Remote on-line real-time monitoring and early warning system for mine site slope

Abstract

The invention relates to the technical field of side slope monitoring and early warning, in particular to a mine site side slope remote online real-time monitoring and early warning system which comprises a multi-source data monitoring unit, a data processing and analyzing unit, a side slope instability judging unit, an instability grading early warning unit and an early warning verification pushing unit. The method comprises the following steps: collecting monitoring data of a plurality of data sources to obtain time sequence monitoring data of each data source, carrying out noise reduction on the time sequence monitoring data, then decomposing the time sequence monitoring data, calculating a time-frequency feature vector through an energy density function, and optimizing the time-frequency feature vector through feature screening; calculating a slope stability evaluation index by optimizing the time-frequency feature vector, constructing an instability risk index to judge a slope instability risk level, constructing an early warning index, and combining the early warning index with the slope instability risk level to establish a graded early warning mechanism to obtain a slope instability early warning level; and verifying the slope instability early warning grade, and performing early warning according to the slope instability early warning grade passing the verification. According to the invention, the false alarm rate of slope monitoring and early warning can be reduced.

Description

Technical Field

[0001] The invention relates to the technical field of slope monitoring and early warning, and in particular to a mine site slope remote online real-time monitoring and early warning system. Background Art

[0002] With the development of mining technology, open-pit mining has been widely used in mineral resource development due to its advantages in economic benefits and mining efficiency. However, the slope safety problem in open-pit mining is becoming increasingly prominent, which not only threatens production safety, but also restricts the improvement of mining efficiency. Therefore, strengthening the research on slope monitoring and early warning technology has important practical significance.

[0003] In recent years, with the rapid development of new generation information technologies such as the Internet of Things, big data and artificial intelligence, slope monitoring technology has also been continuously innovating and improving. From the initial manual inspections and simple measurement marks to modern automated monitoring systems, especially the application of remote online monitoring and early warning systems, mining companies can grasp the slope conditions in real time.

[0004] However, the current slope monitoring and early warning system still has the problem of low early warning accuracy in practical applications. Specifically, the noise interference in the monitoring data collection process is serious, which makes the system unable to accurately judge the true state of the slope; at the same time, the system's lack of fusion analysis capabilities for multi-source data exacerbates the inaccuracy of early warning judgments, ultimately resulting in frequent false alarms.

[0005] Therefore, a remote online real-time monitoring and early warning system for mine site slopes was proposed. Summary of the invention

[0006] The purpose of the present invention is to provide a remote online real-time monitoring and early warning system for slopes on site in a mine, including a multi-source data monitoring unit, a data processing and analysis unit, a slope instability judgment unit, an instability graded early warning unit and an early warning verification and push unit. First, the monitoring data of multiple data sources are used to obtain the time series monitoring data of each data source, and the time series monitoring data is decomposed after noise reduction, and then the time-frequency feature vector is calculated by the energy density function, and the time-frequency feature vector is optimized by feature screening; the slope stability assessment index is calculated by optimizing the time-frequency feature vector, and the instability risk index is constructed to determine the slope instability risk level, and the early warning index is constructed to establish a graded early warning mechanism in combination with the slope instability risk level to obtain the slope instability early warning level; finally, the slope instability early warning level is verified, and an early warning is issued according to the verified slope instability early warning level.

[0007] To achieve the above object, the present invention provides the following technical solutions: A mine site slope remote online real-time monitoring and early warning system, comprising: The multi-source data monitoring unit collects monitoring data from multiple data sources of the mine site slope according to the sampling period, obtains the monitoring data within the sliding time window, and obtains the time series monitoring data of each data source; the data sources include deep displacement, surface displacement, inclination, earth pressure and groundwater level; The data processing and analysis unit includes a noise reduction module, a decomposition module and an optimization module; the noise reduction module performs noise reduction on the time series monitoring data; the decomposition module decomposes the noise-reduced time series monitoring data and calculates the time-frequency feature vector through an energy density function; the optimization module optimizes the time-frequency feature vector through feature screening to obtain an optimized time-frequency feature vector; A slope instability judgment unit is configured to calculate a slope stability evaluation index according to the optimized time-frequency eigenvector, construct an instability risk index based on the slope stability evaluation index, and judge the slope instability risk level according to the instability risk index; An instability graded warning unit is configured to construct a warning index according to the instability risk index and the monitoring data, establish a graded warning mechanism according to the slope instability risk level and the warning index, and obtain a slope instability warning level; The early warning verification and pushing unit calculates the verification index and verifies the slope instability warning level according to the verification index, issues an early warning according to the verified slope instability warning level, and pushes the early warning information to the mobile terminal of the slope site personnel for processing.

[0008] Preferably, the sliding time window size is , the sampling period is The sliding time window is slid in real time with the sampling period as the step length to obtain the length of Time series monitoring data.

