A device and system for preventing and controlling landslide disasters in open-pit mines

Abstract

The present application relates to the technical field of disaster prevention and control, in particular to a device and system for open-pit mine landslide disaster prevention and control, which comprises a data acquisition module, a key disaster-causing factor analysis module, a risk quantification characterization module and a scenario prevention and control deduction module; the data acquisition module is used for collecting historical case data of open-pit mine landslide disaster, legal and regulatory texts, real-time slope images and multi-source heterogeneous monitoring data; the key disaster-causing factor analysis module is used for identifying and screening key disaster-causing factors; the risk quantification characterization module is used for uniformly quantifying quantitative risk prevention and control indicators and qualitative risk prevention and control indicators; the scenario prevention and control deduction module generates landslide disaster-causing mechanism and prevention and control strategies. The present application evaluates the open-pit mine landslide disaster risk prevention and control capability through image recognition, fuzzy theory and logic gate operation, analyzes the key disaster-causing nodes and disaster-causing paths, and effectively improves the comprehensiveness and reliability of the mine area in disaster prevention and control and disaster tracing.

Description

Technical Field

[0001] This invention relates to the field of disaster prevention and control technology, specifically to a device and system for preventing and controlling landslide disasters in open-pit mines. Background Technology

[0002] Open-pit mining, as an important method of coal mining, has advantages such as rapid construction, high production efficiency, and high resource recovery rate. However, high-intensity mining and complex engineering geological environments lead to frequent geological disasters in open-pit mines. In particular, as the scale of mining continues to expand and the slope height increases with mining production, landslides are becoming increasingly prominent, posing a great threat to the life, health, and safety of workers in the mining area, as well as the safe and stable operation of equipment, facilities, and intelligent equipment systems.

[0003] In existing technologies, such as the "Theoretical Basis and Key Technologies for Prevention and Control of Coal and Rock Dynamic Disasters in Deep Coal Mining" (MaQ), a generalized "three-factor" theory for uniformly describing the occurrence mechanism of coal mine disasters has been established. It has refined multi-scale source-specific prevention and control technologies for coal and rock dynamic disasters from the aspects of ideological cognition, principles and methods and core technologies. By analyzing the physical characteristics, stress changes and structural changes of the mining area, it analyzes the possibility and causes of disasters, thereby improving the disaster analysis and prevention and control capabilities of the mining area.

[0004] However, in practical applications, the causes of open-pit mine landslides are numerous, often resulting from a combination of defects such as unclear risks, inadequate monitoring, poor early warning, ineffective evacuation, and improper handling. Existing source tracing and prevention technologies often only focus on one or a few stages of the accident's development, making it difficult to form a comprehensive disaster prevention theory that covers the entire process. This seriously affects the effectiveness of landslide source tracing and prevention. Therefore, it is necessary to propose a prevention and control device and system for open-pit mine landslides to solve the above problems. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a prevention and control device and system for landslide disasters in open-pit mines. Through image recognition, fuzzy theory, and logic gate operations, it achieves quantitative characterization of key disaster-causing factors, thereby assessing the risk prevention and control capabilities of open-pit mine landslide disasters. By constructing a disaster risk scenario prevention and control model through Bayesian mapping, it analyzes key disaster-causing nodes and disaster-causing paths, effectively improving the comprehensiveness and reliability of disaster prevention and control and disaster source tracing in mining areas.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows: a prevention and control system for landslide disasters in open-pit mines, comprising a data acquisition module, a key disaster-causing factor analysis module, a risk quantification and characterization module, a risk dynamic assessment module, and a scenario prevention and control simulation module.

[0007] The data acquisition module is used to collect historical case data, legal and regulatory texts, real-time slope images, and multi-source heterogeneous monitoring data of open-pit mine landslide disasters through web crawling technology, text mining technology, and on-site monitoring equipment, and transmits the historical case data, legal and regulatory texts, real-time slope images, and multi-source heterogeneous monitoring data to the key disaster-causing factor analysis module.

