A goaf coal spontaneous combustion dynamic early warning method and system based on optical fiber temperature measurement

Abstract

This application relates to the field of coal mining technology and provides a method and system for dynamic early warning of spontaneous combustion of coal in goaf areas based on fiber optic temperature measurement. The method preprocesses the raw time-series temperature data of different preset temperature measurement points within the goaf area, which are acquired in real time by distributed temperature-sensing optical fibers deployed in the goaf. Based on the standard time-series temperature data of the goaf and the STGCN goaf temperature prediction model, the method predicts and evaluates the coal temperature at each temperature measurement point in the goaf for future time periods. Next, based on the predicted coal temperature values ​​for each temperature measurement point in the goaf for future time periods, a temperature cloud map of the goaf is generated. The temperature cloud map is iteratively updated based on a gradient optimization algorithm to determine the highest coal temperature value in the goaf. Finally, the trend fluctuation index of the temperature point corresponding to the highest coal temperature value in the goaf is calculated. Based on the trend fluctuation index and the corresponding highest coal temperature value, and according to pre-determined early warning level triggering conditions, dynamic early warning of spontaneous combustion of coal in the goaf is provided.

Description

Technical Field

[0001] This application relates to the field of coal mining technology, and in particular to a dynamic early warning method and system for spontaneous combustion of coal in goaf based on fiber optic temperature measurement. Background Technology

[0002] Goaf refers to the area in a coal mine where all coal has been mined and the support equipment has been removed. Because these areas accumulate a large amount of unmined coal or coal dust, this coal or coal dust is prone to spontaneous combustion under certain conditions. Spontaneous combustion not only leads to the waste of coal resources, but may also cause disasters such as fires and explosions, seriously threatening the safety of the mine.

[0003] Spontaneous combustion of coal in goaf is a complex and dangerous process. As the depth and intensity of coal mining increase, the geological conditions become more complex, and spontaneous combustion accidents in goaf are showing a relative upward trend, seriously threatening the safe production of coal mines.

[0004] Therefore, there is an urgent need to provide a technical solution that addresses the shortcomings of the existing technology. Summary of the Invention

[0005] The purpose of this application is to provide a dynamic early warning method and system for spontaneous combustion of coal in goaf based on fiber optic temperature measurement, so as to solve or alleviate the problems existing in the above-mentioned prior art.

[0006] To achieve the above objectives, this application provides the following technical solution:

[0007] This application provides a dynamic early warning method for spontaneous combustion of coal in goaf areas based on fiber optic temperature measurement, comprising: step S101, preprocessing the raw time-series temperature data of different preset temperature measurement points in the goaf area, which are acquired in real time based on distributed temperature sensing fibers deployed in the goaf area, to obtain time-series temperature standard data of the goaf area; wherein, the preprocessing includes: missing value filling, outlier removal, and data standardization; step S102, constructing a goaf temperature prediction model based on STGCN, and predicting the coal temperature of each temperature measurement point in the goaf area for future time periods according to the time-series temperature standard data. The prediction results are evaluated based on preset evaluation indicators; Step S103: Based on the predicted coal temperature values ​​of each temperature measuring point in the goaf for a future time period, a temperature cloud map of the goaf is generated, and the temperature cloud map is iteratively updated based on a gradient optimization algorithm to determine the highest coal temperature value in the goaf; Step S104: The trend fluctuation index of the temperature point corresponding to the highest coal temperature value in the goaf is calculated, and based on the trend fluctuation index and the corresponding highest coal temperature value, a dynamic early warning is issued for spontaneous combustion of coal in the goaf based on a predetermined early warning level triggering condition.

[0008] Preferably, in step S101, according to the formula:

[0009]

[0010] The distributed temperature-sensing optical fibers are deployed within the goaf area; where 'a' represents the number of distributed temperature-sensing optical fibers deployed within the goaf area, 'D' is the rounded width of the goaf area, 'm' represents the number of temperature measurement points, and 'L' represents the distance between two adjacent temperature measurement points on the distributed temperature-sensing optical fibers.

