Fire monitoring method based on optical fiber temperature measurement system for mines

Abstract

The present invention relates to the field of data processing technology, and in particular to a fire monitoring method based on a mine optical fiber temperature measurement system, including: obtaining a temperature monitoring data sequence at multiple positions in a mine tunnel and the distance between any two positions; obtaining the degree of change of the temperature monitoring data sequence at each sampling moment at each position; obtaining the confidence at each sampling moment; obtaining the connectivity threshold at each sampling moment, and combining the confidence at each sampling moment to obtain the optimal connectivity threshold; according to the optimal connectivity threshold, through cluster analysis based on dynamic splitting of a connectivity graph, obtaining the reference of each temperature monitoring data, and then obtaining the predicted temperature value of each position at the next moment, and comparing it with the early warning value to realize accurate monitoring of the fire in the mine tunnel. The present invention realizes accurate monitoring of the fire based on the mine optical fiber temperature measurement system by adjusting the connectivity graph node weight coefficient according to the regional variation characteristics of the temperature data.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a fire monitoring method based on a mine optical fiber temperature measurement system. Background Art

[0002] When deploying fiber optic sensors in mines, the optical fibers are usually wrapped around the walls of mine tunnels or buried underground. Fiber optic sensors can sense temperature changes through the thermal sensitivity of optical fibers. Through continuously monitored temperature data, the development trend of the fire can be predicted and analyzed, and corresponding measures can be taken in advance. However, due to the complex environmental factors in the mine, the pipelines and equipment in the local area will affect the temperature data measured by the fiber optic sensor, which is not conducive to the accurate extraction of the temperature data change characteristics caused by the fire, and may reduce the accuracy of fire monitoring. Summary of the invention

[0003] The invention provides a fire monitoring method based on a mine optical fiber temperature measurement system to solve the existing problems.

[0004] The fire monitoring method based on the mine optical fiber temperature measurement system of the present invention adopts the following technical scheme:

[0005] An embodiment of the present invention provides a fire monitoring method based on a mine optical fiber temperature measurement system, the method comprising the following steps:

[0006] Acquire a temperature monitoring data sequence at multiple locations in a mine tunnel and the distance between any two locations; each temperature monitoring data in the temperature monitoring data sequence corresponds to a sampling time;

[0007] According to the difference between the temperature monitoring data in the temperature monitoring data sequence at each location and the distribution of the temperature monitoring data, the significance of the change of the temperature monitoring data sequence at each sampling time at each location is obtained;

[0008] According to the difference between the significant degree of changes in the temperature monitoring data series at different sampling times at the same location, and the distribution of the temperature monitoring data at different locations at different sampling times, the confidence of each sampling time is obtained; the connectivity threshold of each sampling time is obtained, and the optimal connectivity threshold is obtained in combination with the confidence of each sampling time;

[0009] According to the optimal connectivity threshold, cluster analysis based on dynamic splitting of the connectivity graph is used to obtain the reference of each temperature monitoring data, and then the predicted temperature value of each location at the next moment is obtained, which is then compared with the warning value to achieve accurate monitoring of fire conditions in mine tunnels.

[0010] Furthermore, the method of obtaining the significance of the change of the temperature monitoring data sequence at each location at each sampling moment according to the difference between the temperature monitoring data in the temperature monitoring data sequence at each location and the distribution of the temperature monitoring data includes the following specific steps:

[0011] The temperature monitoring data sequence at each position is derived using the first-order derivative method to obtain several extreme value points. The temperature monitoring data sequence at each position is divided into several temperature change time periods using all extreme value points as segmentation points.

[0012] According to the temperature monitoring data sequence at each position and several temperature change time periods at each position, the time series change significance of the temperature monitoring data at each sampling moment at each position is obtained;

[0013] According to the time series change significance of the temperature monitoring data at each sampling moment at each position, the change significance of the temperature monitoring data sequence at each sampling moment at each position is obtained.

[0014] Furthermore, the time series change significance of the temperature monitoring data at each sampling moment at each position is obtained according to the temperature monitoring data sequence at each position and a number of temperature change time periods at each position. The corresponding specific formula is:

[0015]

[0016] Among them, TX i，j Indicates the significance of the temporal change of the temperature monitoring data at the jth sampling moment at the ith position, l i，j represents the number of data in the temperature change time period where the temperature monitoring data at the jth sampling moment at the ith position is located, max(l i ) represents the maximum value of the data quantity in all temperature change time periods at the i-th position, c i，j represents the temperature monitoring data at the jth sampling time at the ith position, c i，j-1 represents the temperature monitoring data at the j-1st sampling moment at the ith position, and || represents the absolute value function.

