Tailing dam disaster-causing factor identification and safety evaluation method based on multi-source heterogeneous data fusion

Abstract

The invention discloses a mine tailing dam disaster-causing factor identification and safety evaluation method based on multi-source heterogeneous data fusion, and relates to the technical field of mine tailing dam safety monitoring and risk management. Fusion of'air-sky-ground-underground 'multi-source heterogeneous monitoring data comprises the steps that GNSS and InSAR technologies are cooperatively fused, GNSS three-dimensional deformation is converted into InSAR sight-direction one-dimensional deformation, calibration is completed, a high-temporal-spatial-resolution deformation field is constructed, and large-range deformation scanning and precise capture of local hidden dangers are achieved; according to the method, unmanned aerial vehicle oblique photography and laser radar scanning operation are carried out through UAV-LiDAR, a three-dimensional digital model is produced, a deformation area is obtained by using multi-stage DEM difference, and a complete and high-precision deformation area is generated in combination with boundary information of a high-resolution deformation field; data such as pore water pressure and deep displacement are collected through an underground sensor, and disaster-causing factors and mechanisms are qualitatively analyzed; and carrying out safety evaluation by adopting an AHP-EWM combined weighting method, and dividing into four grades of safety, yellow early warning, orange early warning and red early warning according to comprehensive scores.

Description

Technical Field

[0001] This invention relates to a method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion, belonging to the technical field of safety monitoring and risk management. Background Technology

[0002] Tailings dams are characterized by complex structures and dynamic stability changes over time. They are essential infrastructure in mining production but also significant sources of high potential energy hazard. Dam failures can cause major casualties and property damage, and severely impact the surrounding environment. Therefore, effectively fusing multi-source heterogeneous data from real-time tailings dam safety monitoring to provide timely warnings and safety assessments of abnormal events is of significant practical importance for addressing various problems arising during tailings dam construction. Currently, tailings dam safety monitoring mainly relies on various sensors, video surveillance, and manual inspections. While each method has its advantages, data obtained using a single method is insufficient to accurately grasp the tailings dam accumulation situation. Therefore, developing a tailings dam disaster factor identification and safety assessment method based on multi-source heterogeneous data fusion has become an objective requirement for tailings dam safety monitoring. InSAR technology offers wide coverage, relatively low cost, and all-weather imaging capabilities; UAV-LiDAR technology boasts high flexibility and can fly on demand; GNSS monitoring points and sensor technologies offer high precision and enable continuous, real-time monitoring; online video monitoring systems are intuitive and visual, and can be cross-verified with other monitoring technologies. This invention comprehensively applies these five monitoring methods to conduct precise monitoring, featuring comprehensiveness, reliability, and efficiency. It provides methods and ideas for tailings dam safety monitoring and offers technical support for achieving inherently safe and green mines. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a method for identifying and assessing the disaster-causing factors of tailings dams based on multi-source heterogeneous data fusion. This method comprehensively applies InSAR, UAV-LiDAR, GNSS, sensor technology and online video monitoring system. By effectively fusing the data obtained by the above five monitoring methods, it achieves accurate monitoring of tailings dams. It has the characteristics of comprehensiveness, reliability and efficiency, and provides a method and idea for tailings dam safety monitoring, and provides technical support for achieving inherently safe and green mines.

