A method for intelligent early warning of online hydrological monitoring in coal mines

Abstract

This invention relates to the field of monitoring and early warning technology, and in particular to an intelligent early warning method for online monitoring of hydrology in coal mines. The method includes: collecting raw multidimensional hydrological data from coal mines and preprocessing it to obtain preprocessed hydrological data; based on the preprocessed hydrological data, introducing a behavior mapping structure to obtain normalized behavior mapping values ​​for the preprocessed hydrological data; based on the normalized behavior mapping values ​​of the preprocessed hydrological data, performing anomaly index evolution identification to obtain anomaly behavior indices; based on the anomaly behavior indices, introducing a dynamic clustering algorithm to perform cluster analysis on the preprocessed hydrological data and generate clustering results; based on the clustering results, performing quantitative risk calculation to obtain a comprehensive score; and based on the comprehensive score, identifying abnormal behavior and issuing early warnings. This method solves the technical problem that traditional hydrological monitoring methods are inaccurate in processing and analyzing hydrological data from coal mines, leading to untimely and inaccurate anomaly detection.

Description

Technical Field

[0001] This invention relates to the field of monitoring and early warning technology, and in particular to an intelligent early warning method for online monitoring of hydrology in coal mines. Background Technology

[0002] Hydrological hazards have always been a major risk to coal mine safety, especially incidents such as water inrush and seepage, which often pose a significant threat to production and personnel safety. As coal mining depth increases, the complexity of hydrological conditions and the uncertainty of the hydrological environment also increase. Traditional hydrological monitoring methods mainly rely on single hydrological parameters (such as water level, water temperature, and conductivity) and periodically sample and analyze data to determine hydrological changes. However, these methods have several significant shortcomings when facing complex environments: First, changes in hydrological parameters are usually multi-dimensional, dynamic, and interconnected, making it difficult to fully reflect the complexity of the hydrological environment through the monitoring of a single parameter. Second, traditional hydrological monitoring methods have a low sampling frequency, often failing to promptly capture sudden hydrological anomalies, resulting in delayed early warning systems that fail to issue timely alerts before disasters such as water inrush and seepage occur.

[0003] Furthermore, traditional hydrological monitoring methods often suffer from insufficient data processing capabilities, especially when faced with high-frequency, multi-dimensional, and complex hydrological data. Traditional methods are unable to effectively extract valuable information. Many traditional methods still rely on manually setting thresholds for early warning, which can easily lead to false alarms or missed alarms, and lack sufficient intelligence and adaptability. For example, while the K-Means clustering algorithm can perform clustering analysis on hydrological data, it cannot dynamically adjust cluster centers and cannot well cope with the nonlinear characteristics and dynamic changes of the data. As a result, the early warning system often fails to accurately identify potential risks when the hydrological environment changes drastically.

[0004] In summary, traditional hydrological monitoring methods still suffer from technical problems such as inaccurate processing and analysis of hydrological data from coal mines, leading to untimely and inaccurate anomaly detection. Summary of the Invention

[0005] This invention provides an intelligent early warning method for online monitoring of hydrology in coal mines, which solves the technical problem that traditional hydrological monitoring methods are inaccurate in processing and analyzing hydrological data in coal mines, resulting in untimely and inaccurate anomaly detection.

[0006] The present invention provides an intelligent early warning method for online monitoring of hydrology in coal mines, which specifically includes the following technical solutions:

[0007] A method for intelligent early warning of online hydrological monitoring in coal mines includes the following steps:

[0008] S1. Collect raw multidimensional hydrological data from the coal mine and preprocess it to obtain preprocessed hydrological data; based on the preprocessed hydrological data, introduce a behavior mapping structure to obtain the normalized behavior mapping value of the preprocessed hydrological data; based on the normalized behavior mapping value of the preprocessed hydrological data, perform anomaly index evolution identification to obtain anomaly behavior index.

[0009] S2. Based on the abnormal behavior index, a dynamic clustering algorithm with a central clustering function is introduced to perform cluster analysis on the preprocessed hydrological data and generate clustering results; based on the clustering results, quantitative risk calculation is performed to obtain a comprehensive score; based on the comprehensive score, abnormal behavior is identified and an early warning is issued.

[0010] Preferably, S1 specifically includes:

[0011] The behavior mapping structure introduces a time window and is constructed by combining the current volatility, historical trend, and potential periodic disturbance of the preprocessed hydrological data to obtain the normalized behavior mapping value of the preprocessed hydrological data.

