Ring network data transmission method and diagnosis system applied to underground coal mines

Abstract

The present invention relates to the technical field of mine data transmission and supervision, and particularly relates to a ring network data transmission method and a diagnosis system applied to underground coal mines. The present invention first obtains gas concentration data, humidity data, and the mining start time; further, in the current monitoring period, according to the fluctuation stability characteristics of the target data, combined with the time difference between the sampling time of each data point and the nearest mining start time, obtains the reference coefficient of the target data; according to the similarity characteristics of the gas concentration data and the similarity characteristics of the humidity data between two adjacent monitoring nodes, combined with the reference coefficient, obtains the corrected effective coefficient corresponding to the two adjacent monitoring nodes; further, according to the sum value of all the corrected effective coefficients between any two monitoring nodes, obtains the mutual monitoring coefficient corresponding to the two monitoring nodes, combines the corresponding covariance, and uses a preset dimensionality reduction algorithm to obtain a two-dimensional data set and transmit it. Finally, a safety diagnosis is performed according to the two-dimensional data set.

Description

Technical Field

[0001] The present invention relates to the technical field of mine data transmission and supervision, and particularly relates to a ring network data transmission method and a diagnosis system applied to coal mine underground. Background Art

[0002] Due to the complex environment in the coal mine underground, the gas concentrations in each area may vary. Therefore, multiple sensors need to be installed underground to transmit data to ensure the monitoring of the entire mine environment. And for the data obtained by multiple sensors, in order to reduce the influence of electromagnetic interference and environmental changes, ring network transmission is generally used to ensure the accuracy of the data.

[0003] In the prior art, it is necessary to circularly transmit the gas concentration data obtained by all sensors in the ring network until it finally reaches the control center, and then perform dimensionality reduction analysis through these data to improve the analysis efficiency and conduct safety diagnosis on the coal mine underground. However, affected by the complex environment, the data obtained by some sensors has low availability, and the reference value of the correlation between the data of two different sensors also fluctuates, ultimately resulting in inaccurate dimensionality reduction results and inaccurate safety diagnosis results. Summary of the Invention

[0004] In order to solve the technical problem that inaccurate data dimensionality reduction affects the safety diagnosis result, the purpose of the present invention is to provide a ring network data transmission method and a diagnosis system applied to coal mine underground, and the specific technical solutions adopted are as follows:

[0005] A ring network data transmission method applied to coal mine underground, the method includes:

[0006] Obtain the gas concentration data and humidity data of each monitoring node in the coal mine underground during a preset monitoring period; obtain the starting moment of mining;

[0007] In the current monitoring period, select the gas concentration data of any one of the monitoring nodes as the target data, and according to the fluctuation stability characteristics of the target data, combine the time difference between the sampling moment of each data point in the target data and the nearest starting moment of mining to obtain the reference coefficient of the target data; according to the similarity characteristics of the gas concentration data between two adjacent monitoring nodes, combine the reference coefficient to obtain the effective coefficient corresponding to the two adjacent monitoring nodes; according to the similarity characteristics of the humidity data between two adjacent monitoring nodes, combine the effective coefficient to obtain the corrected effective coefficient corresponding to the two adjacent monitoring nodes;

[0008] Under the current monitoring period, obtain the mutual monitoring coefficient of any two of the monitoring nodes according to the sum value of the modified effective coefficients of all adjacent monitoring nodes between the two monitoring nodes; according to the covariance of the gas concentration data between all any two of the monitoring nodes, combine the mutual monitoring coefficient, and use a preset dimensionality reduction algorithm to obtain a two-dimensional data set and transmit it.

[0009] Further, the method for obtaining the reference coefficient includes:

[0010] In the target data, obtain the local stability parameter of each data point according to the difference characteristics between each data point and its adjacent data point in time series; the difference characteristics between the data point and its adjacent data point in time series are negatively correlated with the local stability parameter; according to the time difference between the collection time of each data point and the nearest mining start time, obtain the fluctuation reference weight of each data point; the time difference between the collection time of the data point and the nearest mining start time is negatively correlated with the fluctuation reference weight;

[0011] According to the local stability parameter and the fluctuation reference weight of all data points in the target data, obtain the reference coefficient of the target data; both the local stability parameter and the fluctuation reference weight of the data point are positively correlated with the reference coefficient of the target data.

[0012] Further, the method for obtaining the mutual monitoring coefficient includes:

[0013] Each of the monitoring nodes transmits data to the adjacent monitoring nodes on both sides to obtain the data transmission path between any two of the monitoring nodes;

[0014] On each data transmission path between any two of the monitoring nodes, take the sum value of the modified effective coefficients of all adjacent monitoring nodes as the monitoring sub-coefficient of each data transmission path between the corresponding two monitoring nodes;

[0015] According to all the monitoring sub-coefficients between the two monitoring nodes, obtain the mutual monitoring coefficient of the corresponding two monitoring nodes; the monitoring sub-coefficient and the mutual monitoring coefficient are positively correlated.