[0008] The key disaster-causing factor analysis module is used to construct a disaster-causing factor network based on historical case data, legal and regulatory texts, real-time slope images, and multi-source heterogeneous monitoring data, using complex network theory. Based on the disaster-causing factor network, it analyzes the three types of centrality and hierarchical structure of each data point. Based on the analysis results, it identifies and screens key disaster-causing factors and transmits them to the risk quantification and characterization module. The three types of centrality include degree centrality, betweenness centrality, and proximity centrality.

[0009] The risk quantification and characterization module is used to analyze and quantify key disaster-causing factors through computer vision image recognition technology and fuzzy set theory, to obtain qualitative and quantitative risk prevention and control indicators, and then transmit these indicators to the risk dynamic assessment module and the scenario prevention and control simulation module. The specific quantification method is to map the data obtained by the data acquisition module to the key disaster-causing factors.

[0010] The dynamic risk assessment module is used to dynamically assess and determine the landslide risk level based on qualitative and quantitative risk prevention and control indicators.

[0011] The scenario prevention and control simulation module is used to construct a scenario prevention and control model for open-pit mine landslide disaster risks using a Bayesian network model. Quantitative risk prevention and control indicators are input into the scenario prevention and control model, and causal reasoning, analytical reasoning, sensitivity judgment, and critical path analysis are performed. Based on the results of dynamic risk assessment, the module identifies the weak links of key indicators, compares historical accident cases with legal and regulatory texts from the data acquisition module, and outputs targeted disaster-causing mechanisms and prevention and control strategies for landslide disasters.

[0012] Furthermore, the data acquisition module includes a web crawler unit, a text mining unit, an image acquisition unit, and a sensor access unit.

[0013] The web crawler unit is used to crawl historical open-pit mine landslide disaster cases and safety regulations texts from public data sources to form a text database.

[0014] The text mining unit is used to extract keywords, perform co-occurrence analysis, and semantic feature modeling on historical open-pit mine landslide disaster cases and safety regulations texts in the text database.

[0015] The image acquisition unit is used to acquire images of slope monitoring experiments and historical landslide accidents.

[0016] The sensor access unit is used to collect multi-source heterogeneous monitoring data, including slope displacement, stress, and rainfall.

[0017] Furthermore, the key disaster-causing factor analysis module is also used to construct a disaster-causing factor database based on historical open-pit mine landslide disaster cases and legal documents; the disaster-causing factor database includes qualitative risk prevention and control indicators and quantitative risk prevention and control indicators.

[0018] Furthermore, the risk quantification and characterization module includes an image recognition unit and a fuzzy quantization unit.

[0019] The image recognition unit is used to extract image features from slope images using computer vision algorithms, and to set a fluctuation threshold for the image features. When the feature value of an image feature exceeds the fluctuation threshold, the image recognition unit marks the image feature as a quantitative risk control indicator. The image features include crack morphology, displacement pattern, and structural surface inclination angle.

[0020] The fuzzy quantization unit is used to convert qualitative risk control indicators into quantitative values ​​through fuzzy set membership functions and logic gate operations.

[0021] Furthermore, the fuzzy quantization unit is also used to combine multi-source heterogeneous data obtained through monitoring and surveillance technology with slope image features extracted by computer vision intelligent algorithms. Fuzzy set membership functions and cluster analysis are used to set the evaluation level and grading threshold of each quantitative risk prevention and control indicator, forming a multi-modal acquisition framework for multi-source heterogeneous data information and its unified quantitative representation.

[0022] Furthermore, the scenario prevention and control simulation module also uses a Bayesian network model to configure the assessment level, grading threshold, and correlation between quantitative risk prevention and control indicators as the prior probability of the root node and the conditional probability table between nodes of the Bayesian network, respectively.

[0023] Furthermore, the scenario prevention and control simulation module is also used to input the initial conditions, evolution process and emergency response measures of historical landslide disasters into the constructed Bayesian network model as input parameters to reproduce the occurrence, development and emergency response process of the disaster.

[0024] Based on Bayesian network-based reverse reasoning and sensitivity analysis, the probability of disasters and accidents is quantitatively calculated, the causal path with the highest probability of occurrence is identified, and the weak links in the emergency response process are analyzed.

[0025] The accident occurrence probability, key causal paths, and weak links in emergency response obtained from the simulation are compared and verified item by item with the recorded results in the actual accident investigation report. Based on the comparison differences, the prior probability, conditional probability table, and network structure of the Bayesian network model are iteratively corrected until the simulation results are consistent with the key conclusions of the investigation report.