[0011] Preferably, in step S102, the time-series temperature standard data at different locations are transformed based on the adjacency matrix to obtain graph structure data; a graph convolutional layer based on a graph convolutional neural network is constructed, and the spatial features of each data node in the graph structure data are extracted through graph convolution operations, and the extracted spatial features of the data nodes are updated; a time-series convolutional layer based on a gated convolutional neural network is constructed, and the convolution kernel slides on the time series of the time-series temperature standard data to perform dot product operations on the subsequences of the time series of the time-series temperature standard data, extracting the time features of each data node, and capturing the evolution pattern of each data node based on the time-series convolution operations; a global pooling operation is performed on the extracted spatial and temporal features of all the data nodes to obtain the global spatiotemporal features of the graph structure data, and the predicted coal temperature value for each temperature measurement point in the goaf is determined for the future time period based on the evolution pattern of the data nodes.

[0012] Preferably, updating the spatial features of the extracted data nodes includes: according to the formula:

[0013]

[0014] The spatial features of the extracted data nodes are updated; wherein, The spatial features of the data node i obtained in the l-th layer graph convolution operation of the graph convolution layer; The adjacency matrix represents the connection relationship between data node i and data node j in the l-th layer graph convolution operation of the graph convolution layer; This refers to the spatial convolution weight matrix in the graph convolution operation of the l-th layer of the graph convolutional layer. σ represents the updated spatial features of data node i in the l-th layer graph convolution operation of the graph convolution layer; σ is the activation function.

[0015] Preferably, the step of capturing the evolution pattern of each data node based on temporal convolution operations includes: according to the formula:

[0016]

[0017] Capture the evolution pattern of the data nodes;

[0018] in, This refers to the temporal features obtained for data node i during the (l+1)th layer temporal convolution operation of the temporal convolutional layer; σ is the input feature of data node i in the l-th temporal convolution operation of the temporal convolutional layer; σ is the activation function; k is the temporal convolution kernel in the l-th temporal convolution operation of the temporal convolutional layer. The size of , p is the stride of the l-th temporal convolution operation of the temporal convolutional layer; d is the dimension of the input feature of the data node i in the l-th temporal convolution operation of the temporal convolutional layer; D is the number of data channels in the l-th temporal convolution operation of the temporal convolutional layer; K is the maximum value of the convolution kernel in the l-th temporal convolution operation of the temporal convolutional layer; P is the maximum stride of the l-th temporal convolution operation of the temporal convolutional layer.

[0019] Preferably, according to the formula:

[0020]

[0021] Spatial features of all extracted data nodes Time characteristics Perform global pooling; where z (l+1) These are the global spatiotemporal features obtained after global pooling operations; This is a feature representation of the data node i. N is the number of data nodes, and N is a positive integer.

[0022] Preferably, in step S103, based on the preset coordinate system of the goaf, the spatial coordinates of each temperature measuring point are defined; the predicted coal temperature values ​​of each temperature measuring point in the goaf for a future time period are interpolated based on the inverse distance weighting method to construct a continuous temperature field of the goaf, so as to generate a temperature cloud map of the goaf; based on the preset objective function, the coal temperature distribution in the goaf is updated according to the temperature gradient of each temperature measuring point in the temperature cloud map, so as to determine the highest coal temperature in the goaf.

[0023] Preferably, according to the formula:

[0024]

[0025] Interpolate the predicted coal temperature values ​​for future time periods at each of the temperature measurement points within the goaf to obtain the coal temperature value z0 at the target location within the goaf; where z i b is the value of the temperature measuring point b around the target location; B is the number of temperature measuring points around the target location, where b and B are both positive integers; d b pis the distance between the target location and the b-th temperature measuring point; p is a constant.

[0026] Preferably, step S104 includes:

[0027] According to the formula:

[0028]

[0029] Determine the trend fluctuation index of the temperature point corresponding to the highest coal temperature in the goaf area.

[0030] Where A is the rate of change of the temperature point corresponding to the highest coal temperature in the goaf, A t A is the coal temperature value at time t, corresponding to the highest coal temperature value within the goaf; t-1 The coal temperature value at time t-1, which corresponds to the highest coal temperature value in the goaf area; For A t With A t-1 The time interval between them; B is the deviation rate of the temperature point corresponding to the highest coal temperature in the goaf; B t The value at time t corresponding to the highest coal temperature within the goaf; R is the value at time t-1 corresponding to the highest coal temperature in the goaf; OSC The time interval corresponding to the highest coal temperature value in the goaf. The amplitude of vibration due to internal temperature change; C Tmax The temperature point corresponding to the highest coal temperature in the goaf at a time interval Maximum coal temperature inside, C Tmin The temperature point corresponding to the highest coal temperature in the goaf at a time interval Minimum coal temperature inside; R OFC f(C) represents the number of temperature changes at the temperature point corresponding to the highest coal temperature within the goaf; t″ C t″ The ) indicates the temperature point corresponding to the highest coal temperature in the goaf within the time interval. The number of times the internal temperature value is different; C t′ C t″ These are the temperature points corresponding to the highest coal temperature in the goaf at different time intervals. Temperature value C at time t′ t′ The temperature value C at time t″ t″ .