[0016] Further, the method for obtaining a two-dimensional data set and transmitting it by using a preset dimensionality reduction algorithm according to the covariance of the gas concentration data between all any two of the monitoring nodes and combining the mutual monitoring coefficient includes:

[0017] Multiply the covariance of the gas concentration data between any two of the monitoring nodes by the mutual monitoring coefficient as the modified covariance between the corresponding two monitoring nodes; wherein the mutual monitoring coefficient is also normalized;

[0018] Based on all the corrected covariances, a two-dimensional data set is obtained using a preset dimensionality reduction algorithm and transmitted.

[0019] Furthermore, the method for obtaining the effective coefficient includes:

[0020] Under the current monitoring period, according to the similarity characteristics of the gas concentration data of two adjacent monitoring nodes, the similarity coefficient corresponding to the two adjacent monitoring nodes is obtained; the product of the similarity coefficient and the sum of the reference coefficients of the corresponding two adjacent monitoring nodes is used as the effective coefficient of the corresponding two adjacent monitoring nodes.

[0021] Furthermore, the method for obtaining the corrected effective coefficient includes:

[0022] Under the current monitoring period, among the humidity data of two adjacent monitoring nodes, according to the difference characteristics of all the humidity data with the same time sequence, the humidity consistency coefficient corresponding to the two adjacent monitoring nodes is obtained; the difference characteristics of the humidity data are negatively correlated with the humidity consistency coefficient;

[0023] According to the humidity consistency coefficient and the effective coefficient of two adjacent monitoring nodes, the corrected effective coefficient corresponding to the two adjacent monitoring nodes is obtained; both the humidity consistency coefficient and the effective coefficient are positively correlated with the corrected effective coefficient.

[0024] Furthermore, the preset dimensionality reduction algorithm is the PCA dimensionality reduction algorithm.

[0025] The present invention also proposes a ring network data diagnosis system applied to coal mine underground, and the system includes:

[0026] A data acquisition module, configured to obtain the gas concentration data and humidity data of each monitoring node in the coal mine underground during the current preset monitoring period; obtain the mining start time;

[0027] A data analysis module, configured to, under the current monitoring period, select the gas concentration data of any one of the monitoring nodes as the target data, and according to the fluctuation stability characteristics of the target data, combine the time difference between the sampling time of each data point in the target data and the nearest mining start time to obtain the reference coefficient of the target data; according to the similarity characteristics of the gas concentration data between two adjacent monitoring nodes, combine the reference coefficient to obtain the effective coefficient corresponding to the two adjacent monitoring nodes; according to the similarity characteristics of the humidity data between two adjacent monitoring nodes, combine the effective coefficient to obtain the corrected effective coefficient corresponding to the two adjacent monitoring nodes;

[0028] A data transmission module, which is used to select the correction effective coefficients of all adjacent monitoring nodes between any two of the monitoring nodes in the current monitoring period, obtain the mutual monitoring coefficients of the corresponding two monitoring nodes; and obtain a two-dimensional data set by using a preset dimensionality reduction algorithm based on the covariance of the gas concentration data between all any two monitoring nodes and combine with the mutual monitoring coefficients, and then transmit the two-dimensional data set.

[0029] A safety diagnosis module, which is used to perform safety diagnosis according to the two-dimensional data set in the current monitoring period.

[0030] Further, the method for performing safety diagnosis includes:

[0031] Obtain a scatter plot of the two-dimensional data set in the current monitoring period, and obtain the outlier degree of the data points in the scatter plot; when the outlier degree is greater than a preset danger threshold, issue a safety alarm.

[0032] Further, the method for obtaining the outlier degree of the data points includes:

[0033] Use the LOF algorithm to obtain the outlier degree of the data points in the scatter plot.

[0034] The present invention has the following beneficial effects:

[0035] The present invention first obtains the gas concentration data and humidity data of each monitoring node in the coal mine underground during a preset monitoring period and obtains the starting moment of mining, thereby obtaining the basis for data analysis. Further, in the current monitoring period, according to the fluctuation stability characteristics of the target data and in combination with the time difference between the sampling moment of each data point and the nearest starting moment of mining, the reference coefficient of the target data is obtained to characterize the reference of the target data and reduce the interference of the gas concentration data with low reference and more environmental noise. Further, according to the similarity characteristics of the gas concentration data between two adjacent monitoring nodes and in combination with the reference coefficient, the effective coefficient corresponding to the two adjacent monitoring nodes is obtained to characterize the effectiveness of the mutual monitoring between the adjacent monitoring nodes and reduce the interference of the data containing noise or distortion on the dimensionality reduction analysis. Further, from the perspective of the similarity of humidity data, the effective coefficient is further corrected to obtain the corrected effective coefficient corresponding to the two adjacent monitoring nodes, improving the accuracy of the effective coefficient and further improving the accuracy of the dimensionality reduction result. Further, in the current monitoring period, according to the sum value of the corrected effective coefficients of all adjacent monitoring nodes between any two monitoring nodes, the mutual monitoring coefficient corresponding to the two monitoring nodes is obtained to characterize the effectiveness of the mutual monitoring between any two monitoring nodes and also reflect the effectiveness of the correlation between the gas concentration data of any two monitoring nodes, preparing for finally obtaining an accurate dimensionality reduction result. Further, according to the covariance of the gas concentration data between all any two monitoring nodes and in combination with the mutual monitoring coefficient, a two-dimensional data set is obtained by using a preset dimensionality reduction algorithm and transmitted, reducing data redundancy and the data transmission pressure, and at the same time providing data support for safety diagnosis. Finally, safety diagnosis is performed according to the two-dimensional data set of the current monitoring period. By analyzing the reference of the data itself and the effectiveness of the data correlation between adjacent monitoring nodes, the present invention determines the effectiveness of the mutual monitoring between monitoring nodes during ring network transmission, corrects the covariance between the data, improves the accuracy of the dimensionality reduction result, and makes the safety diagnosis result more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0037] Figure 1 The flowchart of a ring network data transmission method applied to the coal mine underground provided by an embodiment of the present invention;