[0026] Furthermore, it also includes an early warning and emergency response linkage module, which is used to set up tiered early warning and emergency response procedures; when the Bayesian network model deduces quantitative risk prevention and control indicators that exceed their thresholds, the early warning and emergency response linkage module automatically initiates tiered early warning and triggers the emergency response procedure.

[0027] The technical principle of the above solution is as follows:

[0028] The data acquisition module automatically obtains historical disaster cases and safety regulations from publicly available data sources using web crawling and text mining techniques, constructing a text database. Simultaneously, it acquires structured and image data of real-time slope conditions through on-site monitoring and image acquisition equipment. The key disaster-causing factor analysis module, based on complex network theory, treats disaster-causing factors as network nodes and the relationships between them as edges, constructing a disaster-causing factor network. By calculating the degree centrality, betweenness centrality, and proximity centrality of nodes, combined with hierarchical structure analysis, it identifies nodes that play a crucial connecting role or are easily affected, thus screening key disaster-causing factors from numerous options. The risk quantification and characterization module uses computer vision image recognition technology to automatically extract crack morphology, displacement patterns, and structural surface inclination angles from slope images, forming quantitative risk prevention and control indicators, and setting assessment levels and thresholds. The scenario prevention and control simulation module maps the indicator assessment levels, thresholds, and logical relationships between indicators output by the risk quantification and characterization module to the prior and conditional probabilities of Bayesian network nodes, revealing the shortest or most likely propagation path from the initial disaster-causing factor to the occurrence of the disaster. When the risk indicators derived by the Bayesian network exceed the preset threshold, the scenario prevention and control simulation module automatically triggers the corresponding level of early warning information and initiates the preset emergency response process.

[0029] The above approach has the following beneficial effects:

[0030] 1. This invention, through causal reasoning and critical path analysis using Bayesian networks, can identify high-risk disaster-causing paths in advance before disaster-causing factors have fully developed to the threshold, thereby achieving early warning, improving the timeliness and reliability of disaster warning, and buying valuable time for personnel evacuation.

[0031] 2. This invention objectively analyzes key disaster-causing factors through complex network theory, avoiding interference from redundant indicators, and making prevention and control strategies more consistent with the actual conditions of the mining area. At the same time, by analyzing and reasoning back the disaster-causing paths, it provides a clear direction for on-site rectification, effectively improving the rationality and pertinence of prevention and control strategies.

[0032] 3. This invention conducts dynamic simulation and deduction of risk prevention and control throughout the entire process before and after a landslide disaster, thereby enabling a more in-depth analysis of the disaster-causing mechanism of open-pit mine landslide disasters, comprehensively simulating the dynamic deduction process of disaster risk prevention and control, and thus effectively improving the comprehensiveness and reliability of disaster prevention and control and disaster source tracing in mining areas.

[0033] Furthermore, a device for preventing and controlling landslide disasters in open-pit mines includes a processor and a storage device.

[0034] Beneficial effects: This solution enables stable storage and analysis of various data through the processor and storage. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the structure of the prevention and control system for landslide disasters in open-pit mines according to the present invention.

[0036] Figure 2 This is a flowchart illustrating the analysis of key disaster-causing factors in the prevention and control system for landslide disasters in open-pit mines, as described in this invention.

[0037] Figure 3 This is a flowchart illustrating the analysis of quantitative risk control indicators in the open-pit mine landslide disaster prevention and control system of this invention.

[0038] Figure 4 This is a flowchart illustrating the analysis of qualitative risk control indicators in the open-pit mine landslide disaster prevention and control system of this invention. Detailed Implementation

[0039] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0041] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0042] The following detailed description illustrates the specific implementation method:

[0043] Example 1:

[0044] like Figure 1 As shown, a prevention and control system for landslide disasters in open-pit mines includes a data acquisition module, a key disaster-causing factor analysis module, a risk quantification and characterization module, a risk dynamic assessment module, a scenario prevention and control simulation module, and an early warning and emergency response linkage module; each module is connected to the other via signals.

