Intelligent identification method and platform for tailings pond disaster hazards

Abstract

The application discloses a tailing pond disaster hidden danger intelligent identification method and platform, and relates to the technical field of intelligent monitoring, which comprises the following steps: based on the global monitoring data of the tailing pond, automatically partitioning and constructing a layered data acquisition system; establishing a physical and mechanical database, combining consolidation and mechanical data to construct a mechanical parameter inversion model; identifying the physical and mechanical characteristics of the dam body through the model, determining the disaster type and performing grade division; deducing the disaster evolution process, analyzing the disaster chain and constructing a structure tree; judging the disaster occurrence rate according to the disaster chain, and if the disaster occurrence rate exceeds the early warning threshold, generating a risk early warning report. The technical problem of insufficient precision of the tailing pond disaster hidden danger identification system in the prior art is solved, and through accurate identification of the tailing pond disaster type, dynamic adjustment of the disaster grade and real-time deduction of the disaster evolution process, the technical effect of improving the accuracy of tailing pond disaster hidden danger identification is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent monitoring, in particular to an intelligent identification method and platform for tailing pond disaster hazards. BACKGROUND

[0002] Tailing ponds are important facilities for storing waste materials during the process of mineral exploitation, but their safety problems have always been a key challenge in the mining industry. If tailing pond disaster hazards are not discovered and addressed in a timely manner, they often lead to serious environmental pollution, resource waste, and personnel casualties. Traditional tailing pond disaster monitoring usually relies on manual patrols and limited monitoring means, such as ground sensors and remote sensing technology. These technologies have problems such as incomplete data collection, poor accuracy in disaster identification, lack of disaster evolution process deduction, and delayed early warning. The above shortcomings make the existing technology unable to effectively cope with the dynamics and complexity of tailing pond disasters, so there is an urgent need for a new method and system to improve the identification ability of tailing pond disaster hazards and ensure accurate assessment and timely warning of disaster risks.

[0003] At the present stage, there is a technical problem of insufficient precision of tailing pond disaster hazard identification systems in related technologies. SUMMARY

[0004] The present application provides an intelligent identification method and platform for tailing pond disaster hazards, solving the technical problem of insufficient precision of tailing pond disaster hazard identification systems in existing technologies.

[0005] The present application provides an intelligent identification method for tailing pond disaster hazards, comprising:

[0006] Based on the global monitoring feature data of the target tailing pond, the target monitoring area is automatically partitioned, and a hierarchical data collection system is constructed. A tailing physical and mechanical database is established, and combined with the consolidation data and mechanical feature data of the target tailing pond fed back by the hierarchical data collection system, a tailing pond mechanical parameter inversion model is constructed. Through the tailing pond mechanical parameter inversion model, the physical and mechanical feature data of the target tailing pond dam body are identified, the disaster type is determined and the disaster grade is divided. Based on the tailing pond mechanical parameter inversion model, the disaster evolution process is deduced, the disaster chain is analyzed and a disaster chain structure tree is constructed. Combined with the disaster chain structure tree, the disaster occurrence rate of the target tailing pond is judged, and if the disaster occurrence rate exceeds the preset disaster warning threshold, a risk warning report is generated.

[0007] The present application provides an intelligent identification platform for tailing pond disaster hazards, comprising:

[0008] The hierarchical data acquisition system construction module is configured to automatically partition a target monitoring area and construct a hierarchical data acquisition system based on global monitoring feature data of a target tailing pond; the tailing pond mechanics parameter inversion model construction module is configured to establish a tailing pond physical mechanics database, combine consolidation data and mechanics feature data of the target tailing pond fed back by the hierarchical data acquisition system, and construct a tailing pond mechanics parameter inversion model; the disaster type determination module is configured to identify physical mechanics feature data of a dam body of the target tailing pond through the tailing pond mechanics parameter inversion model, determine a disaster type, and perform disaster grade division; the disaster chain structure tree construction module is configured to deduce a disaster evolution process, analyze a disaster chain, and construct a disaster chain structure tree based on the tailing pond mechanics parameter inversion model; and the risk early warning report generation module is configured to judge a disaster occurrence rate of the target tailing pond in combination with the disaster chain structure tree, and generate a risk early warning report if the disaster occurrence rate exceeds a preset disaster early warning threshold.

[0009] The intelligent tailing pond disaster hidden danger identification method and platform provided in the present application first automatically partitions and constructs a hierarchical data acquisition system based on global monitoring data of a tailing pond; a physical mechanics database is established, a mechanics parameter inversion model is constructed in combination with consolidation and mechanics data; physical mechanics features of a dam body are identified through the model, a disaster type is determined, and grade division is performed; a disaster evolution process is deduced, a disaster chain is analyzed, and a structure tree is constructed; a disaster occurrence rate is judged according to the disaster chain, and a risk early warning report is generated if the disaster occurrence rate exceeds a preset early warning threshold. Through accurate identification of a tailing pond disaster type, dynamic adjustment of a disaster grade, and real-time deduction of a disaster evolution process, the technical effect of improving the accuracy of tailing pond disaster hidden danger identification is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. The flowcharts are used in the present application to illustrate the operations performed by the platform according to the embodiments of the present application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously according to needs. Meanwhile, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0011] Figure 1 A flowchart of the intelligent tailing pond disaster hidden danger identification method provided by the embodiments of the present application;

[0012] Figure 2 A structure diagram of the intelligent tailing pond disaster hidden danger identification platform provided by the embodiments of the present application.

[0013] The reference signs are explained as follows: a hierarchical data acquisition system construction module 10, a tailing pond mechanical parameter inversion model construction module 20, a disaster type determination module 30, a disaster chain structure tree construction module 40, and a risk early warning report generation module 50. DETAILED DESCRIPTION

[0014] The above description is only a summary of the technical scheme of the present application. In order to make the technical means of the present application more clear, the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described.

[0015] In order to make the purposes, technical schemes and advantages of the present application more clear, the present application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting the present application, and all other embodiments obtained by those skilled in the art without making creative efforts fall within the scope of protection of the present application.