[0038] Figure 2 The schematic diagram of the relationship between adjacent monitoring sub-stations in a ring network structure provided by an embodiment of the present invention;

[0039] Figure 3 System block diagram of a ring network data diagnosis system applied to underground coal mines provided by an embodiment of the present invention;

[0040] Figure 4 Scatter plot of a two-dimensional data set provided by an embodiment of the present invention. Detailed implementation manners

[0041] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following combines the accompanying drawings and preferred embodiments to specifically describe a ring network data transmission method and diagnosis system applied to underground coal mines according to the present invention, including its specific implementation manners, structures, features, and effects in detail. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0043] The following specifically describes the specific solutions of a ring network data transmission method and diagnosis system applied to underground coal mines provided by the present invention with reference to the accompanying drawings.

[0044] Please refer to Figure 1 , which shows a flowchart of a ring network data transmission method applied to underground coal mines provided by an embodiment of the present invention, specifically including:

[0045] Step S1: Obtain the gas concentration data and humidity data of each monitoring node in the underground coal mine during a preset monitoring period; obtain the starting time of mining.

[0046] In the embodiment of the present invention, considering the influence of gas release and the ventilation system, during the coal mining process in the mine, the coal seam will gradually release gas, and this process is usually relatively stable. At the same time, the mine ventilation system can effectively regulate the gas concentration by introducing fresh air and discharging waste gas to avoid drastic fluctuations. Therefore, the gas concentration data changes relatively smoothly within a certain period of time. Therefore, the data is analyzed at a preset monitoring period. Considering that the humidity data reflects the humidity characteristics of the environment, the gas concentration data and humidity data of each monitoring node in the underground coal mine during the preset monitoring period are obtained. Considering that during coal mining, the rupture of rocks may cause the rapid release of gas, affecting the change of gas concentration data, the starting time of mining is also obtained.

[0047] As an example, multiple monitoring sub - stations are set up underground in a coal mine. Electrochemical sensors and humidity sensors are installed to be responsible for collecting gas concentration data and humidity data in real - time. Mechanical vibration sensors are installed and thresholds are set to monitor the mining state (whether mining is in progress), obtain the starting moment of mining, and record the positions of these sub - stations. The electrochemical sensors and humidity sensors collect data at a frequency of 1 hz at each monitoring sub - station. The multiple monitoring sub - stations are connected through a network to form a closed ring network structure. The length of the preset monitoring period is 5 minutes, and the current monitoring period is the latest and complete monitoring period; each monitoring sub - station is a monitoring node.

[0048] It should be noted that the number of monitoring nodes and the set thresholds of the mechanical vibration sensors are both related to the actual operation scenario. The installation and management of the monitoring system can refer to the "Summary and Arrangement of 24 Systems of Coal Mine Gas Monitoring System" on the International Coal Network. In other embodiments of the present invention, implementers can set other data collection frequencies and monitoring period lengths.

[0049] Step S2: In the current monitoring period, select the gas concentration data of any monitoring node as the target data. According to the fluctuation stability characteristics of the target data, combined with the time difference between the sampling moment of each data point in the target data and the nearest mining start moment, obtain the reference coefficient of the target data; according to the similarity characteristics of the gas concentration data between two adjacent monitoring nodes, combined with the reference coefficient, obtain the effective coefficient corresponding to the two adjacent monitoring nodes; according to the similarity characteristics of the humidity data between two adjacent monitoring nodes, combined with the effective coefficient, obtain the corrected effective coefficient corresponding to the two adjacent monitoring nodes.

[0050] For the convenience of analyzing the data of each monitoring node one by one, in the current monitoring period, select the gas concentration data of any monitoring node as the target data; considering the influence of gas release and ventilation systems, the change of gas concentration data fluctuates relatively smoothly. At the same time, since coal mining will affect the change of gas concentration data, therefore, combined with the time difference between the sampling moment of each data point in the target data and the nearest mining start moment, analyze the fluctuation stability characteristics of the target data, obtain the reference coefficient of the target data, characterize the reference of the target data, and reduce the interference of gas concentration data with low reference and more environmental noise.

[0051] Preferably, in an embodiment of the present invention, considering that the smaller the difference between a data point and its adjacent data point in time series, the smoother the data fluctuation and the stronger the local stability; at the same time, during mining, it may cause rapid gas release, resulting in large fluctuations in the data points adjacent in time domain at the mining moment. At this time, the data fluctuation is the fluctuation during normal operation, not affected by environmental noise, and still has a high reference value;

[0052] Based on this, in the target data, according to the difference characteristics between each data point and its adjacent data points in time series, the local stability parameter of each data point is obtained; the difference characteristics between a data point and its adjacent data points in time series are negatively correlated with the local stability parameter; according to the time difference between the acquisition time of each data point and the most recent mining start time, the fluctuation reference weight of each data point is obtained; the time difference between the acquisition time of a data point and the most recent mining start time is negatively correlated with the fluctuation reference weight.