[0045] The specific functions of each module are as follows:

[0046] like Figure 2 As shown, the data acquisition module is used to collect historical case data, legal documents, real-time slope images, and multi-source heterogeneous monitoring data of open-pit mine landslide disasters through web crawling technology, text mining technology, and on-site monitoring equipment. This data is then transmitted to the key disaster-causing factor analysis module. In this embodiment, Figure 4 The research mainly refers to the collection of historical case data and legal documents related to landslide disasters in open-pit mines.

[0047] The data acquisition module includes a web crawler unit, a text mining unit, an image acquisition unit, and a sensor access unit.

[0048] The web crawler unit is used to crawl historical open-pit mine landslide disaster cases and safety regulations texts from public data sources to form a text database.

[0049] Specifically, in this embodiment, the user sets keywords through the web crawler unit, such as "open-pit mine landslide", "slope collapse" and "open-pit mine accident". The web crawler unit is based on the BeautifulSoup framework and periodically crawls historical disaster case texts and safety regulation texts from public data sources such as the official website of the Ministry of Emergency Management, CNKI and Wanfang, and stores them in the text library after deduplication and cleaning.

[0050] The text mining unit is used to extract keywords, perform co-occurrence analysis, and semantic feature modeling on historical open-pit mine landslide disaster cases and safety regulations texts in the text database.

[0051] Specifically, in this embodiment, the text mining unit uses the LDA topic model and the Word2Vec word vector method to extract keywords, construct co-occurrence matrices, and model semantic features from text data in the text library, outputting the frequency of disaster-causing factors and providing the relational weights between nodes and edges for the construction of complex networks.

[0052] The image acquisition unit is used to acquire images of slope monitoring experiments and historical landslide accidents.

[0053] Specifically, in this embodiment, the image acquisition unit includes a camera deployed on the top of the open-pit mine slope and a drone equipped with an image recognition device. The image acquisition unit is used to periodically acquire images of the slope surface and historical landslide accident sites using the camera and the drone, and transmits the image data to the risk quantification and characterization module after annotating the image data with timestamps and location information.

[0054] The sensor access unit is used to collect multi-source heterogeneous monitoring data, including slope displacement, stress, and rainfall.

[0055] Specifically, in this embodiment, the sensor access unit includes displacement gauges, stress gauges, pore water pressure gauges, and rainfall sensors from a weather station, all installed inside the slope. The displacement, stress, and rainfall of the slope are detected in real time using these sensors.

[0056] like Figure 3 As shown, the key disaster-causing factor analysis module is used to construct a disaster-causing factor network based on historical case data, legal and regulatory texts, real-time slope images, and multi-source heterogeneous monitoring data, using complex network theory. Based on the disaster-causing factor network, it analyzes the three types of centrality and hierarchical structure of each data point. Based on the analysis results, it identifies and screens key disaster-causing factors and transmits the key disaster-causing factors to the risk quantification and characterization module. The three types of centrality include degree centrality, betweenness centrality, and proximity centrality.

[0057] The key disaster-causing factor analysis module is also used to construct a disaster-causing factor database based on historical open-pit mine landslide disaster cases and legal regulations; the database includes qualitative and quantitative risk prevention and control indicators.

[0058] Specifically, the key disaster-causing factor analysis module performs the following operations:

[0059] First, the historical case data, legal and regulatory texts, real-time slope images and multi-source heterogeneous monitoring data transmitted by the data acquisition module are structured and processed to extract the disaster-causing factor entities and the co-occurrence, causal or temporal correlation relationships between factors, and to construct an initial set of disaster-causing factors.

[0060] Secondly, based on complex network theory, each disaster-causing factor is treated as a network node, and the relationship between factors is treated as an edge. An undirected weighted network of disaster-causing factors G=(V,E,W) is constructed, where V is the set of nodes, E is the set of edges, and W is the set of weights. The weights are determined by a combination of co-occurrence frequency, semantic similarity, and empirical knowledge output by the text mining unit.

[0061] Then, calculate the three types of centrality indicators for each node in the network:

[0062] Degree centrality measures the number of direct connections a node has in a network, reflecting the degree of direct correlation between this factor and other factors.

[0063] Betweenness centrality measures the frequency with which a node appears in all shortest paths, reflecting the bridging effect of this factor in the network.

