk matrix, extract the main risk feature vectors, select the feature vector corresponding to the maximum singular value as the main risk indicator, and map the feature vectors to three main risk dimensions: roof instability risk vector, gas over-limit risk vector, and flood risk vector.

[0013] S60. Based on the obtained roof instability risk vector, gas over-limit risk vector, and flood risk vector, cosine similarity calculation is performed with the preset roof stability vector, gas safety vector, and flood safety vector to obtain the roof instability risk index, gas over-limit risk index, and flood risk index.

[0014] S70. Calculate the comprehensive risk assessment index and compare it with the preset risk threshold to obtain the mining risk assessment results of steeply inclined coal mines.

[0015] Specifically, step S10 includes collecting multimodal data from steeply inclined coal mines, including geological data, mining parameters, and monitoring data. First, geological data such as stratigraphic structure and fault distribution are obtained from geological exploration and mining practices, using methods such as drilling and geological surveying. Second, mining parameters such as mining methods and depth are obtained from mining operation records. Third, real-time monitoring data such as gas concentration, roof displacement, and microseismic signals are monitored using gas concentration monitoring equipment, roof displacement sensors, and microseismic meters. This multimodal data covers key factors such as geological conditions, mining activities, and operational status during the mining process of steeply inclined coal mines, providing a necessary information foundation for subsequent risk assessment.

[0016] Specifically, step S20 includes: standardizing the collected multimodal data to eliminate the influence of different dimensions and magnitudes, ensuring the data falls within the same dimension. Then, the standardized data is organized into a three-dimensional matrix according to different data types, time points, and spatial locations. This multidimensional matrix can comprehensively depict the spatiotemporal distribution characteristics of various key data during steeply inclined coal mine mining, laying the foundation for subsequent time series analysis and spatial correlation analysis.

[0017] Specifically, step S30 includes: performing time series analysis and spatial correlation analysis on the multidimensional matrix to obtain a two-dimensional risk analysis matrix reflecting the correlation between data and potential risk factors. First, an autoregressive moving average (ARMA) model is fitted to each data type over time to analyze its temporal evolution and identify the temporal correlation characteristics of various data types. Second, Moran's I index is used to analyze the correlation between different spatial units in the multidimensional matrix, quantifying the spatial correlation strength of various data types. Finally, by combining the results of time series analysis and spatial correlation analysis, a two-dimensional risk analysis matrix is ​​constructed, whose elements reflect the correlation between different data types and potential risk factors.

[0018] Specifically, step S40 includes: decomposing the risk analysis matrix using Singular Value Decomposition (SVD) to obtain a stable matrix and a risk matrix. First, Singular Value Decomposition is performed on the risk analysis matrix to obtain an orthogonal matrix, a diagonal matrix of singular values, and left and right singular vectors. Then, based on the magnitude of the singular values, a threshold is set to decompose the original matrix into a stable matrix reflecting stable characteristics and a risk matrix containing potential risks. This decomposition method can effectively separate stable factors and risk factors in the original data, laying the foundation for subsequent risk feature extraction and analysis.

[0019] Specifically, step S50 includes: extracting the main risk feature vectors and mapping the risk dimensions of the risk matrix. First, the right singular vector corresponding to the largest singular value is selected as the main risk indicator; this vector reflects the most important risk feature in the risk matrix. Then, this main risk feature vector is mapped to three main risk dimensions: roof instability risk vector, gas exceedance risk vector, and flood risk vector. This risk dimension mapping can better identify the most critical risk factors in the mining of steeply inclined coal mines.

[0020] Specifically, step S60 includes: calculating the roof instability risk index, gas exceedance risk index, and flood risk index based on the three obtained risk vectors. This method of calculating risk indices using cosine similarity can quantify the degree of proximity between various risk factors and safety objectives, providing a basis for comprehensive risk assessment.

[0021] Specifically, step S70 includes: calculating a comprehensive risk assessment index and comparing it with a preset risk threshold to obtain the risk assessment result for the steeply inclined coal mine. The calculation of the comprehensive risk assessment index considers three main risk factors: roof instability, gas exceeding limits, and flooding, and also includes correction terms for other potential risk factors. By comparing it with the preset risk threshold, the final risk assessment result can be obtained, providing a basis for formulating reasonable risk management strategies.