[0009] Preferably, the data processing and analysis unit specifically includes: The noise reduction module uses wavelet basis functions to denoise the time series monitoring data of each data source. Decomposing the layers by wavelet to obtain an approximate coefficient vector and a detail coefficient vector of each layer, performing soft threshold processing on the detail coefficient vector, and performing inverse wavelet transform on the approximate coefficient vector to obtain the time series monitoring data after noise reduction; The decomposition module performs noise reduction on the time series monitoring data of each data source. Decomposing the layer wavelet packet to obtain a decomposition coefficient matrix of each layer, and calculating the time-frequency eigenvector according to the decomposition coefficient matrix through an energy density function; The optimization module calculates the importance score of each feature in the time-frequency feature vector of each data source, and performs feature screening on the time-frequency feature vector according to the importance score to obtain an optimized time-frequency feature vector.

[0010] Preferably, the specific process of the noise reduction module includes: The first-layer wavelet decomposition decomposes and downsamples the time series monitoring data through low-pass filters and high-pass filters to obtain the approximate coefficient vector and detail coefficient vector of the first layer; No. The wavelet decomposition is performed by low-pass filter and high-pass filter. The approximate coefficient vector of the layer is decomposed and downsampled to obtain the The approximate coefficient vector and detail coefficient vector of the layer, ; Perform soft threshold processing on the detail coefficient vector of each layer, and use the detail coefficient vector after soft threshold processing of each layer and the The approximate coefficient vector of the layer is inversely transformed to obtain the denoised time series monitoring data.

[0011] Preferably, the specific process of the decomposition module includes: The first-layer wavelet packet decomposition decomposes and downsamples the denoised time series monitoring data through low-pass filters and high-pass filters to obtain the first-layer decomposition coefficient matrix; No. The first layer of wavelet packet decomposition is processed by low-pass filter and high-pass filter. The decomposition coefficient matrix of the layer is decomposed and downsampled to obtain the The decomposition coefficient matrix of the layer, ;No. The decomposition coefficient matrix of the layer includes decomposition coefficient vector; After the first layer of wavelet packet decomposition, the energy density characteristics of each decomposition coefficient vector are calculated to construct the time-frequency feature vector; The time-frequency feature vector of a data source is: ; in, Indicates The time-frequency feature vector of each data source; Indicates The first layer of the decomposition coefficient matrix The energy density characteristics of the decomposition coefficient vector.

[0012] Preferably, the specific process of the optimization module includes: Calculate the energy concentration and energy proportion of each energy density feature in the time-frequency feature vector; The importance score of each energy density feature is calculated according to the energy concentration and the energy proportion, and the calculation formula is: ; in, Indicates Energy density features in the time-frequency feature vector of the data source Importance rating; express Energy concentration; express The energy proportion of and represents the weight coefficient; The energy density features in the time-frequency feature vector are sorted from large to small according to the importance score, and the top The optimized time-frequency feature vector is constructed based on the energy density features.

[0013] Preferably, calculating the slope stability assessment index according to the optimized time-frequency eigenvector specifically includes: Calculating a weighted energy index of each data source according to the optimized time-frequency feature vector; The slope stability evaluation index is calculated according to the weighted energy index and the spatial distance of the data source monitoring point, and the calculation formula is: ; in, represents the stability assessment index; Indicates the number of data sources; Indicates Importance weight coefficient of each data source; Indicates Weighted energy index of each data source; Indicates and The spatial distance between the monitoring points of the data source; Indicates the reference distance; represents the natural exponential function.

[0014] Preferably, determining the slope instability risk level specifically includes: The instability time-frequency characteristic vector of each data source is constructed based on the historical monitoring data of the slopes that have experienced instability; The Euclidean distance between the optimized time-frequency feature vector and the instability time-frequency feature vector of each data source is calculated to obtain an instability difference degree set, and the smallest instability difference degree in the instability difference degree set is selected to construct an instability risk index, and the formula is: ; in, represents the instability risk index; represents the weight coefficient; represents the regulating factor; represents the minimum instability difference in the instability difference set; represents the slope stability assessment index; The slope instability risk level is determined according to the instability risk index in the following way: ; in, Indicates the risk level of slope instability; Indicates the instability risk index threshold.

[0015] Preferably, the implementation process of the instability graded warning unit includes: Select the previous sampling point sampling points as the evaluation window, and obtaining the Assessment monitoring data; Select the slope stability stage historical sampling points as a reference window, and obtain the Reference monitoring data; Calculate the average cumulative deviation of all data sources within the evaluation window according to the evaluation monitoring data and the reference monitoring data; ; ; in, It represents the average cumulative deviation of all data sources; Indicates the number of data sources; Indicates the first The cumulative deviation of the data sources; express The first Assessment monitoring data; and Respectively The mean and standard deviation of the reference monitoring data; The early warning index is constructed according to the average cumulative deviation, and the formula is: ; in, represents the early warning index; represents the instability risk index; represents the weight coefficient; express The maximum cumulative deviation among the cumulative deviations of the data sources; Establish a graded early warning mechanism based on the early warning index and slope instability risk level: ; in, represents the warning threshold, and .