[0064] Proximity centrality measures the average shortest path length from a node to all other nodes, reflecting the efficiency of information propagation in the network.

[0065] By analyzing the above three types of centrality indicators and combining them with hierarchical structure analysis (in this embodiment, hierarchical clustering or community detection algorithms are used for analysis), nodes that have significant influence at each level (the higher the sum of the three types of centrality indicators, the stronger the influence) are identified as key disaster-causing factors.

[0066] Finally, based on the theory of full-chain, multi-level disaster prevention and control, the identified key disaster-causing factors are classified and stored according to nine links: risk classification and control, hidden danger investigation and management, timely reporting of warning signs, emergency prevention and preparedness, monitoring, evaluation and early warning, efficient rescue and disposal, and rapid recovery and reconstruction. A structured disaster-causing factor database is constructed, and the selected key disaster-causing factors and their network parameters are transmitted to the risk quantification and characterization module.

[0067] like Figure 4 As shown, the risk quantification and characterization module is used to analyze and quantify key disaster-causing factors through computer vision image recognition technology and fuzzy set theory, to obtain qualitative and quantitative risk prevention and control indicators, and then transmits these indicators to the risk dynamic assessment module and the scenario prevention and control simulation module. The specific quantification method is to map the data obtained by the data acquisition module to the key disaster-causing factors.

[0068] The risk quantification and characterization module includes an image recognition unit and a fuzzy quantization unit.

[0069] The image recognition unit is used to extract image features from slope images using computer vision algorithms, and to set a fluctuation threshold for the image features. When the feature value of an image feature exceeds the fluctuation threshold, the image recognition unit marks the image feature as a quantitative risk control indicator. The image features include crack morphology, displacement pattern, and structural surface inclination angle.

[0070] Specifically, the image recognition unit first receives slope monitoring experiment images and historical landslide accident images from the data acquisition module, and uses a deep learning-based computer vision algorithm (YOLO deep learning model is selected in this embodiment) to preprocess the images, including image denoising, enhancement and scale normalization operations.

[0071] Then, the crack morphology in the slope image is extracted using a trained semantic segmentation model, including the crack length, width, direction, density, and distribution area; the displacement pattern is extracted using a feature point matching algorithm, including the displacement vector field, displacement rate, and cumulative displacement of the slope surface; and the dip angle of the structural plane is extracted using Hough transform, including the attitude, dip direction, and dip angle variation trend of the main structural planes.

[0072] Finally, the image recognition unit compares the extracted instability features with the preset fluctuation thresholds. When any feature exceeds its fluctuation threshold (the user sets the fluctuation thresholds for each image feature through the risk quantification characterization module), the key disaster-causing factor is marked as a qualitative risk prevention and control indicator, and the abnormality level (such as mild abnormality, moderate abnormality, and severe abnormality) is labeled. The abnormality is then transmitted to the fuzzy quantization unit along with the original image feature data.

[0073] The fuzzy quantization unit is used to transform qualitative risk control indicators into quantitative values ​​through fuzzy set membership functions and logic gate operations. The fuzzy quantization unit is also used to set the evaluation level and grading threshold for each quantitative risk control indicator using fuzzy set membership functions and cluster analysis, based on multi-source heterogeneous data obtained through monitoring technology and slope image features extracted by computer vision intelligent algorithms. This forms a multi-modal acquisition framework for multi-source heterogeneous data information and its unified quantitative representation.

[0074] Specifically, firstly, for each qualitative risk prevention and control indicator, its domain and corresponding fuzzy set are defined, and a fuzzy set membership function is established. For example, for the indicator of "crack development degree", a fuzzy set {slight, moderate, and severe} is defined, and a trapezoidal membership function is used to map the qualitative description to a membership value in the interval [0,1]. For indicators such as "safety management level", fuzzification is performed using a Gaussian membership function based on questionnaire surveys or expert scoring results.

[0075] Secondly, for multiple qualitative risk control indicators that have logical combination relationships, fuzzy logic gates are used for comprehensive calculation to obtain the comprehensive fuzzy evaluation value of the composite qualitative risk control indicators.

[0076] Then, the maximum membership method is used to transform the fuzzy evaluation value into a precise quantitative risk prevention and control indicator value, which is between 0 and 1 or mapped according to a preset level range.