[0022] Furthermore, the multidimensional matrix of the multimodal data can be represented as:

[0023] M ijk =f(D i ,T j ,S k )+ε ijk ;

[0024] In the formula, M ijk For multidimensional matrix elements; D i T represents the i-th data type; j S represents the j-th time point; k Represents the k-th spatial location; f is the data mapping function; ε ijk This is the random error term.

[0025] Parameter acquisition method:

[0026] D i Obtained through data collection and classification, including geological data, mining parameters, and monitoring data;

[0027] T j Obtained through timestamp records;

[0028] S k Obtained through spatial coordinate system recording.

[0029] 2. Time series analysis and spatial correlation analysis:

[0030] Time series analysis can employ the autoregressive moving average (ARMA) model:

[0031]

[0032] In the formula, X t φ represents the time series value; p is the number of autoregressive terms; q is the number of moving average terms; φ i θ is the autoregressive coefficient. j ε is the moving average coefficient; t It is white noise.

[0033] Spatial correlation analysis can use Moran's I index:

[0034]

[0035] In the formula, n is the number of spatial units; w ij For spatial weights; x i and x j These are the attribute values ​​for spatial units i and j; This represents the attribute mean.

[0036] 3. Risk Analysis Matrix:

[0037] The risk analysis matrix R can be constructed using the results of time series analysis and spatial correlation analysis:

[0038] R mn =α·ARMA(M ijk )+β·I(M ijk )+γ mn ;

[0039] In the formula, R mn For risk analysis matrix elements; ARMA(M ijk ) represents the results of time series analysis; I(M ijk ) represents the spatial correlation analysis results; α and β are weighting coefficients; γ mn This is a random disturbance term.

[0040] 4. Singular Value Decomposition:

[0041] Perform singular value decomposition on the risk analysis matrix R:

[0042]

[0043] In the formula, U and V are orthogonal matrices; ∑ is a singular value diagonal matrix; σ i It is a singular value; u i and v i are the left and right singular vectors; r is the rank of the matrix.

[0044] The stability matrix S and the risk matrix H can be expressed as:

[0045]

[0046] In the formula, This is a preset threshold used to distinguish between stable and risky features.

[0047] 5. Extraction of key risk feature vectors:

[0048] The right singular vector v1 corresponding to the maximum singular value σ1 is selected as the main risk indicator:

[0049] v1 = [v 11 ,v 12 ,...,v 1n ] T ;

[0050] 6. Risk Dimension Mapping:

[0051] Map the main risk feature vector v1 to three main risk dimensions:

[0052]

[0053] In the formula, R roof R gas and R water These are the risk vectors for roof instability, gas exceedance, and flooding, respectively. and ...

[0054] 7. Risk Index Calculation:

[0055] Calculate the risk index using cosine similarity:

[0056]

[0057] In the formula, RI roof RI gas and RI water These are the roof instability risk index, gas over-limit risk index, and flood risk index, respectively; V roof V ags and V water These are the preset roof stability vector, gas safety vector, and flood safety vector, respectively. This indicates taking the modulus.

[0058] 8. Calculation of Comprehensive Risk Assessment Index:

[0059] The comprehensive risk assessment index can be calculated using a weighted summation method:

[0060] RI total =λ1RI roof +λ2RI gas +λ3RI water +δ;

[0061] In the formula, RI total λ1, λ2, and λ3 are the weighting coefficients, and satisfy λ1 + λ2 + λ3 = 1; δ is a correction term used to consider other potential risk factors.

[0062] Ultimately, RI total With the preset risk threshold RIthreshold By comparison, the risk assessment results for mining steeply inclined coal mines are obtained:

[0063] If RI total <RI threshold If so, the mining risk is considered to be within an acceptable range;

[0064] If RI total ≥RI threshold If the risk exceeds acceptable limits, then corresponding risk control measures need to be taken.

[0065] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, which, when executed in a computer, are used to perform the above-described method for risk assessment in steeply inclined coal mines.

[0066] A third aspect of the present invention provides a risk assessment system for steeply inclined coal mine mining, wherein the system includes the aforementioned computer-readable storage medium.

[0067] Compared with existing technologies, the beneficial effects of the risk assessment method, medium, and system for steeply inclined coal mine mining provided by this invention are:

[0068] 1. Employing multimodal data fusion, it fully reflects the geological, mining, and operational status of steeply inclined coal mines. It not only collects conventional data from geological exploration, mining operations, and monitoring, but also integrates real-time monitoring data from various sensors, such as gas concentration, roof displacement, and microseismic signals, significantly enhancing the information content and coverage of the data.

