A coal mine tunnel gas outburst monitoring and early warning method

Abstract

The present invention discloses a coal mine tunnel gas outburst monitoring and early warning method, comprising: collecting monitoring data of the coal mine tunnel through a sensor group, monitoring the monitoring data in real time to obtain detection results, constructing a prediction model, predicting the monitoring data through the prediction model to obtain prediction data, judging the prediction data to obtain early warning results; constructing a connection model, the connection model is used to map the correlation between the coal mine tunnel position and the monitoring data, obtain an initial position, optimize the initial position through an optimization algorithm, a connection model and a prediction model, and obtain a predicted gas outburst position.

Description

Technical Field

[0001] The invention belongs to the technical field of coal mine tunnel gas outburst monitoring, and in particular relates to a coal mine tunnel gas outburst monitoring and early warning method. Background Art

[0002] Coal mine tunnels are a general term for various underground spaces that serve coal mining. They are excavated in different rocks along different directions, at different inclinations, in different sections and lengths, and are used for ventilation, drainage, pedestrians, transportation, and mining operations. Gas outbursts in coal mine tunnels can affect the safety of underground personnel in coal mines. The coal mine tunnel gas outburst monitoring and early warning can effectively improve the safety of underground personnel in coal mines. However, the existing technology cannot provide an effective early warning method for coal mine tunnel gas outbursts, and its early warning method cannot effectively predict the location where gas outbursts may occur. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a coal mine tunnel gas outburst monitoring and early warning method to solve the above problems existing in the prior art.

[0004] To achieve the above object, the present invention provides a coal mine tunnel gas outburst monitoring and early warning method, comprising:

[0005] Collect monitoring data of coal mine tunnels through the sensor group, monitor the monitoring data in real time, obtain detection results, build a prediction model, predict the monitoring data through the prediction model to obtain prediction data, judge the prediction data, and obtain early warning results;

[0006] A connection model is constructed, and the connection model is used to map the correlation between the coal mine tunnel position and the monitoring data, obtain the initial position, optimize the initial position through the optimization algorithm, the connection model and the prediction model, and obtain the predicted gas outburst position.

[0007] Optionally, the monitoring data includes gas concentration information, air pressure information, and wind speed and direction information.

[0008] Optionally, the prediction model adopts a deep learning model, wherein the deep learning model adopts a fully connected neural network model, and the prediction model is trained through historical monitoring data of coal mine tunnels.

[0009] Optionally, the connection model adopts a deep learning model, wherein the deep learning model adopts a fully connected neural network model, and the connection model is trained in chronological order using the location information of the sensor group and the monitoring data corresponding to the sensor group at a previous fixed time and the monitoring data at the current moment.

[0010] Optionally, the process of predicting the monitoring data through the prediction model includes:

[0011] The monitoring data of the current sensor group and several adjacent sensors at the current time and the previous fixed time are obtained, the monitoring data are encoded according to the number and time of the sensor group, a coding matrix is ​​generated, and the coding matrix is ​​input into the prediction model to obtain the predicted monitoring data at the future interval time.

[0012] Optionally, the input of the connection model is the location information of a test point in the coal mine tunnel, and the output is the connection gas concentration and air pressure of the test point.

[0013] Optionally, the process of optimizing the initial position includes:

[0014] The goal is to maximize the predicted monitoring data obtained by processing the location through the connection model and the prediction model in sequence;

[0015] The initial position is updated, and the updated position is input into the connection model to obtain updated connection monitoring data. According to the updated position, the adjacent monitoring data of several closest sensor groups are screened. According to the updated connection monitoring data and the adjacent monitoring data, an updated coding matrix is ​​encoded. The updated coding matrix is ​​predicted by the prediction model to obtain predicted monitoring data. The predicted monitoring data is judged, and the updated position is updated again. The above process is repeated until the maximum number of iterations is reached to obtain the predicted gas outburst position.

[0016] Optionally, the optimization algorithm adopts a moth-to-flame algorithm.

