Early sign identification system and method for tailing pond safety risk based on data mining

Abstract

The application discloses a tailing pond safety risk early sign identification system based on data mining, comprising: an acquisition module for acquiring tailing pond multi-source monitoring data in real time; a feature extraction module for extracting features of the tailing pond multi-source monitoring data to obtain target features; an identification module for identifying the target features with a preset tailing pond safety risk early sign database to determine an identification result; an evaluation module for determining target key indicators of a disaster type when it is determined that the identification result shows a risk; determining a tailing pond safety risk evaluation value based on the target key indicators of the disaster type and the tailing pond multi-source monitoring data; a warning module for comparing the tailing pond safety risk evaluation value with a preset safety risk evaluation threshold, and issuing a warning prompt when it is determined that the tailing pond safety risk evaluation value is greater than or equal to the preset safety risk evaluation threshold; and the target features are compared with the database to discover the risk earlier.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of tailing pond safety risk identification, and in particular to a tailing pond safety risk early sign identification system and method based on data mining. BACKGROUND

[0002] Tailing pond is an important facility for storing the waste residue after the screening of concentrate. Long-term stacking is easy to become a high potential man-made debris flow hazard source. Once a dam break accident occurs, it will cause huge property losses and casualties. In order to protect the lives and property safety of downstream residents, the state continues to establish and perfect the standard system to strengthen the safety operation monitoring of tailing ponds.

[0003] At present, there are the following problems in the safety risk assessment of tailing ponds: (1) The data source used for the safety risk assessment of tailing ponds is relatively single, which cannot prepare to judge the overall trend of the safety of tailing ponds, resulting in relatively low assessment accuracy; (2) The safety risk assessment of tailing ponds at present is based on monitoring data for direct calculation and evaluation, resulting in relatively large calculation pressure and very low evaluation efficiency; (3) The safety risk assessment of tailing ponds at present cannot identify the early signs of tailing pond safety risk, and can only evaluate the risk degree of tailing ponds according to the indicators when the risk is about to occur, resulting in no sufficient time for emergency risk avoidance when tailing pond accidents occur, missing the best risk avoidance opportunity, and causing personnel and property losses.

[0004] Therefore, it is urgent to provide a tailing pond safety risk early sign identification system and method based on data mining to solve the above problems. SUMMARY

[0005] The present application aims to at least partly solve one of the above technical problems. To this end, the first aspect of the present application aims to provide a tailing pond safety risk early sign identification system based on data mining, which realizes accurate identification of early signs by acquiring multi-source monitoring data and performing feature extraction, uses target features for comparison with a database to discover potential risks earlier, and performs evaluation after identifying the existence of risks, so as to make the evaluation more targeted, improve the evaluation efficiency and reliability.

[0006] The second aspect of the present application aims to provide a tailing pond safety risk early sign identification method based on data mining.

[0007] To achieve the above-mentioned purpose, the first aspect of the present application provides a tailing pond safety risk early sign identification system based on data mining, comprising:

[0008] An acquisition module for acquiring real-time multi-source monitoring data of tailing ponds;

[0009] The feature extraction module is configured to extract features from the tailing pond multi-source monitoring data to obtain target features.

[0010] The identification module is configured to identify the target features with a preset tailing pond safety risk early sign database to determine an identification result.

[0011] The evaluation module is configured to determine a target key indicator of a disaster type when the identification result indicates that there is a risk, and determine a tailing pond safety risk evaluation value based on the target key indicator of the disaster type and the tailing pond multi-source monitoring data.

[0012] The early warning module is configured to compare the tailing pond safety risk evaluation value with a preset safety risk evaluation threshold, and issue a warning prompt when the tailing pond safety risk evaluation value is greater than or equal to the preset safety risk evaluation threshold.

[0013] Preferably, the acquisition module comprises:

[0014] The first acquisition submodule is configured to acquire satellite remote sensing images of the tailing pond area to obtain first data.

[0015] The second acquisition submodule is configured to acquire meteorological data of the tailing pond area to obtain second data.

[0016] The third acquisition submodule is configured to acquire geological disaster data of the tailing pond area to obtain third data.

[0017] The fourth acquisition submodule is configured to acquire tailing pond sensor data based on a sensor device to obtain fourth data.

[0018] The fifth acquisition submodule is configured to acquire tailing pond video monitoring data based on a monitoring device to obtain fifth data.

[0019] The first determination submodule is configured to take the first data, the second data, the third data, the fourth data, and the fifth data as the tailing pond multi-source monitoring data.

[0020] Preferably, the system further comprises a preprocessing module configured to preprocess the tailing pond multi-source monitoring data to obtain preprocessed tailing pond multi-source monitoring data.

[0021] The preprocessing module comprises:

[0022] The data cleaning submodule is configured to clean the tailing pond multi-source monitoring data to obtain cleaned tailing pond multi-source monitoring data.

[0023] The second determination submodule is configured to take the cleaned tailing pond multi-source monitoring data as the preprocessed tailing pond multi-source monitoring data.