[0069] 2. By employing mathematical models such as time series analysis, spatial correlation analysis, and singular value decomposition, we can deeply explore the inherent correlation patterns and potential risk characteristics in multimodal data. Compared with simple statistical analysis or empirical models, this method can better capture the dynamic interactions between various factors in complex nonlinear systems such as steeply inclined coal mines.

[0070] 3. The extracted key risk feature vectors are mapped to three major risk dimensions: roof instability, gas exceedance, and flooding, enabling refined analysis and quantitative assessment of key risk factors. This analytical framework more closely reflects the main safety hazards faced in actual production at steeply inclined coal mines, providing a basis for developing targeted prevention and control measures.

[0071] 4. A comprehensive risk assessment index was constructed to fully reflect the impact of various risk factors on mining safety, and compared with preset thresholds to provide clear risk assessment results. This assessment method is objective and quantitative, and can provide scientific decision support for mine managers.

[0072] In summary, the risk assessment method for steeply inclined coal mines proposed in this invention makes full use of multi-source heterogeneous data, adopts advanced mathematical analysis techniques, deeply explores key risk factors, and achieves accurate identification and comprehensive assessment of risks through innovative analytical frameworks. This solves the technical problem that existing risk assessment methods still lack in-depth exploration and detailed analysis of key risk factors, making it difficult to derive targeted risk control measures. Attached Figure Description

[0073] Figure 1 A flowchart of the method provided by the present invention. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0075] like Figure 1 The diagram shown is a flowchart of a risk assessment method for steeply inclined coal mines provided by this invention. The method includes the following steps:

[0076] S10. Collect multimodal data of steeply inclined coal mines, including geological data, mining parameters, and monitoring data; among which, geological data includes stratigraphic structure and fault distribution; mining parameters include coal mining methods and mining depth; and monitoring data includes gas concentration, roof displacement, and microseismic signals.

[0077] S20. Standardize and organize these multimodal data into a multidimensional matrix of multimodal data, where the first dimension of the multidimensional matrix represents different data types, the second dimension represents different time points, and the third dimension represents spatial location.

[0078] S30. Perform time series analysis and spatial correlation analysis on the multidimensional matrix to obtain a two-dimensional risk analysis matrix. The elements of the risk analysis matrix represent the correlation between different data types and potential risk factors.

[0079] S40. The risk analysis matrix is ​​decomposed into a stable matrix and a risk matrix using the singular value decomposition method. The stable matrix reflects the stability characteristics of steeply inclined coal mines, while the risk matrix contains the potential risk information of steeply inclined coal mines.

[0080] S50. Conduct a detailed analysis of the risk matrix, extract the main risk feature vectors, select the feature vector corresponding to the maximum singular value as the main risk indicator, and map the feature vectors to three main risk dimensions: roof instability risk vector, gas over-limit risk vector, and flood risk vector.

[0081] S60. Based on the obtained roof instability risk vector, gas over-limit risk vector, and flood risk vector, cosine similarity calculation is performed with the preset roof stability vector, gas safety vector, and flood safety vector to obtain the roof instability risk index, gas over-limit risk index, and flood risk index.

[0082] S70. Calculate the comprehensive risk assessment index and compare it with the preset risk threshold to obtain the mining risk assessment results of steeply inclined coal mines.

[0083] The specific implementation methods of the above steps are described in detail below:

[0084] The specific implementation of step S10 involves collecting multimodal data from steeply inclined coal mines, including geological data, mining parameters, and monitoring data. First, geological data such as stratigraphic structure and fault distribution are obtained from geological exploration and mining practices. This geological data can be represented as a set D. geo ={d geo,1 ,d geo,2 ,...,d geo,m}, where d geo,i Let represent the i-th geological data indicator. Secondly, mining parameters such as mining methods and depths can be obtained from mining operation records, which can be represented by set D. op ={d op,1 ,d op,2 ,...,d op,n} represents, where d op,j Let represent the j-th mining parameter index. Furthermore, using gas concentration monitoring equipment, roof displacement sensors, and microseismometers, real-time monitoring data such as gas concentration, roof displacement, and microseismic signals can be obtained using set D. mon ={d mon,1 ,d mon,2 ,...,d mon,p} represents, where d mon,k Let represent the k-th monitoring data indicator. Combining the above three types of data, we can obtain the multimodal dataset D = Dk for steeply inclined coal mines. geo ∪D op ∪D mon Where |D|=m+n+p represents the total number of data indicators. These multimodal data cover key factors such as geological conditions, mining activities, and operational status during the mining process of steeply inclined coal mines, providing a necessary information foundation for subsequent risk assessment.