[0017] Furthermore, the significant degree of change of the temperature monitoring data sequence at each sampling moment at each position is obtained according to the significant degree of time series change of the temperature monitoring data at each sampling moment at each position, and the corresponding specific formula is:

[0018]

[0019] Among them, WX i，j Indicates the significant degree of change in the temperature monitoring data sequence at the jth sampling moment at the ith position, TXi，j Indicates the significance of the temporal change of the temperature monitoring data at the jth sampling moment at the ith position, represents the mean value of the time series change significance of the temperature monitoring data at the jth sampling moment at all locations, c i，j represents the temperature monitoring data at the jth sampling time at the ith position, max(c j ) represents the maximum value of the temperature monitoring data at the jth sampling moment at all locations, d i，j，max is the distance between the i-th position and the position of the maximum value of all temperature monitoring data at the j-th sampling moment, || represents the absolute value function, and norm represents the normalization function.

[0020] Furthermore, the confidence level of each sampling moment is obtained according to the difference between the significant degrees of changes in the temperature monitoring data sequences at different sampling moments at the same location and the distribution of the temperature monitoring data at different locations at different sampling moments, and the specific steps include the following:

[0021] The PCA principal component analysis algorithm is used to process the temperature monitoring data at each sampling moment at all positions to obtain the principal component direction angle corresponding to the temperature monitoring data at each sampling moment at all positions;

[0022] The sampling moments other than the j-th sampling moment are recorded as the reference moments of the j-th sampling moment;

[0023] The sequence consisting of the change significance of the temperature monitoring data at any sampling time at all positions is recorded as the change significance sequence at the sampling time;

[0024] Use the DTW algorithm to match the change significance sequences of any two sampling moments, and obtain the DTW distance of the change significance sequences at different sampling moments;

[0025] The confidence level at each sampling moment is obtained based on the principal component direction angle corresponding to the temperature monitoring data at each sampling moment at all locations and the DTW distance of the change significance sequence at different sampling moments.

[0026] Furthermore, the confidence at each sampling moment is obtained according to the principal component direction angle corresponding to the temperature monitoring data at each sampling moment at all positions, and the DTW distance of the change significance sequence at different sampling moments. The corresponding specific formula is:

[0027]

[0028] Among them, LB j Indicates the confidence level at the jth sampling moment; DTW j，yrepresents the DTW distance of the sequence of significant changes between the jth sampling moment and its yth reference moment; Y represents the number of reference moments; θ j represents the principal component direction angle corresponding to the temperature monitoring data at the jth sampling moment; θ j，y It represents the principal component direction angle corresponding to the temperature monitoring data at the yth reference time at the jth sampling time; || is the absolute value function; exp() is an exponential function with a natural constant as the base.

[0029] Furthermore, the method of obtaining the connectivity threshold at each sampling moment and combining the confidence at each sampling moment to obtain the optimal connectivity threshold includes the following specific steps:

[0030] The temperature monitoring data at each sampling moment at all locations are processed by the CABDDCG algorithm to obtain the connectivity threshold at each sampling moment;

[0031] The sampling time with a confidence level greater than a preset confidence threshold is recorded as the target sampling time;

[0032] The optimal connectivity domain value is obtained according to the number of target sampling moments, the connectivity threshold and the confidence of the target sampling moment.

[0033] Furthermore, the optimal connectivity domain value is obtained according to the number of target sampling moments, the connectivity threshold and the confidence of the target sampling moment, and the corresponding specific formula is:

[0034]

[0035] in, represents the optimal connectivity threshold, R represents the number of target sampling moments, LB r represents the confidence level at the rth target sampling moment, θ r It represents the connectivity threshold at the rth target sampling moment, and softmax() is the s0ftmax normalization function.