[0053] According to the local stability parameters and fluctuation reference weights of all data points in the target data, the reference coefficient of the target data is obtained; both the local stability parameter and the fluctuation reference weight of a data point are positively correlated with the reference coefficient of the target data.

[0054] As an example, the calculation formula of the reference coefficient includes:

[0055] ;

[0056] In the formula, represents the reference coefficient of the target data; represents the time series serial number of the data point in the target data; represents the number of data points in the target data; represents the acquisition time of the th data point in the target data; represents the most recent mining start time; represents the th data value of the data point in the target data; represents the th data value of the data point in the target data; represents the th data value of the data point in the target data; represents the first preset non-zero positive parameter; represents the th local stability parameter of the data point in the target data; represents the th fluctuation reference weight of the data point in the target data; represents taking the absolute value.

[0057] In the calculation formula of the reference coefficient, the difference characteristics between data points are represented by the absolute value of the difference, and then the negative correlation mapping is performed by taking the reciprocal to adjust the correlation relationship. The smaller The smaller it is, it indicates that the time difference between the acquisition time of the data point and the nearest mining start time is smaller, and the local fluctuation of the data is more likely to be the normal fluctuation caused by the mining operation, with higher reference value and larger reference coefficient.

[0058] It should be noted that in the calculation formula of the reference coefficient, when and , there is only one adjacent data point in time series for the data point. At this time, only the absolute value of the numerical difference between the data point and one adjacent data point in time series is calculated to obtain the local stability parameter; in an embodiment of the present invention, ; in other embodiments of the present invention, the implementer can also use a negative correlation mapping function such as an exponential function with the natural constant as the base for negative correlation mapping, represents the independent variable, and other first preset non-zero positive parameters can also be set.

[0059] Please refer to Figure 2 , which shows a schematic diagram of the relationship between adjacent monitoring sub-stations in a ring network structure provided by an embodiment of the present invention. Figure 2 It includes monitoring sub-station 1 and monitoring sub-station 2. Each monitoring sub-station is a monitoring node, and data processing is performed within the monitoring sub-station. The arrow represents the data transmission direction; from Figure 2 it can be seen that in the ring network structure, adjacent monitoring nodes are directly connected and can transmit data to each other. The data sent by one node will be received and processed by the next adjacent node, and then forwarded to the next node. If a node fails, the design of the ring network structure allows data to be transmitted through the reverse path, thus ensuring the reliability and continuity of the transmission. Therefore, adjacent nodes can monitor the validity of the gas concentration data of each other.

[0060] To determine the validity of adjacent nodes monitoring the gas concentration data of each other, first, it is necessary to reflect the validity of mutual monitoring from the perspective of the referenceability of their respective data based on the reference coefficient characterizing the referenceability of the gas concentration data; considering that the ventilation system will make the air flow evenly and the gas diffuses in the air, and the gas concentrations of adjacent monitoring nodes are similar, so according to the similarity characteristics of the gas concentration data between two adjacent monitoring nodes, combined with the reference coefficient, the effective coefficient corresponding to the two adjacent monitoring nodes is obtained, which characterizes the validity of adjacent monitoring nodes monitoring each other, reduces the interference of data containing noise or distortion on the dimensionality reduction analysis, and prepares for obtaining an accurate dimensionality reduction result.

[0061] Preferably, in an embodiment of the present invention, considering that in the current monitoring period, the greater the similarity characteristics of the gas concentration data of two adjacent monitoring nodes, the stronger the consistency of the gas concentration data of the adjacent monitoring nodes and the stronger the effectiveness of mutual monitoring; at the same time, the greater the reference coefficient, the higher the reference value of the gas concentration data, the higher the reference value of the similarity characteristics of the gas concentration data of adjacent monitoring nodes, and the stronger the effectiveness of mutual monitoring. Based on this, according to the similarity characteristics of the gas concentration data of two adjacent monitoring nodes, the similarity coefficient corresponding to the two adjacent monitoring nodes is obtained; the product of the similarity coefficient and the sum of the reference coefficients of the corresponding two adjacent monitoring nodes is used as the effective coefficient of the corresponding two adjacent monitoring nodes.

[0062] As an example, the Pearson correlation coefficient of the gas concentration data of two adjacent monitoring nodes is obtained, and the Pearson correlation coefficient is linearly normalized, and the normalization result is used as the similarity coefficient corresponding to the two adjacent monitoring nodes, so as to obtain the effective coefficient.

[0063] In another embodiment of the present invention, the DTW distance of the gas concentration data of two adjacent monitoring nodes can also be obtained, and the DTW distance is used through function negative correlation mapping as the similarity coefficient; the cosine similarity of the gas concentration data of two adjacent monitoring nodes can also be obtained, and after linearly normalizing the cosine similarity, the normalization result is used as the similarity coefficient corresponding to the two adjacent monitoring nodes. Among them, the Pearson correlation coefficient, DTW distance and cosine similarity are all prior arts and will not be elaborated here.