[0077] Finally, the fuzzy quantization unit uses historical data statistics results from monitoring and surveillance technology combined with computer vision intelligent algorithms and cluster analysis to divide the value range of quantitative risk prevention and control indicators into several levels (including low risk, medium risk and high risk), and sets threshold boundaries for each level. This forms a multimodal acquisition and quantitative characterization result of multi-source heterogeneous data information, and transmits the final quantitative risk prevention and control indicators and their level and threshold information to the scenario prevention and control simulation module.

[0078] The dynamic risk assessment module is used to dynamically assess and determine the landslide risk level based on qualitative and quantitative risk prevention and control indicators.

[0079] The scenario-based prevention and control simulation module is used to construct a scenario-based prevention and control model for open-pit mine landslide disaster risks using a Bayesian network model. Quantitative risk prevention and control indicators are input into the model, and causal reasoning, analytical reasoning, sensitivity assessment, and critical path analysis are performed to generate landslide disaster causal mechanisms and prevention and control strategies. The module also uses the Bayesian network model to configure the assessment levels, grading thresholds, and inter-indicator relationships of quantitative risk prevention and control indicators as prior probabilities of the root node and conditional probability tables between nodes. Furthermore, the module uses the initial conditions, evolution process, and emergency response measures of historical landslide disasters as input parameters into the constructed Bayesian network model to reproduce the occurrence, development, and emergency response process of the disaster.

[0080] Based on Bayesian network-based reverse reasoning and sensitivity analysis, the probability of disasters and accidents is quantitatively calculated, the causal path with the highest probability of occurrence is identified, and the weak links in the emergency response process are analyzed.

[0081] The accident occurrence probability, key causal paths, and weak links in emergency response obtained from the simulation are compared and verified item by item with the recorded results in the actual accident investigation report. Based on the comparison differences, the prior probability, conditional probability table, and network structure of the Bayesian network model are iteratively corrected until the simulation results are consistent with the key conclusions of the investigation report.

[0082] Specifically, firstly, using the quantitative risk prevention and control indicators output by the risk quantification and characterization module as the root node of the Bayesian network, the occurrence, development, and emergency response effectiveness of landslide disasters as leaf nodes, and intermediate disaster-causing events as intermediate nodes, a Bayesian network topology framework with a directed acyclic graph structure is constructed. Directed edges between nodes represent causal relationships, with the direction pointing from cause to effect.

[0083] Secondly, the assessment levels and thresholds of quantitative risk prevention and control indicators are mapped to the prior probabilities of the root node: For each quantitative risk prevention and control indicator, the level range of its current quantitative value is mapped to the prior probability distribution. For example, when the crack width indicator is at a high risk level, its prior probability distribution is set as P(severe) = 0.7, P(moderate) = 0.2, and P(minor) = 0.1.

[0084] Then, the conditional probability table is determined as follows:

[0085] Frequency statistics were performed on the co-occurrence frequency of historical cases extracted by the text mining unit, and conditional probabilities were calculated; a Noisy-OR model was used to supplement this calculation, taking into account user experience.

[0086] Given the prior probability distribution of each root node, the scenario prevention and control simulation module uses the chain rule and conditional independence assumption of the Bayesian network model to calculate the posterior probability of the leaf node (the occurrence of landslide disaster).

[0087] When a landslide occurs or abnormal signs are observed, the scenario prevention and control simulation module inputs the landslide event as evidence into the Bayesian network model and calculates the posterior probability of each disaster-causing factor in reverse.

[0088] When performing sensitivity assessment, the scenario prevention and control simulation module uses mutual information entropy or variance reduction methods to calculate the sensitivity coefficient of each quantitative risk prevention and control indicator to the probability of landslide disaster occurrence.

[0089] The scenario prevention and control simulation module is based on the conditional dependencies between nodes in the Bayesian network model. It uses the maximum probability path search algorithm to identify the most likely propagation path from the initial disaster-causing factor to the occurrence of the landslide disaster, forming a key disaster-causing path map. At the same time, it identifies the vulnerable links from the source of risk to emergency failure, which serve as the key directions for optimizing prevention and control strategies.