[0036] Furthermore, according to the optimal connectivity threshold, the reference of each temperature monitoring data is obtained through cluster analysis based on dynamic splitting of the connectivity graph, and then the predicted temperature value of each location at the next moment is obtained, including the following specific steps:

[0037] According to the optimal connectivity threshold, the CABDDCG algorithm is used to divide the temperature monitoring data of all locations at each sampling time into several clusters;

[0038] Calculate the mean of the number of temperature monitoring data of all clusters at any sampling time, and use the quotient of the temperature monitoring data of each cluster at the sampling time and the mean as the reference of each temperature monitoring data in all clusters at the time;

[0039] According to the reference and temperature value of each temperature monitoring data in the temperature monitoring data sequence at each position, an updated temperature value sequence at each position is obtained;

[0040] The updated temperature value sequence is input into the ARIMA model to obtain the predicted temperature value for each location at the next moment in the future.

[0041] Further, the step of obtaining an updated temperature value sequence at each position according to the reference and temperature value of each temperature monitoring data in the temperature monitoring data sequence at each position includes the following specific steps:

[0042] According to the reference and temperature value of each temperature monitoring data in the temperature monitoring data sequence, the updated temperature value of each temperature monitoring data in the temperature monitoring data sequence at each position is obtained. The corresponding specific formula is:

[0043]

[0044] in, represents the updated temperature value of the kth temperature monitoring data in the temperature monitoring data sequence at the ith position, c i，k+1 represents the k+1th temperature monitoring data in the temperature monitoring data sequence at the ith position; c i，k-1 represents the k-1th temperature monitoring data in the temperature monitoring data sequence at the ith position, μ i，k represents the reference of the kth temperature monitoring data in the temperature monitoring data sequence at the ith position, c i，k represents the temperature value of the kth temperature monitoring data in the monitoring data sequence at the i-th position;

[0045] A sequence consisting of updated temperature values ​​of all temperature monitoring data in a temperature monitoring data sequence at any position is recorded as an updated temperature value sequence at the position.

[0046] The beneficial effects of the technical solution of the present invention are:

[0047] By analyzing and processing the temperature monitoring data obtained by the multi-position optical fiber temperature sensor with the CABDDCG algorithm, combined with the structural distribution of the connectivity graph, the connectivity threshold is adjusted, and finally the temperature monitoring data is accurately processed. A temperature prediction model based on the temperature change trend is constructed, which makes the prediction results of the temperature prediction model based on the temperature change trend more accurate, and realizes the accurate monitoring of the fire situation based on the mine optical fiber temperature measurement system, effectively avoiding the influence of temperature data fluctuations caused by environmental factors on the temperature data feature extraction. Specifically, firstly, several temperature change time periods are divided by the derivative value of the temperature monitoring data change curve on a single optical fiber temperature sensor to improve the sensitivity to data changes and capture the slight fluctuations of temperature changes more carefully; then, based on the significant performance of the time series changes of the data points in the temperature change time period, the time series change significance of the data points is obtained, which helps to identify the change trend of each data point in different time periods and improve the recognition ability of abnormal temperature data; according to the consistency of the time series change significance of multiple data points at the same sampling time, combined with the distribution information on the regional position, the change significance of the data points is obtained, and the time series and The spatial distribution improves the accuracy of the significance of changes and reduces the impact of fluctuations in a single data point. Then, according to the structural characteristics of the significance of changes of multiple data points at the same sampling moment and the structural similarity at multiple moments, the connectivity threshold confidence of the connectivity threshold is obtained. The structural characteristics and moment similarity are used to enhance the stability and reliability of the connectivity threshold and avoid the interference of environmental factors. Finally, based on the obtained connectivity threshold confidence of the connectivity threshold and combined with the confidence threshold, the connectivity threshold in the CABDDCG algorithm is adjusted. The connectivity threshold is adjusted dynamically to make the algorithm more flexible and ensure that abnormal and normal temperature data can be accurately distinguished under different environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 The present invention is a flowchart of the steps of the fire monitoring method based on the optical fiber temperature measurement system for mines. DETAILED DESCRIPTION

[0050] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the fire monitoring method based on the optical fiber temperature measurement system for mining proposed by the present invention, its specific implementation method, structure, characteristics and effects are described in detail as follows in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0051] Unless defined otherwise, 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 belongs.

[0052] The specific scheme of the fire monitoring method based on the optical fiber temperature measurement system for mining provided by the present invention is described in detail below with reference to the accompanying drawings.

[0053] See also Figure 1 , which shows a flow chart of the steps of a fire monitoring method based on a mine optical fiber temperature measurement system provided by an embodiment of the present invention, the method comprising the following steps:

[0054] Step S001: Acquire temperature monitoring data sequences at multiple locations in a mine tunnel and the distance between any two locations.