[0012] Preferably, S1 specifically includes:

[0013] Based on the normalized behavior mapping value of the preprocessed hydrological data, and combined with the current hydrological behavior and historical trends, an anomaly index evolution function is constructed to identify the anomaly index evolution.

[0014] Preferably, S1 specifically includes:

[0015] The anomaly index evolution function calculates the anomaly behavior index by calculating the first and second differences of the normalized behavior mapping values ​​of the preprocessed hydrological data, and by introducing square terms and logarithmic operations.

[0016] Preferably, S2 specifically includes:

[0017] In the implementation of the dynamic clustering algorithm, in the initial stage, cluster centers are set, and the difference between the abnormal behavior index and the cluster centers is calculated. The fuzzy membership weights of the cluster centers are introduced with the preprocessed hydrological data, and the cluster centers are updated by combining nonlinear adjustment and logarithmic adjustment to obtain the updated cluster centers.

[0018] Preferably, S2 specifically includes:

[0019] When the deviation between two consecutive updated cluster centers is less than a preset threshold or the preset number of updates is reached, the update stops, the optimal cluster centers are obtained, and the clustering results are generated.

[0020] Preferably, S2 specifically includes:

[0021] Based on the abnormal behavior index, the deviation between the abnormal behavior index and the cluster center mean is calculated, and the standard deviation and arctangent function of the abnormal behavior index are combined to conduct a quantitative risk score and obtain a comprehensive score.

[0022] Preferably, S2 specifically includes:

[0023] The comprehensive score is compared with the preset warning threshold to identify abnormal behavior: when the comprehensive score is greater than the warning threshold, it indicates that there is abnormal behavior and risk. The risk level is classified and a strategy is provided based on the preset risk level threshold; when the comprehensive score is less than or equal to the warning threshold, it indicates that there is no abnormal behavior.

[0024] The beneficial effects of the technical solution of the present invention are:

[0025] 1. By introducing a behavior mapping structure and combining weighted averaging, exponential decay weighting, and periodic perturbation models, the volatility, trend, and periodic perturbations of hydrological data are uniformly encoded into dimensionless normalized behavior mapping values. This behavior mapping structure not only makes hydrological data comparable, monitorable, and fittable, but also effectively captures potential abnormal behavior patterns, providing strong support for subsequent anomaly detection and risk assessment.

[0026] 2. By introducing an anomaly index evolution function, the degree of mutation of each type of hydrological data is quantified to obtain an anomaly behavior index. The anomaly behavior index can dynamically characterize the degree of mutation of hydrological data and can identify potential hydrological anomalies in the early stage, providing precise technical support for the prevention of hydrological disasters in coal mines.

[0027] 3. To address the shortcomings of traditional clustering methods (such as K-Means) in processing dynamic, variable, and nonlinear data, a dynamic clustering algorithm with cluster centers is proposed. This algorithm can not only dynamically adjust cluster centers to adapt to changes in hydrological data, but also effectively identify potential clustering patterns in the data, significantly improving the ability to cope with complex hydrological environments. Attached Figure Description

[0028] Figure 1 This is a flowchart of an intelligent early warning method for online hydrological monitoring in coal mines, as described in this invention. Detailed Implementation

[0029] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0031] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent early warning method for online monitoring of coal mine hydrology provided by this invention.

[0032] See attached document Figure 1 The diagram illustrates a flowchart of an intelligent early warning method for online monitoring of hydrology in coal mines, provided by an embodiment of the present invention. The method includes the following steps:

[0033] S1. Collect raw multidimensional hydrological data from the coal mine and preprocess it to obtain preprocessed hydrological data; based on the preprocessed hydrological data, introduce a behavior mapping structure to obtain the normalized behavior mapping value of the preprocessed hydrological data; based on the normalized behavior mapping value of the preprocessed hydrological data, perform anomaly index evolution identification to obtain anomaly behavior index.