[0017] Compared with the prior art, the present invention has the following advantages and technical effects:

[0018] Through the above technical scheme, the present invention collects monitoring data of coal mine tunnels through a sensor group, monitors the monitoring data in real time, obtains monitoring results, and monitors the coal mine tunnels in real time. At the same time, a prediction model is constructed, and the monitoring data is predicted by the prediction model to obtain prediction data. Based on the prediction data, the possible gas outbursts in the future are monitored; as an extension, the sensor group collects data from limited points, and a connection model can be constructed to extend it to all tunnel point data. Through the optimization algorithm, the connection model and the prediction model, the predicted gas outburst position is obtained to realize the prediction of the possible gas outburst position. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0020] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0021] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0022] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0023] like Figure 1 As shown, this embodiment provides a coal mine tunnel gas outburst monitoring and early warning method, including:

[0024] Collect monitoring data of coal mine tunnels through the sensor group, monitor the monitoring data in real time, obtain detection results, build a prediction model, predict the monitoring data through the prediction model to obtain prediction data, judge the prediction data, and obtain early warning results;

[0025] Construct a connection model, and obtain the predicted gas outburst location through optimization algorithm, connection model and prediction model.

[0026] As some embodiments, sensor groups are arranged at equal intervals in coal mine lanes, wherein the sensor groups include gas concentration sensors, air pressure sensors, and wind speed and direction sensors, etc., and different types of sensors are used to detect gas concentration information, air pressure information, and wind speed and direction information in the coal mine lanes. The above sensor groups are arranged at equal intervals in the coal mine lanes as needed, and the sensor groups are numbered in sequence to detect relevant information in the coal mine lanes, so as to provide a basis for monitoring data for subsequent related monitoring and early warning.

[0027] As some embodiments, after collecting the above monitoring data, the monitoring data is detected in real time, and the gas concentration information, air pressure information, wind direction and wind speed information are judged in real time. The tunnel is monitored in real time according to the judgment result, and the gas concentration information and air pressure information are judged by threshold. When a gas concentration information exceeds a certain threshold, and the air pressure information exceeds a certain threshold, it is considered that a gas outburst has occurred, and a timely warning is given. At the same time, the coal and gas outburst monitoring method can be combined with the wind direction to monitor the gas outburst in the coal mine. The threshold of the gas concentration can be set to 3%, and the air pressure can be set to 1200Pa. To obtain relevant monitoring results.

[0028] As some embodiments, after collecting the above-mentioned monitoring data, the present invention also predicts the gas concentration information in a subsequent period of time based on the above-mentioned monitoring data, and issues an early warning in the future based on the predicted gas concentration information. Specifically, the present invention makes relevant predictions on the above-mentioned monitoring data through a prediction model, wherein the prediction model adopts a deep learning model, and collects the gas concentration information, air pressure information and wind speed and direction information of the current moment and the previous period of time from the current sensor group and several adjacent or closest sensor groups, such as 6, as input data of the deep learning model, and the input data is identified by the deep learning model to obtain the predicted gas concentration information and air pressure information in the next period of time, and encode the input data into three matrices, wherein each row of the matrix represents the corresponding sensor collecting data at different times, the sensor currently to be predicted is taken as the first row, and the remaining sensors are set in sequence according to the number, and each column represents the corresponding time, wherein the sampling interval for collecting data can be set, such as 1h, 2h, etc. The deep learning model adopts a fully connected neural network model, wherein the structure of the fully connected neural network model includes an input layer, several hidden layers, two fully connected layers and an output layer. The number of hidden layers is 5, and each hidden layer contains 200 neurons. The corresponding input data is identified by the above-mentioned deep learning model to obtain the predicted data corresponding to the current sensor group. The interval time between the predicted data and the input data, such as 6h, 12h or 24h, or the time of the predicted data, is determined according to the training data. The training data includes historical data collected by the above-mentioned sensor group. The data of the previous period of time in the historical data is used as training data, and the data collected in the subsequent period of time that needs to be predicted is used as a label. The deep learning model is trained by the training data to obtain predicted gas concentration information and air pressure information.

[0029] As some embodiments, based on the above-mentioned deep learning model, the information of the sensor group at a certain time interval in the future can be predicted, and the future time period can be used as the basis, that is, the data of different sensors at different times are predicted through the above-mentioned deep learning model to obtain the data of different sensor groups in the future period, and combined with the wind direction and wind speed information at the current moment, on this basis, the data of the subsequent time in the future period is predicted, and so on, the data in the future time at a longer time can be predicted. Through this solution, the sensor monitoring prediction related results can be obtained. When the prediction related results exceed the threshold, it is considered that there is a risk of gas outburst at the location of the sensor.