[0055] Due to the complex environmental factors in mine tunnels and the presence of a variety of mining equipment, heat will be lost from equipment components and cables, resulting in varying degrees of changes in the temperature data of multiple fiber optic temperature sensors. Fire monitoring depends on the rapid and significant changes in the temperature of local areas and nearby locations.

[0056] Specifically, this embodiment sets up multiple optical fiber temperature measuring devices at different positions in the mine tunnel, determines the installation positions of the optical fiber temperature measuring sensors according to the layout of the coal mine tunnel and the specific implementation environment, and records the relative distances between these positions. Thus, the temperature monitoring data of the optical fiber temperature measuring sensors at each position can be collected at different sampling time points.

[0057] At this point, a temperature monitoring data sequence is obtained at each position. Each position corresponds to a three-dimensional coordinate, that is, each position corresponds to the length, width and height in the mine tunnel. The distance between positions is calculated based on the three-dimensional coordinates. Each temperature monitoring data in the temperature monitoring data sequence corresponds to a moment.

[0058] Step S002: Obtain the significance of the change of the temperature monitoring data sequence at each location at each sampling moment according to the difference between the temperature monitoring data in the temperature monitoring data sequence at each location and the distribution of the temperature monitoring data.

[0059] It should be noted that in the process of clustering and analyzing the temperature monitoring data of the fiber optic temperature sensors at multiple locations using the CABDDCG algorithm, a connectivity graph can be constructed based on the fiber optic temperature measurement data at multiple locations at a single sampling moment to perform abnormal data analysis. However, when measuring the similarity between temperature monitoring data sequences, the similarity of the temperature data of the temperature monitoring data sequence at a single sampling moment is not only related to the data difference at its position, but also to the change status of its respective time series. Therefore, by combining the significant characteristics of the time series changes of the temperature monitoring data sequence at a single optical position, and combining the significant change consistency characteristics at multiple positions, the significant degree of change of the temperature monitoring data sequence can be obtained. The full name of the CABDDCG algorithm in English is Clustering Algorithm Based on Dynamic Division of Connected Graph, and its Chinese name is Clustering Algorithm Based on Dynamic Division of Connected Graph. This is a well-known technology, and the specific method will not be introduced here.

[0060] It should be further explained that, for the temperature change curve formed by the temperature monitoring data of any optical fiber temperature sensor changing with time, the temperature change curve is divided into time periods by the data points that are 0 after the derivation of the temperature change curve, and several temperature change time periods are obtained; and the significance of the time series change of the temperature monitoring data sequence is obtained thereby.

[0061] Specifically, the first-order derivative method is used to derive the temperature monitoring data sequence at any position to obtain several extreme points. The temperature monitoring data sequence at each position is divided into several temperature change time periods using all extreme points as segmentation points. Taking the i-th position as an example, according to the several temperature change time periods divided by the temperature monitoring data sequence at the i-th position, the specific calculation formula corresponding to the significance of the time series change of the temperature monitoring data at the j-th sampling moment at the i-th position can be obtained:

[0062]

[0063] Among them, TX i，j Indicates the significance of the temporal change of the temperature monitoring data at the jth sampling moment at the ith position, l i，j represents the number of data in the temperature change time period where the temperature monitoring data at the jth sampling moment at the ith position is located, max(l i ) represents the maximum value of the data quantity in all temperature change time periods at the i-th position, c i，j represents the temperature monitoring data at the jth sampling time at the ith position, c i，j-1 represents the temperature monitoring data at the j-1th sampling time at the ith position, and || represents the absolute value function.

[0064] It should be noted that the smaller the difference in temperature monitoring data between adjacent moments, the smaller the time series change of the temperature monitoring data, the larger the amount of data in the temperature change time period, the less obvious the corresponding time series change, and the smaller the significance of the corresponding time series change.

[0065] Furthermore, according to the above method, the significance of the time series change of the temperature monitoring data at each sampling moment at each position is obtained.

[0066] It should be noted that, for multiple temperature monitoring data sequences at a single sampling moment, the more similar the significance of the time series changes of the temperature monitoring data sequences are and the larger the temperature data difference is, the greater the significance of the change of the temperature monitoring data sequences is, thereby obtaining the significance of the change of the temperature monitoring data sequences.