[0031] This application embodiment also provides a dynamic early warning system for spontaneous combustion of coal in goaf areas based on fiber optic temperature measurement, including: a data preprocessing unit configured to preprocess the raw time-series temperature data of different preset temperature measurement points in the goaf area, which are acquired in real time based on distributed temperature-sensing optical fibers deployed in the goaf area, to obtain time-series temperature standard data of the goaf area; wherein, the preprocessing includes: missing value filling, outlier removal, and data standardization; and a coal temperature prediction unit configured to construct a goaf temperature prediction model based on STGCN, and to predict the coal temperature of each temperature measurement point in the goaf area for a future time period based on the time-series temperature standard data. The system predicts and evaluates the prediction results based on preset evaluation indicators. A maximum temperature unit is configured to generate a temperature cloud map of the goaf based on the predicted coal temperature values ​​of each temperature measurement point within the goaf over a future time period, and iteratively update the temperature cloud map based on a gradient optimization algorithm to determine the maximum coal temperature value within the goaf. A trend warning unit is configured to calculate the trend fluctuation index of the temperature point corresponding to the maximum coal temperature value within the goaf, and based on the trend fluctuation index and the corresponding maximum coal temperature value, dynamically warn of spontaneous combustion of coal within the goaf based on predetermined warning level triggering conditions.

[0032] Beneficial effects:

[0033] The dynamic early warning method for spontaneous combustion of coal in goaf based on fiber optic temperature measurement provided in this application firstly preprocesses the raw time-series temperature data of different preset temperature measurement points in the goaf, which are acquired in real time based on distributed temperature-sensing optical fibers deployed in the goaf, to obtain time-series temperature standard data of the goaf. Then, a goaf temperature prediction model based on STGCN is constructed, and the coal temperature of each temperature measurement point in the goaf is predicted for the future time period based on the time-series temperature standard data. The prediction results are evaluated based on preset evaluation indicators. Next, a temperature cloud map of the goaf is generated based on the predicted coal temperature values ​​of each temperature measurement point in the goaf for the future time period. The temperature cloud map is iteratively updated based on a gradient optimization algorithm to determine the highest coal temperature value in the goaf. Finally, the trend fluctuation index of the temperature point corresponding to the highest coal temperature value in the goaf is calculated, and based on the trend fluctuation index and the corresponding highest coal temperature value, a dynamic early warning for spontaneous combustion of coal in the goaf is issued based on predetermined early warning level triggering conditions.

[0034] Therefore, by collecting temperature time-series data from a limited number of temperature measurement points, the coal temperature at each temperature measurement point in the goaf can be dynamically predicted and evaluated in the future time period through a spatiotemporal graph convolutional neural network. The prediction results are then dynamically updated in real time through a gradient optimization algorithm, thereby determining the maximum coal temperature value in the entire goaf in the future time period. This enables effective monitoring and early warning of coal spontaneous combustion accidents in the goaf, providing technical support for safe production in coal mines. Attached Figure Description

[0035] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application.

[0036] in:

[0037] Figure 1 This is a flowchart illustrating a dynamic early warning method for spontaneous combustion of coal in goaf based on fiber optic temperature measurement, according to some embodiments of this application.

[0038] Figure 2 This is a schematic diagram illustrating graph data conversion according to some embodiments of this application;

[0039] Figure 3 This is a schematic diagram of the structure of a dynamic early warning system for spontaneous combustion of coal in a goaf based on fiber optic temperature measurement, according to some embodiments of this application. Detailed Implementation

[0040] The present application will now be described in detail with reference to the accompanying drawings and embodiments. Various examples are provided by way of explanation and not by way of limitation. In fact, those skilled in the art will understand that modifications and variations can be made to the present application without departing from the scope or spirit of the present application. For example, a feature shown or described as part of one embodiment may be used in another embodiment to produce yet another embodiment. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention should fall within the scope of protection of the embodiments of the present invention.