[0016] Preferably, the early warning verification push unit includes: Calculate the verification index. When the verification index is equal to 1, it means that the slope instability warning level has passed the verification, otherwise it means a false alarm. The calculation formula of the verification index is: ; in, represents the verification indicator; Indicates the maximum rate of change among all the evaluation and monitoring data; Represents the average cumulative deviation of all data sources within the evaluation window; Indicates the change rate threshold; Indicates the cumulative deviation threshold.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The monitoring data from multiple data sources are collected through the multi-source data monitoring unit, and the time series monitoring data is obtained in real time through the sliding time window; in the data processing and analysis unit, the wavelet basis function is first used to perform multi-layer wavelet decomposition and soft threshold noise reduction to effectively remove the noise in the monitoring data; then the decomposition coefficient matrix is ​​obtained through multi-layer wavelet packet decomposition, and the energy density function is introduced to calculate the time-frequency feature vector; the importance of the feature is evaluated by calculating the energy concentration and energy proportion, and the most representative features are selected to construct the optimized time-frequency feature vector; the multi-level data processing method can not only effectively extract the key features of the slope state, but also reduce data redundancy and improve the representativeness of the features, providing high-quality feature data for subsequent analysis, and effectively reducing the probability of false alarms due to data noise and redundant features.

[0018] 2. A slope stability assessment method considering the spatial distribution of monitoring points is proposed. First, the weighted energy index of each data source is calculated based on the optimized time-frequency feature vector, so that the importance of different frequency band features can be automatically adjusted according to their impact on slope stability; secondly, not only the time-frequency characteristics of the monitoring data are considered, but also the spatial position information is integrated, so that there is a stronger correlation between monitoring points with similar distances, and abnormal deformations in local areas can be effectively identified, making the assessment results more comprehensive and accurate; the slope stability assessment method of the present invention can not only accurately reflect the overall stability of the slope, but also promptly discover signs of local instability, effectively avoiding false warnings caused by abnormalities in a single monitoring point.

[0019] 3. A complete early warning mechanism has been built, including two dimensions: instability risk assessment and early warning index calculation. In the instability risk assessment, by comparing the difference between the current monitoring data and the historical stable state data, the abnormal changes in the slope state can be accurately captured; in the early warning index calculation, by comparing the evaluation window and the reference window, combined with the calculation of the cumulative deviation, the interference of short-term fluctuations can be effectively filtered out; the early warning verification link adopts a dual verification mechanism based on the change rate and the cumulative deviation, avoiding the one-sidedness that may be caused by a single indicator; the multi-level and multi-dimensional early warning mechanism combined with the additional verification process effectively controls the overall false alarm rate of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a structural schematic diagram of a mine site slope remote online real-time monitoring and early warning system of the present invention; Figure 2 It is a schematic diagram of the process of a mine site slope remote online real-time monitoring and early warning system of the present invention; Figure 3 The figure is a schematic diagram of the process flow of calculating and optimizing the time-frequency feature vector of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] See also Figures 1 to 3 The present invention provides a remote online real-time monitoring and early warning system for slopes on site of a mine, and the technical scheme is as follows: Embodiment 1 In order to verify the early warning effect of the present invention under the condition of deep rock softening, a mine site slope remote online real-time monitoring and early warning system is applied to the northern slope of a certain open-pit coal mine. Figure 1 ,include: The multi-source data monitoring unit collects monitoring data from multiple data sources of the mine site slope according to the sampling period, obtains the monitoring data within the sliding time window, and obtains the time series monitoring data of each data source; the data sources include deep displacement, surface displacement, inclination, earth pressure and groundwater level; The data processing and analysis unit includes a noise reduction module, a decomposition module and an optimization module; the noise reduction module performs noise reduction on the time series monitoring data; the decomposition module decomposes the noise-reduced time series monitoring data and calculates the time-frequency feature vector through an energy density function; the optimization module optimizes the time-frequency feature vector through feature screening to obtain an optimized time-frequency feature vector; A slope instability judgment unit is configured to calculate a slope stability evaluation index according to the optimized time-frequency eigenvector, construct an instability risk index based on the slope stability evaluation index, and judge the slope instability risk level according to the instability risk index; An instability graded warning unit is configured to construct a warning index according to the instability risk index and the monitoring data, establish a graded warning mechanism according to the slope instability risk level and the warning index, and obtain a slope instability warning level; The early warning verification and pushing unit calculates the verification index and verifies the slope instability warning level according to the verification index, issues an early warning according to the verified slope instability warning level, and pushes the early warning information to the mobile terminal of the slope site personnel for processing.

[0023] Furthermore, the sliding time window size is , the sampling period is The sliding time window is slid in real time with the sampling period as the step length to obtain the length of Time series monitoring data.

[0024] By setting the sliding time window size and sampling period, and sliding with the sampling period as the step length, it is possible to obtain fixed-length time series monitoring data in real time, ensuring the continuity and real-time nature of the data. In addition, the sliding window method can capture the dynamic changes of the slope state in a timely manner, providing a reliable data basis for subsequent analysis.

[0025] Furthermore, the data processing and analysis unit specifically includes: The noise reduction module uses wavelet basis functions to denoise the time series monitoring data of each data source. Decomposing the layers by wavelet to obtain an approximate coefficient vector and a detail coefficient vector of each layer, performing soft threshold processing on the detail coefficient vector, and performing inverse wavelet transform on the approximate coefficient vector to obtain the time series monitoring data after noise reduction; The decomposition module performs noise reduction on the time series monitoring data of each data source. Decomposing the layer wavelet packet to obtain a decomposition coefficient matrix of each layer, and calculating the time-frequency eigenvector according to the decomposition coefficient matrix through an energy density function; The optimization module calculates the importance score of each feature in the time-frequency feature vector of each data source, and performs feature screening on the time-frequency feature vector according to the importance score to obtain an optimized time-frequency feature vector.

