Coal mine risk analysis method, equipment and medium based on Internet of Things monitoring data

Abstract

The present invention discloses a coal mine risk analysis method, equipment and medium based on Internet of Things monitoring data, which relates to the technical field of coal mine risk analysis, and comprises the following steps: obtaining a thermal imaging image of a coal mine goaf in real time, marking the thermal image of the goaf, and performing spontaneous combustion risk monitoring and analysis based on the thermal image of the goaf; performing a first division processing on the thermal image of the goaf, calculating the regional average temperature based on the processed thermal image of the goaf, and obtaining temperature data of the first goaf block; performing data processing on the temperature data of the first goaf block, and obtaining temperature data of the second goaf block; and predicting and analyzing the spontaneous combustion risk of the coal mine goaf based on the temperature data of the second goaf block. The present invention is used to solve the problem that the existing coal mine risk analysis technology cannot perform comprehensive monitoring and analysis of all goafs through thermal imaging images when analyzing the spontaneous combustion risk of coal mine goafs, and at the same time, cannot perform predictive analysis of the spontaneous combustion risk of all goafs by using thermal imaging images.

Description

Technical Field

[0001] The present invention relates to the technical field of coal mine risk analysis, and in particular to a coal mine risk analysis method, equipment and medium based on Internet of Things monitoring data. Background Art

[0002] Coal mine risk analysis technology refers to the technology that uses system theory, modern information technology and various advanced technical means to identify, analyze, evaluate and manage the safety risks existing in the coal mine production process; in the field of coal mining, with the continuous expansion of mining depth and scale, the risk problem of coal mine goaf has become a key factor that seriously threatens coal mine production safety and efficient use of resources.

[0003] However, due to the large amount of coal fragments left in the goaf, spontaneous combustion is very likely to occur under suitable ventilation, temperature and humidity conditions; therefore, the risk of spontaneous combustion is the most important risk issue in coal mine goafs; the existing analysis of the risk of spontaneous combustion in coal mine goafs mainly relies on thermal imaging equipment to collect thermal imaging images or temperature sensors to collect temperature information for analysis; thermal imaging images can quickly, intuitively and comprehensively obtain the temperature distribution of the entire surface and shallow layer of the goaf in a non-contact manner; although through the precise analysis of thermal imaging images, temperature abnormalities can be clearly found, but when temperature abnormalities are detected, it often means that risks have already occurred. At this time, there is very little time left for relevant personnel to deal with the risks, because thermal imaging images are not predictive and can only be used as monitoring. and analysis tools; while the temperature information collected by temperature sensors is predictive, and temperature changes can be predicted, and risks can be foreseen in advance, but the temperature sensor can only be arranged in a certain place in the goaf, and can only obtain the temperature information of this place. For example, in the patent application with publication number CN117993559A, a method for predicting spontaneous combustion of coal in goaf is disclosed. This method obtains the temperature information of the goaf through a temperature sensor, and cannot comprehensively obtain the temperature distribution of the entire surface and shallow layer of the goaf; therefore, when analyzing the spontaneous combustion risk of coal mine goaf, the existing coal mine risk analysis technology cannot use thermal imaging images to comprehensively monitor and analyze all goafs, and at the same time use thermal imaging images to predict and analyze the spontaneous combustion risk of all goafs. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems in the prior art to a certain extent, by acquiring thermal imaging images of coal mine goafs in real time, marking the thermal images of the goafs for spontaneous combustion risk monitoring and analysis; performing a first division processing on the thermal images of the goafs, and then calculating the regional average temperature to obtain the temperature data of the first goaf block for data processing, obtaining the temperature data of the second goaf block, and then predicting and analyzing the spontaneous combustion risk of the coal mine goafs; so as to solve the problem that the existing coal mine risk analysis technology cannot use thermal imaging images to comprehensively monitor and analyze all goafs when analyzing the spontaneous combustion risk of coal mine goafs, and at the same time use thermal imaging images to predict and analyze the spontaneous combustion risk of all goafs.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a coal mine risk analysis method based on Internet of Things monitoring data, comprising the following steps:

[0006] Acquire thermal images of coal mine goafs in real time, mark the goaf thermal maps, and conduct spontaneous combustion risk monitoring and analysis based on the goaf thermal maps;

[0007] Performing a first division process on the goaf thermal map, calculating the regional average temperature based on the processed goaf thermal map, and obtaining the first goaf block temperature data;

[0008] Processing the temperature data of the first goaf block to obtain the temperature data of the second goaf block;

[0009] Based on the temperature data of the second goaf block, the spontaneous combustion risk of the coal mine goaf was predicted and analyzed.