[0041] like Figure 1 , Figure 2 As shown in the embodiments of this application, the dynamic early warning method for spontaneous combustion of coal in goaf based on fiber optic temperature measurement includes:

[0042] Step S101: Preprocess the raw time-series temperature data of different preset temperature measurement points in the goaf area, which are acquired in real time based on the distributed temperature sensing optical fiber deployed in the goaf area, to obtain the time-series temperature standard data of the goaf area.

[0043] In this application, a distributed temperature-sensing optical fiber network is deployed in the goaf to monitor the temperature of a limited number of temperature measurement points in real time. Specifically, parallel temperature-sensing optical fibers are arranged at 5-meter intervals on the return air and intake air sides of the goaf, with temperature measurement points spaced 1 meter apart on each fiber, enabling real-time temperature monitoring within the goaf. The temperature is monitored according to the formula:

[0044]

[0045] Distributed temperature-sensing optical fibers are deployed within the goaf. Here, 'a' represents the number of distributed temperature-sensing optical fibers deployed within the goaf, 'D' is the rounded width of the goaf, 'm' represents the number of temperature measurement points, and 'L' represents the distance between two adjacent temperature measurement points on the distributed temperature-sensing optical fibers.

[0046] Real-time time-series temperature data from different temperature measurement points within the goaf area is acquired via distributed temperature-sensing optical fibers and transmitted to edge devices via transmission optical fibers. The edge devices then perform missing value imputation, outlier removal, and data standardization on the raw time-series temperature data from each measurement point. Specifically, missing values ​​are imputed using an approximate averaging method, and rows are removed from the imputed data using box plot analysis.

[0047] Then, the original time-series temperature data after data removal is processed according to the formula:

[0048]

[0049] Data standardization is performed. Here, u represents the standardized data, i.e., the time-series temperature standard data; x represents the original time-series temperature data after data removal; μ represents the average value of the original time-series temperature data after data removal; and s represents the variance of the original time-series temperature data after data removal.

[0050] Step S102: Construct a goaf temperature prediction model based on STGCN, and predict the coal temperature at each temperature measurement point in the goaf for future time periods based on time-series temperature standard data, and evaluate the prediction results based on preset evaluation indicators.

[0051] In this application, temperature time-series standard data at different locations are transformed based on an adjacency matrix to generate graph-structured data. In this graph-structured data, nodes represent the spatial locations of temperature measurement points corresponding to each temperature time-series standard data point, and edges represent the connections between nodes, i.e., the relationships between the temperature measurement points corresponding to each temperature time-series standard data point. In other words, each temperature monitoring point in the mine is used as a node, and an adjacency matrix is ​​used to represent the connections between nodes, capturing the spatial relationships between each temperature monitoring point. The adjacency matrix is ​​combined with the standard data, and combined with the node feature matrix, to convert the temperature time-series standard data into graph data.

[0052] A temperature prediction model based on Spatio-Temporal Graph Convolutional Networks (STGCN) is used. This model comprises two spatio-temporal convolutional blocks and a global pooling layer. The two spatio-temporal convolutional blocks fuse the spatial and temporal features of each data node in the graph-structured data, respectively. The global pooling layer then performs global pooling on the spatial and temporal features of all nodes, thereby obtaining the global spatio-temporal characteristics of the temperature within the goaf.

[0053] Each spatiotemporal convolutional block contains two spatial graph convolutional layers and one temporal gated convolutional layer, with the temporal convolutional layer located between the two spatial graph convolutional layers. The spatial graph convolutional layers are built upon a spatial graph convolutional neural network and extract the spatial features of each data node in the graph-structured data through graph convolution operations. These extracted spatial features are then updated using graph convolution operations to reflect the spatial relationships between temperature measurement points within the goaf. Specifically, according to the formula:

[0054]

[0055] Update the spatial features of the extracted data nodes;

[0056] in, This refers to the spatial features obtained for data node i in the l-th layer of the graph convolution operation. Let be the adjacency matrix, representing the connection relationship between data node i and data node j in the l-th layer graph convolution operation of the graph convolution layer; This is the spatial convolution weight matrix in the graph convolution operation of the l-th layer of the graph convolutional layer; Let σ represent the spatial features of data node i after the graph convolution operation at the l-th layer of the graph convolution layer; σ is the Leaky ReLU activation function.

[0057] The temporal convolutional layer in the spatiotemporal convolutional block is constructed based on a gated convolutional neural network. It slides a convolutional kernel across the time series of the time-series temperature standard data, performing dot product operations on subsequences of the time series to extract the temporal features of each data node. Simultaneously, it captures the evolutionary pattern of each data node based on the temporal convolution operation. Specifically, the temporal convolution operation selects a continuous time step from the time series of the time-series temperature standard data to obtain a subsequence of the time series.