[0085] The specific implementation of step S20 involves standardizing the collected multimodal data and organizing it into a multidimensional matrix representation. First, the data indicators in D are normalized to eliminate the influence of different dimensions and magnitudes, ensuring the data falls within the same dimension range. The normalized data can be represented as follows: Where min(D) and max(D) represent the minimum and maximum values ​​of the data in D, respectively. Then, the standardized data are categorized according to different data types D. i Time point T j and spatial location S k Organized into a three-dimensional matrix M ijk Where i = 1, 2, ..., |D|, j = 1, 2, ..., J, k = 1, 2, ..., K. This multidimensional matrix can be represented as:

[0086]

[0087] In the formula, M ijk For matrix elements; T represents the i-th type of normalized data; j S represents the j-th time point; k Represents the k-th spatial location; f is the data mapping function; ε ijk This represents the random error term. This multidimensional matrix representation can more comprehensively depict the spatiotemporal distribution characteristics of various key data during the mining process of steeply inclined coal mines, laying the foundation for subsequent time series analysis and spatial correlation analysis.

[0088] The specific implementation of step S30 is based on the multidimensional matrix M. ijk Time series analysis and spatial correlation analysis were performed to obtain a two-dimensional risk analysis matrix R that reflects the correlation between data and potential risk factors. mn First, for each data metric In time series T j An autoregressive moving average (ARMA) model was fitted to the data, and its temporal evolution was analyzed. The ARMA model can be expressed as:

[0089]

[0090] In the formula, Let be the value of the i-th data at time t; P and Q are the number of autoregressive terms and the number of moving average terms, respectively; φ p and θ q For model parameters; ε t This is white noise. The temporal correlation characteristics of various data can be identified through ARMA model analysis, denoted as ARMA(M). ijk Secondly, Moran's I index was used to analyze the multidimensional matrix M. ijk Different spatial units S k The correlation between data points quantifies the spatial strength of the association between different types of data. Moran's I index can be expressed as:

[0091]

[0092] In the formula, n is the number of spatial units; w kl Spatial weights; and Let i be the value of the i-th type of data in cells k and l; Let I be the average value of the i-th data type. This Moran's I index reflects the degree of correlation in the spatial distribution of various data types, denoted as I(M). ijk Finally, the results of time series analysis and spatial correlation analysis are combined to construct a two-dimensional risk analysis matrix R. mn Its elements can be represented as:

[0093] R mn =α·ARMA(M ijk )+β·I(M ijk )+γ mn ;

[0094] In the formula, R mn These are the elements of the risk analysis matrix; α and β are weighting coefficients; γ mn This is a random disturbance term. The risk analysis matrix R... mn This reflects the correlation between different data types and potential risk factors, providing a foundation for subsequent risk feature extraction and analysis.

[0095] The specific implementation of step S40 is to use the Singular Value Decomposition (SVD) method to transform the risk analysis matrix R. mn It is decomposed into a stability matrix S and a risk matrix H. First, the risk analysis matrix R... mn Singular value decomposition can be represented as:

[0096]

[0097] In the formula, U and V are orthogonal matrices; ∑ is a singular value diagonal matrix; σ i It is a singular value; u i and v i Let r be the left and right singular vectors; r is the rank of the matrix. Then, a threshold is set based on the magnitude of the singular values. matrix R mn Decomposed into a stability matrix S and a risk matrix H, where:

[0098]

[0099] In the formula, The threshold is a preset value used to distinguish between stable and risky features. This decomposition method can effectively separate stable and risky factors in the original data. S reflects the stable characteristics of steeply inclined coal mines, while H contains potential risk information, laying the foundation for subsequent risk feature extraction and analysis.