[0010] Furthermore, real-time acquisition of thermal imaging images of coal mine goafs, marking of thermal images of goafs, and spontaneous combustion risk monitoring and analysis based on thermal images of goafs include the following sub-steps:

[0011] Based on a thermal imager connected to the Internet of Things, the relative position of the thermal imager and the coal mine goaf in space is kept unchanged, and the thermal image of the coal mine goaf is continuously collected at a first time interval, which is marked as a goaf thermal map, denoted as TUi, i represents the time when the goaf thermal map is collected, and the first time interval is tj;

[0012] Based on the heat map of the goaf, the temperature values ​​corresponding to each pixel point of the heat map of the goaf are extracted, and the temperatures are compared to obtain the highest temperature and the pixel point corresponding to the highest temperature, which are marked as the hottest pixel point, and the position of the hottest pixel point in the heat map of the goaf is determined; the first radius is set to r1, the second radius is set to r2, and the third radius is set to r3, r3>r2>r1; with the hottest pixel point as the center, circles are drawn on the heat map of the goaf with r1, r2 and r3 as the radii respectively, and the obtained circular area with a radius of r1 is marked as the first circular area; the annular area enclosed by the boundary of the circle with a radius of r2 and the boundary of the circle with a radius of r1 is marked as the second annular area; the annular area enclosed by the boundary of the circle with a radius of r3 and the boundary of the circle with a radius of r2 is marked as the third annular area.

[0013] Furthermore, the spontaneous combustion risk monitoring and analysis based on the thermal map of the goaf also includes the following sub-steps:

[0014] The temperature values ​​corresponding to the pixels in the first circular area, the second annular area, and the third annular area are extracted respectively; the weight of the first circular area is set to b1, the weight of the second annular area is set to b2, and the weight of the third annular area is set to b3; the average value TPi of the temperature values ​​corresponding to the pixels in the first circular area, the second annular area, and the third annular area is calculated by the weighted average formula, and the weighted average formula is as follows: , where T1 is the average temperature value corresponding to all pixels in the first circular area; T2 is the average temperature value corresponding to all pixels in the second annular area; T3 is the average temperature value corresponding to all pixels in the third annular area;

[0015] Determine whether TPi exceeds the safety temperature threshold. When TPi exceeds the safety temperature threshold, it is determined that the coal mine goaf has a risk of spontaneous combustion. The safety temperature threshold is T0.

[0016] Furthermore, the first division processing of the goaf thermal map includes the following sub-steps:

[0017] Set the first side length to a1 and the moving step length to a2, a2 <a1;

[0018] Starting from the lower left corner of the goaf heat map with a square with a side length of a1, the goaf heat map is divided from left to right and from bottom to top. The square with a side length of a1 moves horizontally or vertically by a2 steps each time, and the divided goaf heat map is marked as a goaf block map, denoted as TFi, and the blocks in the goaf block map are denoted as TFi (x, y), where x represents the xth block from left to right, and y represents the yth block from bottom to top.

[0019] Further, the regional average temperature is calculated based on the processed thermal map of the goaf to obtain the temperature data of the first goaf block, including the following sub-steps:

[0020] For all blocks in the goaf block map, the temperature values ​​corresponding to all pixels in the block are extracted respectively, and the temperature values ​​are sorted from large to small, which is recorded as TFi (x, y) block temperature sequence;

[0021] Set the first ratio to n%, extract the first n% of the temperature values ​​of the TFi (x, y) block temperature sequence, and calculate the average value, which is marked as the average high temperature value of the TFi (x, y) block, recorded as W(i, x, y);

[0022] The average high temperature value of the TFi (x, y) block is stored according to the goaf block map and the TFi (x, y) position to which it belongs, and marked as the first goaf block temperature data; each goaf block map includes the average high temperature value of the TFi (x, y) blocks of all blocks in the goaf block map, and each TFi (x, y) block contains the average high temperature value of the TFi (x, y) blocks collected and processed at all acquisition times.

[0023] Furthermore, performing data processing on the first goaf block temperature data to obtain the second goaf block temperature data includes the following sub-steps:

[0024] Based on the first goaf block temperature data, for all blocks in the goaf block diagram TFi, extract the average high temperature value W(i, x, y) of the block TFi (x, y) respectively, then extract the goaf block diagram TF(i-1) obtained by the previous acquisition time, and W(i-1, x, y) of the block corresponding to TF(i-1); set the first temperature weight to c1, the second temperature weight to c2, c1>c2; use [c1*W(i, x, y)+c2*W(i-1, x, y)] to replace W(i, x, y), and record the replaced TFi (x, y) block average high temperature value W(i, x, y) as TFi (x, y) block average high temperature value V(i, x, y); repeat the replacement of W(i, x, y) of all blocks in all goaf block diagrams, and obtain the second goaf block temperature data after completion.