[0058] Specifically, according to the formula:

[0059]

[0060] Capture the evolution patterns of data nodes. Among them, This refers to the temporal features obtained for data node i during the (l+1)th temporal convolution operation in the temporal convolutional layer. Let σ be the input feature of data node i in the l-th temporal convolution operation of the temporal convolutional layer; σ is the Leaky ReLU activation function; and k is the temporal convolution kernel in the l-th temporal convolution operation of the temporal convolutional layer. The size of , p is the stride of the l-th temporal convolution operation of the temporal convolutional layer; d is the dimension of the input feature of data node i in the l-th temporal convolution operation of the temporal convolutional layer; D is the number of data channels in the l-th temporal convolution operation of the temporal convolutional layer; K is the maximum value of the convolution kernel in the l-th temporal convolution operation of the temporal convolutional layer; P is the maximum stride of the l-th temporal convolution operation of the temporal convolutional layer.

[0061] Next, a global pooling layer is used to perform global pooling operations on the spatial and temporal features of all extracted data nodes, thus aggregating all information from the graph structure data and obtaining its global spatiotemporal features. Specifically, according to the formula:

[0062]

[0063] Spatial features of all extracted data nodes Time characteristics Perform global pooling. Where z (l+1) These are the global spatiotemporal features obtained after global pooling operations; For the feature representation of data node i, N is the number of data nodes, and N is a positive integer.

[0064] Finally, based on the temperature changes within the goaf over time, the coal temperature at each temperature measuring point is inferred from the extracted global spatiotemporal features. Specifically, the predicted coal temperature for each temperature measuring point within the goaf is determined based on the evolution pattern of the data nodes, and this predicted value is output through the output layer of the temperature prediction model. This allows for real-time dynamic visualization of the spatial and temporal changes in coal temperature at each temperature measuring point within the goaf.

[0065] In this application, after outputting the predicted coal temperature values ​​for future time periods at each temperature measurement point through the output layer, the mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R²) are used. 2 The prediction results of the goaf temperature prediction model are evaluated. Specifically, according to the formula:

[0066]

[0067] The evaluation metrics for the prediction results of the goaf temperature prediction model are determined as follows: Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of Decision (R). 2 Among them, Y h Let y be the h-th predicted coal temperature value from the goaf temperature prediction model. h To be with Y h The corresponding actual coal temperature value within the goaf; H is a positive integer. This represents the average of H predicted coal temperatures.

[0068] Step S103: Based on the predicted coal temperature values ​​of each temperature measuring point in the goaf for a future time period, generate a temperature cloud map of the goaf, and iteratively update the temperature cloud map based on the gradient optimization algorithm to determine the highest coal temperature value in the goaf.

[0069] In this application, the spatial coordinates of each temperature measurement point are defined based on a pre-defined coordinate system of the goaf. Then, the predicted coal temperature values ​​for future time periods at each temperature measurement point within the goaf are interpolated using the inverse distance weighting method to construct a continuous temperature field for the goaf. The continuous temperature field data is then plotted into a temperature cloud map of the goaf using the contour function. The temperature cloud map of the goaf allows for the division of the goaf into heat dissipation zone, oxidation heating zone, and asphyxiation zone, revealing the dynamic development patterns of these three zones.

[0070] Specifically, according to the formula:

[0071]

[0072] Interpolate the predicted coal temperature values ​​for future time periods at various temperature monitoring points within the goaf to obtain the coal temperature value z0 at the target location within the goaf. Where z i b is the temperature value at the b-th temperature measurement point around the target location; B is the number of temperature measurement points around the target location, where b and B are both positive integers; d b p is the distance between the target location and the b-th temperature measurement point; p is an adjustable parameter, with a value of constant 1 or constant 2.

[0073] Finally, based on the preset objective function, the coal temperature distribution within the goaf is updated according to the temperature gradient at each temperature measurement point in the temperature cloud map, in order to determine the highest coal temperature within the goaf. Specifically, the preset objective function is:

[0074]

[0075] Among them, T′ eu The temperature T is the temperature at the temperature measurement point with spatial coordinates (e, u) within the goaf. eu Updated temperature; The temperature gradient is the temperature gradient at the temperature measurement point with spatial coordinates (e, u) within the goaf; T is a variable representing the temperature distribution within the goaf.