[0026] The data processing and analysis unit adopts a three-level processing architecture, realizes data noise reduction through multi-layer decomposition of wavelet basis functions and soft threshold processing, extracts time-frequency features through wavelet packet decomposition, and finally obtains optimized time-frequency feature vectors through feature screening, which can not only effectively remove noise in monitoring data, but also extract the most representative features, improving the accuracy and efficiency of subsequent analysis. In addition, the time-frequency feature vector is calculated through the energy density function, which can better characterize the time-frequency characteristics of the data.

[0027] The northern slope of an open-pit coal mine is composed of Quaternary loess, sandstone and mudstone, and there are multiple groups of weak structural surfaces. 14 monitoring points are arranged on the slope, including 5 deep displacement monitoring points, 5 surface displacement monitoring points, 2 inclination monitoring points, 1 earth pressure monitoring point and 1 groundwater level monitoring point. Table 1 lists some key parameter settings of the system.

[0028] Table 1 System parameter settings

[0029] Furthermore, the specific process of the noise reduction module includes: The first-layer wavelet decomposition decomposes and downsamples the time series monitoring data through low-pass filters and high-pass filters to obtain the approximate coefficient vector and detail coefficient vector of the first layer; No. The wavelet decomposition is performed by low-pass filter and high-pass filter. The approximate coefficient vector of the layer is decomposed and downsampled to obtain the The approximate coefficient vector and detail coefficient vector of the layer, ; Perform soft threshold processing on the detail coefficient vector of each layer, and use the detail coefficient vector after soft threshold processing of each layer and the The approximate coefficient vector of the layer is inversely transformed to obtain the time series monitoring data after noise reduction; The soft threshold processing formula is: ; in, Represents the detail coefficient vector Detail coefficients; Express Detail coefficient after soft threshold processing; represents a symbolic function; represents the adaptive threshold, represents the noise standard deviation of the detail coefficient vector, Indicates the length of the time series monitoring data.

[0030] The soft threshold processing method with adaptive threshold can automatically adjust the threshold size according to the characteristics of the data. It can not only effectively remove noise but also maintain the characteristics of the original monitoring data to the greatest extent. It is more flexible and effective than the fixed threshold method and can adapt to the characteristics of different monitoring data.

[0031] Furthermore, the specific process of the decomposition module includes: The first-layer wavelet packet decomposition decomposes and downsamples the denoised time series monitoring data through low-pass filters and high-pass filters to obtain the first-layer decomposition coefficient matrix; No. The first layer of wavelet packet decomposition is processed by low-pass filter and high-pass filter. The decomposition coefficient matrix of the layer is decomposed and downsampled to obtain the The decomposition coefficient matrix of the layer, ;No. The decomposition coefficient matrix of the layer includes decomposition coefficient vector; After the first layer of wavelet packet decomposition, the energy density characteristics of each decomposition coefficient vector are calculated to construct the time-frequency feature vector; The time-frequency feature vector of a data source is: ; in, Indicates The time-frequency feature vector of each data source; Indicates The first layer of the decomposition coefficient matrix Energy density characteristics of the decomposition coefficient vector; No. The energy density feature calculation formula in the time-frequency feature vector of a data source is: ; Indicates The first layer of the decomposition coefficient matrix Decomposition coefficient vector Energy density characteristics; Indicates The length of the decomposition coefficient vector; express The A decomposition coefficient.

[0032] Through layer-by-layer wavelet packet decomposition, the time series monitoring data can be divided into more detailed frequency bands to obtain more detailed time-frequency information; at the same time, the use of energy density function to calculate the characteristics can not only reflect the energy distribution of data in different frequency bands and effectively capture the key characteristics of slope deformation, but also reduce the data dimension and improve the calculation efficiency.

[0033] Furthermore, the specific process of the optimization module includes: Calculate the energy concentration and energy proportion of each energy density feature in the time-frequency feature vector; the energy concentration and energy proportion calculation formula is: ; ; in, express Energy concentration; Indicates the maximum time-frequency feature vector ; express The energy proportion of Indicates the number of layers of wavelet packet decomposition; The importance score of each energy density feature is calculated according to the energy concentration and the energy proportion, and the calculation formula is: ; in, Indicates Energy density features in the time-frequency feature vector of the data source Importance rating; express Energy concentration; express The energy proportion of and represents the weight coefficient; The energy density features in the time-frequency feature vector are sorted from large to small according to the importance score, and the top The optimized time-frequency feature vector is constructed based on the energy density features.

[0034] By calculating the energy concentration and energy proportion to evaluate the importance of features and performing feature screening accordingly, the most representative features can be effectively identified and retained, while redundant and irrelevant features can be eliminated. This can significantly reduce the data dimension while maintaining feature representativeness and improve the efficiency of subsequent calculations.