[0016] In the following description, "some embodiments" are referred to, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict, and the term "first\second" referred to is only to distinguish similar objects, and does not represent a specific order of the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, platform, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0017] The embodiments of the present application provide an intelligent identification method for tailing pond disaster hazards, as shown in the method, the method comprises: Figure 1 The method comprises the following steps:

[0018] At step S100, based on the global monitoring feature data of the target tailings pond, the target monitoring area is automatically partitioned, and a layered data acquisition system is constructed. Specifically, when constructing the tailings pond layered data acquisition system, first determine the monitoring content, covering topography, dam deformation, seepage, and weather conditions and other data, select satellite remote sensing, unmanned aerial vehicle, ground sensor network and other equipment for global monitoring feature data acquisition. The collected data needs to be preprocessed to remove noise and errors, then use data clustering algorithms such as K-Means algorithm, analyze the data according to the topography, dam deformation, seepage and other multi-dimensional features, divide the tailings pond into multiple monitoring sub-regions, and form a monitoring sub-region set. Then analyze the feature data of each sub-region in depth, determine the monitoring level according to the dam stability, distance from residential areas or important facilities, historical disaster occurrence and other factors, and then develop a layered acquisition strategy according to the monitoring level, and set the monitoring frequency, method and accuracy differently. High monitoring level area has high monitoring frequency, uses multiple high-precision monitoring methods and requires millimeter-level accuracy, and low monitoring level area is the opposite. Finally, use GPS time device and other devices to time synchronize the monitoring data obtained by the layered acquisition strategy, and use ground control points and coordinate conversion algorithms for spatial calibration to ensure the consistency of the data in time and space, and provide a reliable data foundation for subsequent tailings pond disaster hidden danger analysis.

[0019] In one possible implementation, based on the global monitoring feature data of the target tailings pond, the target monitoring area is automatically partitioned, and a layered data acquisition system is constructed, step S100 further includes step S110, collecting the global monitoring feature data of the target tailings pond. Specifically, collecting global monitoring feature data of the target tailings pond is the key beginning of intelligent identification of tailings pond disaster hidden dangers. First, determine the monitoring content, covering topography, dam structure, seepage and weather data. Satellite remote sensing is used to obtain macro topographic information of the tailings pond, total station, GPS monitoring station and other devices are used to monitor dam displacement and cracks, seepage pressure, flow and water quality are monitored by using seepage pressure gauge, and rainfall, wind speed and other weather data are collected by using weather station. Then select appropriate monitoring equipment, satellite remote sensing is used for periodic observation of a large area, ground sensors such as displacement, seepage pressure, rain gauge and other devices are responsible for accurate measurement, and unmanned aerial vehicle realizes flexible high-resolution low-altitude monitoring. Then develop a monitoring plan, determine the monitoring frequency according to the risk level of the tailings pond, measure the displacement of the high-risk dam body every day and the seepage every hour, and appropriately reduce the frequency for the low-risk pond, at the same time, establish a perfect data recording and storage system and regularly backup. Finally, do a good job in data quality control, calibrate the equipment before monitoring, regularly maintain and check, audit the data in real time, set threshold alarm and perform consistency check, so as to ensure that comprehensive and accurate data is collected, and lay a solid foundation for subsequent tailings pond disaster hidden danger analysis.

[0020] Step S120, analyze the global monitoring feature data, and automatically partition the target tailings pond by a data clustering algorithm to obtain a monitoring sub-region set. Specifically, after completing the collection of the global monitoring feature data of the target tailings pond, the data needs to be analyzed and automatically partitioned with the aid of a data clustering algorithm. First, data preprocessing is performed to clean up abnormal values generated by sensors and to process noise, radiation error, etc. of satellite remote sensing images, and then different dimensions and numerical range data are standardized by methods such as minimum-maximum normalization or Z-score normalization. Then, according to the data characteristics of the tailings pond, a suitable clustering algorithm is selected. If the data is spherical cluster and the number of partitions can be estimated, the K-Means algorithm is more suitable, which is simple to calculate and efficient. If the data distribution is complex and sensitive to noise, the DBSCAN algorithm is more optimal, which can discover clusters of arbitrary shape and identify noise points. Taking the K-Means algorithm as an example, the number of clusters K is determined first, which is selected by multiple tests combined with the "elbow rule" figure, and the maximum number of iterations is set. Then, K cluster centers are randomly initialized, and the data point assignment and cluster center update steps are iterated until the conditions are met. Finally, the clustering results are evaluated by indicators such as the silhouette coefficient and the Calinski-Harabasz index. If it is not ideal, the initial cluster center can be reselected or the clustering algorithm can be replaced, so as to obtain a high-quality monitoring sub-region set, which provides strong support for subsequent work.

[0021] Step S130, based on the feature data of the monitoring sub-region set, different hierarchical acquisition strategies are taken. Specifically, after obtaining the monitoring sub-region set, a hierarchical acquisition strategy needs to be developed according to the feature data of each sub-region, in order to improve monitoring efficiency, reduce cost and ensure monitoring accuracy. First, evaluate the characteristics of each region, evaluate the risk level according to topography, dam stability, seepage conditions and surrounding environment, analyze the sensitivity of different disaster types, and comprehensively consider the risk level and disaster sensitivity, sort out the actual monitoring needs, and clarify the monitoring focus of each region. Then, according to the evaluation results, divide the key and secondary regions, set the key parts of the dam, the water accumulation area and the dam breach risk area as the key regions, and the periphery and stable area as the secondary regions. For key areas, high-frequency real-time monitoring is used, high-precision ground displacement sensors are used to collect key part displacement data every 5 minutes, and unmanned aerial vehicles are used to patrol water accumulation areas every 1 hour; the secondary region is sampled regularly, satellite remote sensing monitoring is performed once a day, and seepage conditions are detected once a week. Key areas require high-precision monitoring, with displacement sensor accuracy reaching sub-millimeter level, and seepage monitoring accuracy reaching two decimal places; the accuracy requirement of the secondary region is lower, with displacement accuracy in centimeter level and seepage accuracy in integer level. At the same time, combined with multiple monitoring methods, aerial remote sensing is used for large-area rapid evaluation, ground sensors provide long-term high-precision local data, and unmanned aerial vehicles are used for local high-precision monitoring. According to the needs of different regions, the acquisition frequency and accuracy are adjusted according to the data requirements, the monitoring frequency and accuracy of some regions are increased during the flood period, and they are appropriately reduced during the stable period. In addition, a real-time data feedback mechanism is established, and once the data fluctuates abnormally, it is fed back in time, and the acquisition strategy is dynamically adjusted according to the actual situation, such as upgrading the secondary region with unstable factors to the key region and strengthening the monitoring, so as to realize the comprehensive, efficient and accurate monitoring of the tailings pond.