[0076] By differentiating the preset objective function, the temperature gradient at each temperature measurement point and the impact of temperature changes on the objective function are calculated. The direction of the temperature gradient is along the direction of the maximum temperature increase, specifically according to the formula:

[0077]

[0078] The temperature gradient of the temperature measurement point with spatial coordinates (e, u) in the goaf is determined, and then the coal temperature distribution in the goaf is updated using the temperature gradient to determine the highest coal temperature f(T) in the goaf.

[0079] Through the above steps, the temperature cloud map is continuously iterated and updated, and the temperature of each point in the temperature cloud map is adjusted to generate a continuous final temperature field for the goaf. Through the final temperature field, the coordinates and temperature values ​​of the highest coal temperature point in the entire goaf over a future time period can be accurately determined.

[0080] Step S104: Calculate the trend fluctuation index of the temperature point corresponding to the highest coal temperature in the goaf, and based on the trend fluctuation index and the corresponding highest coal temperature, dynamically warn of spontaneous combustion of coal in the goaf based on the predetermined warning level triggering conditions.

[0081] Specifically, according to the formula:

[0082]

[0083] Determine the trend fluctuation index of the temperature point corresponding to the highest coal temperature in the goaf.

[0084] Where A is the rate of change of the temperature point corresponding to the highest coal temperature in the goaf, A t A represents the coal temperature at time t, corresponding to the highest coal temperature within the goaf; t-1 The coal temperature at time t-1 corresponds to the highest coal temperature in the goaf. For A t With A t-1 The time interval between them; B is the deviation rate of the temperature point corresponding to the highest coal temperature in the goaf; B t This is the value at time t corresponding to the highest coal temperature in the goaf. R is the value at time t-1 corresponding to the highest coal temperature in the goaf; OSC The time interval corresponding to the highest coal temperature in the goaf. The amplitude of vibration due to internal temperature change; C Tmax The temperature point corresponding to the highest coal temperature in the goaf at a time interval Maximum coal temperature inside, C Tmin The temperature point corresponding to the highest coal temperature in the goaf at a time interval Minimum coal temperature inside; R OFC f(C) represents the number of temperature changes at the temperature point corresponding to the highest coal temperature within the goaf; t′ C t″ This indicates the temperature point corresponding to the highest coal temperature in the goaf within a time interval. The number of times the internal temperature value is different; C t′ C t″ These are the temperature points corresponding to the highest coal temperature in the goaf at different time intervals. Temperature value C at time t′ t′ The temperature value C at time t″ t″ .

[0085] By constructing trends for the highest temperature points within goaf areas, the potential development trends of the current goaf status can be effectively reflected, thereby enabling early detection of goaf fire hazards. Simultaneously, based on coal sample heating tests, the warning level trigger thresholds for different coal qualities in different coal mines are determined. By comparing trend fluctuation indicators with these warning level trigger thresholds, different levels of warnings can be issued for the future state of goaf areas, allowing for timely and effective adjustments to response measures.

[0086] like Figure 3 As shown in the figure, this application embodiment also provides a dynamic early warning system for spontaneous combustion of coal in goaf based on fiber optic temperature measurement, including:

[0087] The data preprocessing unit 301 is configured to preprocess the raw time-series temperature data of different preset temperature measurement points in the goaf, which are acquired in real time based on the distributed temperature sensing optical fiber deployed in the goaf, to obtain the time-series temperature standard data of the goaf; wherein, the preprocessing includes: missing value filling, outlier removal and data standardization.

[0088] The coal temperature prediction unit 302 is configured to construct a goaf temperature prediction model based on STGCN, and predict the coal temperature of each temperature measurement point in the goaf for a future time period based on time-series temperature standard data, and evaluate the prediction results based on preset evaluation indicators.

[0089] The highest point temperature unit 303 is configured to generate a temperature cloud map of the goaf based on the predicted coal temperature values ​​of each temperature measuring point in the goaf over a future time period, and to iteratively update the temperature cloud map based on a gradient optimization algorithm to determine the highest coal temperature value in the goaf.

[0090] The trend warning unit 304 is configured to calculate the trend fluctuation index of the temperature point corresponding to the highest coal temperature in the goaf, and to provide dynamic early warning of spontaneous combustion of coal in the goaf based on the trend fluctuation index and the corresponding highest coal temperature, and based on the predetermined early warning level triggering conditions.