[0035] Further, calculating the slope stability evaluation index according to the optimized time-frequency eigenvector specifically includes: The weighted energy index of each data source is calculated according to the optimized time-frequency feature vector, and the calculation formula is: ; ; in, Indicates Weighted energy index of each data source; Energy density characteristics The weight coefficient of The slope stability evaluation index is calculated according to the weighted energy index and the spatial distance of the data source monitoring point, and the calculation formula is: ; in, represents the stability assessment index; Indicates Importance weight coefficient of each data source; Indicates and The spatial distance between the monitoring points of the data source; Indicates the reference distance; represents the natural exponential function.

[0036] When calculating the slope stability assessment index, the feature weight and spatial distance weight are combined; the importance of different features is taken into account through the weighted energy index calculation, while the spatial distance weight reflects the spatial correlation between monitoring points. The assessment method that considers both feature importance and spatial distribution makes the stability assessment more comprehensive and accurate, and can better reflect the actual state of the slope.

[0037] Furthermore, the determination of the risk level of slope instability specifically includes: The instability time-frequency characteristic vector of each data source is constructed based on the historical monitoring data of the slopes that have experienced instability; The Euclidean distance between the optimized time-frequency feature vector and the instability time-frequency feature vector of each data source is calculated to obtain an instability difference degree set, and the smallest instability difference degree in the instability difference degree set is selected to construct an instability risk index, and the formula is: ; in, represents the instability risk index; represents the weight coefficient; represents the regulating factor; represents the minimum instability difference in the instability difference set; represents the slope stability assessment index; The slope instability risk level is determined according to the instability risk index in the following way: ; in, Indicates the risk level of slope instability; Indicates the instability risk index threshold.

[0038] By constructing the instability time-frequency feature vector and calculating the Euclidean distance between the current state and the historical instability cases, the instability risk of the slope can be effectively assessed. Combining the instability difference with the stability assessment index to construct the instability risk index and make a risk level judgment, the instability risk of the slope can be identified more accurately and the reliability of the early warning can be improved.

[0039] Furthermore, the implementation process of the instability graded warning unit includes: Select the previous sampling point sampling points as the evaluation window, and obtaining the Assessment monitoring data; Select the slope stability stage historical sampling points as a reference window, and obtain the Reference monitoring data; Calculate the average cumulative deviation of all data sources within the evaluation window according to the evaluation monitoring data and the reference monitoring data; ; ; in, It represents the average cumulative deviation of all data sources; Indicates the number of data sources; Indicates the first The cumulative deviation of the data sources; express The first Assessment monitoring data; and Respectively The mean and standard deviation of the reference monitoring data; The early warning index is constructed according to the average cumulative deviation, and the formula is: ; in, represents the early warning index; represents the instability risk index; represents the weight coefficient; express The maximum cumulative deviation among the cumulative deviations of the data sources; Establish a graded early warning mechanism based on the early warning index and slope instability risk level: ; in, represents the warning threshold, and .

[0040] By establishing a comparison mechanism between the evaluation window and the reference window and combining the calculation of the average cumulative deviation from multiple data sources, accurate identification of abnormal slope conditions is achieved. By weighted combination of the instability risk index and the cumulative deviation and considering the maximum deviation in the data source, both the assessment of the overall risk is guaranteed and the sensitivity to local anomalies is improved. At the same time, a progressive three-level warning threshold mechanism is adopted to achieve refined classification of warning levels and improve the accuracy of warnings.

[0041] Furthermore, the warning verification push unit includes: Calculate the verification index. When the verification index is equal to 1, it means that the slope instability warning level has passed the verification, otherwise it means a false alarm. The calculation formula of the verification index is: ; in, represents the verification indicator; Indicates the maximum rate of change among all the evaluation and monitoring data; Represents the average cumulative deviation of all data sources within the evaluation window; Indicates the change rate threshold; Indicates the cumulative deviation threshold.

[0042] No. Data source The rate of change of the evaluation monitoring data , Indicates Evaluation and monitoring data, express Evaluation monitoring data of the previous sampling point, Indicates the sampling period.

[0043] Table 2 Early warning process data record

[0044] Table 2 records the changes in key indicators of the slope from normal to unstable, verifying the effectiveness of the early warning system. Represents an example of a data sampling point. Four hours later, the system issued a level 3 warning.

[0045] The dual verification mechanism of change rate and cumulative deviation effectively avoids false alarms caused by instantaneous fluctuations in data or cumulative slow changes. It can not only detect sudden rapid deformations in a timely manner, but also does not ignore slow and continuous deformations. It significantly improves the reliability of early warnings and provides a reliable verification basis for the push of early warning information.

[0046] The present invention collects multi-dimensional data of slopes in real time through a multi-source data monitoring unit, and uses a noise reduction module, a decomposition module and an optimization module in a data processing and analysis unit to perform multi-level processing and analysis on the time series monitoring data, and extracts the most representative optimized time-frequency feature vector; in a slope instability judgment unit, a slope stability evaluation index is constructed in combination with spatial distance, and the instability risk level is judged by comparing historical instability cases; in an instability graded warning unit, a warning index is constructed based on comparative analysis of an evaluation window and a reference window, and a graded warning mechanism is established; finally, through the double verification mechanism of a warning verification and push unit, the warning result is strictly verified and pushed in time; through the combination of multi-level data processing, multi-dimensional risk assessment and double verification, the system realizes accurate warning of slope instability risks, effectively reduces the false alarm rate of warnings, and provides reliable technical support for real-time online monitoring of mine slopes.