[0067] Specifically, according to the time series change significance of the temperature monitoring data at each sampling moment at each position, the specific calculation formula for the change significance of the temperature monitoring data at each sampling moment at each position is obtained as follows:

[0068]

[0069] Among them, WX i，j Indicates the significant degree of change in the temperature monitoring data sequence at the jth sampling moment at the ith position, TX i，j Indicates the significance of the temporal change of the temperature monitoring data at the jth sampling moment at the ith position, represents the mean value of the time series change significance of the temperature monitoring data at the jth sampling moment at all locations, c i，j represents the temperature monitoring data at the jth sampling time at the ith position, max(c j ) represents the maximum value of the temperature monitoring data at the jth sampling moment at all locations, d i，j，max is the distance between the i-th position and the position of the maximum value of all temperature monitoring data at the j-th sampling time, || represents the absolute value function, norm represents the linear normalization function, and the normalization object is the value of all temperature monitoring data.

[0070] Furthermore, according to the above method, the significance of the change of the temperature monitoring data at each location at each sampling time is obtained.

[0071] Step S003: According to the difference between the change significance of the temperature monitoring data series at different sampling times at the same location, and the distribution of the temperature monitoring data at different locations and different sampling times, the confidence of each sampling time is obtained; the connectivity threshold of each sampling time is obtained, and the optimal connectivity threshold is obtained in combination with the confidence of each sampling time.

[0072] Specifically, the temperature monitoring data at each sampling moment at all locations are processed by the CABDDCG algorithm to obtain the connectivity threshold at each sampling moment.

[0073] It should be noted that the temperature monitoring data at each sampling moment at all positions and the three-dimensional coordinates of all positions constitute the connectivity graph corresponding to each moment. The connectivity threshold at each sampling moment is determined by the CABDDCG algorithm based on the significant degree of change of the acquired data points, so that each sampling moment has a specific connectivity threshold. Then, the corresponding connectivity graph is constructed based on these connectivity thresholds, thereby obtaining multiple different connectivity graphs. The connectivity graph and connectivity threshold are the features obtained during the calculation process of the CABDDCG algorithm.

[0074] It should be further explained that, for the connectivity graph, when it is affected by the structural change of the fire, the change significance of the data point will be manifested as a structural directional change on the connectivity graph. Therefore, by using the PCA principal component analysis algorithm, each data point on the connectivity graph is processed to obtain the principal component direction angle on the connectivity graph, and the change significance of each data point is projected onto its principal component direction to obtain the change significance sequence. If the change degree sequence has a consistent change in time series, the higher the connectivity threshold corresponding to the connectivity graph is, the higher the degree of differentiation of the temperature structural change caused by the fire. By analyzing the temporal consistency of the change degree sequence, the degree of differentiation of the connectivity graph for the temperature structural change caused by the fire can be evaluated. If the change degree sequence shows a consistent change in time series, then the connectivity threshold corresponding to the connectivity graph has a higher confidence in distinguishing the temperature structural change caused by the fire. The principal component direction angle is a feature obtained in the PCA principal component analysis process. The PCA principal component analysis algorithm and the CABDDCG algorithm are common and well-known technologies, and the specific methods are not introduced in detail here.

[0075] Specifically, the PCA principal component analysis algorithm is used to process the temperature monitoring data at each sampling moment at all positions to obtain the principal component direction angle corresponding to the temperature monitoring data at each sampling moment at all positions. The sampling moments other than the j-th sampling moment are recorded as the reference moments of the j-th sampling moment. The sequence consisting of the change significance of the temperature monitoring data at any sampling moment at all positions is recorded as the change significance sequence of the sampling moment.

[0076] Furthermore, the DTW algorithm is used to match the change significance sequences of any two sampling moments to obtain the DTW distance of the change significance sequences at different sampling moments, wherein the DTW algorithm is a well-known technology and the specific method is not introduced here. Then the specific calculation formula of the confidence at each sampling moment is:

[0077]

[0078] Among them, LB j Indicates the confidence level at the jth sampling moment; DTW j，y represents the DTW distance of the sequence of significant changes between the jth sampling moment and its yth reference moment; Y represents the number of reference moments; θ j represents the principal component direction angle corresponding to the temperature monitoring data at the jth sampling moment; θ j，y It represents the principal component direction angle corresponding to the temperature monitoring data at the yth reference moment at the jth sampling moment; || is the absolute value function; exp() is an exponential function with a natural constant as the base. This embodiment adopts the exp(-x) model to present the inverse proportional relationship and normalization processing. x is the input of the model. The implementer can set the inverse proportional function and normalization function according to the actual situation.