[0091] The dynamic early warning system for spontaneous combustion of coal in goaf based on fiber optic temperature measurement provided in this application can realize the steps and processes of the dynamic early warning method for spontaneous combustion of coal in goaf based on fiber optic temperature measurement in any of the above embodiments, and achieve the same technical effect, which will not be described in detail here.

[0092] In the description of this invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0093] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A dynamic early warning method for spontaneous combustion of coal in goaf areas based on fiber optic temperature measurement, characterized in that, include: Step S101: Preprocess the raw time-series temperature data of different preset temperature measurement points in the goaf area, which are acquired in real time based on the distributed temperature sensing optical fiber deployed in the goaf area, to obtain the time-series temperature standard data of the goaf area; wherein, the preprocessing includes: missing value filling, outlier removal and data standardization. Step S102: Construct a goaf temperature prediction model based on STGCN, and predict the coal temperature at each temperature measurement point in the goaf for future time periods based on the time-series temperature standard data, and evaluate the prediction results based on preset evaluation indicators; wherein, The time-series temperature standard data at different locations are transformed based on the adjacency matrix to obtain graph-structured data. A graph convolutional layer based on a graph convolutional neural network is constructed to extract the spatial features of each data node in the graph-structured data through graph convolution operations, and the extracted spatial features of the data nodes are updated. A temporal convolutional layer based on a gated convolutional neural network is constructed. By sliding the convolution kernel on the time series of the time-series temperature standard data, dot product operations are performed on the subsequences of the time series of the time-series temperature standard data to extract the temporal features of each data node, and the evolution pattern of each data node is captured based on the temporal convolution operation. Global pooling is performed on the extracted spatial and temporal features of all data nodes to obtain the global spatiotemporal features of the graph-structured data, and the predicted coal temperature values ​​for each temperature measurement point in the goaf are determined for future time periods based on the evolution patterns of the data nodes. Step S103: Based on the predicted coal temperature values ​​of each temperature measuring point in the goaf for a future time period, generate a temperature cloud map of the goaf, and iteratively update the temperature cloud map based on a gradient optimization algorithm to determine the highest coal temperature value in the goaf. Step S104: Calculate the trend fluctuation index of the temperature point corresponding to the highest coal temperature in the goaf, and based on the trend fluctuation index and the corresponding highest coal temperature, dynamically warn of spontaneous combustion of coal in the goaf based on the predetermined early warning level triggering conditions.

2. The method for dynamic early warning of spontaneous combustion of coal in goaf based on fiber optic temperature measurement according to claim 1, characterized in that, In step S101, According to the formula: The distributed temperature-sensing optical fiber is deployed within the goaf area; in, The number of distributed temperature-sensing optical fibers deployed within the goaf area. The width of the goaf is rounded up; The number of temperature measurement points; The distance between two adjacent temperature measuring points on the distributed temperature sensing fiber is denoted as .

3. The method for dynamic early warning of spontaneous combustion of coal in goaf based on fiber optic temperature measurement according to claim 1, characterized in that, The step of updating the spatial features of the extracted data nodes includes: According to the formula: The spatial features of the extracted data nodes are updated; in, For the data node The first convolutional layer in the graph Spatial features obtained in layer graph convolution operations; The adjacency matrix represents the data nodes. and the data node The first convolutional layer in the graph Connection relationships in layer graph convolution operations; The first convolutional layer of the graph Spatial convolution weight matrix in layer graph convolution operation; For the data node The first convolutional layer in the graph Updated spatial features during layer graph convolution operations; This is the activation function.

4. The method for dynamic early warning of spontaneous combustion of coal in goaf based on fiber optic temperature measurement according to claim 1, characterized in that, The method of capturing the evolution pattern of each data node based on temporal convolution operations includes, according to the formula: Capture the evolution pattern of the data nodes; in, For the data node The first of the temporal convolutional layers Temporal features obtained in layer-time convolution operations; For the data node The first of the temporal convolutional layers Input features in layer-wise temporal convolution operations; For activation functions; The first of the temporal convolutional layers Temporal convolution kernels in layer-by-layer temporal convolution operations Size, The first of the temporal convolutional layers The stride of a layer-wise temporal convolution operation; For the data node The first of the temporal convolutional layers The dimension of the input features in a layer-by-layer temporal convolution operation; The first of the temporal convolutional layers The number of data channels in a layer-wise temporal convolution operation; The first of the temporal convolutional layers The maximum value of the convolution kernel in a layer-time convolution operation; The first of the temporal convolutional layers The maximum stride size during layer-time convolution operations.