[0100] The specific implementation of step S50 involves extracting the main risk feature vectors and mapping the risk dimensions of the risk matrix H. First, the right singular vector v1 = [v1, v1, v1, v1] corresponding to the maximum singular value σ1 is selected. 11 ,v 12 ,...,v 1n ] T As the primary risk indicator, this vector reflects the most significant risk characteristic in the risk analysis matrix. Then, this primary risk characteristic vector v1 is mapped to three main risk dimensions: the roof instability risk vector R. roof Gas over-limit risk vector R gas and flood risk vector R water The formulas for calculating these risk vectors are as follows:

[0101]

[0102] In the formula, and Here, n1 and n2 are the corresponding weighting coefficients, and n1 and n2 are the dividing points of the risk dimension. This risk dimension mapping can better identify the most critical risk factors in the mining of steeply inclined coal mines.

[0103] The specific implementation of step S60 is to calculate the roof instability risk index RI based on the three obtained risk vectors. roof Gas Over-Limit Risk Index (RI) gas and the Flood Risk Index (RI) water The formulas for calculating these three risk indices are as follows:

[0104]

[0105]

[0106] Among them, V roof V gas and V water These are preset roof stability vector, gas safety vector, and flood safety vector, respectively. This method of calculating a risk index using cosine similarity can quantify the degree of proximity between various risk factors and safety objectives, providing a basis for comprehensive risk assessment.

[0107] The specific implementation of step S70 is to calculate the comprehensive risk assessment index RI. total and the preset risk threshold RI threshold By comparing the results, the risk assessment results for steeply inclined coal mines were obtained. The formula for calculating the comprehensive risk assessment index is as follows:

[0108] RI total =λ1RI roof+λ2RI gas +λ3RI water +δ;

[0109] In the formula, λ1, λ2, and λ3 are weighting coefficients, satisfying λ1 + λ2 + λ3 = 1, and δ is a correction term used to consider other potential risk factors. This is compared with a preset risk threshold RI. threshold By comparing the results, the final evaluation can be obtained: if RI total <RI threshold If the RI is within acceptable limits, then the mining risk is considered to be within acceptable limits; if the RI total ≥RI threshold If the risk exceeds acceptable levels, then corresponding risk control measures are required. This comprehensive assessment method can fully reflect various risk factors in the mining process of steeply inclined coal mines, providing an important basis for formulating reasonable risk management strategies.

[0110] In summary, this risk assessment method for steeply inclined coal mines fully utilizes multimodal data and identifies key risk characteristics through mathematical models such as time series analysis, spatial correlation analysis, and singular value decomposition. These characteristics are then mapped to three major risk dimensions: roof instability, gas exceedance, and flooding. Subsequently, various risk indices are calculated and a comprehensive assessment is conducted, providing a scientific and systematic risk assessment method for the safe mining of steeply inclined coal mines.

[0111] A second aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, which, when executed in a computer, are used to perform the above-described method for risk assessment in steeply inclined coal mines.

[0112] A third aspect of the present invention provides a risk assessment system for steeply inclined coal mine mining, wherein the system includes the aforementioned computer-readable storage medium.

[0113] Specifically, the principle of this invention is to extract key characteristics reflecting mine safety risks through in-depth analysis of multimodal data and establish a quantitative risk assessment model. This method effectively solves the problems existing in current technologies primarily based on the following principles:

[0114] First, this invention fully considers the characteristics of geological conditions, mining activities, and production operations during the mining of steeply inclined coal mines, and collects rich multimodal data, including geological data, mining parameters, and monitoring data. This comprehensive data acquisition can more accurately reflect the actual state of the coal mine, laying the foundation for subsequent in-depth analysis.

[0115] Secondly, this invention employs mathematical modeling methods such as time series analysis and spatial correlation analysis to deeply explore the inherent correlations and dynamic changes in multimodal data. Compared to relying solely on empirical models or simple statistical analysis, this modeling method can better capture the interactions between various elements in complex nonlinear systems, creating conditions for accurately identifying key risk factors.

[0116] Furthermore, this invention uses singular value decomposition to decompose the original risk analysis matrix into two parts: stable features and risk features. This decomposition can not only effectively identify potential risk factors in the mine, but also preserve the stable features of the mine itself, providing a basis for comprehensive and accurate risk assessment.

[0117] Finally, this invention constructs a comprehensive assessment model based on risk dimensions, mapping the extracted key risk factors to three major risk dimensions: roof instability, gas over-limit, and flood, and calculating the risk index of each dimension and its comprehensive index.