[0090] The early warning and emergency response linkage module is used to set up tiered early warning and emergency response procedures. When the quantitative risk control indicators derived by the Bayesian network model exceed their thresholds, the early warning and emergency response linkage module automatically initiates tiered early warning and triggers the emergency response procedure.

[0091] Specifically, users can pre-divide multiple warning levels (in this embodiment, blue, yellow, orange, and red warnings, corresponding to low, medium, high, and extremely high risks, respectively) based on the assessment level and threshold of quantitative risk prevention and control indicators through the early warning and emergency linkage module. For each level, users can configure corresponding warning triggering conditions, warning information content, and release scope. For different warning levels, users can pre-set emergency response procedures to be linked to them. In this implementation, the emergency response procedures include on-site inspection, personnel evacuation, equipment shutdown, slope reinforcement, emergency resource allocation, and dispatch of rescue forces.

[0092] When the Bayesian network model calculates in real time that the value of any quantitative risk control indicator exceeds its preset threshold, the early warning and emergency response module automatically determines the corresponding early warning level, immediately initiates a tiered early warning (in this embodiment, early warning information is released through sound and light alarms, SMS push notifications, large screen displays, etc.), and simultaneously triggers the emergency response process corresponding to that level.

[0093] After the emergency response process is executed, the early warning and emergency linkage module will feed back the response results (including evacuation completion rate, reinforcement effect and trend of hazard change in this embodiment) to the Bayesian network model in real time for updating node probabilities and subsequent inferences, forming a dynamic closed loop of monitoring, early warning, response and feedback.

[0094] This embodiment conducts dynamic simulation and deduction of risk prevention and control throughout the entire process before and after a landslide disaster, analyzes the probability and relationship of each disaster-causing factor, and deduces the complete disaster-causing path. This enables a more in-depth analysis of the disaster-causing mechanism of open-pit mine landslide disasters, comprehensively simulates the dynamic deduction process of disaster risk prevention and control, and effectively improves the comprehensiveness and reliability of disaster prevention and control and disaster source tracing in mining areas.

[0095] Example 2:

[0096] Unlike the embodiments described above, a prevention and control device for landslide disasters in open-pit mines includes a processor and a storage device.

[0097] Specifically, the storage unit is used to store historical case data, legal and regulatory texts, real-time slope images, slope displacement, stress, and rainfall data. The processor is used to analyze the above types of data.

[0098] This embodiment, through the inclusion of a processor and storage, enables the stable operation of the prevention and control system for landslide disasters in open-pit mines.

[0099] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A prevention and control system for landslide disasters in open-pit mines, characterized in that, It includes a data acquisition module, a key disaster-causing factor analysis module, a risk quantification and characterization module, a risk dynamic assessment module, and a scenario prevention and control simulation module; The data acquisition module is used to collect historical case data, legal and regulatory texts, real-time slope images and multi-source heterogeneous monitoring data of open-pit mine landslide disasters through web crawling technology, text mining technology and on-site monitoring equipment, and transmit the historical case data, legal and regulatory texts, real-time slope images and multi-source heterogeneous monitoring data to the key disaster-causing factor analysis module. The key disaster-causing factor analysis module is used to construct a disaster-causing factor network based on historical case data, legal and regulatory texts, real-time slope images, and multi-source heterogeneous monitoring data, using complex network theory. Based on the disaster-causing factor network, it analyzes the three types of centrality and hierarchical structure of each data point. Based on the analysis results, it identifies and screens key disaster-causing factors and transmits these factors to the risk quantification and characterization module. The three types of centrality include degree centrality, betweenness centrality, and proximity centrality. The risk quantification and characterization module is used to analyze and quantify key disaster-causing factors through computer vision image recognition technology and fuzzy set theory, to obtain qualitative and quantitative risk prevention and control indicators, and then transmit these indicators to the risk dynamic assessment module and the scenario prevention and control simulation module. The specific quantification method is to map the data obtained by the data acquisition module to the key disaster-causing factors. The risk dynamic assessment module is used to dynamically assess and determine the landslide risk level based on qualitative and quantitative risk prevention and control indicators. The scenario prevention and control simulation module is used to construct a scenario prevention and control model for open-pit mine landslide disaster risks using a Bayesian network model. Quantitative risk prevention and control indicators are input into the scenario prevention and control model, and causal reasoning, analytical reasoning, sensitivity judgment, and critical path analysis are performed. Based on the results of dynamic risk assessment, the module identifies the weak links of key indicators, compares historical accident cases with legal and regulatory texts from the data acquisition module, and outputs targeted disaster-causing mechanisms and prevention and control strategies for landslide disasters.