[0034] Raw, multidimensional hydrological data from the coal mine is collected using data acquisition equipment, including physical and chemical indicators such as water temperature, conductivity, pH value, and metal ion concentration (e.g., Fe²⁺, Mg²⁺). The data acquisition equipment (e.g., hydrological sensors) is deployed at different key locations underground in the coal mine (e.g., water-conducting fractures in goaf areas, roadway corners, and water-bearing faults) based on expert experience. The raw, multidimensional hydrological data is preprocessed to obtain preprocessed hydrological data, eliminating problems such as high noise, different physical dimensions, and asynchronous time sequences. The preprocessing process includes noise reduction, calibration, time synchronization, standardization, and normalization, all employing techniques well-known to those skilled in the art, which will not be elaborated upon here.

[0035] Furthermore, a behavior mapping structure is introduced, enabling each type of hydrological data (i.e., preprocessed hydrological data) to be expressed in a unified behavior tensor space. The technical purpose of introducing the behavior mapping structure is to simultaneously encode the current volatility, historical trend, and potential periodic perturbations of the preprocessed hydrological data into a dimensionless expression, making it comparable, monitorable, and fitable. Combining weighted average, exponential decay weighting, and a periodic perturbation model, a behavior mapping structure is constructed to obtain the normalized behavior mapping value of the preprocessed hydrological data, i.e., the behavior tensor space expression of the multidimensional data. The specific formula is as follows:

[0036] ;

[0037] Among them, among them, Indicates in Time of the first Normalized behavior mapping value of hydrological data after preprocessing; Indicates in Time of the first Preprocessed hydrological data (such as water temperature, conductivity, pH value, etc.), here representing historical hydrological data; This is an exponential decay factor, representing the time decay rate of historical hydrological data, i.e., the rate at which the influence weight of historical hydrological data on the current behavior decays. It is used to control the intensity of the influence of historical hydrological data on hydrological fluctuations at the current moment. It is adjusted using expert experience based on the actual hydrological change cycle in coal mines, with a reference value range of [missing value]. ; This is the periodic disturbance frequency, used to describe the periodic fluctuation characteristics of preprocessed hydrological data. It reflects the frequency of periodic disturbances such as tidal effects and production cycles. The value is determined based on the specific scenario, with a reference range of [value missing]. ; This is the duration of the sliding window, which is determined according to specific needs and is not limited here. This is the total number of categories of preprocessed hydrological data, set according to expert experience. It is a very small positive number, used to prevent division by zero or numerical instability during calculations. The reference range is... ; It is the integral variable in the integration process, representing the time interval. Every moment within; Indicates in Time of the first Preprocessed hydrological data (such as water temperature, conductivity, pH value, etc.); In the past The weighted influence of preprocessed hydrological data at one time point on preprocessed hydrological data at the current time point; It is a periodic disturbance term, representing the periodic fluctuations of hydrological data over time, in order to capture the impact of periodic fluctuations on hydrological changes; By time window Preprocessed hydrological data Weighted integration was performed, taking into account the impact of historical hydrological data on the current moment. The impact reflects the characteristic that the importance of historical hydrological data is gradually decreasing; Used for normalization, by calculating the square mean of each type of preprocessed hydrological data within a time window, to ensure that different types of hydrological data have the same dimensions and to eliminate the differences in the fluctuation range of various types of hydrological data;

[0038] Mapping values ​​to the normalized behavior of preprocessed hydrological data As input for anomaly index evolution identification, anomaly behavior indices are calculated; the core technical objective of anomaly index evolution identification is to dynamically characterize the degree of abrupt change between current hydrological behavior and its historical trends. Based on a dynamic anomaly detection framework, an indexed evaluation model, and an adaptive evolution mechanism, anomaly index evolution functions are constructed to perform anomaly index evolution identification, thereby obtaining anomaly behavior indices; specifically as follows:

[0039] ;

[0040] in, It is the first Preprocessed hydrological data The abnormal behavior index at any given time is used to quantitatively score the degree of change in this type of hydrological data. The larger the abnormal behavior index, the more unstable the data. Is Time of the first The normalization behavior of the preprocessed hydrological data is the second difference of the mapped values. Is Time of the first The normalization behavior mapping value of the preprocessed hydrological data is the first difference. The calculation methods of the first and second differences are well known to those skilled in the art and will not be described in detail here. This is a smoothing factor used to avoid excessively small values ​​during calculations, especially when the difference value is close to zero. It is determined based on expert experience, and the reference range is [range missing]. ; It is the size of the time window, used to limit the smoothed average range of the difference calculation and control the length of the time range during the integration process. It is determined according to specific needs and is not limited here. This is a regularization term used to prevent the influence of excessively small values. It is determined based on expert experience, and the reference range is [range to be specified]. ; This is used to describe the acceleration (degree of abrupt change) of current hydrological data changes, and a smoothing factor is introduced. To prevent potential numerical problems in calculations; The squared mean of the rate of change over a period of time is used to describe the trend of hydrological data. The introduction of the squared term ensures that the value of the anomalous behavior index increases significantly when the rate of change is large. Finally, the logarithmic operation is used to compress the anomalous behavior index to a reasonable range to avoid excessive amplification of extreme values.