[0025] Furthermore, predicting and analyzing the spontaneous combustion risk of each block in the goaf based on the temperature data of the second goaf block includes the following sub-steps:

[0026] The average high temperature value V(i, x, y) of each block in the second goaf block temperature data of all acquisition times is taken, and the second goaf block temperature data is divided into the first block temperature training set and the first block temperature test set in a ratio of 7:3 on the time scale;

[0027] Normalize the data in the first block temperature training set and the first block temperature test set, and normalize the data size to the [0,1] interval. After completion, obtain the second block temperature training set and the second block temperature test set;

[0028] Based on the multilayer perceptron model, the number of neurons in the input layer of the multilayer perceptron model is set to 1, the number of hidden layers is set to 2, the number of neurons in the output layer is set to 1, the activation function is set to the ReLU function, the mean square error is defined as the loss function, and the completed multilayer perceptron model is marked as the original prediction model.

[0029] Furthermore, predicting and analyzing the spontaneous combustion risk of each block in the goaf based on the temperature data of the second goaf block also includes the following sub-steps:

[0030] Set the learning rate to e1, the number of training rounds to e2, and the batch size to e3. Use the second block temperature training set to train the original prediction model. After completion, the model to be tested is obtained.

[0031] The model to be tested is tested using the second block temperature test set, and the calculated mean absolute error of the model is calculated to obtain the mean absolute error value MA of the model. The qualified threshold is set to MB, and it is determined whether MA satisfies MA≤MB. If so, the model to be tested is marked as a block temperature prediction model; if not, the model to be tested is trained again using the second block temperature training set, and after completion, the model to be tested is tested again using the second block temperature test set until a block temperature prediction model is obtained;

[0032] Based on the temperature data of the second goaf block, the block temperature prediction model is used to predict the future temperature changes of the coal mine goaf corresponding to the goaf block diagram to obtain the predicted temperature data, and determine whether there is data in the predicted temperature data that is not less than the safety temperature threshold T0. If so, it is determined that the coal mine goaf has a risk of spontaneous combustion. If not, it is determined that the coal mine goaf has no risk of spontaneous combustion for the time being.

[0033] In a second aspect, the present invention provides an electronic device, comprising a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the above method are performed.

[0034] On the third aspect, the present invention provides a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps in the above method are performed.

[0035] The beneficial effects of the present invention are:

[0036] 1. The present invention obtains the thermal imaging map of the goaf in a coal mine in real time, marks the thermal map of the goaf, and conducts spontaneous combustion risk monitoring and analysis based on the thermal map of the goaf; performs a first partitioning process on the thermal map of the goaf, calculates the average temperature of the area based on the processed thermal map of the goaf to obtain the temperature data of the first goaf block; processes the temperature data of the first goaf block to obtain the temperature data of the second goaf block; predicts and analyzes the spontaneous combustion risk of the coal mine goaf based on the temperature data of the second goaf block; while comprehensively monitoring and analyzing the entire goaf through the thermal imaging image, it can predict and analyze the spontaneous combustion risk of the entire goaf using the thermal imaging image.

[0037] 2. The present invention divides three areas around the pixel point and then calculates the temperature value by weighted average. The advantage is that a high temperature corresponding to a single pixel point does not fully represent the temperature of the corresponding point in the goaf and may also be caused by other environmental factors. This can reduce the error rate during risk monitoring and analysis; divide by a square with side length a1 and moving step length a2, and a2 < a1, so that there is partial overlap between adjacent blocks obtained by the division, which can avoid the special situation where a high-temperature area is evenly divided by the blocks, reduce the boundary error, and improve the accuracy of subsequent prediction and analysis; replace the temperature by weighted average of the temperatures at two moments, which can eliminate the short-term fluctuations caused by irrelevant factors in the temperature data, improve the data quality, and further improve the accuracy of prediction and analysis; the temperature prediction based on the multi-layer perceptron model improves the accuracy, generalization ability, and robustness of the prediction compared with the traditional linear regression fitting. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is the flowchart of the steps of the method of the present invention;

[0039] Figure 2 is the frequency scatter plot of the present invention;

[0040] Figure 3 is the amplitude scatter plot of the present invention;

[0041] Figure 4 is the pulse width scatter plot of the present invention;

[0042] Figure 5 is the structural schematic diagram of the electronic device of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0044] Example 1

[0045] See also Figure 1 As shown, the present invention provides a coal mine risk analysis method based on Internet of Things monitoring data, comprising the following steps:

[0046] Step S1, real-time acquisition of thermal imaging images of coal mine goafs, marking of goaf thermal images, and spontaneous combustion risk monitoring and analysis based on the goaf thermal images; Step S1 includes the following sub-steps:

[0047] Step S101, based on the thermal imager connected to the Internet of Things, the relative position of the thermal imager and the coal mine goaf in space is kept unchanged, the purpose is to make the thermal image map collected by the thermal imager correspond to the coal mine goaf, the pixel points in the thermal image map correspond to the position points of the coal mine goaf one by one, and will not change with time; the thermal image map of the coal mine goaf is continuously collected at a first time interval, marked as the goaf thermal map, recorded as TUi, i represents the time when the goaf thermal map is collected, the first time interval is tj, in this embodiment, the first time interval is tj=6s, that is, the thermal image map of the coal mine goaf is collected every 6s;

[0048] Step S102, based on the heat map of the goaf, extract the temperature value corresponding to each pixel point of the heat map of the goaf, and compare the high and low temperatures to obtain the highest temperature and the pixel point corresponding to the highest temperature, mark it as the hottest pixel point, and determine the position of the hottest pixel point in the heat map of the goaf;

[0049] Step S103, set the first radius to r1, the second radius to r2, and the third radius to r3, where r3>r2>r1; because it is necessary to draw a circle in the figure, it is inconvenient to directly use the length distance as the unit, so in order to facilitate the processing of the radius unit using the number of pixels, in this embodiment, r1 is 2 pixels, r2 is 3 pixels, and r3 is 4 pixels;

[0050] Step S104, see Figure 2 As shown, with the hottest pixel as the center, circles are drawn on the thermal map of the goaf with r1, r2 and r3 as the radius, and the obtained circular area with radius r1 is marked as the first circular area; the annular area surrounded by the boundary of the circle with radius r2 and the boundary of the circle with radius r1 is marked as the second annular area; the annular area surrounded by the boundary of the circle with radius r3 and the boundary of the circle with radius r2 is marked as the third annular area. The purpose of dividing the area is that some irrelevant environmental factors may cause abnormal pixel values ​​at a certain point in the thermal imaging image, resulting in high corresponding temperature values, while there is no abnormal high temperature in the actual goaf, so it is necessary to judge the area around the pixel to prevent misjudgment;

[0051] Step S105, respectively extracting the temperature values ​​corresponding to the pixels in the first circular area, the second annular area, and the third annular area, and extracting the temperature values ​​of all the pixels in the area;

[0052] Step S106, setting the weight of the first circular area to b1, the weight of the second annular area to b2, and the weight of the second annular area to b3;

[0053] Step S107, calculating the average value TPi of the temperature values ​​corresponding to the pixels in the first circular area, the second annular area, and the third annular area by a weighted average formula, the weighted average formula is as follows: , where T1 is the average temperature value corresponding to all pixels in the first circular area; T2 is the average temperature value corresponding to all pixels in the second annular area; T3 is the average temperature value corresponding to all pixels in the third annular area; because the temperature of the high temperature area often decreases from the center of the area to the surrounding area, if the high temperature area is small, and only one area is divided to obtain the average, the average value may be much smaller than the maximum value, so that the high temperature may not be recognized. Therefore, multiple areas are divided and different weights are assigned. The closer to the center, the higher the weight, so that the average value can be close to the maximum value, that is, the high temperature cannot be recognized, and the pixel abnormality caused by irrelevant environmental factors is avoided. Misjudgment as high temperature;

[0054] Step S108, judging whether TPi exceeds the safety temperature threshold. When TPi exceeds the safety temperature threshold, it is judged that the coal mine goaf has the risk of spontaneous combustion. The safety temperature threshold is T0. In this embodiment, T0=55°C, because when the coal temperature reaches 30°C-35°C, the coal oxidation may be accelerated; as the temperature continues to rise, when it reaches about 70-80°C, the coal in the goaf will have a relatively high risk of spontaneous combustion; so for safety considerations, T0=55°C;

[0055] In the specific implementation process, in the grayscale thermal imaging image, the pixel value is usually linearly or approximately linearly corresponding to the temperature; that is, the higher the pixel value, the higher the corresponding temperature; the lower the pixel value, the lower the corresponding temperature; for example, an 8-bit grayscale thermal imaging image has a pixel value range of 0-255. If white represents high temperature and black represents low temperature, then the pixel value of the white pixel is close to 255, corresponding to a higher temperature; the pixel value of the black pixel is close to 0, corresponding to a lower temperature; and the pseudo-color thermal imaging image maps different temperature ranges to different colors, that is, different colors correspond to temperature intervals, so in this implementation, a grayscale thermal imager is used.

[0056] Step S2: Perform a first partitioning process on the goaf thermal map, calculate the regional average temperature based on the processed goaf thermal map, and obtain the temperature data of the first goaf block. Step S2 includes the following sub-steps:

[0057] Step S201: Set the first side length as a1 and the moving step length as a2, where a2 < a1. Both the first side length a1 and the moving step length a2 are in units of the number of pixels. In this embodiment, a1 is 12 pixels and a2 is 8 pixels, with a2 < a1. The purpose is to make the adjacent blocks obtained by partitioning have partial overlap, which can avoid the special situation where a high-temperature area is evenly divided by the blocks. If a high-temperature area is evenly divided by the blocks, the temperature data obtained in subsequent processing will be inaccurate, thus affecting subsequent analysis.