[0022] Step S140, the monitoring data obtained by the hierarchical acquisition strategy is time-synchronized and space-calibrated. Specifically, time-synchronization and space-calibration of the monitoring data obtained by the hierarchical acquisition strategy is a key link to ensure the accuracy and reliability of the analysis of the hidden dangers of the tailings dam disaster. In terms of time synchronization, high-precision GPS time service is selected as the time reference to provide a unified time reference for satellite remote sensing equipment, ground sensor network, unmanned aerial vehicle and other monitoring equipment. The satellite remote sensing equipment compares and calibrates the time stamp with the GPS time service in real time when collecting data; the ground sensor is connected with the GPS time service device through wired (such as RS-485 bus) or wireless (Bluetooth, Wi-Fi, etc.) to realize time synchronization; the unmanned aerial vehicle communicates with the ground GPS time service base station before taking off to set the accurate time. At the same time, the time synchronization accuracy is verified regularly, if the device time deviation is more than 1 millisecond, it will be recalibrated, and the log record operation and deviation condition will be established. In terms of space calibration, a unified geographic coordinate system is first determined, such as WGS84 coordinate system or local plane rectangular coordinate system, satellite remote sensing image uses ground control points to convert pixel coordinates into geographic coordinates; the ground sensor is installed with total station, GPS receiver and other precise measurement of three-dimensional coordinates, and the installation direction and angle are recorded; the unmanned aerial vehicle obtains position and attitude information by GPS module and IMU during flight, and the data is processed for orthographic correction and calibrated by ground control points. After calibration, different monitoring data is fused in the unified coordinate system, the spatial consistency is checked, the same spatial position data is compared, if the difference is beyond the range, further calibration is checked, so as to ensure the accuracy and consistency of the monitoring data in time and space, and lay a solid data foundation for the analysis of the hidden dangers of the tailings dam disaster.

[0023] In a possible implementation, different hierarchical acquisition strategies are adopted based on the feature data of the set of monitoring sub-regions, and step S130 further includes step S131 of analyzing the feature data of a target monitoring sub-region in the set of monitoring sub-regions to determine a monitoring level of the target monitoring sub-region. Specifically, in tailings pond monitoring, analyzing the feature data of a monitoring sub-region and determining a monitoring level is a key to rationally allocate monitoring resources and carry out targeted monitoring. First, data collection and arrangement are performed to clearly define the collection range, covering data such as topography (slope, slope direction, elevation, etc.), dam structure (material type, thickness, stress distribution, crack, etc.), seepage (flow rate, flow, water quality), and surrounding environment (distance from residential areas and important infrastructure), etc. The data are collected through multiple channels such as satellite remote sensing, sensors, non-destructive testing technology, and field survey, and are stored in a relational or non-relational database according to categories and time sequence. Then, an evaluation index system is established, and evaluation indexes such as dam crack length, seepage flow rate change rate, and distance from residential areas that can reflect the risk level are determined, and the analytic hierarchy process or other methods are used to determine the index weight, for example, the dam stability index weight can be set to 0.5. Then, evaluation calculation and level determination are performed. First, the index data of different dimensions and value ranges are standardized, such as minimum-maximum standardization, then the weighted summation method is used to calculate the comprehensive evaluation value, and finally, the monitoring level is determined according to the pre-set level division standard, such as the comprehensive evaluation value ≥ 0.8 for high level, 0.5-0.8 for medium level, and less than 0.5 for low level, to provide a basis for subsequent monitoring strategy development.

[0024] Step S132, based on the monitoring level of the target monitoring sub-region, determine the hierarchical acquisition strategy of the target monitoring sub-region. Specifically, after determining the monitoring level of the target monitoring sub-region, the hierarchical acquisition strategy needs to be determined to improve the monitoring efficiency and rationally use the resources. In terms of monitoring frequency, the high-level monitoring area has high risk, and the displacement monitoring of the dam body is performed every 10 minutes, the seepage monitoring is performed every 30 minutes, and the unmanned aerial vehicle is patrolled every 2 hours; the medium-level monitoring area performs the dam body displacement monitoring once a day and the seepage monitoring in the morning and evening once, and the satellite remote sensing is performed once a week; the low-level monitoring area performs the dam body appearance inspection once every 3 days, the seepage monitoring once a week, and the satellite remote sensing once every two weeks. In terms of monitoring mode, the high-level area adopts the ground distributed optical fiber sensor, high-precision displacement meter and inclination sensor, combined with the unmanned aerial vehicle carrying the thermal imager, laser radar and other equipment for high-frequency patrol; the medium-level area mainly adopts the ground conventional sensor and regular high-resolution satellite remote sensing monitoring; the low-level area mainly relies on the monthly medium-resolution satellite remote sensing monitoring, and occasionally uses the ground simple equipment to verify the anomaly. In terms of monitoring precision, the displacement of the high-level area reaches sub-millimeter level, and the seepage is accurate to three decimal places; the displacement of the medium-level area is millimeter level, and the seepage is accurate to two decimal places; the displacement of the low-level area is centimeter level, and the seepage is accurate to one decimal place, so as to realize the accurate and efficient monitoring of different risk sub-regions and protect the safety of the tailings pond.

[0025] Step S133, the hierarchical collection strategy setting includes: monitoring frequency, monitoring method, monitoring accuracy setting. Specifically, to achieve efficient and accurate monitoring of different risk areas of the tailings pond and reasonably allocate monitoring resources, the hierarchical collection strategy needs to be determined according to the monitoring level of the target monitoring sub-area. For high-level monitoring areas, due to high risk, high-precision MEMS displacement sensors are selected for displacement monitoring, data is collected every 5 minutes; electromagnetic flowmeter and pressure sensor are used for seepage monitoring, data is recorded every 15 minutes; unmanned aerial vehicle equipped with multispectral camera and thermal infrared imager is arranged for inspection 4 times a day. Ground distributed optical fiber sensing technology and high-precision GNSS receiver are used, combined with small aircraft monitoring equipped with SAR, the displacement monitoring accuracy reaches 0.1 millimeter level, the seepage monitoring accuracy requires that the flow measurement error is within ±0.5%, and the pressure measurement error is within ±0.1%. The risk of medium-level monitoring areas is moderate, displacement monitoring of dam body is performed 2 times a day, seepage monitoring is performed 4 times a day, satellite remote sensing monitoring is performed 2 times a week, and unmanned aerial vehicle low-altitude patrol is performed 1 time a month. Ground routine monitoring and satellite remote sensing monitoring are combined, the displacement monitoring accuracy is 1 millimeter level, the seepage monitoring flow error is within ±1%, and the pressure error is within ±0.5%. The risk of low-level monitoring areas is low, dam body appearance inspection is performed 1 time every 3 days, seepage monitoring is performed 2 times a week, and satellite remote sensing monitoring is performed 1 time a month. Mainly rely on satellite remote sensing periodic monitoring and occasional ground simple detection, the displacement monitoring accuracy is 1 centimeter level, the seepage monitoring flow error is within ±2%, and the pressure error is within ±1%, so as to ensure the safe operation of the tailings pond.