[0004] To address the above problems, the specific technical solution of this invention is as follows: A method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion, comprising the following steps: 1) Convert GNSS three-dimensional deformation into line-of-sight one-dimensional deformation, and calibrate InSAR data to the line-of-sight deformation of a GNSS monitoring point; 2) Select a relatively stable GNSS monitoring point with few missing time series data; interpolate the missing values ​​in the time series of the GNSS monitoring point; perform wavelet denoising on the interpolated monitoring point to obtain a complete and accurate GNSS time series. 3) Select the InSAR pixel closest to the GNSS monitoring point and calculate the difference between the pixel and the denoised GNSS line-of-sight time series for each period. Add this difference to the InSAR data for each period to obtain a high spatiotemporal resolution deformation field that integrates GNSS and InSAR technologies. 4) Under the high spatiotemporal resolution deformation field monitoring technology, the monitoring data of the entire tailings dam, upstream and downstream of the dam body, reservoir area and surrounding areas are continuously acquired to form a line-of-sight cumulative deformation rate map. By color distribution and rate magnitude, the area with the most severe deformation in the reservoir area and the boundary of deformation influence are identified. 5) Employing UAV-LiDAR technology, a high-precision 3D model of the tailings dam is rapidly generated; high-resolution aerial images are acquired through periodic aerial surveys to identify dangerous factors such as cracks, gullies, uplifts, and local landslides on the dam slope and crest, while simultaneously establishing 3D models at different times; the deformation amount and range of the tailings dam and surrounding mountains are calculated; and the deformation zone data from UAV-LiDAR is integrated with the boundary information of the high-resolution deformation field to generate a complete and high-precision deformation "target" area. 6) GNSS has the feature of acquiring real-time displacement data of monitoring points in all weather conditions. In the identified deformation "target" area, according to the "three points form a line" principle, the number of devices can be increased or decreased according to the actual deformation "target" area area, slope and other conditions to monitor deformation data in real time. 7) By transmitting data monitored by various surface sensors such as rain gauges, water level gauges, and water measuring weirs to the online video monitoring system, specific data such as rainfall in the reservoir area, reservoir water level, seepage flow, and dam surface displacement can be obtained in real time; 8) By burying underground sensors such as piezometers, microseismic sensors, pore water pressure gauges, and earth pressure gauges inside the dam body and foundation, the location of the dam body seepage line and the data and patterns of changes in the soil and rock mechanics inside the dam body are monitored in real time. Combined with the deformation data of the "target" area, the disaster-causing factors and mechanisms are qualitatively analyzed. 9) Based on the real-time monitoring data of the tailings dam, the weight of each influencing factor in the tailings dam is calculated by using the combined weighting method of AHP-EWM analytic hierarchy process and entropy weighting method. By assigning values ​​to each influencing factor, the safety risk level of the tailings dam is evaluated in real time based on the comprehensive score.

[0005] In step 1), the deformation Z in the north direction of the GNSS coordinate system n Deformation Z in the east direction e The projection is the deformation Z in the direction of ground distance. j Then, the deformation Z in the distance direction.j and vertical deformation Z u The deformation Z projected onto the line of sight l ; Z j =Z n cos(α-270°) + Z e cos(360°-α)=-Z n sin(α) + Z e cos(α) Z l =Z u cos(θ)-Z j sin(θ) Z l =Z u cos(θ) + Z n sin(α)sin(θ)-Z e cos(α)sin(θ) Where θ is the radar incident angle; α is the radar heading azimuth angle, Z n Z e and Z u These are the deformations in the north, east, and vertical directions, respectively; Z j Deformation in the direction of distance from the ground; By performing differential interferometry processing on SAR images of the monitoring area acquired by InSAR technology at different times, the phase signal dominated by the deformation of the dam body itself is extracted, and the cumulative deformation rate and time series deformation map are generated.

[0006] In step 2), based on the missing duration and data trend characteristics, linear interpolation, spline interpolation, and other methods are selected to ensure the consistency of the interpolation result with the trend of the original sequence and the integrity of the monitoring data; high-frequency noise is removed by threshold processing, and then the denoised monitoring sequence is reconstructed.

[0007] In step 5), UAV-LiDAR technology is used to quickly generate a high-precision digital ground model and a real-world 3D model of the tailings dam, and to identify dangerous factors such as cracks, gullies, uplifts and local landslides on the dam slope and top. The DEM difference method is used to calculate the deformation and deformation range of the tailings dam and the surrounding mountains.

[0008] In step 6), the number of devices is increased or decreased according to the actual deformation "target" area, slope, etc., based on the principle of "three points forming a line" in the identified deformation "target" area, and deformation data is monitored in real time; the data is fused with the InSAR monitoring data to achieve collaborative monitoring that combines point and surface monitoring and complements each other's advantages.

[0009] In step 7), an online video monitoring system is used to verify the information using its intuitive and visual images.

[0010] In step 8), the deformation characteristics of the "target" area are integrated with the data and patterns of changes in soil and rock mechanics inside the dam body to qualitatively analyze the disaster-causing factors and mechanisms, and macroscopically predict the future deformation patterns of the deformation area.

[0011] In step 9), based on real-time monitoring data of the tailings dam, the AHP-EWM combined weighting method is used to conduct a safety assessment of the tailings dam. The subjective weight W1 is obtained using the Analytic Hierarchy Process (AHP), and the objective weight W2 is obtained using the Entropy Weight Method (EWM). The formula W is then used to... i =αW1+(1-α)W2, where W i The combined weights are obtained using the AHP-EWM combined weighting method, where α is the combined weight coefficient, α∈(0,1); based on the comprehensive evaluation value, four safety levels of the tailings dam are obtained, W i ∈[90,100) is safe, W i ∈[80,90) is the yellow level, W i ∈[60,80) is orange level, W i ∈(0,60) is the red level, and finally, corresponding preventive measures are taken according to the safety level of the tailings dam.