[0058] Step S202: Start from the lower left corner of the goaf thermal map with a square of side length a1, that is, the lower left vertex of the square coincides with the lower left vertex pixel of the goaf thermal map. Partition the goaf thermal map in the order from left to right and from bottom to top. The square of side length a1 moves a2 step lengths horizontally or vertically each time. When moving to the edge, for example, when moving to the upper edge and the right edge in the order from left to right and from bottom to top, it may occur that the right side or the upper side of the square exceeds the right side or the upper side of the goaf thermal map. At this time, the square can be moved back to keep the right side or the upper side of the square aligned with the right side or the upper side of the goaf thermal map.

[0059] Step S203: Please refer to Figure 3 As shown, mark the partitioned goaf thermal map as the goaf block map, denoted as TFi, and denote the blocks in the goaf block map as TFi(x, y), where x represents the xth block from left to right and y represents the yth block from bottom to top.

[0060] Step S204: For all blocks in the goaf block map, extract the temperature values corresponding to all pixel points in the block respectively, and sort the temperature values from largest to smallest, denoted as the temperature sequence of the TFi(x, y) block.

[0061] Step S205, set the first ratio to n%, in this embodiment, the first ratio n% = 35%; extract the first n% of the temperature values ​​of the TFi (x, y) block temperature sequence, and calculate the average value, marked as the TFi (x, y) block average high temperature value, recorded as W (i, x, y), the purpose is to improve the accuracy of subsequent predictions, because when a high temperature point appears in a certain block, the temperature around the high temperature point may still be normal. If the overall average is taken, it is bound to cause a large difference between the average value and the high temperature, thereby ignoring the high temperature point, that is, ignoring the dangerous factor. Therefore, taking the first proportion of the temperature as an average can avoid the failure to identify the high temperature and avoid the misjudgment of the abnormal pixel point as high temperature due to irrelevant environmental factors; and if the recognition accuracy is improved by reducing the block division method, it will lead to excessive subsequent data, affecting the efficiency of prediction analysis;

[0062] Step S206, storing the average high temperature value of the TFi (x, y) block according to the corresponding goaf block diagram and the corresponding TFi (x, y) position, and marking it as the first goaf block temperature data; each goaf block diagram includes the average high temperature value of the TFi (x, y) blocks of all blocks in the goaf block diagram, and each TFi (x, y) block includes the average high temperature value of the TFi (x, y) blocks collected and processed at all collection times;

[0063] In the specific implementation process, the setting of the first side length a1 and the moving step length a2 needs to be set according to the actual resolution of the thermal image and the area of ​​the goaf collected by each thermal imager. For example, the actual resolution of the thermal image is high, while the area of ​​the goaf collected by the imager is small, which means that the single pixel point of the thermal image corresponds to the actual goaf with high accuracy. At this time, the first side length a1 can be appropriately increased to reduce the amount of data and improve the efficiency of subsequent processing. The setting of the moving step length a2 determines the overlapping area of ​​adjacent blocks. The overlapping area of ​​adjacent blocks should not be too large or too small, and is generally between 25% and 50%. Therefore, the value range of a2 is generally [0.5*a1, 0.75*a1].

[0064] Step S3, processing the temperature data of the first goaf block to obtain the temperature data of the second goaf block; Step S3 includes the following sub-steps:

[0065] Step S301, based on the first goaf block temperature data, for all blocks in the goaf block map TFi, respectively extract the block average high temperature value W(i, x, y) of the block TFi(x, y), and then extract the goaf block map TF(i-1) obtained by the previous acquisition time. For example, in this embodiment, the first time interval is 12s, that is, the goaf block map TF(i-1) obtained by the acquisition and processing 12s ago and W(i-1, x, y) of the block corresponding to TF(i-1) are extracted;

[0066] Step S302, setting the first temperature weight to c1 and the second temperature weight to c2. In this embodiment, the first temperature weight c1=0.65, the second temperature weight c2=0.35, and c1>c2;

[0067] Step S303, using [c1*W(i, x, y)+c2*W(i-1, x, y)] to replace W(i, x, y), and recording the replaced TFi (x, y) block average high temperature value W(i, x, y) as TFi (x, y) block average high temperature value V(i, x, y);

[0068] For example, W(2, 2, 3) = 28.8°C, W(1, 2, 3) = 27.6°C, c1 = 0.65, c2 = 0.35, then c1*W(i, x, y) + c2*W(i-1, x, y) = 0.65*28.8 + 0.35*27.6 = 28.38°C

[0069] Step S304, repeatedly replace W(i, x, y) of all blocks in all goaf block diagrams, and obtain the temperature data of the second goaf block after completion;

[0070] In the specific implementation process, the purpose of replacing W(i, x, y) with [c1*W(i, x, y)+c2*W(i-1, x, y)] is to remove short-term fluctuations in temperature data caused by irrelevant factors, thereby improving data quality. For example, in thermal imaging acquisition, slight changes in the sensor environment, such as changes in air flow rate and ground vibration, will cause some small fluctuations in the data collected by thermal imaging. These fluctuations can be smoothed out by weighted averaging, making the temperature change trend clearer. The calculated smoothed data will better reflect the overall temperature change trend and reduce the impact of single-point data fluctuations. The reason why the first temperature weight c1>the second temperature weight c2 is that the more recent the data, the greater the weight, which can better reflect the timeliness of the data and improve the accuracy of subsequent prediction analysis.