[0079] Furthermore, according to the above method, the confidence level at each sampling moment is obtained.

[0080] It should be noted that LB j It is the degree of consistency of the change characteristics of the connectivity threshold at the j-th sampling moment in the structure of the connectivity graph construction, that is, the degree of discrimination of the connectivity threshold at the j-th sampling moment for the regional temperature change characteristics of the fire situation, and the subsequent optimal connectivity threshold can be determined from this.

[0081] Specifically, the preset confidence threshold ∈=0.81 in this embodiment is described as an example. Other implementations may be set to other values, which are not limited in this embodiment. When the confidence of any sampling moment is greater than the preset confidence threshold, it can be considered that it can be used to obtain the weighted average of the optimal connectivity threshold. The sampling moment with a confidence greater than the preset confidence threshold is recorded as the target sampling moment. According to the number of target sampling moments, the connectivity threshold and the confidence of the target sampling moment, the optimal connectivity domain value is obtained, and the specific calculation formula of the optimal connectivity threshold is:

[0082]

[0083] in, represents the optimal connectivity threshold, R represents the number of target sampling moments, LB r represents the confidence level at the rth target sampling moment, θ r It represents the connectivity threshold of the rth target sampling moment, softmax() is the softmax normalization function, and the normalized object is the confidence of each target sampling moment.

[0084] Step S004: According to the optimal connectivity threshold, through cluster analysis based on dynamic splitting of the connectivity graph, the reference of each temperature monitoring data is obtained, and then the predicted temperature value of each position at the next moment is obtained, and compared with the warning value to realize accurate monitoring of the fire situation in the mine tunnel.

[0085] Specifically, according to the optimal connectivity threshold, the temperature monitoring data at each sampling moment at all locations are processed by the CABDDCG algorithm, and the temperature monitoring data at all locations at each sampling moment are divided into a number of clusters.

[0086] It should be noted that the clustering results of the divided clusters are represented by the distribution of sample points in multiple clusters at each sampling moment, each cluster contains temperature data corresponding to multiple sample points, each cluster can be regarded as a temperature data set, and there are multiple temperature data sets at each sampling moment. The amount of data in its temperature set is used to characterize the size of the data reference when constructing the temperature prediction model. This achieves the adjustment of the construction of the ARIMA time series temperature prediction model, specifically by adjusting the autoregressive order and moving average order of the ARIMA model according to the data reference, to obtain an accurate temperature prediction model. This achieves the temperature prediction of the mine tunnel. When the predicted value exceeds the warning value determined by the temperature limit in the industry standards and safety specifications, a warning message is issued to prompt the corresponding warning operation to complete the accurate monitoring of the fire in the mine tunnel. Among them, the ARIMA model is a well-known technology, and the specific method is not introduced here.

[0087] Specifically, the mean of the number of temperature monitoring data of all clusters at any sampling time is calculated, and the quotient of the temperature monitoring data of each cluster at the sampling time and the mean is used as the reference of each temperature monitoring data in all clusters at the time. According to the reference and temperature value of each temperature monitoring data in the temperature monitoring data sequence at any position, the specific calculation formula for obtaining the updated temperature value of each temperature monitoring data in the temperature monitoring data sequence at the position is:

[0088]

[0089] in, represents the updated temperature value of the kth temperature monitoring data in the temperature monitoring data sequence at the ith position, c i，k+1 represents the k+1th temperature monitoring data in the temperature monitoring data sequence at the ith position; c i，k-1 represents the k-1th temperature monitoring data in the temperature monitoring data sequence at the ith position, μ i，k represents the reference of the kth temperature monitoring data in the temperature monitoring data sequence at the ith position, c i，kRepresents the temperature value of the kth temperature monitoring data in the monitoring data sequence at the ith position.

[0090] What needs to be explained is that: corresponding to the calculation of the updated temperature value of the first and last temperature data in the temperature monitoring data sequence, the formula Replace with c respectively i，k+1 and c i，k-1 .