[0026] In a possible implementation, the hierarchical acquisition strategy setting includes: setting of monitoring frequency, monitoring mode and monitoring accuracy. Step S133 further includes step S1331. The monitoring mode includes: control remote sensing, ground sensor and unmanned aerial vehicle monitoring. Specifically, tailing pond monitoring needs to comprehensively use control remote sensing, ground sensor and unmanned aerial vehicle monitoring to achieve omnibearing and multi-level monitoring. In control remote sensing monitoring, a satellite is selected according to requirements, for example, WorldView with high resolution is used for topographic feature detail monitoring, Landsat with thermal infrared is used for analyzing thermal anomaly and seepage, imaging time and orbit parameters are set to obtain images, the images are preprocessed through radiation correction and geometric correction, and then image classification, change detection algorithm and GIS are used for interpretation and analysis. In ground sensor monitoring, a sensor is selected according to monitoring parameters, for example, a fiber grating displacement sensor and a vibrating string seepage pressure gauge, and then the sensor is reasonably arranged at key parts of a dam body, data is collected at a set frequency, the data is transmitted to a data center through wired or wireless transmission, the data is filtered and denoised, a warning threshold is set, and warning is performed when the threshold is exceeded. In unmanned aerial vehicle monitoring, a small multi-rotor or fixed-wing unmanned aerial vehicle is selected according to monitoring range and accuracy, a high-definition camera, a thermal infrared camera, a laser radar and other equipment are carried, a flight route is planned and parameters are set before flight, data is collected during flight, and then the data is transmitted to a ground processing station, images are spliced, anomalies are identified, thermal infrared data is analyzed to locate leakage points, laser radar data is processed to generate a three-dimensional model, and a comprehensive and efficient tailing pond monitoring system is constructed.

[0027] In a possible implementation, the global monitoring feature data is analyzed, the target tailings pond is automatically partitioned by a data clustering algorithm, and a monitoring sub-region set is obtained, and step S120 further includes step S121 of matching the physical and mechanical feature data with the disaster types of the disaster classification data set. Specifically, to accurately identify the potential disaster risk of the tailings pond, it is crucial to match the physical and mechanical feature data with the disaster types of the disaster classification data set. First, data collection and arrangement are performed, displacement sensors, stress sensors, osmotic pressure meters and other equipment, and geological radar non-destructive testing technology are used to collect physical and mechanical feature data such as dam displacement, stress, pore water pressure and internal structure, and historical tailings pond disaster cases are arranged, key features are extracted, a disaster classification data set is formed, and a database is established. Then, feature extraction and quantification are performed, key features such as displacement rate, stress value and seepage velocity are extracted from the physical and mechanical feature data and quantified, and typical features of disaster types such as the displacement mutation degree of dam landslide and the overall deformation degree of dam collapse are also quantified. Subsequently, a cosine similarity algorithm, which can handle high-dimensional data and is not sensitive to scale changes, is selected as the matching algorithm, the quantified two types of data are substituted into the calculation, and the disaster type with the highest similarity is found as the preliminary matching result. Finally, the matching result is verified, the actual monitoring and historical case checking are compared for rationality, if not, the reasons such as data error, feature extraction and algorithm parameters are analyzed and corrected, and the matching is performed again until the result is reasonable, thereby laying a solid foundation for tailings pond disaster risk assessment and early warning.

[0028] Step S122, based on the mechanical parameters of the physical and mechanical characteristics data and the relationship between the disaster critical condition, determine the disaster type. Specifically, to determine the tailings dam disaster type, according to the relationship between the mechanical parameters of the physical and mechanical characteristics data and the disaster critical condition, it is completed through the following steps. First, data collection and arrangement, use strain gauge, pressure sensor, level gauge, total station and other equipment to collect dam stress, strain, displacement and seepage data, at the same time, widely collect domestic and foreign tailings dam historical disaster cases, analyze the changes of physical and mechanical parameters before and after the disaster. Then build the mechanical inversion model, according to the engineering structure and mechanical principle, select the finite element back analysis method, discrete dam into unit, establish the mechanical equilibrium equation, combine with the boundary and initial condition modeling, get the material parameters through indoor and field test, substitute into the model and compare the actual monitoring data to calibrate the parameters. Then establish the trigger mechanism model, deeply analyze the key mechanical factors of various disasters, such as the shear stress of landslide surface exceeds the shear strength, the collapse of dam is that the compressive stress exceeds the compressive strength, based on the mechanical inversion model and historical data to determine the critical condition, such as the shear stress to shear strength ratio reaches 1.2 when landslide, the compressive stress reaches 80% of the compressive strength when collapse, according to this to build the model, verify and modify the optimization with historical data. Finally, real-time monitoring and determining the disaster type, use wireless sensor network to collect data and transmit to data processing center, input the trigger mechanism model, if the mechanical parameters meet the critical condition, the system judges the disaster type and gives warning, such as shear stress to shear strength ratio reaches 1.2, warning landslide, provide strong support for disaster prevention and emergency treatment.

[0029] Step S123, through data analysis, calculate the disaster grade of the disaster type. Specifically, to calculate the disaster grade of the disaster type, it is necessary to analyze the data strictly, according to the key indexes such as dam deformation rate and pore water pressure change. First, collect data from multiple sources, install displacement and pore water pressure sensor at key parts of the dam, collect data regularly, collect weather data at the same time, then store and label detailed information according to time and type. Then determine the key indicators and threshold, select the dam deformation rate and pore water pressure change as the key indicators, determine the threshold of each grade by analyzing historical data, numerical simulation and expert experience, using statistical method. Then analyze the data and calculate, preprocess the data in real time, remove outliers, calculate the dam deformation rate and pore water pressure change rate, compare with the threshold to determine the disaster grade. Finally, verify the results regularly, compare the actual situation and the results of other monitoring methods, update the threshold in time according to the verification and new data, so as to ensure the scientific and accurate calculation of disaster grade, provide support for tailings dam safety management and disaster warning.