[0071] Step S4, predicting and analyzing the spontaneous combustion risk of the goaf of the coal mine based on the temperature data of the second goaf block; Step S4 includes the following sub-steps:

[0072] Step S401, taking the average high temperature value V(i, x, y) of each block in the second goaf block temperature data at all acquisition times, and dividing the second goaf block temperature data into a first block temperature training set and a first block temperature test set at a ratio of 7:3 on a time scale, or 8:2 or 9:1, to ensure that more data is used for training;

[0073] Step S402, normalize the data in the first block temperature training set and the first block temperature test set, and normalize the data size to the [0,1] interval, and after completion, obtain the second block temperature training set and the second block temperature test set; the normalization formula is as follows: , where V1 is the normalized value, V0 is the original value, Vm can be the maximum value in the data, or it can be set by yourself. For example, in this embodiment, the safety temperature threshold T0=55°C, Vm can be set to 55 / 0.7=78°C;

[0074] Step S403, see Figure 4 As shown in the figure, based on the multi-layer perceptron model, the number of neurons in the input layer of the multi-layer perceptron model is set to 1, because only one temperature data is input; the number of hidden layers is 2, the number of neurons in the output layer is 1, and only one temperature data is output. The activation function is set to the ReLU function, and the mean square error is defined as the loss function. The set multi-layer perceptron model is marked as the original prediction model; the mean square error formula is as follows: , where Tz k is the actual temperature value in the first block temperature training set, Ty k is the predicted temperature value output by the model, MSE is the mean square error, and K is the total number of predicted temperature values ​​output by the model;

[0075] Step S404, setting the learning rate to e1, the number of training rounds to e2, and the batch size to e3, using the second block temperature training set to perform model training on the original prediction model, and obtaining the model to be tested after completion; in this embodiment, the learning rate e1=0.0001, the number of training rounds e2=200, and the batch size e3=32;

[0076] Step S405, using the second block temperature test set to test the model to be tested, and calculating the mean absolute error of the model to obtain the mean absolute error value MA of the model, setting the qualified threshold to MB, and judging whether MA satisfies MA≤MB; the mean absolute error calculation formula is as follows: , where MAE is the mean absolute error;

[0077] Step S406, if the conditions are met, the model to be tested is marked as a block temperature prediction model; if the conditions are not met, the model to be tested is trained again using the second block temperature training set, and after completion, the model to be tested is tested again using the second block temperature test set until a block temperature prediction model is obtained; for example, the qualified threshold MB is set to 0.15, the mean absolute error value of the model output MA=0.17, MA>MB, then the model to be tested is judged to be unqualified, and the model to be tested is trained again using the second block temperature training set;

[0078] Step S407: Based on the temperature data of the second goaf block, the future temperature change of the goaf area of ​​the coal mine corresponding to the goaf block diagram is predicted using the block temperature prediction model to obtain the predicted temperature data. It is necessary to obtain the historical data of the moving time length. For example, if the temperature change in the next 6 hours needs to be predicted, it is necessary to obtain the temperature data of the second goaf block in the previous 3 days.

[0079] Step S408, determining whether there is data in the predicted temperature data that is not less than the safety temperature threshold T0, if yes, determining that the coal mine goaf has a risk of spontaneous combustion, if not, determining that the coal mine goaf has no risk of spontaneous combustion for the time being.

[0080] In the specific implementation process, the multilayer perceptron is a simple feedforward neural network, which consists of an input layer, multiple hidden layers and an output layer; each neuron receives the output of the neurons in the previous layer, and passes it to the next layer after processing through weighted summation and activation function; for example, in time series prediction, the sequence value of the past period of time is used as the input of the input layer, and after being processed by the hidden layer, the value of the next time step is predicted in the output layer; the loss function is used to measure the difference between the model prediction result and the real data; the training set data is input into the model, the predicted output is calculated through forward propagation, and then the value of the loss function is calculated; then the weight parameters of the model are updated according to the loss function through the back propagation algorithm; this process will be repeated many times, called the number of training rounds, until the model converges or reaches the predetermined number of training rounds.