[0024] Preferably, the data cleaning submodule comprises:

[0025] anomaly evaluation unit, configured to:

[0026] read the tailings pond multi-source monitoring data, and determine to-be-cleaned data;

[0027] input the to-be-cleaned data into a pre-trained classification model for classification, to obtain a plurality of sets of to-be-cleaned data of different categories;

[0028] select one of the sets of to-be-cleaned data;

[0029] calculate an anomaly evaluation value of the set of to-be-cleaned data;

[0030] compare the anomaly evaluation value with a preset anomaly evaluation threshold value;

[0031] when it is determined that the anomaly evaluation value is greater than or equal to the preset anomaly evaluation threshold value, take the set of to-be-cleaned data as a target set of to-be-cleaned data;

[0032] a screening unit, configured to divide data in the target set of to-be-cleaned data into a plurality of target to-be-cleaned sub-data, and determine whether each target to-be-cleaned sub-data is an abnormal sub-data; and when it is determined that a target to-be-cleaned sub-data is an abnormal sub-data, calculate a deviation coefficient corresponding to the abnormal sub-data;

[0033] a data cleaning unit, configured to:

[0034] compare the deviation coefficient corresponding to the abnormal sub-data with a preset deviation coefficient threshold value;

[0035] if the deviation coefficient corresponding to the abnormal sub-data is greater than or equal to the preset deviation coefficient threshold value, delete the abnormal sub-data;

[0036] if the deviation coefficient corresponding to the abnormal sub-data is less than the preset deviation coefficient threshold value, obtain a preset data cleaning rule, and clean the abnormal sub-data based on the preset data cleaning rule;

[0037] iterate through all the sets of to-be-cleaned data, to obtain cleaned tailings pond multi-source monitoring data.

[0038] Preferably, the screening unit comprises:

[0039] a segmentation sub-unit, configured to divide data in the target set of to-be-cleaned data into a plurality of target to-be-cleaned sub-data;

[0040] a first calculation sub-unit, configured to calculate a variance of data values in each target to-be-cleaned sub-data, to obtain a variance value corresponding to each target to-be-cleaned sub-data;

[0041] a screening sub-unit, configured to compare the variance value with a preset variance threshold value;

[0042] determining that the variance value is greater than or equal to a preset variance threshold value, taking the target sub-data to be cleaned corresponding to the variance value as abnormal sub-data;

[0043] traversing all the target sub-data to be cleaned to obtain a plurality of abnormal sub-data;

[0044] a second calculation sub-unit configured to calculate a deviation coefficient corresponding to each abnormal sub-data.

[0045] Preferably, the identification module comprises:

[0046] a mapping sub-module configured to map the target feature to a preset space to obtain a target feature vector corresponding to the target feature;

[0047] a similarity calculation sub-module configured to perform similarity calculation on the target feature vector corresponding to the target feature and a feature vector in a preset tailings pond safety risk early warning database;

[0048] an identification result determination sub-module configured to take the similarity calculation result as an identification result.

[0049] Preferably, the evaluation module comprises:

[0050] a risk determination sub-module configured to compare the identification result with a preset similarity threshold value, and determine that the tailings pond is at risk when the identification result is greater than or equal to the preset similarity threshold value;

[0051] a screening sub-module configured to:

[0052] obtain tailings pond historical disaster information when it is determined that the tailings pond is at risk; the tailings pond historical disaster information comprises index information corresponding to different disaster types of the tailings pond and a frequency of abnormality of various indexes;

[0053] construct a disaster index correlation coefficient matrix based on the index information corresponding to different disaster types of the tailings pond;

[0054] determine a correlation degree of each index based on the frequency of abnormality of various indexes and the disaster index correlation coefficient matrix to obtain a plurality of correlation degrees of indexes;

[0055] compare the correlation degrees of the plurality of indexes with a preset correlation degree threshold value;

[0056] take an index corresponding to the correlation degree greater than or equal to the preset correlation degree threshold value as a target key index of a disaster type;

[0057] a sixth acquisition sub-module configured to acquire a data value of the target key index in tailings pond multi-source monitoring data;

[0058] The computing sub-module is configured to determine a tailing pond safety risk assessment value based on the data value corresponding to the target key indicator in the tailing pond multi-source monitoring data and a preset algorithm.

[0059] Preferably, the computing sub-module is configured to determine a tailing pond safety risk assessment value based on the data value corresponding to the target key indicator in the tailing pond multi-source monitoring data and a preset algorithm, including:

[0060]

[0061] wherein, represents the tailing pond safety risk assessment value; represents an average risk value of the tailing pond safety risk assessment value; represents a risk value corresponding to the i-th target key indicator in the i-th disaster type; represents a total number of disaster types; represents a total number of target key indicators; represents a probability density function of the occurrence of the i-th disaster type. Preferably, the early warning module includes:

[0062] The comparison sub-module is configured to compare the tailing pond safety risk assessment value with a preset safety risk assessment threshold value;

[0063] The early warning information generation sub-module is configured to generate early warning information when it is determined that the tailing pond safety risk assessment value is greater than or equal to the preset safety risk assessment threshold value;

[0064] The query sub-module is configured to query a current on-duty information table to determine current shift manager information;

[0065] The early warning sending sub-module is configured to send the early warning information to the current shift manager to make a warning prompt.

[0066] To achieve the above-mentioned purpose, the second aspect of the present application proposes a tailing pond safety risk early sign identification method based on data mining, including:

[0067] Real-time acquisition of tailing pond multi-source monitoring data;

[0068] Extraction of the tailing pond multi-source monitoring data to obtain target features;

[0069] Identification of the target features with a preset tailing pond safety risk early sign database to determine an identification result;

[0070]

[0071] ​​​When it is determined that the identification result shows a risk, a target key indicator of a disaster type is determined; and based on the target key indicator of the disaster type and multi-source monitoring data of the tailing pond, a tailing pond safety risk assessment value is determined.

[0072] The tailing pond safety risk assessment value is compared with a preset safety risk assessment threshold value, and when it is determined that the tailing pond safety risk assessment value is greater than or equal to the preset safety risk assessment threshold value, a warning prompt is issued.

[0073] The present application provides a tailing pond safety risk early sign identification system and method based on data mining, which can more comprehensively and deeply understand the condition of the tailing pond by acquiring multi-source monitoring data and performing feature extraction, and improve the accuracy of risk identification; the early signs are accurately identified by comparing the features with a database, and potential risks are found earlier; after the risk is identified, the evaluation is performed, so that the evaluation is more targeted, and the evaluation efficiency and reliability are improved; the warning is issued in time according to the comparison of the evaluation value and the threshold value, time is gained for taking measures to respond, and the risk harm is reduced; which helps to improve the overall safety guarantee level of the tailing pond and reduce the possibility of accidents.