[0091] Furthermore, the sequence consisting of the updated temperature values ​​of all temperature monitoring data in the temperature monitoring data sequence at any position is recorded as the updated temperature value sequence at the position; the updated temperature value sequence of each position is input into the ARIMA time series temperature prediction model to obtain the predicted temperature value corresponding to each position at the next moment in the future. The preset warning value in this embodiment is 40, which is described as an example. Other implementations may be set to other values, which are not limited in this embodiment. When the predicted temperature value is greater than the preset warning value, it is determined that there is an abnormal temperature at the corresponding position of the temperature monitoring data sequence, an alarm is issued to the corresponding security personnel, and the corresponding warning operation is automatically performed to complete the accurate monitoring of the fire situation in the mine tunnel.

[0092] So far, the present invention is completed.

[0093] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A fire monitoring method based on a mine optical fiber temperature measurement system, characterized in that: The method comprises the following steps: Acquire a temperature monitoring data sequence at multiple locations in a mine tunnel and the distance between any two locations; each temperature monitoring data in the temperature monitoring data sequence corresponds to a sampling time; According to the difference between the temperature monitoring data in the temperature monitoring data sequence at each location and the distribution of the temperature monitoring data, the significance of the change of the temperature monitoring data sequence at each sampling time at each location is obtained; According to the difference between the significant degree of changes in the temperature monitoring data series at different sampling times at the same location, and the distribution of the temperature monitoring data at different locations at different sampling times, the confidence of each sampling time is obtained; the connectivity threshold of each sampling time is obtained, and the optimal connectivity threshold is obtained in combination with the confidence of each sampling time; According to the optimal connectivity threshold, cluster analysis based on dynamic splitting of the connectivity graph is used to obtain the reference of each temperature monitoring data, and then the predicted temperature value of each location at the next moment is obtained, which is compared with the warning value to achieve accurate fire monitoring in the mine tunnel; The specific steps of obtaining the change significance of the temperature monitoring data sequence at each location at each sampling moment according to the difference between the temperature monitoring data in the temperature monitoring data sequence at each location and the distribution of the temperature monitoring data are as follows: The temperature monitoring data sequence at each position is derived using the first-order derivative method to obtain several extreme value points. The temperature monitoring data sequence at each position is divided into several temperature change time periods using all extreme value points as segmentation points. According to the temperature monitoring data sequence at each position and several temperature change time periods at each position, the time series change significance of the temperature monitoring data at each sampling moment at each position is obtained; According to the time series change significance of the temperature monitoring data at each sampling moment at each position, the change significance of the temperature monitoring data sequence at each sampling moment at each position is obtained; The method of obtaining the confidence level of each sampling moment according to the difference between the significant degrees of changes in the temperature monitoring data sequences at different sampling moments at the same location and the distribution of the temperature monitoring data at different locations at different sampling moments includes the following specific steps: The PCA principal component analysis algorithm is used to process the temperature monitoring data at each sampling moment at all positions to obtain the principal component direction angle corresponding to the temperature monitoring data at each sampling moment at all positions; The sampling moments other than the j-th sampling moment are recorded as the reference moments of the j-th sampling moment; The sequence consisting of the change significance of the temperature monitoring data at any sampling time at all positions is recorded as the change significance sequence at the sampling time; Use the DTW algorithm to match the change significance sequences of any two sampling moments, and obtain the DTW distance of the change significance sequences at different sampling moments; The confidence level at each sampling moment is obtained based on the principal component direction angle corresponding to the temperature monitoring data at each sampling moment at all locations and the DTW distance of the change significance sequence at different sampling moments; The method of obtaining the connectivity threshold at each sampling moment and combining the confidence at each sampling moment to obtain the optimal connectivity threshold includes the following specific steps: The temperature monitoring data at each sampling moment at all locations are processed by the CABDDCG algorithm to obtain the connectivity threshold at each sampling moment; The sampling time with a confidence level greater than a preset confidence threshold is recorded as the target sampling time; According to the number of target sampling moments, the connectivity threshold and the confidence of the target sampling moments, the optimal connectivity domain value is obtained; According to the optimal connectivity threshold, the reference of each temperature monitoring data is obtained through cluster analysis based on dynamic splitting of the connectivity graph, and then the predicted temperature value of each location at the next moment is obtained. The specific steps include the following: According to the optimal connectivity threshold, the CABDDCG algorithm is used to divide the temperature monitoring data of all locations at each sampling time into several clusters; Calculate the mean of the number of temperature monitoring data of all clusters at any sampling time, and use the quotient of the temperature monitoring data of each cluster at the sampling time and the mean as the reference of each temperature monitoring data in all clusters at the time; According to the reference and temperature value of each temperature monitoring data in the temperature monitoring data sequence at each position, an updated temperature value sequence at each position is obtained; The updated temperature value sequence is input into the ARIMA model to obtain the predicted temperature value for each location at the next moment in the future.