[0030] Step S124, collect real-time physical and mechanical characteristic data, and dynamically adjust the disaster level of the disaster type. Specifically, to real-time control the safety condition of the tailings pond, a real-time monitoring system needs to be built to dynamically adjust the disaster level. First, build a real-time monitoring system, select high-precision sensors, and arrange displacement, pore water pressure and stress sensors at key parts of the dam body such as dam shoulder, dam slope and dam bottom, build a stable communication network mainly using LoRa, NB-IoT and other wireless transmission technologies, and optical fiber communication as backup, and encrypt data during transmission. Then, collect and preprocess data in real time, set the collection frequency according to the regional risk and actual situation, collect key data every 5 minutes in high-risk areas, and collect data every 15 minutes in stable areas, remove noise using Kalman filtering algorithm, and set threshold to process abnormal values. Then, carry out dynamic assessment and adjustment of disaster level, build an assessment model based on real-time data, input displacement, pore water pressure, stress and other parameters to calculate the risk degree, and compare the assessment results with the current disaster level in real time, and adjust in time if the threshold is exceeded, such as a certain area due to sudden change of displacement rate and pore water pressure, from intermediate to high level and warning. Finally, record the results of each assessment and adjustment in detail, store them in a special database, and feed back the results to the relevant departments, optimize the monitoring system, assessment model and disaster level division standard according to the feedback, and comprehensively ensure the safe operation of the tailings pond.

[0031] In a possible implementation, matching the physical mechanical characteristic data with the disaster type of the disaster classification data set, step S121 further includes step S1211, identifying a potential disaster type by the target tailing pond consolidation data and mechanical characteristic data, and determining the disaster characteristics of the potential disaster type. Specifically, to ensure the safety of the tailing pond, it is very crucial to identify the potential disaster type and determine its characteristics. First, data collection and arrangement are performed, and samples are collected at different depths and positions of the tailing pond by professional drilling equipment, and are sent to a laboratory to measure the compression coefficient and other consolidation data by a high-pressure consolidation instrument; resistance strain gauge type stress sensors, fiber bragg grating strain sensors, total station instruments, and prisms are installed at key stress positions of the dam body, mechanical characteristic data are collected by an automatic collection system every 15 minutes, and a database is established by using MySQL to store the data in categories. Then, a numerical model is established, a three-dimensional model is established by using FLAC3D software according to the actual terrain, physical and mechanical parameters of the tailings are accurately input, mechanical responses under different working conditions are simulated, common disaster types are combined, and potential disasters are judged according to the simulation results, for example, if the shear stress exceeds the shear strength and the displacement has a sliding trend, it may be a dam body landslide, if the strain exceeds the ultimate strain, the stress is abnormally concentrated, or the dam body collapses, it may be a tailing pond leakage. Finally, the disaster characteristics are determined, when the dam body slides, the displacement increases, the direction is consistent with the sliding surface, the shear stress increases, and the surface cracks widen and lengthen; when the dam body collapses, the settlement suddenly increases, the structure deforms, the stress is abnormally concentrated, and the strain exceeds the limit; when the tailing pond leaks, the seepage data are abnormal, the surrounding water level changes, and the surrounding soil and water may be polluted. This set of processes can provide a strong basis for disaster early warning and prevention.

[0032] Step S1212, based on the monitoring sub-region set and disaster type, classified storage is performed. Specifically, in order to realize efficient management of tailings dam monitoring data, it is necessary to perform classified storage based on the monitoring sub-region set and disaster type. First, according to the topography of the tailings dam, dam structure and key monitoring requirements, regions such as dam shoulder, dam slope (subdivided according to slope), dam bottom, and tailings accumulation sub-regions divided according to tailings accumulation time and height are divided, each region is labeled with geographical coordinates and a unique number, and disaster types such as dam landslide, collapse, leakage and debris flow are sorted out, and their key features and triggering factors are clarified. Then, a storage architecture is built, a relational database Oracle is selected to handle a large amount of data, and a distributed file system Ceph is used to store unstructured data. In Oracle, a "monitoring sub-region table" is designed to record sub-region basic information, a "disaster type table" is designed to record disaster related content, and a "monitoring data storage table" is designed to associate the two and monitoring data. When data is stored, various sensors collect data in real time and perform preliminary denoising and filtering, and the data is stored in the corresponding table according to the sub-region and disaster type number. Unstructured data is classified and stored in Ceph according to the sub-region, and key information is recorded in the database. Finally, a full weekly and daily incremental backup strategy is developed, and the backup is stored in a remote location. When new data, region range adjustment or new disaster type is found, the database is updated in time to ensure data consistency and timeliness, and to lay a solid data foundation for tailings dam disaster analysis and early warning.

[0033] Step S1213, analyzing the disaster characteristics and the disaster type, and constructing the disaster classification data set. Specifically, constructing the disaster classification data set is of great significance to tailings dam disaster prevention and control, and the process mainly includes four steps. First, data collection and sorting, from real-time monitoring equipment and historical case materials, a wide range of disaster characteristic data is collected, such as displacement, stress and crack data of dam landslide, seepage and pollution index data of tailings dam leakage, and the definition, principle and inducing factors of each disaster type are sorted out. Then, data analysis and feature extraction, using clustering analysis and other data mining and statistical methods, similar feature clusters are analyzed and found out, and then key features such as dam landslide displacement rate threshold and tailings dam leakage seepage velocity mutation are extracted. Then, the disaster classification data set is constructed, the relational database structure is adopted, the "disaster type table", "disaster feature table" and "association table" are designed, the data after sorting and analysis is filled in according to the structure, and an organic connection is established. Finally, verification and update, the accuracy of the data set is verified by comparing historical and real-time data, and it is updated regularly according to new monitoring data and research results to ensure its reliability at all times and provide strong support for tailings dam disaster prevention and control.

[0034] Step S200, a tailings physical and mechanical database is established, and target tailings pond consolidation data and mechanical characteristic data fed back by the hierarchical data collection system are combined to construct a tailings pond mechanical parameter inversion model. Specifically, the establishment of the tailings physical and mechanical database and the construction of the tailings pond mechanical parameter inversion model can provide scientific support for tailings pond safety management. First, a comprehensive data collection plan covering historical data and field sampling is developed, and the density, particle size distribution, shear strength, and other physical and mechanical parameters of different tailings are collected. The MySQL is selected to design the "tailings basic information table", "physical parameter table" and "mechanical parameter table", and the data are entered and strictly controlled in quality through the primary key and foreign key association to complete the database construction. Then, the hierarchical data collection system is deployed vertically according to the depth and accumulation characteristics of the target tailings pond and horizontally according to the grid spacing, various sensors are installed at each monitoring point to collect consolidation and mechanical characteristic data, which are transmitted to the data processing center through wired or wireless transmission, classified and sorted, and integrated with the database data. Finally, the finite element inversion model is selected according to the actual and data characteristics, the database data are used to initialize the parameters, the boundary and initial conditions are combined, the model is run and the parameters are optimized using genetic algorithm and the like, so that the results fit the monitoring data, and the model is verified with new monitoring data, which is applied to safety assessment and disaster warning.