[0074] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application can be achieved and obtained by the structure particularly pointed out in the written description and the accompanying drawings.

[0075] The technical solutions of the present application will be further described in detail below with the help of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0076] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, and are used to explain the present application together with embodiments of the present application, and do not constitute a limitation on the present application. In the drawings:

[0077] Figure 1 is a block diagram of a tailing pond safety risk early sign identification system based on data mining according to an embodiment of the present application;

[0078] Figure 2 is a block diagram of an acquisition module according to an embodiment of the present application;

[0079] Figure 3 is a flowchart of a tailing pond safety risk early sign identification method based on data mining according to an embodiment of the present application. DETAILED DESCRIPTION

[0080] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0081] Example 1

[0082] like Figure 1 As shown, a data mining-based system for identifying early signs of safety risks in tailings dams includes:

[0083] The acquisition module is used to acquire multi-source monitoring data of the tailings dam in real time;

[0084] The feature extraction module is used to extract features from the multi-source monitoring data of the tailings dam to obtain target features;

[0085] The identification module is used to identify target features against a pre-set database of early signs of tailings dam safety risks and determine the identification result.

[0086] The assessment module is used to determine the target key indicators of the disaster type when the identification results indicate that there is a risk; and to determine the safety risk assessment value of the tailings dam based on the target key indicators of the disaster type and the multi-source monitoring data of the tailings dam.

[0087] The early warning module is used to compare the safety risk assessment value of the tailings dam with a preset safety risk assessment threshold, and to issue an early warning when it is determined that the safety risk assessment value of the tailings dam is greater than or equal to the preset safety risk assessment threshold.

[0088] In this embodiment, the multi-source monitoring data of the tailings dam includes, but is not limited to, satellite remote sensing image data, meteorological data, geological disaster data, tailings dam sensor data, and video surveillance data.

[0089] In this embodiment, a feature extraction model is used to extract features from the multi-source monitoring data of the tailings dam; the feature extraction model includes convolutional neural network models, etc.

[0090] In this embodiment, the pre-set early signs database of tailings dam safety risks is a database of early signs of tailings dam safety risks established based on feature analysis of historical disaster data of tailings dams.

[0091] In this embodiment, the identification result is the similarity calculation value between the feature vector of the target feature and the feature vector in the preset tailings dam safety risk early signs database.

[0092] In this embodiment, the preset security risk assessment threshold is set in advance based on industry experience.

[0093] The beneficial effects of the above technical solutions are: by acquiring multi-source monitoring data and performing feature extraction, the condition of the tailing pond can be more comprehensively and deeply understood, and the accuracy of risk identification is improved; early signs are accurately identified by comparing the features with the database, and potential risks are discovered earlier; after identifying the risks, evaluation is performed, making the evaluation more targeted, improving the evaluation efficiency and reliability; according to the comparison of the evaluation value and the threshold, early warning is issued in time, time is gained for taking measures to reduce the risk harm; it is helpful to improve the overall safety guarantee level of the tailing pond and reduce the possibility of accidents.

[0094] Embodiment 2

[0095] As shown in Figure 2 : an acquisition module comprising:

[0096] a first acquisition sub-module configured to acquire satellite remote sensing images of a tailing pond area to obtain first data;

[0097] a second acquisition sub-module configured to acquire meteorological data of the tailing pond area to obtain second data;

[0098] a third acquisition sub-module configured to acquire geological disaster data of the tailing pond area to obtain third data;

[0099] a fourth acquisition sub-module configured to acquire tailing pond sensor data based on a sensor device to obtain fourth data;

[0100] a fifth acquisition sub-module configured to acquire tailing pond video monitoring data based on a monitoring device to obtain fifth data;

[0101] a first determination sub-module configured to take the first data, the second data, the third data, the fourth data, and the fifth data as multi-source monitoring data of the tailing pond.

[0102] In this embodiment, the second data includes real-time meteorological data, meteorological prediction and forecast data, meteorological statistical data, and meteorological warning data.

[0103] In this embodiment, the third data includes access geological disaster prone risk type data, geological disaster monitoring and early warning data, and geological disaster early warning feedback monthly report data.

[0104] In this embodiment, the fourth data includes displacement data, phreatic line data, reservoir water level data, and seepage water quantity data. A suitable displacement sensor, such as a total station, a GPS displacement monitoring instrument, or the like, is installed at a key position of the tailings reservoir, and regular measurement and data collection are performed, real-time displacement change data are obtained through connection of the sensor and a data collection system, a phreatic line monitoring sensor, such as a pressure measuring pipe, is buried, the sensor monitors water pressure change in real time, and position data of the phreatic line are obtained through conversion, a water level sensor, such as a drop-in type water level gauge, is installed at a suitable position in the reservoir, the sensor continuously monitors water level height and transmits data, seepage water quantity can be indirectly calculated through setting of a flow sensor in a relevant area or in combination with other measurement means, and the collected data are analyzed and processed to obtain accurate seepage water quantity information.

[0105] In this embodiment, the fifth data includes overflow well site monitoring data, beach top ore drawing site monitoring data, tailings discharge pipeline monitoring data, dam body downstream slope monitoring data, flood discharge facility import and export monitoring data, reservoir water level gauge monitoring data, and dry beach marker pole monitoring data.

[0106] The above technical solution has the following beneficial effects: different types of data are obtained through multiple channels, including satellite remote sensing images, meteorological data, geological disaster data, sensor data, and video monitoring data, which can comprehensively reflect various conditions of the tailings reservoir; multiple dimensions such as space, weather, and geology are covered, making the monitoring of the tailings reservoir more three-dimensional and in-depth; real-time data are obtained by using sensor equipment and monitoring equipment, ensuring the timeliness and accuracy of the data, which helps to discover problems in time; multiple source data are integrated together, which is convenient for comprehensive analysis and evaluation, and can better understand the overall state and potential risks of the tailings reservoir; the comprehensive use of multiple data can improve the early warning capability for potential dangers and abnormal conditions, and gain time for taking corresponding measures.