[0064] Considering that the environmental humidity between adjacent monitoring nodes is also relatively consistent, the effective coefficient can be further corrected from the perspective of humidity data similarity. At the same time, humidity is related to air flow. If the humidity consistency between adjacent nodes is strong, it means that the air flow state is better, and the similarity characteristics of the gas concentration of adjacent monitoring nodes can be supplemented and verified. Therefore, according to the similarity characteristics of the humidity data between two adjacent monitoring nodes, combined with the effective coefficient, the corrected effective coefficient corresponding to the two adjacent monitoring nodes is obtained, the accuracy of the effective coefficient is improved, and then the accuracy of the dimensionality reduction result is improved, and finally the accuracy and diagnostic effect of coal mine safety monitoring are improved.

[0065] Preferably, in an embodiment of the present invention, considering that the smaller the difference in humidity data with the same time sequence, the more similar the humidity data of two adjacent monitoring nodes, the more consistent the humidity, the better the air flow state is reflected, the higher the accuracy of the effective coefficient, and the higher the credibility. Therefore, in the current monitoring period, in the humidity data of two adjacent monitoring nodes, according to the difference characteristics of all humidity data with the same time sequence, the humidity consistency coefficient corresponding to the two adjacent monitoring nodes is obtained; the difference characteristics of the humidity data are negatively correlated with the humidity consistency coefficient;

[0066] Obtain the corrected effective coefficient corresponding to two adjacent monitoring nodes according to the humidity consistency coefficient and the effective coefficient of the two adjacent monitoring nodes; both the humidity consistency coefficient and the effective coefficient are positively correlated with the corrected effective coefficient.

[0067] As an example, assume that monitoring node A and monitoring node B are adjacent, and A transmits to B. The calculation formula for the corrected effective coefficient includes:

[0068] ;

[0069] In the formula, is the serial number of the monitoring node; is the serial number of the monitoring node; represents two adjacent monitoring nodes and monitoring node 's corrected effective coefficient; represents the time sequence serial number of the data point in the current monitoring period; represents the number of data points in the current monitoring period; represents monitoring node in the current monitoring period at the th data point's humidity value; represents monitoring node in the current monitoring period at the th data point's humidity value; represents the second preset non-zero positive parameter; represents two adjacent monitoring nodes and monitoring node 's effective coefficient; represents taking the absolute value; represents two adjacent monitoring nodes and monitoring node 's humidity consistency coefficient. In this example, .

[0070] In the calculation formula of the corrected effective coefficient, the difference characteristics of the humidity data with the same time sequence are represented by the absolute value of the difference, and then the negative correlation mapping is performed by taking the reciprocal to adjust the correlation relationship, reflecting the similar characteristics of the humidity data, and obtaining the humidity consistency coefficient; The smaller it is, the smaller the difference of the humidity data with the same time sequence, the stronger the similar characteristics, the better the air flow state is reflected, the higher the accuracy of the effective coefficient, the higher the credibility, the larger the humidity consistency coefficient, and the larger the corrected effective coefficient; The larger it is, the stronger the similar characteristics of the gas concentration data, the greater the effectiveness of mutual monitoring between adjacent monitoring nodes, and the larger the corrected effective coefficient.

[0071] It should be noted that the humidity data is only used for calculating and correcting the effective coefficient, transmitted and processed between adjacent monitoring nodes, and there is no need to transmit the humidity data to other non-adjacent monitoring nodes. Only the corrected effective coefficient is transmitted to reduce transmission redundancy. For example, the transmission order is ABCD. The humidity data of A does not need to be transmitted to C and D, and only is transmitted to C and D.

[0072] In other embodiments of the present invention, the implementer can also use a negative correlation mapping function such as function for negative correlation mapping, and other preset non-zero positive parameters can also be set.

[0073] Step S3: In the current monitoring period, according to the sum of the corrected effective coefficients of all adjacent monitoring nodes between any two monitoring nodes, obtain the mutual monitoring coefficient of the corresponding two monitoring nodes; according to the covariance of the gas concentration data between all any two monitoring nodes, combined with the mutual monitoring coefficient, use a preset dimensionality reduction algorithm to obtain a two-dimensional data set and transmit it.

[0074] Through the ring network transmission, the gas concentration data of each monitoring node in the current monitoring period and the corrected effective coefficient between each monitoring node and its adjacent monitoring node are received. However, the effectiveness of mutual monitoring between non-adjacent transmission nodes is still unknown. Since the ring network transmission is passed layer by layer, each node only forwards the received data to the next adjacent node after receiving the data. This layer-by-layer forwarding mechanism limits the direct communication between nodes and can only be transmitted through adjacent nodes. Therefore, the effectiveness of mutual supervision between two non-adjacent nodes is the cumulative sum of the effectiveness of adjacent nodes in the transmission process. Therefore, in the current monitoring period, according to the sum of the corrected effective coefficients of all adjacent monitoring nodes between any two monitoring nodes, obtain the mutual monitoring coefficient of the corresponding two monitoring nodes, which represents the effectiveness of mutual monitoring between any two monitoring nodes and also reflects the effectiveness of the correlation between the gas concentration data of any two monitoring nodes, preparing for obtaining an accurate dimensionality reduction result.