2. The prevention and control system for landslide disasters in open-pit mines according to claim 1, characterized in that, The data acquisition module includes a web crawler unit, a text mining unit, an image acquisition unit, and a sensor access unit; The web crawler unit is used to crawl historical open-pit mine landslide disaster cases and safety regulations texts from public data sources to form a text database. The text mining unit is used to extract keywords, perform co-occurrence analysis, and semantic feature modeling on historical open-pit mine landslide disaster cases and safety regulations in the text database. The image acquisition unit is used to acquire images from slope monitoring experiments and historical landslide accidents. The sensor access unit is used to collect multi-source heterogeneous monitoring data, including slope displacement, stress, and rainfall.

3. The prevention and control system for landslide disasters in open-pit mines according to claim 2, characterized in that, The key disaster-causing factor analysis module is also used to build a disaster-causing factor database based on historical open-pit mine landslide disaster cases and legal documents; the disaster-causing factor database includes qualitative risk prevention and control indicators and quantitative risk prevention and control indicators.

4. The prevention and control system for landslide disasters in open-pit mines according to claim 3, characterized in that, The risk quantification and characterization module includes an image recognition unit and a fuzzy quantization unit; The image recognition unit is used to extract image features from slope images using computer vision algorithms, set fluctuation thresholds for image features, and when the feature value of an image feature exceeds the fluctuation threshold, the image recognition unit is used to mark the image feature as a quantitative risk control indicator. Image features include crack morphology, displacement pattern, and structural surface dip angle; The fuzzy quantization unit is used to convert qualitative risk control indicators into quantitative values ​​through fuzzy set membership functions and logic gate operations.

5. The prevention and control system for landslide disasters in open-pit mines according to claim 4, characterized in that, The fuzzy quantization unit is also used to obtain multi-source heterogeneous data through monitoring and surveillance technology and slope image features extracted by computer vision intelligent algorithms. It uses fuzzy set membership function and cluster analysis to set the evaluation level and grading threshold of each quantitative risk prevention and control indicator, forming a multi-modal acquisition framework for multi-source heterogeneous data information and its unified quantitative representation.

6. The prevention and control system for landslide disasters in open-pit mines according to claim 5, characterized in that, The scenario prevention and control simulation module also uses a Bayesian network model to configure the assessment level, grading threshold, and correlation between quantitative risk prevention and control indicators as the prior probability of the root node and the conditional probability table between nodes of the Bayesian network, respectively.

7. The prevention and control system for landslide disasters in open-pit mines according to claim 6, characterized in that, The scenario prevention and control simulation module is also used to input the initial conditions, evolution process and emergency response measures of historical landslide disasters into the constructed Bayesian network model as input parameters to reproduce the occurrence, development and emergency response process of the disasters. Based on Bayesian network-based reverse reasoning and sensitivity analysis, we quantitatively calculate the probability of disasters and accidents, identify the causal paths with the highest probability of occurrence, and analyze the weak links in the emergency response process. The accident occurrence probability, key causal paths, and weak links in emergency response obtained from the simulation are compared and verified item by item with the recorded results in the actual accident investigation report. Based on the comparison differences, the prior probability, conditional probability table, and network structure of the Bayesian network model are iteratively corrected until the simulation results are consistent with the key conclusions of the investigation report.

8. The prevention and control system for landslide disasters in open-pit mines according to claim 7, characterized in that, It also includes an early warning and emergency response linkage module, which is used to set up tiered early warning and emergency response procedures; when the Bayesian network model deduces quantitative risk prevention and control indicators that exceed their thresholds, the early warning and emergency response linkage module automatically initiates tiered early warning and triggers the emergency response procedure.

9. A prevention and control device for landslide disasters in open-pit mines, comprising any one of the prevention and control systems for landslide disasters in open-pit mines according to claims 1-8, characterized in that, This includes the processor and storage.

Images