[0041] S2. Based on the abnormal behavior index, a dynamic clustering algorithm with a central clustering function is introduced to perform cluster analysis on the preprocessed hydrological data and generate clustering results; based on the clustering results, quantitative risk calculation is performed to obtain a comprehensive score; based on the comprehensive score, abnormal behavior is identified and an early warning is issued.

[0042] To avoid the problem that traditional clustering methods such as K-Means cannot effectively handle dynamic, variable, and non-linear behavioral data, a dynamic clustering algorithm with dynamic cluster centers is designed. This algorithm can not only effectively identify potential clustering patterns in the data, but also automatically and dynamically adjust the cluster centers according to changes in the data. The specific implementation process is as follows:

[0043] In the initial stage, selection is carried out using a random method. Each cluster center represents an anomalous behavior pattern at the initial moment. The number of cluster centers is determined based on expert experience and is not limited here.

[0044] Furthermore, based on the fuzzy C-means algorithm and Gaussian mixture model, a cluster center update formula is constructed to update the cluster centers, yielding the updated cluster centers. The cluster center update formula is as follows:

[0045] ;

[0046] Among them, the left side of the equal sign Is The latest update Cluster centers, ; Is Time of the first Preprocessed hydrological data and the first The fuzzy membership weights of each cluster center are calculated using the membership function, with a reference range of values. ,and The membership function is a technique well-known to those skilled in the art and will not be elaborated upon here; the right side of the equals sign It is the current number Cluster centers of each cluster; This is a non-linear adjustment factor used to control the influence of distant data points in distance metrics. It is adjusted according to the actual application, and the reference value range is [value range missing]. ; This is a logarithmic adjustment factor used to balance the impact of abnormal fluctuations in hydrological data on clustering. It is determined based on expert experience, and the reference range is [value missing]. ; It is the first Abnormal behavior index of hydrological data after preprocessing The difference between cluster centers, after being processed by cube, represents the nonlinear effect of distance on the result; The change in the square of the anomalous behavior index combined with the logarithmic function can compress the influence of extreme values, so that an excessively large anomalous behavior index will not dominate the distance calculation. By combining nonlinear adjustment and logarithmic adjustment, it can maintain good robustness in environments with large data variations and strong noise.

[0047] During the update process, when the deviation between the cluster centers after two consecutive updates is less than a threshold determined by expert experience or when the preset number of updates is reached, the update is stopped, the optimal cluster centers are obtained, and the clustering result is generated. .

[0048] After clustering, to determine whether the current situation is a high-risk water inrush state, a nonlinear comprehensive scoring formula is constructed based on the abnormal behavior index, combined with the standard deviation and arctangent function of the abnormal behavior index, to perform quantitative risk scoring and obtain a comprehensive score; the nonlinear comprehensive scoring formula is as follows:

[0049] ;

[0050] in, Is The overall score at any given moment, i.e., the current risk score; It is the first The standard deviation of the abnormal behavior index of the preprocessed hydrological data is a well-known technique in the art and will not be elaborated here. Is The mean of the cluster centers at time step 1 is the average of all optimal cluster centers, reflecting the mean of the cluster centers at time step 2. The overall trend of all abnormal behavior indices at any given time; It is the first The mean of the abnormal behavior index of the hydrological data after preprocessing is calculated using a technique well-known to those skilled in the art, and will not be elaborated here. It is a smoothing factor used to prevent the denominator from being zero or the value from being unstable. The reference range for its value is... ; Indicates the first The deviation between the abnormal behavior index of the preprocessed hydrological data and the mean of the cluster centers is used to describe the degree of difference between the current hydrological data and the overall trend. A larger deviation means that the change in the hydrological data has a greater impact on the risk of the entire coal mine, and vice versa. By introducing the arctangent function to nonlinearly adjust the deviation, the impact of excessive deviation on the overall score can be suppressed. This represents the risk score of each type of preprocessed hydrological data. The scores are then summed and averaged to obtain a comprehensive score, ensuring that the contribution of each type of preprocessed hydrological data to the comprehensive score is relatively balanced.