[0035] Step S300, the physical and mechanical characteristic data of the target tailings pond dam body are identified through the tailings pond mechanical parameter inversion model, and the disaster type is determined and the disaster grade is divided. Specifically, to ensure the safety of the tailings pond, it is very important to identify the physical and mechanical characteristic data of the dam body and determine the disaster related conditions by means of the tailings pond mechanical parameter inversion model. First, the displacement, stress, pore water pressure and other sensors are used to collect data in all directions of the dam body in real time, which are sorted according to the monitoring point position and time and input into the inversion model. For example, the finite element inversion model calculates the mechanical state of the dam body according to the principle of mechanical equilibrium and deformation compatibility combined with the input data. Then, the stress, strain, displacement and other key physical and mechanical characteristic data are extracted from the calculation results, and whether they are beyond the normal range is judged by comparing with the history, theory and safety standard to evaluate the stability of the dam body. Then, according to the relationship between the common disaster types of the tailings pond and the physical and mechanical characteristic data and the disaster, a judgment criterion is established, and the data analyzed by the inversion model are judged by considering multiple factors to determine the type of the possible disaster. Finally, the grade division standard is developed according to the severity and consequences of the disaster, such as the dam body landslide, seepage and other disasters are divided into low, medium and high grades according to the displacement rate, seepage flow and other indicators, and the grade of the determined disaster is evaluated according to the standard and the development trend of the disaster to provide strong support for the safety management of the tailings pond.

[0036] Step S400, based on the tailings pond mechanical parameter inversion model, deduce the disaster evolution process, analyze the disaster chain and build the disaster chain structure tree. Specifically, the tailings pond mechanical parameter inversion model is used to deduce the disaster, analyze the chain and build the structure tree, which is of great significance to the tailings pond disaster prevention and mitigation. In implementation, first, according to the actual engineering data of the tailings pond, through laboratory test, field test and reference to historical data, the model basic parameters are determined, such as the elastic modulus of the tailings material, the Poisson's ratio, the internal friction angle, and the boundary conditions of the dam body and the foundation contact, the surface load, etc., and the real-time collected physical and mechanical characteristic data of the dam body displacement, stress, pore water pressure, etc. are input. Then, the model is used to simulate the mechanical response of the tailings pond under normal operation, heavy rain, earthquake and other working conditions, considering the influence of rainwater infiltration and seismic wave input on the stability of the dam body stress, in the process, closely tracking the changes of key indicators such as dam body displacement, stress concentration, pore water pressure, material strength, etc. Then, according to the index change, the disaster causal relationship is sorted out, such as the increase of pore water pressure causing the shear strength of soil body to decrease and triggering the landslide of the dam body, the key nodes such as the stress concentration of the dam body exceeding the material damage strength, the seepage breaking through the key anti-seepage layer are determined. Finally, taking the disaster type as the root node, the key nodes and events are set as branches according to the cause and development order, the disaster chain structure tree framework is built, and the specific data, event description, analysis results in the disaster evolution are filled into each node, making the structure tree become a visual tool for directly presenting the disaster development and internal relations, providing a scientific basis for disaster prevention.

[0037] In one possible implementation, based on the tailings pond mechanical parameter inversion model, the disaster evolution process is deduced, the disaster chain is analyzed and the disaster chain structure tree is built, step S400 further includes step S410, the interaction relationship of the disaster type is identified through the tailings pond mechanical parameter inversion model, and the disaster chain is built. Specifically, the tailings pond mechanical parameter inversion model is crucial in identifying the interaction relationship of the disaster type and building the disaster chain. First, data is deeply mined from the model operation results, in addition to conventional data such as dam body stress and strain, meteorological, geological and other operating environment data are also collected, and are arranged in time sequence according to time sequence, which is convenient for observing the change trend. Then, for common disaster types such as dam body landslide, collapse, seepage, piping, etc., a disaster characteristic index library is established respectively, statistical methods and professional knowledge are used to correlate and analyze the data, compare the index changes before and after the disaster occurs, deduce the cause and effect relationship, and judge the influence of one disaster on other disasters. Finally, according to the interaction relationship, the initial disaster and triggering condition causing a series of disasters are determined, taking the initial disaster as the starting point, connecting the subsequent disasters in turn according to the cause and effect order, building the disaster chain, and at the same time, analyzing and integrating multiple different disaster chains, comparing similarities and differences, finding out common triggering factors and key links, and fully mastering the disaster development law, providing a basis for tailings pond disaster prevention and control.

[0038] Step S420, the disaster type is taken as a tree node, and the disaster chain is taken as an edge. Specifically, taking the disaster type as a tree node and the disaster chain as an edge is the key to constructing the disaster chain structure tree, and can intuitively present the disaster correlation. Specifically, first, the possible disasters of the tailings dam, such as dam landslide, collapse, leakage, piping, and debris flow, are comprehensively combed, historical cases, physical and mechanical principles, and triggering conditions are collected, and then each disaster is defined as a tree node, and attributes such as disaster name, number, detailed description, damage degree, and probability estimation are given. Then, the disaster chain constructed by the tailings dam mechanical parameter inversion model and the disaster interaction relationship analysis is reviewed, and the sequence and causal relationship of each disaster are determined, which is converted into an edge in the tree structure, the edge connects two disaster nodes, the direction reflects the development sequence, and attributes such as causal relationship description and impact strength evaluation are given. Finally, the nodes and edges are integrated, and the disaster chain structure tree is preliminarily constructed, and the node connection, causal relationship, and attribute information are comprehensively checked for whether they are reasonable and accurate, and problems are corrected in time to ensure that the tree structure truly reflects the internal relationship and development context of the disaster, and lays a foundation for subsequent analysis and application based on the tree structure algorithm.

[0039] Step S430, based on the tree structure algorithm, a disaster chain structure tree is constructed. Specifically, in order to clearly present the development context of the tailings dam disaster and provide a visual basis for disaster prevention and control, a disaster chain structure tree can be constructed based on the tree structure algorithm. Taking depth-first search (DFS) and breadth-first search (BFS) algorithms as examples, first, data preparation is done, the starting node is determined according to the previous analysis of the tailings dam disaster, such as setting “heavy rainfall” which often triggers a series of disasters as the root node, and the interaction relationship between each disaster type is combed to determine the child nodes of each node, and an adjacency list or adjacency matrix is arranged. Based on DFS, starting from the root node, a path is explored as deeply as possible, like from “heavy rainfall” to “dam pore water pressure increase”, then to “dam landslide” and “tailings dam leakage”, and backtracking is explored until all nodes are visited, and the tree structure is constructed according to the visiting order. Based on BFS, starting from the root node, the directly connected nodes of the same level are visited first, and then the next level is visited in turn, such as first visiting the direct child nodes of “heavy rainfall”, and then visiting the child nodes of the child nodes, and the tree is constructed according to the level order. Finally, the constructed tree is optimized to remove redundant nodes and unreasonable connections, and verified by actual disaster cases and monitoring data to check whether it is consistent with the actual situation, and corrected in time if there is deviation, so that the structure tree is more reliable and practical, and helps the tailings dam disaster management.