[0107] Embodiment 3

[0108] Further comprising a preprocessing module configured to preprocess the tailings reservoir multi-source monitoring data to obtain preprocessed tailings reservoir multi-source monitoring data.

[0109] The preprocessing module comprises:

[0110] A data cleaning submodule configured to clean the tailings reservoir multi-source monitoring data to obtain cleaned tailings reservoir multi-source monitoring data.

[0111] A second determination submodule configured to determine the cleaned tailings reservoir multi-source monitoring data as the preprocessed tailings reservoir multi-source monitoring data.

[0112] The beneficial effects of the above technical solutions are: through the data cleaning submodule, the noise, errors and outliers in the data can be removed, thereby improving the accuracy and reliability of the data; ensuring that the preprocessed tailing pond multi-source monitoring data is more pure and effective, providing more valuable basis for subsequent analysis and decision-making; clean data is conducive to subsequent accurate analysis and modeling, avoiding incorrect conclusions or decisions due to data problems; high-quality data input helps the stable operation of the entire monitoring system, reducing faults or abnormalities caused by data problems.

[0113] Embodiment 4

[0114] The data cleaning submodule comprises:

[0115] The anomaly evaluation unit is configured to:

[0116] read the tailing pond multi-source monitoring data and determine the data to be cleaned;

[0117] input the data to be cleaned into a pre-trained classification model for classification to obtain a plurality of sets of data to be cleaned of different categories;

[0118] select one set of data to be cleaned;

[0119] calculate the anomaly evaluation value of the set of data to be cleaned;

[0120] compare the anomaly evaluation value with a preset anomaly evaluation threshold;

[0121] determine that the anomaly evaluation value is greater than or equal to the preset anomaly evaluation threshold, and take the set of data to be cleaned as a target set of data to be cleaned;

[0122] The screening unit is configured to divide the data in the target set of data to be cleaned into a plurality of target sub-data to be cleaned, and determine whether each target sub-data to be cleaned is an abnormal sub-data; and when it is determined that the target sub-data to be cleaned is an abnormal sub-data, calculate the deviation coefficient corresponding to the abnormal sub-data.

[0123] The data cleaning unit is configured to:

[0124] compare the deviation coefficient corresponding to the abnormal sub-data with a preset deviation coefficient threshold;

[0125] if the deviation coefficient corresponding to the abnormal sub-data is greater than or equal to the preset deviation coefficient threshold, delete the abnormal sub-data;

[0126] if the deviation coefficient corresponding to the abnormal sub-data is less than the preset deviation coefficient threshold, obtain a preset data cleaning rule; and based on the preset data cleaning rule, clean the abnormal sub-data;

[0127] Traverse all the to-be-cleaned data sets to obtain the cleaned tailings pond multi-source monitoring data.

[0128] In this embodiment, the abnormal evaluation value of the to-be-cleaned data set is calculated:

[0129]

[0130] wherein, represents the abnormal evaluation value of the to-be-cleaned data set; represents an error factor; represents the range of data in the to-be-cleaned data set; represents the variance of data in the to-be-cleaned data set; represents the standard deviation of data in the to-be-cleaned data set; represents a natural constant; represents the mean of data in the to-be-cleaned data set.

[0131] In this embodiment, the classification model includes but is not limited to a decision tree classification model, a naive Bayes classification model, a support vector machine classification model, a neural network classification model, and the like.

[0132] In this embodiment, the preset deviation coefficient threshold is used to measure the deviation degree of data and is determined based on historical data analysis.

[0133] In this embodiment, the preset data cleaning rule includes removing duplicate data, processing missing values, correcting data format, removing outliers, consistency checking, and cleaning invalid characters, and the like.

[0134] The working principle of the above technical solution is: the to-be-cleaned data is classified to obtain a plurality of categories of to-be-cleaned data sets; first, the entire to-be-cleaned data set is evaluated to determine whether the data abnormality in the to-be-cleaned data set exceeds the preset abnormal evaluation threshold, if it exceeds the preset abnormal evaluation threshold, the to-be-cleaned data set is taken as a target to-be-cleaned data set; the target to-be-cleaned data set is secondarily screened, the data in the target to-be-cleaned data set is evenly divided into a plurality of target to-be-cleaned sub-data, and it is determined whether each target to-be-cleaned sub-data is an abnormal sub-data; the abnormal sub-data is thirdly screened, the deviation coefficient is screened, and when it is determined that the target to-be-cleaned sub-data is an abnormal sub-data, the deviation coefficient corresponding to the abnormal sub-data is calculated; if the deviation coefficient corresponding to the abnormal sub-data is greater than or equal to the preset deviation coefficient threshold, the abnormal sub-data is deleted; if the deviation coefficient corresponding to the abnormal sub-data is less than the preset deviation coefficient threshold, the preset data cleaning rule is obtained; the abnormal sub-data is cleaned based on the preset data cleaning rule; through the three screenings of the to-be-cleaned data set, the abnormal sub-data is determined, the abnormal sub-data is cleaned, and the cleaning efficiency and accuracy are improved.

[0135] The beneficial effects of the above technical solutions are: the abnormal evaluation unit can accurately identify the data set that may have abnormalities, improving the pertinence and efficiency of data screening; the screening unit further subdivides the data and judges the abnormal sub-data, and calculates the deviation coefficient, which helps to more accurately evaluate the abnormality degree of the data; different processing methods are adopted according to different deviation coefficients, that is, the data with serious abnormalities is directly deleted, and the data with relatively slight abnormalities is cleaned according to rules, which guarantees the rationality and scientificity of data processing; finally, the tailings pond multi-source monitoring data after cleaning can be obtained, effectively improving the quality and accuracy of the data, and providing more reliable basis for subsequent analysis and decision-making.