2. The fire monitoring method based on the mine optical fiber temperature measurement system according to claim 1 is characterized in that: According to the temperature monitoring data sequence at each position and several temperature change time periods at each position, the time series change significance of the temperature monitoring data at each sampling moment at each position is obtained, and the corresponding specific formula is: Among them, TX i,j Indicates the significance of the temporal change of the temperature monitoring data at the jth sampling moment at the ith position, l i,j represents the number of data in the temperature change time period where the temperature monitoring data at the jth sampling moment at the ith position is located, max(l i ) represents the maximum value of the data quantity in all temperature change time periods at the i-th position, c i,j represents the temperature monitoring data at the jth sampling time at the ith position, c i,j-1 represents the temperature monitoring data at the j-1th sampling time at the ith position, and || represents the absolute value function.

3. The fire monitoring method based on the mine optical fiber temperature measurement system according to claim 1 is characterized in that: According to the time series change significance of the temperature monitoring data at each sampling moment at each position, the change significance of the temperature monitoring data sequence at each sampling moment at each position is obtained, and the corresponding specific formula is: Among them, WX i,j Indicates the significant degree of change in the temperature monitoring data sequence at the jth sampling moment at the ith position, TX i,j Indicates the significance of the temporal change of the temperature monitoring data at the jth sampling moment at the ith position, represents the mean value of the time series change significance of the temperature monitoring data at the jth sampling moment at all locations, c i,j represents the temperature monitoring data at the jth sampling time at the ith position, max(c j ) represents the maximum value of the temperature monitoring data at the jth sampling moment at all locations, d i,j,max is the distance between the i-th position and the position of the maximum value of all temperature monitoring data at the j-th sampling moment, || represents the absolute value function, and norm represents the normalization function.

4. The fire monitoring method based on the mine optical fiber temperature measurement system according to claim 1 is characterized in that: The confidence level at each sampling moment is obtained based on the principal component direction angle corresponding to the temperature monitoring data at each sampling moment at all positions and the DTW distance of the change significance sequence at different sampling moments. The corresponding specific formula is: Among them, LB j Indicates the confidence level at the jth sampling moment; DTW j,y represents the DTW distance of the sequence of significant changes between the jth sampling moment and its yth reference moment; Y represents the number of reference moments; θ j represents the principal component direction angle corresponding to the temperature monitoring data at the jth sampling moment; θ j,y It represents the principal component direction angle corresponding to the temperature monitoring data at the yth reference time at the jth sampling time; || is the absolute value function; exp() is an exponential function with a natural constant as the base.

5. The fire monitoring method based on the mine optical fiber temperature measurement system according to claim 1 is characterized in that: The optimal connectivity domain value is obtained according to the number of target sampling moments, the connectivity threshold and the confidence of the target sampling moment. The corresponding specific formula is: in, represents the optimal connectivity threshold, R represents the number of target sampling moments, LB r represents the confidence level at the rth target sampling moment, represents the connectivity threshold at the rth target sampling moment, and softmax() is the softmax normalization function.

6. The fire monitoring method based on the mine optical fiber temperature measurement system according to claim 1 is characterized in that: The step of obtaining an updated temperature value sequence at each position according to the reference and temperature value of each temperature monitoring data in the temperature monitoring data sequence at each position includes the following specific steps: According to the reference and temperature value of each temperature monitoring data in the temperature monitoring data sequence, the updated temperature value of each temperature monitoring data in the temperature monitoring data sequence at each position is obtained. The corresponding specific formula is: in, represents the updated temperature value of the kth temperature monitoring data in the temperature monitoring data sequence at the ith position, c i,k+1 represents the k+1th temperature monitoring data in the temperature monitoring data sequence at the ith position; c i,k-1 represents the k-1th temperature monitoring data in the temperature monitoring data sequence at the ith position, μ i,k represents the reference of the kth temperature monitoring data in the temperature monitoring data sequence at the ith position, c i,k represents the temperature value of the kth temperature monitoring data in the monitoring data sequence at the i-th position; A sequence consisting of updated temperature values ​​of all temperature monitoring data in a temperature monitoring data sequence at any position is recorded as an updated temperature value sequence at the position.

Images