[0081] Example 2

[0082] See also Figure 5 As shown, Figure 5 The structural schematic diagram of an electronic device is illustrated, and the electronic device may include: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus. The memory stores computer-readable instructions, and the processor may call the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps in the coal mine risk analysis method based on the Internet of Things monitoring data are executed to achieve the following functions: real-time acquisition of thermal imaging images of coal mine goafs, marking of goaf thermal images, and spontaneous combustion risk monitoring and analysis based on the goaf thermal images; first division processing of the goaf thermal images, regional average temperature calculation based on the processed goaf thermal images, and obtaining first goaf block temperature data; data processing of the first goaf block temperature data to obtain second goaf block temperature data; and prediction and analysis of the spontaneous combustion risk of the coal mine goaf based on the second goaf block temperature data.

[0083] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0084] Example 3

[0085] The present invention also provides a computer-readable storage medium.

[0086] The present invention provides a storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps in the above-mentioned coal mine risk analysis method based on Internet of Things monitoring data are executed to realize the following functions: real-time acquisition of thermal imaging images of coal mine goafs, marking of goaf thermal images, and spontaneous combustion risk monitoring and analysis based on the goaf thermal images; performing a first division processing on the goaf thermal images, calculating the regional average temperature based on the processed goaf thermal images, and obtaining first goaf block temperature data; performing data processing on the first goaf block temperature data to obtain second goaf block temperature data; and predicting and analyzing the spontaneous combustion risk of coal mine goafs based on the second goaf block temperature data.

[0087] Through the description of the above implementation methods, the embodiments of the present invention can be provided as methods, systems or computer program products. Based on such an understanding, the above technical solutions can be essentially or partly contributed to the prior art in the form of software products, which can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and include several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0088] In the embodiments provided by the present invention, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of systems, modules and units can be electrical, mechanical or other forms.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A coal mine risk analysis method based on Internet of Things monitoring data, characterized in that: The steps include: Acquire thermal images of coal mine goafs in real time, mark the goaf thermal maps, and conduct spontaneous combustion risk monitoring and analysis based on the goaf thermal maps; Performing a first division process on the goaf thermal map, calculating the regional average temperature based on the processed goaf thermal map, and obtaining the first goaf block temperature data; Processing the temperature data of the first goaf block to obtain the temperature data of the second goaf block; Based on the temperature data of the second goaf block, the spontaneous combustion risk of the goaf in the coal mine is predicted and analyzed; Real-time acquisition of thermal imaging images of coal mine goafs, marking of goaf thermal maps, and spontaneous combustion risk monitoring and analysis based on goaf thermal maps include the following sub-steps: Based on a thermal imager connected to the Internet of Things, the relative position of the thermal imager and the coal mine goaf in space is kept unchanged, and the thermal image of the coal mine goaf is continuously collected at a first time interval, which is marked as a goaf thermal map, denoted as TUi, i represents the time when the goaf thermal map is collected, and the first time interval is tj; Based on the heat map of the goaf, the temperature values ​​corresponding to each pixel point of the heat map of the goaf are extracted, and the temperature is compared to obtain the highest temperature and the pixel point corresponding to the highest temperature, which are marked as the hottest pixel point, and the position of the hottest pixel point in the heat map of the goaf is determined; the first radius is set to r1, the second radius is set to r2, and the third radius is set to r3, r3>r2>r1; with the hottest pixel point as the center, circles are drawn on the heat map of the goaf with r1, r2 and r3 as the radii respectively, and the obtained circular area with a radius of r1 is marked as the first circular area; the annular area enclosed by the boundary of the circle with a radius of r2 and the boundary of the circle with a radius of r1 is marked as the second annular area; the annular area enclosed by the boundary of the circle with a radius of r3 and the boundary of the circle with a radius of r2 is marked as the third annular area; The spontaneous combustion risk monitoring and analysis based on the thermal map of the goaf also includes the following sub-steps: The temperature values ​​corresponding to the pixels in the first circular area, the second annular area, and the third annular area are extracted respectively; the weight of the first circular area is set to b1, the weight of the second annular area is set to b2, and the weight of the third annular area is set to b3; the average value TPi of the temperature values ​​corresponding to the pixels in the first circular area, the second annular area, and the third annular area is calculated by the weighted average formula, and the weighted average formula is as follows: , where T1 is the average temperature value corresponding to all pixels in the first circular area; T2 is the average temperature value corresponding to all pixels in the second annular area; T3 is the average temperature value corresponding to all pixels in the third annular area; Determine whether TPi exceeds the safety temperature threshold. When TPi exceeds the safety temperature threshold, it is determined that the coal mine goaf has a risk of spontaneous combustion, and the safety temperature threshold is T0.