[0136] Embodiment 5

[0137] The screening unit comprises:

[0138] The segmentation sub-unit is configured to divide the data in the target data set to be cleaned into a plurality of target sub-data to be cleaned evenly.

[0139] The first calculation sub-unit is configured to calculate the variance of the data values in each target sub-data to be cleaned, and obtain a variance value corresponding to each target sub-data to be cleaned.

[0140] The screening sub-unit is configured to compare the variance value with a preset variance threshold.

[0141] When it is determined that the variance value is greater than or equal to the preset variance threshold, the target sub-data to be cleaned corresponding to the variance value is regarded as abnormal sub-data.

[0142] All target sub-data to be cleaned are traversed to obtain a plurality of abnormal sub-data.

[0143] The second calculation sub-unit is configured to calculate a deviation coefficient corresponding to each abnormal sub-data.

[0144] In this embodiment, the preset variance threshold is used to measure the fluctuation of the data, which is set in advance based on industry experience.

[0145] In this embodiment, the deviation coefficient comprises:

[0146]

[0147] Q represents the deviation coefficient corresponding to the abnormal sub-data. represents the data value of the pth data in the abnormal sub-data. represents the mean value of the data in the abnormal sub-data; c represents the standard deviation of the abnormal sub-data; and s represents the total number of data in the abnormal sub-data.

[0148] The beneficial effects of the above technical solutions are: the data is subdivided by the segmentation subunit, which facilitates more accurate analysis of the data characteristics of each small part, the first calculation subunit calculates the variance to effectively reflect the dispersion degree of the data, thereby accurately identifying the sub-data with a higher abnormality possibility; the screening subunit judges based on the variance threshold, improving the accuracy and reliability of screening abnormal sub-data; traversing all target sub-data to be cleaned ensures that there is no omission and potential abnormal conditions can be found comprehensively; the second calculation subunit calculates the deviation coefficient, providing a quantitative basis for further processing of abnormal sub-data, so that subsequent processing decisions are more scientific and reasonable.

[0149] Embodiment 6

[0150] The identification module comprises:

[0151] The mapping sub-module is configured to map the target feature to a preset space to obtain a target feature vector corresponding to the target feature.

[0152] The similarity calculation sub-module is configured to calculate the similarity between the target feature vector corresponding to the target feature and a feature vector in the preset tailing pond safety risk early warning database.

[0153] The identification result determination sub-module is configured to take the similarity calculation result as an identification result.

[0154] In this embodiment, the calculation method of the similarity includes but is not limited to the Euclidean distance and the cosine similarity algorithm.

[0155] In this embodiment, the specific implementation of mapping the target feature to a preset space to obtain a target feature vector corresponding to the target feature includes the following steps: first, determine the dimension and feature representation method of the preset space. This can be a multi-dimensional vector space. Then, through a certain mapping algorithm or function, the values of the target feature are converted and processed so as to correspond to the dimension of the preset space. For example, linear transformation, nonlinear transformation, etc. can be used to calculate and combine the attribute values of the target feature to obtain the coordinate values in the preset space, thereby forming the target feature vector.

[0156] In this embodiment, similarity calculation is performed between the target feature vector corresponding to the target feature and the feature vectors in the preset tailings pond safety risk early sign database: the target feature vector and the feature vectors in the database are standardized or normalized to eliminate the influence of factors such as dimension; common methods such as cosine similarity, Euclidean distance, and Pearson correlation coefficient can be used; the target feature vector is calculated with each feature vector in the database one by one, and the corresponding similarity values are obtained according to the selected similarity measurement method; the calculated similarity values are compared and sorted to determine those feature vectors that are most similar to the target feature vector; a suitable similarity threshold is set according to actual needs, and when the similarity reaches or exceeds the threshold, it is considered that there is a certain degree of similarity; according to the results of similarity calculation and threshold judgment, relevant information such as similar feature vectors and their corresponding similarity values is output.

[0157] The beneficial effects of the above technical solutions are: the mapping submodule converts the target feature into a feature vector, facilitating digital processing and comparison and improving the efficiency of data processing; the similarity calculation submodule can accurately determine the similarity degree of the target feature and the known safety risk early sign by comparing with the feature vectors in the preset database; the identification result determination submodule can quickly obtain the identification result, which helps to timely discover possible safety risks of the tailings pond and realize early warning; with the help of feature vectors and similarity calculation, the accuracy and reliability of identification are improved.

[0158] Embodiment 7

[0159] The evaluation module comprises:

[0160] The risk determination submodule is configured to compare the identification result with a preset similarity threshold, and determine that the tailings pond has a risk when the identification result is greater than or equal to the preset similarity threshold.

[0161] The screening submodule is configured to:

[0162] When it is determined that the tailings pond has a risk, the historical disaster information of the tailings pond is obtained; the historical disaster information of the tailings pond comprises index information corresponding to different disaster types of the tailings pond and the frequency of abnormal occurrence of various indexes.

[0163] A disaster index correlation coefficient matrix is constructed based on the index information corresponding to different disaster types of the tailings pond.

[0164] The correlation degree of each index is determined based on the frequency of abnormal occurrence of various indexes and the disaster index correlation coefficient matrix, and the correlation degrees corresponding to a plurality of indexes are obtained.

[0165] The correlation degrees corresponding to the plurality of indexes are compared with a preset correlation degree threshold.

[0166] The indicators corresponding to the correlation degree being greater than or equal to the preset correlation threshold are used as the target key indicators for disaster types;

[0167] The sixth acquisition submodule is used to acquire the data values ​​corresponding to the target key indicators in the multi-source monitoring data of the tailings dam;

[0168] The calculation submodule is used to determine the safety risk assessment value of the tailings dam based on the data values ​​corresponding to the target key indicators in the multi-source monitoring data of the tailings dam and the preset algorithm.