5. The method for dynamic early warning of spontaneous combustion of coal in goaf based on fiber optic temperature measurement according to claim 1, characterized in that, According to the formula: Spatial features of all extracted data nodes Time characteristics Perform global pooling operation; in, These are the global spatiotemporal features obtained after global pooling operations; For the data node Feature representation, ; The number of data nodes. It is a positive integer.

6. The method for dynamic early warning of spontaneous combustion of coal in goaf based on fiber optic temperature measurement according to claim 1, characterized in that, In step S103, Based on the preset coordinate system of the goaf area, the spatial coordinates of each temperature measuring point are defined; The predicted coal temperature values ​​for each temperature measurement point in the goaf are interpolated for future time periods based on the inverse distance weighting method to construct a continuous temperature field of the goaf and generate a temperature cloud map of the goaf. Based on a preset objective function, the coal temperature distribution in the goaf is updated according to the temperature gradient of each temperature measurement point in the temperature cloud map, so as to determine the highest coal temperature in the goaf.

7. The method for dynamic early warning of spontaneous combustion of coal in goaf based on fiber optic temperature measurement according to claim 6, characterized in that, According to the formula: Interpolate the predicted coal temperature values ​​for future time periods at each temperature measurement point within the goaf to obtain the coal temperature value at the target location within the goaf. ; in, For the target location surrounding the first The value of the temperature measurement point; The number of temperature measuring points around the target location. All are positive integers; For the target position and the first The distance between the temperature measurement points; It is a constant.

8. The method for dynamic early warning of spontaneous combustion of coal in goaf based on fiber optic temperature measurement according to claim 1, characterized in that, Step S104 includes: According to the formula: Determine the trend fluctuation index of the temperature point corresponding to the highest coal temperature in the goaf area. in, This represents the rate of change of the temperature point corresponding to the highest coal temperature within the goaf. The temperature point corresponding to the highest coal temperature in the goaf. The coal temperature value at that moment; The temperature point corresponding to the highest coal temperature in the goaf. The coal temperature value at that moment; for and The time interval between; The deviation rate of the temperature point corresponding to the highest coal temperature value in the goaf area; The temperature point corresponding to the highest coal temperature in the goaf. The value at time; The temperature point corresponding to the highest coal temperature in the goaf. The value at time; The time interval is the time point corresponding to the highest coal temperature in the goaf. The amplitude of vibrations due to changes in internal temperature; The temperature point corresponding to the highest coal temperature in the goaf at a time interval The maximum coal temperature inside, The temperature point corresponding to the highest coal temperature in the goaf at a time interval Minimum coal temperature inside; The number of temperature changes at the temperature point corresponding to the highest coal temperature in the goaf; This indicates the temperature point corresponding to the highest coal temperature in the goaf within a time interval. The number of times the internal temperature value is different; These are the temperature points corresponding to the highest coal temperature in the goaf at different time intervals. Inner Time temperature value and time Temperature value .

9. A dynamic early warning system for spontaneous combustion of coal in goaf areas based on fiber optic temperature measurement, characterized in that, The dynamic early warning method for spontaneous combustion of coal in goaf areas based on fiber optic temperature measurement, as described in any one of claims 1-8, is used to provide dynamic early warning for spontaneous combustion of coal in the goaf area. The early warning system includes: The data preprocessing unit is configured to preprocess the raw time-series temperature data of different preset temperature measurement points in the goaf, which are acquired in real time based on the distributed temperature sensing optical fiber deployed in the goaf, to obtain the time-series temperature standard data of the goaf; wherein, the preprocessing includes: missing value filling, outlier removal and data standardization. The coal temperature prediction unit is configured to construct a goaf temperature prediction model based on STGCN, and predict the coal temperature of each temperature measurement point in the goaf for a future time period according to the time series temperature standard data, and evaluate the prediction results based on preset evaluation indicators. The highest point temperature unit is configured to generate a temperature cloud map of the goaf based on the predicted coal temperature values ​​of each temperature measuring point in the goaf over a future time period, and to iteratively update the temperature cloud map based on a gradient optimization algorithm to determine the highest coal temperature value in the goaf. The trend warning unit is configured to calculate the trend fluctuation index of the temperature point corresponding to the highest coal temperature in the goaf, and based on the trend fluctuation index and the corresponding highest coal temperature, dynamically warn of spontaneous combustion of coal in the goaf according to a predetermined warning level triggering condition.

Images