[0118] To better understand and implement this invention, a specific application scenario is provided below: A certain mining area is operating a steeply inclined coal mine. The mine has complex geological conditions, with a dip angle exceeding 60 degrees, posing numerous safety hazards such as roof instability, excessive gas levels, and flooding. To comprehensively assess the mining risks of this coal mine, the mine management decided to adopt the risk assessment method based on multimodal data analysis proposed in this invention.

[0119] I. Data Collection

[0120] First, the mine management organized relevant technical personnel to conduct a comprehensive data collection operation on the steeply inclined coal mine. This included:

[0121] Geological data acquisition: Through drilling and geological surveying, geological information such as the stratigraphic structure and fault distribution of the coal mine was obtained. Key data indicators include: stratigraphic lithology (d... geo,1 ), stratigraphic dip angle (d) geo,2 ), fault development density (d geo,3 ), fault strike (d) geo,4 ), fault dip angle (d) geo,5 )wait.

[0122] Mining Parameter Collection: Key parameters such as mining methods and mining depth were collected from mining operation records. Main data indicators include: mining method (d...). op,1 ), mining technology (d op,2 ), mining depth (d) op,3 ), coal mining intensity (d) op,4 )wait.

[0123] Data Acquisition: Using gas concentration monitoring equipment, roof displacement sensors, and microseismometers, key operational data such as gas concentration, roof displacement, and microseismic signals were collected in real time at the coal mine. Key data indicators include: gas concentration (d... mon,1 ), gas emission rate (d mon,2 ), Top plate displacement (d) mon,3 ), microseismic signal amplitude (d mon,4 ), microseismic signal frequency (d mon,5 )wait.

[0124] These geological data, mining parameters, and monitoring data constitute the multimodal dataset of this steeply inclined coal mine:

[0125] D={d geo,1 ,d geo,2 ,...,d geo,5 ,d op,1 ,d op,2 ,...,d op,4 ,d mon,1 ,d mon,2 ,...,d mon,5}

[0126] II. Data Standardization and Matrix Construction

[0127] After collecting the raw multimodal data, it is first standardized to eliminate the influence of dimensions and magnitudes. Specifically, the min-max standardization method is used to normalize each index value to the [0,1] interval:

[0128]

[0129] in, Let be the standardized value of the i-th data index, and min(D) and max(D) be the minimum and maximum values ​​of the data in D, respectively.

[0130] The standardized multimodal data is organized into a three-dimensional matrix M according to data type, time point, and spatial location. ijk Where i = 1, 2, ..., 14 represents 14 data indicators, j = 1, 2, ..., 30 represents 30 time points, and k = 1, 2, ..., 50 represents 50 spatial locations. This three-dimensional matrix can be represented as:

[0131]

[0132] In the formula, M ijk For matrix elements; T represents the i-th type of standardized data; j S represents the j-th time point; k Represents the k-th spatial location; f is the data mapping function; εijk This is the random error term.

[0133] Below is the three-dimensional matrix M constructed from the multimodal data of this steeply inclined coal mine. ijk Partial sample data:

[0134] serial number index Spatial location 1 Spatial location 2 Spatial location 3 Spatial location 4 Spatial location 5 1 <![CDATA[d geo,1 ]]> 0.28 0.33 0.41 0.39 0.35 2 <![CDATA[d geo,2 ]]> 0.57 0.62 0.68 0.59 0.63 3 <![CDATA[d geo,3 ]]> 0.42 0.38 0.47 0.43 0.4 4 <![CDATA[d geo,4 ]]> 0.51 0.46 0.55 0.49 0.53 5 <![CDATA[d geo,5 ]]> 0.64 0.59 0.71 0.66 0.62 6 <![CDATA[d geo,6 ]]> 0.76 0.72 0.81 0.79 0.74 7 <![CDATA[d geo,7 ]]> 0.53 0.48 0.58 0.55 0.51 8 <![CDATA[d geo,8 ]]> 0.69 0.65 0.73 0.71 0.67 9 <![CDATA[d geo,9 ]]> 0.82 0.78 0.86 0.84 0.8 10 <![CDATA[d geo,10 ]]> 0.35 0.31 0.42 0.38 0.33 11 <![CDATA[d geo,11 ]]> 0.46 0.41 0.52 0.48 0.43 12 <![CDATA[d geo,12 ]]> 0.27 0.22 0.34 0.29 0.24 13 <![CDATA[d geo,13 ]]> 0.58 0.53 0.63 0.61 0.55 14 <![CDATA[d geo,14 ]]> 0.49 0.44 0.54 0.51 0.46

[0135] III. Time Series Analysis and Spatial Correlation Analysis

[0136] Based on the constructed three-dimensional matrix M ijk Next, time series analysis and spatial correlation analysis will be performed to uncover the inherent correlation patterns between the data.