[0169] In this embodiment, when tailings inventory is at risk, there may be one or more types of disasters. Therefore, it is necessary to determine the target key indicators of the disaster type based on historical data.

[0170] In this embodiment, a disaster index correlation coefficient matrix is ​​constructed based on the index information corresponding to different disaster types of tailings ponds, including:

[0171] The correlation coefficient matrix of disaster indicators is as follows:

[0172]

[0173] in, Represents the correlation coefficient matrix of disaster indicators; Indicates the first The disaster type and the first The correlation between the indicators, and satisfying the first... The disaster type and the first When there is a correlation between the indicators =1; No. The disaster type and the first When there is no correlation between the indicators =0.

[0174] In this embodiment, the correlation degree of each indicator is determined based on the frequency of anomalies in various indicators and the correlation coefficient matrix of disaster indicators, including:

[0175]

[0176] in, A matrix representing the correlation between indicators; = , These represent the correlation degree of each indicator; A frequency matrix representing the occurrence of anomalies in various indicators. = , This indicates the frequency of abnormal occurrences for each indicator.

[0177] In this embodiment, the preset correlation threshold is set in advance based on industry experience.

[0178] The beneficial effects of the above technical solutions are: the risk determination sub-module accurately determines whether the tailing pond is at risk based on the identification result and the threshold value, improving the objectivity and accuracy of risk determination; the screening sub-module can in-depth analyze key indicators related to disasters by obtaining historical disaster information and constructing a correlation coefficient matrix, etc., making the evaluation more targeted; determining the target key indicators helps to focus on key points and more efficiently perform subsequent evaluation and monitoring; the calculation sub-module determines the safety risk assessment value by combining multi-source monitoring data and a preset algorithm, making the evaluation result more scientific and reliable; overall, it is conducive to comprehensively controlling the safety risk of the tailing pond, discovering potential risks in time and realizing early warning, and ensuring the safe operation of the tailing pond.

[0179] Embodiment 8

[0180] The calculation sub-module is configured to determine a tailing pond safety risk assessment value based on a data value corresponding to a target key indicator in the multi-source monitoring data of the tailing pond and a preset algorithm, and the calculation sub-module includes:

[0181]

[0182] wherein, represents the tailing pond safety risk assessment value; represents an average risk value of the tailing pond safety risk assessment value; represents a risk value corresponding to the i-th target key indicator in the i-th disaster type; represents a total number of disaster types; represents a total number of target key indicators; represents a probability density function of the occurrence of the i-th disaster type;

[0183] The beneficial effects of the above technical solutions are: through specific formula calculation, the safety risk assessment value of the tailing pond can be accurately quantified, making the evaluation result more objective and comparable; the formula comprehensively considers the average risk value, the risk values of different disaster types and target key indicators, the number of disaster types, the number of target key indicators, and the probability density function of the occurrence of risk, etc., fully reflecting the risk status of the tailing pond; providing a scientific decision basis for relevant departments and personnel, which is helpful for formulating more reasonable risk response strategies and measures; the evaluation value can be dynamically adjusted according to real-time monitoring data and changes, reflecting the changes of the safety state of the tailing pond in time.

[0184] Embodiment 9

[0185] The early warning module includes:

[0186] ​​​The comparison submodule is configured to compare the tailing pond safety risk assessment value with a preset safety risk assessment threshold value;

[0187] The early warning information generation submodule is configured to generate early warning information when it is determined that the tailing pond safety risk assessment value is greater than or equal to the preset safety risk assessment threshold value.

[0188] The query submodule is configured to query a current on-duty information table to determine current shift manager information.

[0189] The early warning sending submodule is configured to send the early warning information to the current shift manager to make a warning prompt.

[0190] The above technical solution has the beneficial effects that the situation that the tailing pond safety risk exceeds the threshold value can be found in time, and early warning information can be generated quickly, thereby improving the timeliness of the response; the manager information is determined by querying the on-duty information table, so that the early warning information can be accurately sent to the relevant person in charge, and information confusion and omission can be avoided; the responsibility of the current shift manager is clear, which helps the current shift manager to take timely response measures, and helps to strengthen the management and supervision of the tailing pond; the method helps to prevent possible risks in advance, ensures the safety of the tailing pond and the surrounding environment, and reduces the probability of potential hazards.

[0191] To achieve the above purpose, as shown in Figure 3 The second aspect of the present application provides a tailing pond safety risk early sign identification method based on data mining, which comprises steps S1-S5:

[0192] S1: Real-time acquisition of tailing pond multi-source monitoring data;

[0193] S2: Extraction of the tailing pond multi-source monitoring data to obtain target features;

[0194] S3: Identification of the target features with a preset tailing pond safety risk early sign database to determine an identification result;

[0195] S4: Determination of a target key indicator of a disaster type when it is determined that the identification result shows that there is a risk; determination of a tailing pond safety risk assessment value based on the target key indicator of the disaster type and the tailing pond multi-source monitoring data;

[0196] S5: Comparison of the tailing pond safety risk assessment value with a preset safety risk assessment threshold value; issuance of a warning prompt when it is determined that the tailing pond safety risk assessment value is greater than or equal to the preset safety risk assessment threshold value.

[0197] The above technical scheme has the beneficial effects that: by acquiring multi-source monitoring data and performing feature extraction, the condition of the tailing pond can be more comprehensively and deeply understood, and the accuracy of risk identification is improved; early signs are accurately identified by comparing the features with the database, and potential risks are discovered earlier; after the risk is identified, evaluation is performed, so that the evaluation is more targeted, and the evaluation efficiency and reliability are improved; a warning is timely issued according to the comparison of the evaluation value and the threshold value, time is won for taking response measures, and the risk harm is reduced; and the overall safety guarantee level of the tailing pond is improved, and the possibility of accidents is reduced.