[0047] Embodiment 2 In order to achieve the early warning effect under complex geological conditions, a mine site slope remote online real-time monitoring and early warning system was applied to the eastern slope of a metal mine. Figure 1 For the specific implementation of the system unit in Figure 2 , the specific steps include: S10. Collect monitoring data from multiple data sources of the mine site slope according to the sampling period, obtain the monitoring data within the sliding time window, and obtain the time series monitoring data of each data source; the data source includes deep displacement, surface displacement, inclination, earth pressure and groundwater level; S20. Denoise the time series monitoring data, decompose the denoised time series monitoring data, calculate the time-frequency feature vector through the energy density function, optimize the time-frequency feature vector through feature screening, and obtain an optimized time-frequency feature vector; S30. Calculating a slope stability evaluation index according to the optimized time-frequency eigenvector, constructing an instability risk index based on the slope stability evaluation index, and determining a slope instability risk level according to the instability risk index; S40. constructing a warning index according to the instability risk index and the monitoring data, establishing a graded warning mechanism according to the slope instability risk level and the warning index, and obtaining a slope instability warning level; S50. Calculate the verification index and verify the slope instability warning level according to the verification index, issue a warning according to the verified slope instability warning level, and push the warning information to the mobile terminal of the slope site personnel for processing.

[0048] Furthermore, the sliding time window size is , the sampling period is The sliding time window is slid in real time with the sampling period as the step length to obtain the length of Time series monitoring data.

[0049] Further, see Figure 3 , step S20 specifically includes: S21. Use wavelet basis function to analyze the time series monitoring data of each data source. Decomposing the layers by wavelet to obtain an approximate coefficient vector and a detail coefficient vector of each layer, performing soft threshold processing on the detail coefficient vector, and performing inverse wavelet transform on the approximate coefficient vector to obtain the time series monitoring data after noise reduction; S22. The time series monitoring data after noise reduction of each data source is processed Decomposing the layer wavelet packet to obtain a decomposition coefficient matrix of each layer, and calculating the time-frequency eigenvector according to the decomposition coefficient matrix through an energy density function; S23. Calculate the importance score of each feature in the time-frequency feature vector of each data source, perform feature screening on the time-frequency feature vector according to the importance score, and obtain an optimized time-frequency feature vector.

[0050] The lithology of the slope is mainly metamorphic sandstone and gneiss. 17 monitoring points are arranged on the slope, including 6 deep displacement monitoring points, 6 surface displacement monitoring points, 3 inclination monitoring points, 1 earth pressure monitoring point and 1 groundwater level monitoring point. Table 3 lists some key parameter settings of the system.

[0051] Table 3 System parameter settings

[0052] Furthermore, the specific process of step S21 includes: The first-layer wavelet decomposition decomposes and downsamples the time series monitoring data through low-pass filters and high-pass filters to obtain the approximate coefficient vector and detail coefficient vector of the first layer; No. The wavelet decomposition is performed by low-pass filter and high-pass filter. The approximate coefficient vector of the layer is decomposed and downsampled to obtain the The approximate coefficient vector and detail coefficient vector of the layer, ; Perform soft threshold processing on the detail coefficient vector of each layer, and use the detail coefficient vector after soft threshold processing of each layer and the The approximate coefficient vector of the layer is inversely transformed to obtain the denoised time series monitoring data.

[0053] Furthermore, the specific process of step S22 includes: The first-layer wavelet packet decomposition decomposes and downsamples the denoised time series monitoring data through low-pass filters and high-pass filters to obtain the first-layer decomposition coefficient matrix; No. The first layer of wavelet packet decomposition is processed by low-pass filter and high-pass filter. The decomposition coefficient matrix of the layer is decomposed and downsampled to obtain the The decomposition coefficient matrix of the layer, ;No. The decomposition coefficient matrix of the layer includes decomposition coefficient vector; After the first layer of wavelet packet decomposition, the energy density characteristics of each decomposition coefficient vector are calculated to construct the time-frequency feature vector; The time-frequency feature vector of a data source is: ; in, Indicates The time-frequency feature vector of each data source; Indicates The first layer of the decomposition coefficient matrix The energy density characteristics of the decomposition coefficient vector.

[0054] Furthermore, the specific process of step S23 includes: Calculate the energy concentration and energy proportion of each energy density feature in the time-frequency feature vector; The importance score of each energy density feature is calculated according to the energy concentration and the energy proportion, and the calculation formula is: ; in, Indicates Energy density features in the time-frequency feature vector of the data source Importance rating; express Energy concentration; express The energy proportion of and represents the weight coefficient; The energy density features in the time-frequency feature vector are sorted from large to small according to the importance score, and the top The optimized time-frequency feature vector is constructed based on the energy density features.