[0137] Time series analysis: An ARMA model is fitted to each data indicator over time to identify its temporal evolution characteristics. For example, gas concentration (d...) mon,1 For example, its ARMA model can be represented as:

[0138]

[0139] Analysis using the ARMA model revealed that the gas concentration in this coal mine exhibits significant autocorrelation and moving average characteristics, meaning that the current gas concentration is significantly related to the value at the previous moment and the random disturbance term. This reflects the dynamic variation pattern of gas emission in the mine.

[0140] Spatial correlation analysis: Moran's I index was used to analyze the matrix M. ijk The correlation between different spatial units is used to quantify the correlation strength of various data in spatial distribution. This is exemplified by the top plate displacement (d). mon,3 For example, Moran's I index is calculated as follows:

[0141]

[0142] The Moran's I index results indicate a strong positive spatial correlation between roof displacement data from different mining faces in this coal mine, meaning that the displacement trends of adjacent faces are similar. This reflects the spatial distribution characteristics of roof instability risk in this mine.

[0143] By combining the results of time series analysis and spatial correlation analysis, a two-dimensional risk analysis matrix R is constructed to reflect the correlation of various data and risk factors in steeply inclined coal mines. mn .

[0144] IV. Singular Value Decomposition and Risk Feature Extraction

[0145] Based on the constructed risk analysis matrix R mnNext, the singular value decomposition method is used to decompose it into a stable feature matrix S and a risk feature matrix H.

[0146] First, regarding R mn Perform singular value decomposition:

[0147]

[0148] Where U and V are orthogonal matrices, ∑ is a singular value diagonal matrix, and σ i For singular values, u i and v i Let r be the left and right singular vectors, respectively, and r be the rank of the matrix.

[0149] Then, a threshold is set based on the magnitude of the singular values. R mn Decomposed into a stable feature matrix S and a risk feature matrix H:

[0150]

[0151] Here, S reflects the stability characteristics of the coal mine, while H contains potential risk information.

[0152] Next, select the right singular vector v1 = [v1, v ... 11 ,v 12 ,...,v 1n ] T As a primary risk indicator, this vector reflects R mn The most important risk characteristic. Then, v1 is mapped to three main risk dimensions:

[0153] Roof instability risk vector:

[0154] R roof =0.43v 11 +0.27v 12 +0.30v 13 ;

[0155] Gas over-limit risk vector:

[0156] R gas =0.35v 14 +0.31v 15 +0.34v 16 ;

[0157] Flood risk vector:

[0158] R water =0.41v 17 +0.29v 18 +0.30v 19 ;

[0159] This risk dimension mapping helps identify the most critical risk factors during the mining of this steeply inclined coal mine.

[0160] V. Risk Index Calculation and Comprehensive Assessment

[0161] Based on the above three risk dimensions, the roof instability risk index, gas exceedance risk index, and flood risk index are calculated respectively:

[0162] Roof instability risk index:

[0163]

[0164] Gas Over-Limit Risk Index:

[0165]

[0166] Flood Risk Index:

[0167]

[0168] Among them, V roof V gas and V water These are the preset roof stability vector, gas safety vector, and flood safety vector, respectively.

[0169] Finally, the comprehensive risk assessment index for this steeply inclined coal mine was calculated:

[0170] RI total =0.4·RI roof +0.3·RI gas +0.3·RI water +0.05 = 0.71;

[0171] By comparing with the preset risk threshold RI threshold When compared with a value of 0.65, the results indicate that the mining risk of this coal mine exceeds the acceptable range, and necessary risk control measures need to be taken.