[0012] The advantages of this invention are: 1. InSAR technology has a wide coverage area, can monitor tailings dams in multiple areas simultaneously, has good long-term continuity, relatively low data cost, and is not limited by ground conditions; UAV-LiDAR technology is flexible, efficient, and has high resolution; GNSS and sensor measurement accuracy is high, reaching the millimeter level, and can achieve real-time, continuous monitoring with high reliability; online video monitoring systems can provide visual information about the dam body. 2. InSAR technology can monitor deformations from millimeters to centimeters; UAV-LiDAR technology can monitor deformations from centimeters to decimeters; GNSS can provide relative positioning accuracy at the centimeter or even millimeter level, and can continuously acquire three-dimensional displacement data of monitoring points in real time. The combination of InSAR, GNSS, and UAV-LiDAR can obtain a complete and high-precision deformation "target" area, and the above monitoring results can be intuitively verified through an online video monitoring system. 3. By integrating monitoring data obtained from multiple monitoring methods, we can gain a comprehensive, in-depth, and real-time understanding of the tailings dam dynamics, making up for the shortcomings and deficiencies of single monitoring data and providing more accurate comprehensive monitoring results for tailings dam safety monitoring. 4. By conducting safety assessments of the real-time monitoring of tailings dam dynamics, effective support can be provided for timely implementation of preventative measures in the dam area. Attached Figure Description

[0013] Figure 1The flowchart shows the method for identifying disaster-causing factors and evaluating the safety of tailings dams based on multi-source heterogeneous data fusion.

[0014] Figure 2a A three-dimensional view of SAR imaging.

[0015] Figure 2b This is a vertical view of SAR imaging.

[0016] Figure 2c This is a side view of a SAR image. Detailed Implementation

[0017] The present invention will now be described in detail with reference to the accompanying drawings.