[0075] Preferably, in an embodiment of the present invention, considering that when the ring network structure is intact, there are two transmission paths between any two monitoring nodes, and both transmission paths need to be analyzed. Therefore, each monitoring node transmits data to the adjacent monitoring nodes on both sides to obtain the data transmission path between any two monitoring nodes;

[0076] On each data transmission path between any two monitoring nodes, take the sum of the corrected effective coefficients of all adjacent monitoring nodes as the monitoring sub-coefficient of each data transmission path of the corresponding two monitoring nodes;

[0077] Obtain the mutual monitoring coefficient corresponding to two monitoring nodes according to all the monitoring sub - coefficients between the two monitoring nodes; the monitoring sub - coefficients and the mutual monitoring coefficient are positively correlated.

[0078] As an example, take the average value of all the monitoring sub - coefficients between two monitoring nodes as the mutual monitoring coefficient corresponding to the two monitoring nodes.

[0079] For example, the ring network structure is ABCDE, where E is adjacent to A. At this time, there are two data transmission paths between A and C, namely ABC and AEDC, and the monitoring sub - coefficients are respectively 、 ,Take the average value of and as the mutual monitoring coefficient corresponding to A and C.

[0080] It should be noted that in the ring network structure, a failure may occur at one point. At this time, there is only 1 data transmission path between two monitoring nodes. For example, the transmission line between B and C is interrupted. At this time, there is only the data transmission path of AEDC between A and C. Take as the mutual monitoring coefficient corresponding to A and C.

[0081] Since the ring network equipment and lines are regularly maintained and inspected, the probability of failure is relatively low. In the embodiments of the present invention, only the situation of at most 1 failure occurring simultaneously is considered. When 2 or more failures occur simultaneously, it means that the ring network transmission equipment cannot perform safety monitoring on the gas concentration in the coal mine underground, and shutdown maintenance is required in time.

[0082] After characterizing the effectiveness of the correlation between the gas concentration data of two monitoring nodes through the mutual monitoring coefficient, the covariance of the gas concentration data between all any two monitoring nodes can be combined with the mutual monitoring coefficient, and a preset dimensionality reduction algorithm can be used to obtain a two - dimensional data set and transmit it, reducing data redundancy and the data transmission pressure.

[0083] Preferably, in an embodiment of the present invention, considering that the covariance represents the correlation of the gas concentration data of two monitoring nodes, and the mutual monitoring coefficient reflects the transmission reliability between them. By multiplying the covariance by the mutual monitoring coefficient, the corrected covariance not only reflects the similarity degree between the data, but also comprehensively considers the transmission reliability of the data. This correction method can effectively reduce the influence of noise or low - reliability transmission on the final dimensionality reduction result, ensuring that the data input into the dimensionality reduction algorithm is more accurate. Based on this, take the product of the covariance and the mutual monitoring coefficient of the gas concentration data between any two monitoring nodes as the corrected covariance corresponding to the two monitoring nodes; where the mutual monitoring coefficient has also been normalized;

[0084] Based on all the corrected covariances, use a preset dimensionality reduction algorithm to obtain a two - dimensional data set and transmit it.

[0085] As an example, based on the mutual monitoring coefficients between all pairs of monitoring nodes, linear normalization is performed on the mutual monitoring coefficients; the product of the normalized mutual monitoring coefficient and the covariance of the gas concentration data between the corresponding pair of monitoring nodes is used as the corrected covariance between the corresponding pair of monitoring nodes. Based on all the corrected covariances, an optimized covariance matrix is obtained, and then the principal component analysis algorithm (PCA) is used to obtain the two-dimensional dataset after dimensionality reduction and transmit it.

[0086] It should be noted that in other embodiments of the present invention, the implementer can also perform dimensionality reduction based on other dimensionality reduction algorithms of the covariance matrix, such as the linear discriminant analysis algorithm (LDA).

[0087] The embodiment of the present invention also proposes a ring network data diagnosis system applied to coal mines. Please refer to Figure 3 , which shows the system block diagram of a ring network data diagnosis system applied to coal mines provided by an embodiment of the present invention. The system includes: a data acquisition module 101, a data analysis module 102, a data transmission module 103, and a safety diagnosis module 104, specifically including:

[0088] The data acquisition module 101 is used to obtain the gas concentration data and humidity data of each monitoring node in the coal mine during the currently preset monitoring period; obtain the start time of mining.

[0089] The data analysis module 102 is used to select the gas concentration data of any monitoring node as the target data during the current monitoring period. According to the fluctuation stability characteristics of the target data, combined with the time difference between the sampling time of each data point in the target data and the nearest mining start time, the reference coefficient of the target data is obtained; according to the similarity characteristics of the gas concentration data between two adjacent monitoring nodes, combined with the reference coefficient, the effective coefficient of the corresponding two adjacent monitoring nodes is obtained; according to the similarity characteristics of the humidity data between two adjacent monitoring nodes, combined with the effective coefficient, the corrected effective coefficient of the corresponding two adjacent monitoring nodes is obtained.