[0172] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for risk assessment in steeply inclined coal mines, characterized in that, Includes the following steps: S10. Collect multimodal data of steeply inclined coal mines, including geological data, mining parameters, and monitoring data; wherein, the geological data includes stratigraphic structure and fault distribution; the mining parameters include mining methods and mining depth; and the monitoring data includes gas concentration, roof displacement, and microseismic signals. S20. Standardize and organize the multimodal data into a multidimensional matrix of multimodal data, wherein the first dimension of the multidimensional matrix represents different data types, the second dimension represents different time points, and the third dimension represents spatial location; S30. Perform time series analysis and spatial correlation analysis on the multidimensional matrix to obtain a two-dimensional risk analysis matrix. The elements of the risk analysis matrix represent the correlation between different data types and potential risk factors. S40. The risk analysis matrix is ​​decomposed into a stable matrix and a risk matrix using the singular value decomposition method, wherein the stable matrix reflects the stability characteristics of the steeply inclined coal mine, and the risk matrix contains the potential risk information of the steeply inclined coal mine. S50. Perform a detailed analysis of the risk matrix, extract the main risk feature vectors, select the feature vector corresponding to the maximum singular value as the main risk indicator, and map the feature vectors to three main risk dimensions: roof instability risk vector, gas over-limit risk vector, and flood risk vector. S60. Based on the obtained roof instability risk vector, gas over-limit risk vector, and flood risk vector, cosine similarity calculation is performed with the preset roof stability vector, gas safety vector, and flood safety vector to obtain the roof instability risk index, gas over-limit risk index, and flood risk index. S70. Calculate the comprehensive risk assessment index and compare it with the preset risk threshold to obtain the mining risk assessment result of steeply inclined coal mines; The risk analysis matrix is ​​calculated as follows: ; In the formula, For risk analysis matrix elements; The results are from time series analysis. The results are from spatial correlation analysis; and These are the weighting coefficients; For random disturbance terms; The risk analysis matrix is ​​decomposed using the singular value decomposition method, as shown below: ; In the formula, and It is an orthogonal matrix; It is a singular value diagonal matrix; It is a singular value; and These are singular vectors on the left and right; Let be the rank of the matrix; a stable matrix. and risk matrix They are represented as follows: ; ; In the formula, This is a preset threshold used to distinguish between stable and risky characteristics; The feature vectors are mapped to three main risk dimensions, as shown below: ; ; ; In the formula, , and These are the risk vectors for roof instability, gas exceedance, and flooding, respectively. , and These are the corresponding weighting coefficients; and This serves as the dividing point for risk dimensions.

2. The method for risk assessment in steeply inclined coal mines according to claim 1, characterized in that, The multidimensional matrix representation of the multimodal data is as follows: ; In the formula, Elements of a multidimensional matrix; Indicates the first Types of data; Indicates the first A point in time; Indicates the first A spatial location; For data mapping functions; This is the random error term.

3. The method for risk assessment in steeply inclined coal mines according to claim 2, characterized in that, The time series analysis used an autoregressive moving average model, as shown below: ; In the formula, These are time series values; This represents the number of autoregressive terms. The number of terms in the moving average; These are the autoregressive coefficients; The moving average coefficient; It is white noise.

4. The risk assessment method for steeply inclined coal mine mining according to claim 3, characterized in that, Spatial correlation analysis uses Moran's I index, specifically represented as follows: ; In the formula, This refers to the number of spatial units; Spatial weights; and spatial unit and The attribute value; This represents the attribute mean.

5. The method for risk assessment in steeply inclined coal mines according to claim 4, characterized in that, The formula for calculating the risk index is: ; ; ; In the formula, , and These are the roof instability risk index, the gas over-limit risk index, and the flood risk index; , and These are the preset roof stability vector, gas safety vector, and flood safety vector, respectively. The modulo operation is used; the comprehensive risk assessment index is calculated using a weighted summation method. ; In the formula, It is a comprehensive risk assessment index; , and Let be the weight coefficient, and satisfy... ; This is a correction item used to consider other potential risk factors; Ultimately, With preset risk threshold By comparison, the risk assessment results for mining steeply inclined coal mines are obtained: if If so, the mining risk is considered to be within an acceptable range; if If the risk exceeds acceptable limits, then corresponding risk control measures need to be taken.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program instructions, which, when executed in a computer, are used to perform the risk assessment method for steeply inclined coal mine mining as described in any one of claims 1-5.

7. A risk assessment system for steeply inclined coal mine mining, characterized in that, It includes the computer-readable storage medium of claim 6.

Images