[0055] Further, calculating the slope stability evaluation index according to the optimized time-frequency eigenvector specifically includes: Calculating a weighted energy index of each data source according to the optimized time-frequency feature vector; The slope stability evaluation index is calculated according to the weighted energy index and the spatial distance of the data source monitoring point, and the calculation formula is: ; in, represents the stability assessment index; Indicates the number of data sources; Indicates Importance weight coefficient of each data source; Indicates Weighted energy index of each data source; Indicates and The spatial distance between the monitoring points of the data source; Indicates the reference distance; represents the natural exponential function.

[0056] Furthermore, the determination of the risk level of slope instability specifically includes: The instability time-frequency characteristic vector of each data source is constructed based on the historical monitoring data of the slopes that have experienced instability; The Euclidean distance between the optimized time-frequency feature vector and the instability time-frequency feature vector of each data source is calculated to obtain an instability difference degree set, and the smallest instability difference degree in the instability difference degree set is selected to construct an instability risk index, and the formula is: ; in, represents the instability risk index; represents the weight coefficient; represents the regulating factor; represents the minimum instability difference in the instability difference set; SI represents the slope stability assessment index; The slope instability risk level is determined according to the instability risk index in the following way: ; in, Indicates the risk level of slope instability; Indicates the instability risk index threshold.

[0057] Furthermore, the implementation process of step S40 includes: Select the previous sampling point sampling points as the evaluation window, and obtaining the Assessment monitoring data; Select the slope stability stage historical sampling points as a reference window, and obtain the Reference monitoring data; Calculate the average cumulative deviation of all data sources within the evaluation window according to the evaluation monitoring data and the reference monitoring data; ; ; in, It represents the average cumulative deviation of all data sources; Indicates the number of data sources; Indicates the first The cumulative deviation of the data sources; express The first Assessment monitoring data; and Respectively The mean and standard deviation of the reference monitoring data; The early warning index is constructed according to the average cumulative deviation, and the formula is: ; in, represents the early warning index; represents the instability risk index; represents the weight coefficient; express The maximum cumulative deviation among the cumulative deviations of the data sources; Establish a graded early warning mechanism based on the early warning index and slope instability risk level: ; in, represents the warning threshold, and .

[0058] Furthermore, step S50 specifically includes: Calculate the verification index. When the verification index is equal to 1, it means that the slope instability warning level has passed the verification, otherwise it means a false alarm. The calculation formula of the verification index is: ; in, represents the verification indicator; Indicates the maximum rate of change among all the evaluation and monitoring data; Represents the average cumulative deviation of all data sources within the evaluation window; Indicates the change rate threshold; Indicates the cumulative deviation threshold.

[0059] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote online real-time monitoring and early warning system for mine site slopes, characterized in that: include: The multi-source data monitoring unit collects monitoring data from multiple data sources of the mine site slope according to the sampling period, obtains the monitoring data within the sliding time window, and obtains the time series monitoring data of each data source; the data sources include deep displacement, surface displacement, inclination, earth pressure and groundwater level; The data processing and analysis unit includes a noise reduction module, a decomposition module and an optimization module; the noise reduction module performs noise reduction on the time series monitoring data; the decomposition module decomposes the noise-reduced time series monitoring data and calculates the time-frequency feature vector through an energy density function; the optimization module optimizes the time-frequency feature vector through feature screening to obtain an optimized time-frequency feature vector; A slope instability judgment unit is configured to calculate a slope stability evaluation index according to the optimized time-frequency eigenvector, construct an instability risk index based on the slope stability evaluation index, and judge the slope instability risk level according to the instability risk index; An instability graded warning unit is configured to construct a warning index according to the instability risk index and the monitoring data, establish a graded warning mechanism according to the slope instability risk level and the warning index, and obtain a slope instability warning level; The early warning verification and pushing unit calculates the verification index and verifies the slope instability warning level according to the verification index, issues an early warning according to the verified slope instability warning level, and pushes the early warning information to the mobile terminal of the slope site personnel for processing.

2. A mine site slope remote online real-time monitoring and early warning system according to claim 1, characterized in that: The sliding time window size is , the sampling period is The sliding time window is slid in real time with the sampling period as the step length to obtain the length of The timing monitoring data.

3. A mine site slope remote online real-time monitoring and early warning system according to claim 1, characterized in that: The data processing and analysis unit specifically includes: The noise reduction module uses wavelet basis functions to denoise the time series monitoring data of each data source. Decomposing the layers by wavelet to obtain an approximate coefficient vector and a detail coefficient vector of each layer, performing soft threshold processing on the detail coefficient vector, and performing inverse wavelet transform on the approximate coefficient vector to obtain the time series monitoring data after noise reduction; The decomposition module performs noise reduction on the time series monitoring data of each data source. Decomposing the layer wavelet packet to obtain a decomposition coefficient matrix of each layer, and calculating the time-frequency feature vector according to the decomposition coefficient matrix through an energy density function; The optimization module calculates the importance score of each feature in the time-frequency feature vector of each data source, and performs feature screening on the time-frequency feature vector according to the importance score to obtain the optimized time-frequency feature vector.