[0051] Ultimately, the overall score will be determined. Compared with the warning threshold set based on expert experience Compare and identify abnormal behavior: when When this occurs, it indicates the presence of abnormal behavior and poses a risk, and a risk level classification is established, such as when... At that time, the risk level was Level 1; when At that time, the risk level was level two. At that time, the risk level was level three, among which, and These are risk level thresholds preset based on expert experience; and corresponding strategies are given based on the classified risk levels, such as closing the mining area, starting drainage pumps, and issuing audible and visual alarms; when... If the time is right, it means there is no abnormal behavior and no risk.

[0052] In summary, a method for intelligent early warning of online hydrological monitoring in coal mines has been developed.

[0053] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0054] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0055] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for intelligent early warning of online hydrological monitoring in coal mines, characterized in that, Includes the following steps: S1. Collect raw multidimensional hydrological data from the coal mine and preprocess it to obtain preprocessed hydrological data; Based on the preprocessed hydrological data, a time window is introduced, and a behavior mapping structure is constructed by combining the current volatility, historical trend, and potential periodic disturbances of the preprocessed hydrological data. The normalized behavior mapping value of the preprocessed hydrological data is obtained, and the calculation formula is as follows: ； in, Indicates in Time of the first Normalized behavior mapping value of hydrological data after preprocessing; Indicates in Time of the first Preprocessed hydrological data; It is an exponential decay factor; It is the frequency of the periodic disturbance; It is the duration of the sliding window; This refers to the total number of categories of preprocessed hydrological data; It is a positive number, and its range of values ​​is... ; It is the integral variable in the integration process, representing the time interval. Every moment within; Indicates in Time of the first Preprocessed hydrological data; Based on the normalized behavior mapping value of the preprocessed hydrological data, combined with the current hydrological behavior and historical trends, an anomaly index evolution function is constructed to identify the anomaly index evolution and obtain the anomaly behavior index. The calculation formula is as follows: ； in, It is the first Preprocessed hydrological data Abnormal behavior index at any given time; Is Time of the first The normalization behavior of preprocessed hydrological data is mapped to the second difference of the values. Is Time of the first The normalization behavior of preprocessed hydrological data maps the first difference of the values. It is a smoothing factor; It is the size of the time window; It is a regularization term; S2. Based on the abnormal behavior index, a dynamic clustering algorithm with cluster centers is introduced to perform cluster analysis on the preprocessed hydrological data and generate clustering results. In the initial stage, cluster centers are set, and the difference between the abnormal behavior index and the cluster centers is calculated. Fuzzy membership weights between the preprocessed hydrological data and the cluster centers are introduced. Based on the fuzzy C-means algorithm and Gaussian mixture model, combined with nonlinear adjustment and logarithmic adjustment, the cluster centers are updated to obtain the updated cluster centers. Based on the clustering results and the abnormal behavior index, quantitative risk calculation is performed to obtain a comprehensive score. Based on the comprehensive score, abnormal behavior is identified and early warning is issued.

2. The intelligent early warning method for online monitoring of hydrology in coal mines according to claim 1, characterized in that, S2 specifically includes: When the deviation between two consecutive updated cluster centers is less than a preset threshold or the preset number of updates is reached, the update stops, the optimal cluster centers are obtained, and the clustering results are generated.

3. The intelligent early warning method for online monitoring of hydrology in coal mines according to claim 2, characterized in that, S2 specifically includes: Based on the abnormal behavior index, the deviation between the abnormal behavior index and the cluster center mean is calculated, and the standard deviation and arctangent function of the abnormal behavior index are combined to conduct a quantitative risk score and obtain a comprehensive score.

4. The intelligent early warning method for online monitoring of hydrology in coal mines according to claim 3, characterized in that, S2 specifically includes: The comprehensive score is compared with the preset warning threshold to identify abnormal behavior: when the comprehensive score is greater than the warning threshold, it indicates that there is abnormal behavior and risk. The risk level is classified and a strategy is provided based on the preset risk level threshold; when the comprehensive score is less than or equal to the warning threshold, it indicates that there is no abnormal behavior.

Images