[0040] Step S500: Based on the disaster chain structure tree, determine the disaster occurrence rate of the target tailings dam. If the disaster occurrence rate exceeds a preset disaster warning threshold, generate a risk warning report. Specifically, using the disaster chain structure tree to determine the tailings dam disaster occurrence rate and generate a warning report is an important means of preventing tailings dam disasters. In implementation, first, the disaster chain structure tree is analyzed to clarify the disaster characteristics of each node, such as the geological conditions, mechanical parameters, and triggering factors of dam landslides. The disaster propagation path is then traced, such as the process where heavy rainfall causes an increase in pore water pressure in the dam, leading to dam landslides and tailings dam leakage. Next, a disaster occurrence rate assessment model is constructed. Based on the analysis of the structure tree, key assessment indicators such as rainfall and dam stress are determined, and weights are reasonably allocated to each indicator. A mathematical model is established using regression analysis and neural network algorithms. Then, data is collected in real time through the tailings dam monitoring system. After preprocessing such as cleaning, denoising, and normalization, the data is substituted into the model to calculate the disaster occurrence rate. Finally, the calculation results are compared with preset thresholds set based on safety standards, historical data, and expert experience. If the incidence rate of a certain disaster exceeds the threshold, a risk warning report is immediately generated, covering the type of disaster exceeding the threshold, the incidence rate, possible consequences, and emergency response recommendations. This report is then promptly sent to relevant departments and responsible persons to assist in the safety management of tailings ponds.

[0041] This application's embodiments employ a tailings dam-wide monitoring system that automatically partitions and constructs a hierarchical data acquisition system; establishes a physical and mechanical database, and constructs a mechanical parameter inversion model by combining consolidation and mechanical data; identifies the physical and mechanical characteristics of the dam body through this model, determines the disaster type, and classifies its level; extrapolates the disaster evolution process, analyzes the disaster chain, and constructs a structure tree; and judges the disaster occurrence rate based on the disaster chain. If the rate exceeds the warning threshold, a risk warning report is generated. This achieves the technical effect of improving the accuracy of tailings dam disaster hazard identification through precise identification of tailings dam disaster types, dynamic adjustment of disaster levels, and real-time extrapolation of disaster evolution processes.

[0042] In the above text, refer to Figure 1 A method for intelligent identification of tailings dam disaster hazards according to embodiments of the present invention is described in detail. Next, reference will be made to... Figure 2 This invention describes an intelligent identification platform for tailings dam disaster hazards according to an embodiment of the present invention.

[0043] The intelligent identification platform for tailing pond disaster hidden danger according to the embodiment of the present application is used to solve the technical problem of insufficient precision of the tailing pond disaster hidden danger identification system in the prior art, and through accurate identification of the tailing pond disaster type, dynamic adjustment of the disaster grade and real-time deduction of the disaster evolution process, the technical effect of improving the accuracy of tailing pond disaster hidden danger identification is achieved. The intelligent identification platform for tailing pond disaster hidden danger comprises: a hierarchical data acquisition system construction module 10, a tailing pond mechanical parameter inversion model construction module 20, a disaster type determination module 30, a disaster chain structure tree construction module 40 and a risk early warning report generation module 50.

[0044] The hierarchical data acquisition system construction module 10 is used to automatically partition the target monitoring area based on the global monitoring feature data of the target tailing pond, and construct a hierarchical data acquisition system.

[0045] The tailing pond mechanical parameter inversion model construction module 20 is used to establish a tailing physical mechanics database, combine the consolidation data and mechanical feature data of the target tailing pond fed back by the hierarchical data acquisition system, and construct a tailing pond mechanical parameter inversion model.

[0046] The disaster type determination module 30 is used to identify the physical and mechanical feature data of the target tailing pond dam body through the tailing pond mechanical parameter inversion model, determine the disaster type and perform disaster grade division.

[0047] The disaster chain structure tree construction module 40 is used to deduce the disaster evolution process, analyze the disaster chain and construct a disaster chain structure tree based on the tailing pond mechanical parameter inversion model.

[0048] The risk early warning report generation module 50 is used to judge the disaster occurrence rate of the target tailing pond in combination with the disaster chain structure tree, and generate a risk early warning report if the disaster occurrence rate exceeds a preset disaster early warning threshold.

[0049] Next, the specific configuration of the hierarchical data acquisition system construction module 10 will be described in detail. As described above, the hierarchical data acquisition system construction module 10 further comprises: a target tailing pond data acquisition unit, which is used to acquire the global monitoring feature data of the target tailing pond; a monitoring sub-region set acquisition unit, which is used to analyze the global monitoring feature data, automatically partition the target tailing pond through a data clustering algorithm, and obtain a monitoring sub-region set; a hierarchical acquisition strategy determination unit, which is used to adopt different hierarchical acquisition strategies based on the feature data of the monitoring sub-region set; and a data synchronization calibration unit, which is used to perform time synchronization and space calibration on the monitoring data obtained by the hierarchical acquisition strategy.

[0050] The different hierarchical acquisition strategies are adopted based on feature data of the monitoring sub-region set, and the hierarchical acquisition strategy determination unit further includes: a monitoring level determination sub-unit, configured to analyze feature data of a target monitoring sub-region in the monitoring sub-region set, and determine a monitoring level of the target monitoring sub-region; a hierarchical acquisition strategy determination sub-unit, configured to determine a hierarchical acquisition strategy of the target monitoring sub-region based on the monitoring level of the target monitoring sub-region; and a hierarchical acquisition strategy composition sub-unit, configured to set the hierarchical acquisition strategy to include settings of monitoring frequency, monitoring mode, and monitoring accuracy.

[0051] The hierarchical acquisition strategy settings include settings of monitoring frequency, monitoring mode, and monitoring accuracy, and the hierarchical acquisition strategy composition sub-unit further includes a monitoring mode composition micro-unit, configured to set the monitoring mode to include control remote sensing, ground sensor monitoring, and unmanned aerial vehicle monitoring.