[0198] Obviously, various modifications and changes can be made to the present application by those skilled in the art without departing from the spirit and scope of the present application. Thus, it is intended that the present application cover the modifications and changes as long as they come within the scope of the claims and their equivalents.

Claims

1. A data mining-based system for identifying early signs of safety risks in tailings dams, characterized in that, include: The acquisition module is used to acquire multi-source monitoring data of the tailings dam in real time; The feature extraction module is used to extract features from the multi-source monitoring data of the tailings dam to obtain target features; The identification module is used to identify target features against a pre-set database of early signs of tailings dam safety risks and determine the identification result. The assessment module is used to determine the target key indicators for the disaster type when the identification results indicate that a risk exists; Based on the target key indicators of the disaster type and the multi-source monitoring data of the tailings dam, the safety risk assessment value of the tailings dam is determined. The early warning module is used to compare the safety risk assessment value of the tailings dam with the preset safety risk assessment threshold, and to issue an early warning when it is determined that the safety risk assessment value of the tailings dam is greater than or equal to the preset safety risk threshold. The recognition module includes: The mapping submodule is used to map the target features to a preset space to obtain the target feature vector corresponding to the target features; The similarity calculation submodule is used to perform similarity calculation between the target feature vector corresponding to the target feature and the feature vector in the preset tailings dam safety risk early signs database; The identification result determination submodule is used to take the similarity calculation result as the identification result; The evaluation module includes: The risk determination submodule is used to compare the identification result with a preset similarity threshold, and to determine that the tailings inventory has a risk when the identification result is greater than or equal to the preset similarity threshold. The filtering submodule is used for: Obtain historical disaster information of tailings dams; the historical disaster information of tailings dams includes indicator information corresponding to different disaster types of tailings dams and the frequency of abnormal occurrence of various indicators; Construct a disaster index correlation coefficient matrix based on the index information corresponding to different disaster types of tailings ponds; The correlation degree of each indicator is determined based on the frequency of anomalies in various indicators and the correlation coefficient matrix of disaster indicators, thus obtaining the correlation degree of several indicators. The correlation degree corresponding to the aforementioned indicators is compared with a preset correlation degree threshold; The indicators corresponding to the correlation degree being greater than or equal to the preset correlation threshold are used as the target key indicators for disaster types; The sixth acquisition submodule is used to acquire the data values ​​corresponding to the target key indicators in the multi-source monitoring data of the tailings dam; The calculation submodule is used to determine the safety risk assessment value of the tailings dam based on the data values ​​of the target key indicators in the multi-source monitoring data of the tailings dam and the preset algorithm. A disaster index correlation coefficient matrix is ​​constructed based on the index information corresponding to different disaster types in tailings ponds, including: The correlation coefficient matrix of disaster indicators is as follows: Where G represents the disaster index correlation coefficient matrix; α mn Let α represent the correlation between the m-th disaster type and the n-th indicator, and satisfy the condition that there is a correlation between the m-th disaster type and the n-th indicator. mn =1; when there is no correlation between the m-th disaster type and the n-th indicator, α mn =0; The correlation degree of each indicator is determined based on the frequency of anomalies in various indicators and the correlation coefficient matrix of disaster indicators, including: Where W represents the correlation matrix of the indicators; W = [W1, W2, ..., W... n ],W1,W2…,W n These represent the correlation degree of each indicator; F represents the frequency matrix of anomalies in various indicators, F = [F1, F2, ..., F]. n ],F1,F2…,F n This indicates the frequency of abnormal occurrences for each indicator.

2. The tailings dam safety risk early warning system based on data mining as described in claim 1, characterized in that, The acquisition module includes: The first acquisition submodule is used to acquire satellite remote sensing images of the tailings dam area to obtain the first data; The second acquisition submodule is used to acquire meteorological data of the tailings dam area to obtain the second data. The third acquisition submodule is used to acquire geological disaster data in the tailings dam area to obtain third data. The fourth acquisition submodule is used to acquire sensor data from the tailings dam based on sensor devices to obtain the fourth data. The fifth acquisition submodule is used to acquire video monitoring data of the tailings dam based on the monitoring equipment to obtain the fifth data; The first determining submodule is used to use the first data, the second data, the third data, the fourth data, and the fifth data as multi-source monitoring data for the tailings dam.

3. The tailings dam safety risk early warning system based on data mining as described in claim 1, characterized in that, It also includes a preprocessing module, used to preprocess the multi-source monitoring data of the tailings dam to obtain preprocessed multi-source monitoring data of the tailings dam; The preprocessing module includes: The data cleaning submodule is used to clean the multi-source monitoring data of the tailings dam to obtain cleaned multi-source monitoring data of the tailings dam. The second determination submodule is used to use the cleaned tailings pond multi-source monitoring data as the pre-processed tailings pond multi-source monitoring data.

4. The data mining-based early warning system for identifying safety risks in tailings dams as described in claim 3, characterized in that, The data cleaning submodule includes: Anomaly evaluation unit, used for: Read multi-source monitoring data from the tailings dam to determine the data to be cleaned; The data to be cleaned is input into a pre-trained classification model for classification, resulting in a set of data to be cleaned in several categories. Choose any set of data to be cleaned; Calculate the anomaly evaluation value of the dataset to be cleaned; The abnormal evaluation value is compared with a preset abnormal evaluation threshold; When the abnormal evaluation value is determined to be greater than or equal to the preset abnormal evaluation threshold, the dataset to be cleaned is combined into a target dataset to be cleaned. The filtering unit is used to divide the data in the target dataset to be cleaned into several target sub-data sets to be cleaned, determine whether each target sub-data set to be cleaned is an abnormal sub-data set, and calculate the deviation coefficient corresponding to the abnormal sub-data set when the target sub-data set to be cleaned is determined to be an abnormal sub-data set. The data cleaning unit is used for: The deviation coefficient corresponding to the abnormal sub-data is compared with a preset deviation coefficient threshold. If the deviation coefficient corresponding to the abnormal sub-data is greater than or equal to the preset deviation coefficient threshold, the abnormal sub-data will be deleted. If the deviation coefficient corresponding to the abnormal sub-data is less than a preset deviation coefficient threshold, a preset data cleaning rule is obtained; and the abnormal sub-data is cleaned based on the preset data cleaning rule. By iterating through all the datasets to be cleaned, we obtain the cleaned multi-source monitoring data of the tailings dam.