[0018] like Figure 1 The flowchart shown illustrates a method for identifying disaster-causing factors and conducting safety assessments of tailings dams based on multi-source heterogeneous data fusion, comprising the following steps: Step 1: Monitor dam area displacement using InSAR (Inductive Aperture Radar Interferometry) technology. Differential interferometry is performed on SAR images of the monitoring area acquired by InSAR technology at different times to extract the phase signal dominated by the dam's own deformation, generating cumulative deformation rate and time-series deformation maps. Step 2: Convert the GNSS (Global Navigation Satellite System) three-dimensional deformation into a one-dimensional deformation along the line of sight, and calibrate the InSAR data to the line-of-sight deformation of a specific GNSS monitoring point; Deformation Z in the north direction in the GNSS coordinate system n Deformation Z in the east direction e The projection is the deformation Z in the direction of ground distance. j Then, the deformation Z in the distance direction. j and vertical deformation Z u The deformation Z projected onto the line of sight l ; in accordance with: Z j =Z n cos(α-270°) + Z e cos(360°-α)=-Z n sin(α) + Z e cos(α) Z l =Z u cos(θ)-Z j sin(θ) Z l =Z u cos(θ) + Z n sin(α)sin(θ)-Z e cos(α)sin(θ) The one-dimensional deformation along the line of sight at a given GNSS monitoring point is calculated, where θ is the radar incident angle; α is the radar azimuth angle (starting from north, clockwise is positive); and Z... n Z e and Z u These are the deformations in the north, east, and vertical directions, respectively; Z j This refers to deformation along the distance from the ground. For example... Figure 2a , Figure 2b , Figure 2c As shown, where Figure 2a For SAR imaging three-dimensional view, Figure 2b This is a vertical view of SAR imaging. Figure 2c This is a side view of a SAR image. Step 3: Select a relatively stable GNSS monitoring point with few missing time series data. Based on the duration of the missing data and data trend characteristics, use methods such as linear interpolation or spline interpolation to impute the missing values ​​in the time series of this GNSS monitoring point, ensuring the consistency of the imputed result with the original sequence and the integrity of the monitoring data. High-frequency noise is removed through thresholding, and then the denoised, complete, and accurate GNSS time monitoring sequence is reconstructed. Step 4: Select the InSAR pixel closest to the GNSS monitoring point. Calculate the difference between this pixel and the denoised GNSS line-of-sight time series for each period. Add this difference to the InSAR data for each period to obtain a high spatiotemporal resolution deformation field that integrates GNSS and InSAR technologies. Under the high spatiotemporal resolution deformation field monitoring technology, continuously acquire monitoring data of the entire tailings dam, upstream and downstream of the dam body, reservoir area and surrounding areas to form a line-of-sight cumulative deformation rate map. Identify the area with the most severe deformation in the reservoir area and the boundary of deformation impact through color distribution and rate magnitude. Step 5: Employ UAV-LiDAR (Unmanned Aerial Vehicle LiDAR) technology to rapidly generate a high-precision 3D model of the tailings dam. Through periodic aerial surveys, a high-precision Digital Earth Model (DEM) is generated. Then, using automated modeling software, a 3D model of the tailings dam is created, identifying hazardous factors such as cracks, gullies, uplift, and localized landslides on the dam slope and crest. Simultaneously, 3D models of the tailings dam at different stages are established. Using the DEM difference method, the deformation and range of the tailings dam and surrounding mountains are calculated by subtracting the DEMs from those from two different periods. Finally, the deformation zone data from UAV-LiDAR is integrated with the boundary information of the high-resolution deformation field to generate a complete and high-precision deformation "target" region. In the identified deformation "target" area, the number of devices is increased or decreased according to the "three points form a line" principle, based on the actual area, slope, and other conditions of the deformation "target" area, and deformation data is monitored in real time. An online video monitoring system was used to verify the accuracy of the above real-time monitoring results using its intuitive and visual images. Various surface sensors, such as rain gauges, water level gauges, and water measuring weirs, are used to transmit the monitored data to an online video monitoring system, enabling real-time acquisition of data such as rainfall, water level, seepage flow, and dam surface displacement in the reservoir area. By burying underground sensors such as piezometers, microseismic sensors, pore water pressure gauges, and earth pressure gauges inside the dam body and foundation, the location of the dam body's seepage line and the data and patterns of changes in the soil and rock mechanics inside the dam body are monitored in real time. By integrating the deformation characteristics of the "target" area with the data and patterns of changes in the soil and rock mechanics inside the dam body, the disaster-causing factors and mechanisms are qualitatively analyzed, thereby macroscopically predicting the future deformation patterns of the deformation area. Step 6: Based on real-time monitoring data of the tailings dam, a safety assessment of the tailings dam is conducted using the AHP-EWM (Analytic Hierarchy Process-Entropy Weighting) combined weighting method. The subjective weight W1 is obtained using the AHP method, and the objective weight W2 is obtained using the EWM method. The formula W... i =αW1+(1-α)W2, where W i The combined weights are obtained using the AHP-EWM combined weighting method, where α is the combined weight coefficient, α∈(0,1). Based on the comprehensive evaluation value, the tailings dam safety level, W, is obtained. i ∈[90, 100) is safe, W i ∈[80, 90) is the yellow level, W i ∈[60, 80) is orange level, W i ∈(0, 60) is the red level, and finally, corresponding preventive measures are taken according to the safety level of the tailings dam.

Claims

1. A method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion, characterized in that, The steps are as follows: 1) Convert GNSS three-dimensional deformation into line-of-sight one-dimensional deformation, and calibrate InSAR data to the line-of-sight deformation of a GNSS monitoring point; 2) Select a relatively stable GNSS monitoring point with few missing time series data; The missing values ​​of the time series of the GNSS monitoring point are imputed; the imputed monitoring point is then subjected to wavelet denoising to obtain a complete and accurate GNSS time series. 3) Select the InSAR pixel closest to the GNSS monitoring point and calculate the difference between the pixel and the denoised GNSS line-of-sight time series for each period. Add this difference to the InSAR data for each period to obtain a high spatiotemporal resolution deformation field that integrates GNSS and InSAR technologies. 4) Under the high spatiotemporal resolution deformation field monitoring technology, the monitoring data of the entire tailings dam, upstream and downstream of the dam body, reservoir area and surrounding areas are continuously acquired to form a line-of-sight cumulative deformation rate map. By color distribution and rate magnitude, the area with the most severe deformation in the reservoir area and the boundary of deformation influence are identified. 5) Employing UAV-LiDAR technology, a high-precision 3D model of the tailings dam is rapidly generated; high-resolution aerial images are acquired through periodic aerial surveys to identify dangerous factors such as cracks, gullies, uplifts, and local landslides on the dam slope and crest, while simultaneously establishing 3D models at different times; the deformation and deformation range of the tailings dam and surrounding mountains are calculated; and the deformation zone data from UAV-LiDAR is integrated with the boundary information of the high-resolution deformation field to generate a complete and high-precision deformation "target" area. 6) GNSS has the feature of acquiring real-time displacement data of monitoring points in all weather conditions. In the identified deformation "target" area, according to the "three points form a line" principle, the number of devices can be increased or decreased according to the actual deformation "target" area area, slope and other conditions to monitor deformation data in real time. 7) By transmitting data monitored by various surface sensors such as rain gauges, water level gauges, and water measuring weirs to the online video monitoring system, specific data such as rainfall in the reservoir area, reservoir water level, seepage flow, and dam surface displacement can be obtained in real time; 8) By burying underground sensors such as piezometers, microseismic sensors, pore water pressure gauges, and earth pressure gauges inside the dam body and foundation, the location of the dam body seepage line and the data and patterns of changes in soil and rock mechanics inside the dam body are monitored in real time. Combined with deformation data of the "target" area, the disaster-causing factors and mechanisms are qualitatively analyzed. 9) Based on the real-time monitoring data of the tailings dam, the weight of each influencing factor in the tailings dam is calculated by using the combined weighting method of AHP-EWM analytic hierarchy process and entropy weighting method. By assigning values ​​to each influencing factor, the safety risk level of the tailings dam is evaluated in real time based on the comprehensive score.