4. A mine site slope remote online real-time monitoring and early warning system according to claim 3, characterized in that: The specific process of the noise reduction module includes: The first layer of wavelet decomposition decomposes and downsamples the time series monitoring data through a low-pass filter and a high-pass filter to obtain an approximate coefficient vector and a detail coefficient vector of the first layer; No. The wavelet decomposition is performed by low-pass filter and high-pass filter. The approximate coefficient vector of the layer is decomposed and downsampled to obtain the The approximate coefficient vector and detail coefficient vector of the layer, ; The detail coefficient vector of each layer is subjected to soft threshold processing, and the detail coefficient vector after soft threshold processing of each layer and the first The approximate coefficient vector of the layer is subjected to inverse wavelet transform to obtain the time series monitoring data after noise reduction.

5. A mine site slope remote online real-time monitoring and early warning system according to claim 3, characterized in that: The specific process of decomposing the module includes: The first-layer wavelet packet decomposition decomposes and downsamples the denoised time series monitoring data through low-pass filters and high-pass filters to obtain the first-layer decomposition coefficient matrix; No. The first layer of wavelet packet decomposition is processed by low-pass filter and high-pass filter. The decomposition coefficient matrix of the layer is decomposed and downsampled to obtain the The decomposition coefficient matrix of the layer, ;No. The decomposition coefficient matrix of the layer includes decomposition coefficient vector; After the layer wavelet packet decomposition, the energy density characteristics of each decomposition coefficient vector are calculated to construct the time-frequency feature vector; The time-frequency feature vector of a data source is: ； in, Indicates The time-frequency feature vector of the data source; Indicates The first layer decomposition coefficient matrix The energy density characteristics of the decomposition coefficient vector.

6. A mine site slope remote online real-time monitoring and early warning system according to claim 3, characterized in that: The specific process of the optimization module includes: Calculating the energy concentration and energy proportion of each energy density feature in the time-frequency feature vector; The importance score of each energy density feature is calculated according to the energy concentration and the energy proportion, and the calculation formula is: ； in, Indicates Energy density features in the time-frequency feature vector of the data source Importance rating; express Energy concentration; express The energy proportion of and represents the weight coefficient; According to the importance score, the energy density features in the time-frequency feature vector are sorted from large to small, and the front The optimized time-frequency feature vector is constructed based on the energy density features.

7. A mine site slope remote online real-time monitoring and early warning system according to claim 1, characterized in that: Calculating the slope stability evaluation index according to the optimized time-frequency eigenvector specifically includes: Calculating a weighted energy index of each data source according to the optimized time-frequency feature vector; The slope stability evaluation index is calculated according to the weighted energy index and the spatial distance between the data source monitoring point, and the calculation formula is: ； in, represents the stability assessment index; Indicates the number of data sources; Indicates Importance weight coefficient of each data source; Indicates Weighted energy index of each data source; Indicates and The spatial distance between the monitoring points of the data source; Indicates the reference distance; represents the natural exponential function.

8. The mine site slope remote online real-time monitoring and early warning system according to claim 1 is characterized in that: Determining the slope instability risk level specifically includes: The instability time-frequency characteristic vector of each data source is constructed based on the historical monitoring data of the slopes that have experienced instability; The Euclidean distance between the optimized time-frequency feature vector and the instability time-frequency feature vector of each data source is calculated to obtain an instability difference degree set, and the smallest instability difference degree in the instability difference degree set is selected to construct an instability risk index, and the formula is: ； in, represents the instability risk index; represents the weight coefficient; represents the regulating factor; represents the minimum instability difference in the instability difference set; represents the slope stability assessment index; The slope instability risk level is determined according to the instability risk index in the following way: ； in, Indicates the risk level of slope instability; Indicates the instability risk index threshold.

9. A mine site slope remote online real-time monitoring and early warning system according to claim 1, characterized in that: The implementation process of the instability graded warning unit includes: Select the previous sampling point sampling points as the evaluation window, and obtaining the Assessment monitoring data; Select the slope stability stage historical sampling points as a reference window, and obtain the Reference monitoring data; Calculate the average cumulative deviation of all data sources within the evaluation window according to the evaluation monitoring data and the reference monitoring data; ； ； in, It represents the average cumulative deviation of all data sources; Indicates the number of data sources; Indicates the first The cumulative deviation of the data sources; express The first Assessment monitoring data; and Respectively The mean and standard deviation of the reference monitoring data; The early warning index is constructed according to the average cumulative deviation, and the formula is: ； in, represents the early warning index; represents the instability risk index; represents the weight coefficient; express The maximum cumulative deviation among the cumulative deviations of the data sources; Establish a graded early warning mechanism based on the early warning index and slope instability risk level: ； in, represents the warning threshold, and .

10. A mine site slope remote online real-time monitoring and early warning system according to claim 1, characterized in that: The early warning verification push unit includes: Calculate the verification index. When the verification index is equal to 1, it means that the slope instability warning level has passed the verification, otherwise it means a false alarm. The calculation formula of the verification index is: ； in, represents the verification indicator; Indicates the maximum rate of change among all the evaluation and monitoring data; Represents the average cumulative deviation of all data sources within the evaluation window; Indicates the change rate threshold; Indicates the cumulative deviation threshold.

Images