5. The data mining-based early warning system for identifying safety risks in tailings dams as described in claim 4, characterized in that, The filtering unit includes: The segmentation unit is used to divide the data in the target dataset to be cleaned into several target sub-data units of the target dataset to be cleaned; The first calculation subunit is used to calculate the variance of the data values ​​in each target sub-data to be cleaned, and to obtain the variance value corresponding to each target sub-data to be cleaned. A filtering subunit is used to compare the variance value with a preset variance threshold; When the variance value is determined to be greater than or equal to a preset variance threshold, the target sub-data to be cleaned corresponding to the variance value is regarded as abnormal sub-data. Iterate through all the target sub-data to be cleaned to obtain several abnormal sub-data; The second calculation subunit is used to calculate the deviation coefficient corresponding to each abnormal sub-data.

6. The tailings dam safety risk early warning system based on data mining as described in claim 1, characterized in that, The calculation submodule is used to determine the safety risk assessment value of the tailings dam based on the data values ​​of the target key indicators in the multi-source monitoring data of the tailings dam and a preset algorithm, including: Where P represents the tailings dam safety risk assessment value; χ represents the average risk value of the tailings dam safety risk assessment value; Λ i,j Let f(x) represent the risk value corresponding to the j-th target key indicator in the i-th disaster type; m represents the total number of disaster types; n represents the total number of target key indicators; and f(x)dx represents the probability density function of the risk of the i-th disaster type.

7. The tailings dam safety risk early warning system based on data mining as described in claim 1, characterized in that, The early warning module includes: The comparison submodule is used to compare the tailings dam safety risk assessment value with a preset safety risk assessment threshold. The early warning information generation submodule is used to generate early warning information when it is determined that the safety risk assessment value of the tailings dam is greater than or equal to a preset safety risk assessment threshold. The query submodule is used to query the current duty information table and determine the management personnel information for the current shift. The early warning sending submodule is used to send the early warning information to the current shift manager to provide an early warning reminder.

8. A method for identifying early signs of safety risks in tailings dams based on data mining, characterized in that, include: Real-time acquisition of multi-source monitoring data from tailings ponds; The target features are obtained by extracting the multi-source monitoring data of the tailings dam. The target characteristics are identified against a pre-set database of early signs of tailings dam safety risks, and the identification results are determined. When the identification results indicate the presence of risk, the target key indicators for the disaster type are determined; Based on the target key indicators of the disaster type and the multi-source monitoring data of the tailings dam, the safety risk assessment value of the tailings dam is determined. The tailings dam safety risk assessment value is compared with a preset safety risk assessment threshold. When the tailings dam safety risk assessment value is determined to be greater than or equal to the preset safety risk threshold, an early warning is issued. The target characteristics are identified against a pre-set database of early signs of tailings dam safety risks, and the identification results are determined, including: The target features are mapped to a preset space to obtain the target feature vector corresponding to the target features; The similarity between the target feature vector corresponding to the target feature and the feature vector in the preset tailings dam safety risk early signs database is calculated. The similarity calculation result is used as the recognition result; When the identification results indicate the presence of risk, the target key indicators for the disaster type are determined; based on the target key indicators for the disaster type and multi-source monitoring data of the tailings dam, the safety risk assessment value of the tailings dam is determined, including: The identification result is compared with a preset similarity threshold. If the identification result is greater than or equal to the preset similarity threshold, it is determined that the tailings inventory is at risk. Obtain historical disaster information of tailings dams; the historical disaster information of tailings dams includes indicator information corresponding to different disaster types of tailings dams and the frequency of abnormal occurrence of various indicators; Construct a disaster index correlation coefficient matrix based on the index information corresponding to different disaster types of tailings ponds; The correlation degree of each indicator is determined based on the frequency of anomalies in various indicators and the correlation coefficient matrix of disaster indicators, thus obtaining the correlation degree of several indicators. The correlation degree corresponding to the aforementioned indicators is compared with a preset correlation degree threshold; The indicators corresponding to the correlation degree being greater than or equal to the preset correlation threshold are used as the target key indicators for disaster types; Obtain the data values ​​corresponding to the target key indicators in the multi-source monitoring data of the tailings dam; Based on the data values ​​of the target key indicators in the multi-source monitoring data of the tailings dam and the preset algorithm, the safety risk assessment value of the tailings dam is determined. A disaster index correlation coefficient matrix is ​​constructed based on the index information corresponding to different disaster types in tailings ponds, including: The correlation coefficient matrix of disaster indicators is as follows: Where G represents the disaster index correlation coefficient matrix; α mn Let α represent the correlation between the m-th disaster type and the n-th indicator, and satisfy the condition that there is a correlation between the m-th disaster type and the n-th indicator. mn =1; when there is no correlation between the m-th disaster type and the n-th indicator, α mn =0; The correlation degree of each indicator is determined based on the frequency of anomalies in various indicators and the correlation coefficient matrix of disaster indicators, including: Where W represents the correlation matrix of the indicators; W = [W1, W2, ..., W... n ],W1,W2…,W n These represent the correlation degree of each indicator; F represents the frequency matrix of anomalies in various indicators, F = [F1, F2, ..., F]. n ],F1,F2…,F n This indicates the frequency of abnormal occurrences for each indicator.

Images