2. The method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion according to claim 1, characterized in that, In step 1), the deformation Z in the north direction of the GNSS coordinate system n Deformation Z in the east direction e The projection is the deformation Z in the direction of ground distance. j Then, the deformation Z in the distance direction. j and vertical deformation Z u The deformation Z projected onto the line of sight l ; WITH j =Z n cos(α-270°)+Z e cos(360°-α)=-Z n sin(α)+Z e cos(α) WITH l =Z u cos(θ)-Z j sin(θ) Z l =Z u cos(θ)+Z n sin(α)sin(θ)-Z e cos(α)sin(θ) Where θ is the radar incident angle; α is the radar heading azimuth angle, Z n Z e and Z u These are the deformations in the north, east, and vertical directions, respectively; Z j Deformation in the direction of distance from the ground; By performing differential interferometry processing on SAR images of the monitoring area acquired by InSAR technology at different times, the phase signal dominated by the deformation of the dam body itself is extracted, and the cumulative deformation rate and time series deformation map are generated.

3. The method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion according to claim 1, characterized in that, In step 2), based on the missing duration and data trend characteristics, linear interpolation, spline interpolation, and other methods are selected to ensure the consistency of the interpolation result with the trend of the original sequence and the integrity of the monitoring data; high-frequency noise is removed by threshold processing, and then the denoised monitoring sequence is reconstructed.

4. The method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion according to claim 1, characterized in that, In step 5), UAV-LiDAR technology is used to quickly generate a high-precision digital ground model and a real-world 3D model of the tailings dam, and to identify dangerous factors such as cracks, gullies, uplifts and local landslides on the dam slope and top. The deformation and range of the tailings dam and surrounding mountains were calculated using the DEM difference method.

5. The method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion according to claim 1, characterized in that, In step 6), the number of devices is increased or decreased according to the actual deformation "target" area, slope, etc., based on the principle of "three points forming a line" in the identified deformation "target" area, and deformation data is monitored in real time; the data is fused with the InSAR monitoring data to achieve collaborative monitoring that combines point and surface monitoring and complements each other's advantages.

6. The method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion according to claim 1, characterized in that, In step 7), an online video monitoring system is used to verify the information using its intuitive and visual images.

7. The method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion according to claim 1, characterized in that, In step 8), the deformation characteristics of the "target" area are integrated with the data and patterns of changes in soil and rock mechanics inside the dam body to qualitatively analyze the disaster-causing factors and mechanisms, and macroscopically predict the future deformation patterns of the deformation area.

8. The method for identifying disaster-causing factors and assessing the safety of tailings dams based on multi-source heterogeneous data fusion according to claim 1, characterized in that, In step 9), based on real-time monitoring data of the tailings dam, the AHP-EWM combined weighting method is used to conduct a safety assessment of the tailings dam. The subjective weight W1 is obtained using the Analytic Hierarchy Process (AHP), and the objective weight W2 is obtained using the Entropy Weight Method (EWM). The formula W is then used to... i =αW1+(1-α)W2, where W i The combined weights are obtained using the AHP-EWM combined weighting method, where α is the combined weight coefficient, α∈(0,1); based on the comprehensive evaluation value, four safety levels of the tailings dam are obtained, W i ∈[90,100) is safe, W i ∈[80,90) is the yellow level, W i ∈[60,80) is orange level, W i ∈(0,60) is the red level, and finally, corresponding preventive measures are taken according to the safety level of the tailings dam.

Images