[0090] The data transmission module 103 is used to select the corrected effective coefficients of all adjacent monitoring nodes between any two monitoring nodes during the current monitoring period, and obtain the mutual monitoring coefficients of the corresponding two monitoring nodes; according to the covariance of the gas concentration data between all pairs of any two monitoring nodes, combined with the mutual monitoring coefficients, a two-dimensional dataset is obtained using a preset dimensionality reduction algorithm and transmitted.

[0091] The safety diagnosis module 104 is used to perform safety diagnosis based on the two-dimensional dataset of the current monitoring period.

[0092] Among them, the implementation processes of the data acquisition module 101, the data analysis module 102, and the data transmission module 103 have been described in a ring network data transmission method applied to underground coal mines, and will not be elaborated here.

[0093] After obtaining the two-dimensional data set after dimensionality reduction, safety diagnosis can be performed based on the two-dimensional data set of the current monitoring period to ensure the safety of coal mine operations.

[0094] Preferably, in an embodiment of the present invention, considering that a scatter plot can easily and intuitively analyze the distribution characteristics of data points, thereby identifying abnormal distributions, identifying abnormal gas concentrations underground in coal mines, and giving safety alerts, a scatter plot of the two-dimensional data set of the current monitoring period is obtained, and the outlier degree of the data points in the scatter plot is obtained; when the outlier degree is greater than a preset danger threshold, a safety alert is given.

[0095] As an example, please refer to Figure 4 , which shows a scatter plot of the two-dimensional data set provided by an embodiment of the present invention. The horizontal axis is dimension 1, and the vertical axis is dimension 2. The scatter plot is obtained by using the t-SNE (t-distributed Stochastic Neighbor Embedding) method, and the outlier degree of the data points in the scatter plot is obtained by using the LOF algorithm. The preset danger threshold is set to 10. When the outlier degree is greater than 10, it is considered that there is an abnormality, and a safety alert is automatically given to remind the staff to check for safety hazards in time and improve the safety of underground coal mine operations.

[0096] In other embodiments of the present invention, the implementer can also use a density-based clustering algorithm, such as the DBSCAN algorithm, to identify the number of outliers, set a safety threshold for the number of outliers, and give a safety alert.

[0097] In summary, in view of the technical problem that inaccurate dimensionality reduction affects the safety diagnosis result, the present invention proposes a ring network data transmission method and diagnosis system applied to underground coal mines. The present invention first obtains gas concentration data, humidity data, and the mining start time; further, in the current monitoring period, according to the fluctuation stability characteristics of the target data, combined with the time difference between the sampling time of each data point and the nearest mining start time, the reference coefficient of the target data is obtained; according to the similarity characteristics of the gas concentration data and the similarity characteristics of the humidity data between two adjacent monitoring nodes, combined with the reference coefficient, the correction effective coefficient corresponding to the two adjacent monitoring nodes is obtained; further, according to the sum value of all the correction effective coefficients between any two monitoring nodes, the mutual monitoring coefficient corresponding to the two monitoring nodes is obtained, combined with the corresponding covariance, and a two-dimensional data set is obtained and transmitted by using a preset dimensionality reduction algorithm. Finally, safety diagnosis is performed based on the two-dimensional data set.

[0098] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0099] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the key point of each embodiment is to illustrate the differences from other embodiments.

Claims

1. A ring network data transmission method used in coal mines, characterized in that: The method comprises: Obtain the gas concentration data and humidity data of each monitoring node in the coal mine in the preset monitoring period; obtain the mining start time; In the current monitoring cycle, the gas concentration data of any of the monitoring nodes is selected as the target data, and the reference coefficient of the target data is obtained according to the fluctuation stability characteristics of the target data and the time difference between the sampling time of each data point in the target data and the most recent mining start time; according to the similar characteristics of the gas concentration data between two adjacent monitoring nodes and the reference coefficient, the effective coefficient corresponding to the two adjacent monitoring nodes is obtained; according to the similar characteristics of the humidity data between two adjacent monitoring nodes and the effective coefficient, the corrected effective coefficient corresponding to the two adjacent monitoring nodes is obtained; In the current monitoring cycle, the mutual monitoring coefficient of the corresponding two monitoring nodes is obtained according to the sum of the modified effective coefficients of all adjacent monitoring nodes between any two monitoring nodes; according to the covariance of the gas concentration data between all any two monitoring nodes, combined with the mutual monitoring coefficient, a two-dimensional data set is obtained and transmitted using a preset dimensionality reduction algorithm; The method for obtaining the mutual monitoring coefficient includes: Each of the monitoring nodes transmits data to the adjacent monitoring nodes on both sides to obtain a data transmission path between any two of the monitoring nodes; On each of the data transmission paths between any two of the monitoring nodes, the sum of the modified effective coefficients of all adjacent monitoring nodes is used as the monitoring sub-coefficient of each of the data transmission paths corresponding to the two monitoring nodes; According to all the monitoring sub-coefficients between the two monitoring nodes, a mutual monitoring coefficient corresponding to the two monitoring nodes is obtained; the monitoring sub-coefficients and the mutual monitoring coefficient are positively correlated.