[0052] The global monitoring feature data is analyzed, the target tailing pond is automatically partitioned through a data clustering algorithm, and a monitoring sub-region set is obtained, and the monitoring sub-region set acquisition unit further includes: a disaster type matching sub-unit, configured to match the physical and mechanical feature data with disaster types in a disaster classification data set; a disaster type determination sub-unit, configured to determine a disaster type based on a relationship between mechanical parameters of the physical and mechanical feature data and disaster critical conditions; a disaster level calculation sub-unit, configured to calculate a disaster level of the disaster type through data analysis; and a disaster level adjustment sub-unit, configured to collect real-time physical and mechanical feature data, and dynamically adjust the disaster level of the disaster type.

[0053] The physical and mechanical feature data is matched with disaster types in a disaster classification data set, and the disaster type matching sub-unit further includes: a disaster feature determination micro-unit, configured to identify a potential disaster type through the target tailing pond consolidation data and mechanical feature data, and determine disaster features of the potential disaster type; a disaster type classification storage micro-unit, configured to classify and store the monitoring sub-region set and the disaster type; and a disaster classification data set construction micro-unit, configured to analyze the disaster features and the disaster type, and construct the disaster classification data set.

[0054] Below, the specific configuration of the disaster chain structure tree construction module 40 will be described in detail. As described above, based on the tailings pond mechanical parameter inversion model, the disaster evolution process is deduced, the disaster chain is analyzed, and the disaster chain structure tree is constructed. The disaster chain structure tree construction module 40 further comprises: a disaster chain construction unit, which is configured to identify the interaction relationship of the disaster types by the tailings pond mechanical parameter inversion model, and construct the disaster chain; a tree structure determination unit, which is configured to take the disaster types as tree nodes and the disaster chain as edges; and a disaster chain structure tree construction unit, which is configured to construct a disaster chain structure tree based on a tree structure algorithm.

[0055] The intelligent identification platform for tailings pond disaster hidden dangers provided in the embodiments of the present application can execute the intelligent identification method for tailings pond disaster hidden dangers provided in any of the embodiments of the present application, and has the function modules and beneficial effects corresponding to the execution method.

[0056] Although the present application makes various references to certain modules in the platform according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, and each unit and module included is only divided according to the functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for the convenience of mutual differentiation, and do not limit the protection scope of the present application.

[0057] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for intelligent identification of potential hazards in tailings dams, characterized in that, include: Based on the full-area monitoring feature data of the target tailings dam, the target monitoring area is automatically divided into zones to construct a hierarchical data acquisition system; A tailings physical and mechanical database was established, and a tailings dam mechanical parameter inversion model was constructed by combining the target tailings dam consolidation data and mechanical characteristic data fed back by the layered data acquisition system. By using the tailings dam mechanical parameter inversion model, the physical and mechanical characteristics of the target tailings dam body are identified, the disaster type is determined, and the disaster level is classified. Based on the tailings dam mechanical parameter inversion model, the disaster evolution process is deduced, the disaster chain is analyzed, and a disaster chain structure tree is constructed. Based on the disaster chain structure tree, the disaster occurrence rate of the target tailings dam is determined. If the disaster occurrence rate exceeds the preset disaster warning threshold, a risk warning report is generated. The automatic partitioning of the target monitoring area to construct a hierarchical data acquisition system includes: The target tailings dam is subjected to the collection of the full-area monitoring feature data; Analyze the full-domain monitoring feature data, and automatically partition the target tailings pond using a data clustering algorithm to obtain a set of monitoring sub-regions; Based on the characteristic data of the monitored sub-region set, different hierarchical acquisition strategies are adopted; The monitoring data acquired by the hierarchical acquisition strategy is synchronized in time and calibrated in space. Based on the feature data of the monitored sub-region set, different hierarchical acquisition strategies are adopted, including: Analyze the characteristic data of the target monitoring sub-region in the monitoring sub-region to determine the monitoring level of the target monitoring sub-region; Based on the monitoring level of the target monitoring sub-region, determine the hierarchical data acquisition strategy for the target monitoring sub-region; The hierarchical data acquisition strategy settings include: monitoring frequency, monitoring method, and monitoring accuracy settings; The construction of the disaster chain structure tree includes: By using the tailings dam mechanical parameter inversion model, the interaction relationships of the disaster types are identified, and the disaster chain is constructed; The disaster type is used as a tree node, and the disaster chain is used as an edge; A disaster chain structure tree is constructed based on a tree structure algorithm.

2. The intelligent identification method for tailings dam disaster hazards according to claim 1, characterized in that, The monitoring methods include: remote sensing, ground sensors, and drone monitoring.

3. The intelligent identification method for tailings dam disaster hazards according to claim 1, characterized in that, Determine the type of disaster and classify its severity, including: The physical and mechanical feature data are matched with the disaster types in the disaster classification dataset; Based on the relationship between the mechanical parameters and the critical conditions of the disaster according to the aforementioned physical and mechanical characteristic data, the type of disaster is determined; The disaster level of the disaster type is calculated through data analysis; Collect real-time physical and mechanical characteristic data to dynamically adjust the disaster level of the disaster type.

4. The intelligent identification method for tailings dam disaster hazards according to claim 3, characterized in that, The disaster classification dataset includes: By using the consolidation data and mechanical characteristic data of the target tailings dam, potential disaster types are identified, and the disaster characteristics of the potential disaster types are determined. The data is classified and stored based on the monitored sub-region set and disaster type; The disaster characteristics and disaster types are analyzed to construct the disaster classification dataset.

5. An intelligent identification platform for tailings dam disaster hazards, characterized in that, The platform is used to implement the intelligent identification method for tailings dam disaster hazards according to any one of claims 1-4, and the platform includes: A hierarchical data acquisition system construction module is used to automatically partition the target monitoring area based on the full-area monitoring feature data of the target tailings dam and construct a hierarchical data acquisition system. The tailings dam mechanical parameter inversion model construction module is used to establish a tailings physical and mechanical database, and to construct a tailings dam mechanical parameter inversion model by combining the target tailings dam consolidation data and mechanical characteristic data fed back by the hierarchical data acquisition system. The disaster type determination module is used to identify the physical and mechanical characteristic data of the target tailings dam body through the tailings dam mechanical parameter inversion model, determine the disaster type, and classify the disaster level. The disaster chain structure tree construction module is used to deduce the disaster evolution process, analyze the disaster chain, and construct the disaster chain structure tree based on the tailings dam mechanical parameter inversion model. A risk warning report generation module is used to determine the disaster occurrence rate of the target tailings dam by combining the disaster chain structure tree. If the disaster occurrence rate exceeds a preset disaster warning threshold, a risk warning report is generated.

Images