2. The method for analyzing coal mine risks based on Internet of Things monitoring data according to claim 1 is characterized in that: The first division process of the thermal map of the goaf includes the following sub-steps: Set the first side length to a1 and the moving step length to a2, a2 <a1; Starting from the lower left corner of the goaf heat map with a square with a side length of a1, the goaf heat map is divided from left to right and from bottom to top. The square with a side length of a1 moves horizontally or vertically by a2 steps each time, and the divided goaf heat map is marked as a goaf block map, denoted as TFi, and the blocks in the goaf block map are denoted as TFi (x, y), where x represents the xth block from left to right, and y represents the yth block from bottom to top.

3. The method for analyzing coal mine risks based on Internet of Things monitoring data according to claim 2 is characterized in that: Calculating the regional average temperature based on the processed thermal map of the goaf to obtain the temperature data of the first goaf block includes the following sub-steps: For all blocks in the goaf block map, the temperature values ​​corresponding to all pixels in the block are extracted respectively, and the temperature values ​​are sorted from large to small, which is recorded as TFi (x, y) block temperature sequence; Set the first ratio to n%, extract the first n% of the temperature values ​​of the TFi (x, y) block temperature sequence, and calculate the average value, which is marked as the average high temperature value of the TFi (x, y) block, recorded as W(i, x, y); The average high temperature value of the TFi (x, y) block is stored according to the goaf block map and the TFi (x, y) position to which it belongs, and marked as the first goaf block temperature data; each goaf block map includes the average high temperature value of the TFi (x, y) blocks of all blocks in the goaf block map, and each TFi (x, y) block contains the average high temperature value of the TFi (x, y) blocks obtained by all acquisition times.

4. The method for analyzing coal mine risks based on Internet of Things monitoring data according to claim 3 is characterized in that: Processing the temperature data of the first goaf block to obtain the temperature data of the second goaf block includes the following sub-steps: Based on the first goaf block temperature data, for all blocks in the goaf block diagram TFi, extract the average high temperature value W(i, x, y) of the block TFi (x, y) respectively, then extract the goaf block diagram TF(i-1) obtained by the previous acquisition time, and W(i-1, x, y) of the block at the corresponding position of TF(i-1); set the first temperature weight to c1, the second temperature weight to c2, c1>c2; use [c1*W(i, x, y)+c2*W(i-1, x, y)] to replace W(i, x, y), and record the replaced TFi (x, y) block average high temperature value W(i, x, y) as TFi (x, y) block average high temperature value V(i, x, y); repeat the replacement of W(i, x, y) of all blocks in all goaf block diagrams, and obtain the second goaf block temperature data after completion.

5. The method for coal mine risk analysis based on Internet of Things monitoring data according to claim 4 is characterized in that: The prediction and analysis of the spontaneous combustion risk of each block in the goaf based on the temperature data of the second goaf block includes the following sub-steps: The average high temperature value V(i, x, y) of each block in the second goaf block temperature data of all acquisition times is taken, and the second goaf block temperature data is divided into the first block temperature training set and the first block temperature test set in a ratio of 7:3 on the time scale; Normalize the data in the first block temperature training set and the first block temperature test set, and normalize the data size to the [0,1] interval. After completion, obtain the second block temperature training set and the second block temperature test set; Based on the multilayer perceptron model, the number of neurons in the input layer of the multilayer perceptron model is set to 1, the number of hidden layers is set to 2, the number of neurons in the output layer is set to 1, the activation function is set to the ReLU function, the mean square error is defined as the loss function, and the completed multilayer perceptron model is marked as the original prediction model.

6. The method for coal mine risk analysis based on Internet of Things monitoring data according to claim 5 is characterized in that: The prediction and analysis of the spontaneous combustion risk of each block in the goaf based on the temperature data of the second goaf block also includes the following sub-steps: Set the learning rate to e1, the number of training rounds to e2, and the batch size to e3. Use the second block temperature training set to train the original prediction model. After completion, the model to be tested is obtained. The model to be tested is tested using the second block temperature test set, and the calculated mean absolute error of the model is calculated to obtain the mean absolute error value MA of the model. The qualified threshold is set to MB, and it is determined whether MA satisfies MA≤MB. If so, the model to be tested is marked as a block temperature prediction model; if not, the model to be tested is trained again using the second block temperature training set, and after completion, the model to be tested is tested again using the second block temperature test set until a block temperature prediction model is obtained; Based on the temperature data of the second goaf block, the block temperature prediction model is used to predict the future temperature changes of the coal mine goaf corresponding to the goaf block diagram to obtain the predicted temperature data, and to determine whether there is data in the predicted temperature data that is not less than the safety temperature threshold T0. If so, it is determined that the coal mine goaf has a risk of spontaneous combustion; if not, it is determined that the coal mine goaf has no risk of spontaneous combustion for the time being.

7. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in the method according to any one of claims 1 to 6 are executed.

8. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method according to any one of claims 1 to 6 are executed.

Images