2. The ring network data transmission method used in coal mines according to claim 1, characterized in that: The method for obtaining the reference coefficient includes: In the target data, the local stability parameter of each data point is obtained according to the difference characteristics between each data point and the adjacent data points in time series; the difference characteristics between the data point and the adjacent data points in time series are negatively correlated with the local stability parameter; the fluctuation reference weight of each data point is obtained according to the time difference between the collection time of each data point and the most recent mining start time; the time difference between the collection time of the data point and the most recent mining start time is negatively correlated with the fluctuation reference weight; The reference coefficient of the target data is obtained according to the local stability parameters and the fluctuation reference weights of all data points in the target data; the local stability parameters and the fluctuation reference weights of the data points are positively correlated with the reference coefficient of the target data.

3. The ring network data transmission method used in coal mines according to claim 1, characterized in that: The method of obtaining and transmitting a two-dimensional data set by using a preset dimensionality reduction algorithm based on the covariance of the gas concentration data between any two of the monitoring nodes and in combination with the mutual monitoring coefficient includes: The covariance of the gas concentration data between any two of the monitoring nodes and the mutual monitoring coefficient are multiplied to obtain a corrected covariance between the corresponding two monitoring nodes; wherein the mutual monitoring coefficient is also normalized; Based on all the modified covariances, a two-dimensional data set is acquired and transmitted using a preset dimensionality reduction algorithm.

4. The ring network data transmission method used in coal mines according to claim 1, characterized in that: The method for obtaining the effective coefficient includes: In the current monitoring cycle, based on the similar characteristics of the gas concentration data of the two adjacent monitoring nodes, the similarity coefficients corresponding to the two adjacent monitoring nodes are obtained; the product of the similarity coefficient and the reference coefficient and value corresponding to the two adjacent monitoring nodes is used as the effective coefficient corresponding to the two adjacent monitoring nodes.

5. The ring network data transmission method used in coal mines according to claim 1, characterized in that: The method for obtaining the modified effective coefficient includes: In the current monitoring cycle, in the humidity data of two adjacent monitoring nodes, according to the difference characteristics of all the humidity data with the same time series, the humidity consistency coefficient corresponding to the two adjacent monitoring nodes is obtained; the difference characteristics of the humidity data are negatively correlated with the humidity consistency coefficient; According to the humidity consistency coefficient and the effective coefficient of two adjacent monitoring nodes, the corrected effective coefficient corresponding to the two adjacent monitoring nodes is obtained; the humidity consistency coefficient and the effective coefficient are both positively correlated with the corrected effective coefficient.

6. The ring network data transmission method used in coal mines according to claim 1, characterized in that: The preset dimensionality reduction algorithm is the PCA dimensionality reduction algorithm.

7. A ring network data diagnosis system used in coal mines, characterized in that: The system comprises: The data acquisition module is used to obtain the gas concentration data and humidity data of each monitoring node in the coal mine in the current preset monitoring period; obtain the mining start time; A data analysis module is used to select the gas concentration data of any of the monitoring nodes as the target data in the current monitoring cycle, and obtain the reference coefficient of the target data based on the fluctuation stability characteristics of the target data and the time difference between the sampling time of each data point in the target data and the most recent mining start time; obtain the effective coefficient corresponding to the two adjacent monitoring nodes based on the similar characteristics of the gas concentration data between the two adjacent monitoring nodes and the reference coefficient; obtain the corrected effective coefficient corresponding to the two adjacent monitoring nodes based on the similar characteristics of the humidity data between the two adjacent monitoring nodes and the effective coefficient; A data transmission module, used to select the corrected effective coefficients of all adjacent monitoring nodes between any two monitoring nodes in the current monitoring cycle, and obtain the mutual monitoring coefficient of the corresponding two monitoring nodes; according to the covariance of the gas concentration data between all any two monitoring nodes, combined with the mutual monitoring coefficient, a preset dimensionality reduction algorithm is used to obtain and transmit a two-dimensional data set; A safety diagnosis module, used for performing safety diagnosis based on the two-dimensional data set of the current monitoring period; The method for obtaining the mutual monitoring coefficient includes: Each of the monitoring nodes transmits data to the adjacent monitoring nodes on both sides to obtain a data transmission path between any two of the monitoring nodes; On each of the data transmission paths between any two of the monitoring nodes, the sum of the modified effective coefficients of all adjacent monitoring nodes is used as the monitoring sub-coefficient of each of the data transmission paths corresponding to the two monitoring nodes; According to all the monitoring sub-coefficients between the two monitoring nodes, a mutual monitoring coefficient corresponding to the two monitoring nodes is obtained; the monitoring sub-coefficients and the mutual monitoring coefficient are positively correlated.

8. The ring network data diagnosis system used in coal mines according to claim 7, characterized in that: The method for performing safety diagnosis comprises: A scatter plot of the two-dimensional data set in the current monitoring period is obtained, and the outlier degree of the data points in the scatter plot is obtained; when the outlier degree is greater than a preset danger threshold, a safety alarm is issued.

9. The ring network data diagnosis system used in coal mines according to claim 8, characterized in that: The method for obtaining the outlier degree of the data point includes: The outlier degree of the data points in the scatter plot is obtained using the LOF algorithm.

Images