[0030] As some embodiments, in addition to predicting the monitoring data at a future time, the present invention also provides a related method for predicting the location of a possible gas outburst accident. Specifically, based on the data collected by each sensor group, a related connection model can be used to simulate the mapping relationship between the sensor position and the data collected by the sensor at the current moment, wherein the connection model also adopts a deep learning model, and the deep learning model can adopt a fully connected neural network, wherein the structure of the fully connected neural network model includes an input layer, a plurality of hidden layers, two fully connected layers and an output layer, the number of hidden layers is 3, and each hidden layer contains 100 neurons, and its position information is the three-dimensional position information of the current sensor group in the coal mine tunnel, and its output is the corresponding collected gas concentration information and air pressure information. The connection model is optimized in real time by sequentially training the information collected by the sensor for a period of time before, such as 4h, and the current moment. Through this model, the monitoring of a limited position under the tunnel can be expanded to the prediction of the current data collection at each position in the tunnel. It should be noted that the above is only an exemplary display of the relevant deep learning model category, and other more accurate models can also be selected for relevant predictions and connection fitting.

[0031] As some embodiments, based on the above-mentioned connection model and prediction model, an optimization algorithm is used to find the possible location of gas outburst. Specifically, the connection model is trained through monitoring data collected by sensors at the current time and a period of time before, such as within the previous 4 hours, to generate an optimized connection model.

[0032] An initial population is set, wherein the initial population includes individuals representing the location,

[0033] The initial population is input into the optimized connection model to generate monitoring data of the current position. The monitoring data of the current position is combined with several nearest sensor groups and the current wind speed and direction information to predict the monitoring data of the current position in the future to obtain predicted data. The predicted data is judged to determine whether it is the predicted data with the largest value. If it is not the largest predicted data, the initial population is optimized and the above process is repeated until the maximum number of iterations is reached to generate the final optimized population. The optimized population predicts the future data of the current position through the connection model to obtain predicted gas concentration and air pressure data. The gas concentration and air pressure data of the optimized population position are judged to obtain the final early warning result.

[0034] As some embodiments, the specific steps of the optimization algorithm are:

[0035] 1) Initializing a population, wherein the population includes a number of different individuals or moth positions, and the individuals include the position of a certain point in the coal mine tunnel, wherein the position selected according to artificial experience or randomly is used as the initialization individual.

[0036] Set the relevant objective function, constraint conditions and corresponding fitness, where the objective function predicts the model output value to maximize the weighted sum of the predicted gas concentration and air pressure. The constraint conditions include the boundary conditions of the location to ensure that the location search does not exceed the scope of the coal mine tunnel, and the fitness function is the weighted sum of the predicted gas concentration and air pressure.

[0037] 2) In the improved moth-to-fire algorithm, the moth needs to update and iterate the flames around it. The flames represent the local optimal solution, and the number of flames is determined. At the same time, during the iteration process, the number of flames needs to be adaptively reduced. The adaptive reduction mechanism formula is as follows:

[0038] FN = round(Nt*(N-1) / T)

[0039] Where: N is the initial number of flames; t is the current number of iterations; T is the maximum number of iterations, and FN represents the number of flames during the tth iteration.

[0040] 3) The fitness of different individuals is calculated according to the objective function. The greater the fitness, the better the quality of the individual. During the calculation process, the data corresponding to the individual, that is, the position data of a certain point in the tunnel, is input into the connection model, and the collected data of several nearest sensors are obtained. The wind direction data is the average value of several nearest sensors and is combined into a coding matrix. The above coding matrix is ​​predicted by the prediction model to obtain the predicted gas concentration and air pressure at the current position.

[0041] The fitness is sorted from large to small. After sorting, the corresponding number of flames is selected according to the sorting result, and the individuals in the fitness sorting are used as flames. The number of flames is calculated according to the above fitness reduction mechanism formula, and the current best flame is stored. The best flame is the individual with the highest number of flames in the fitness sorting.

[0042] 4) According to the above individual update method, or the moth position update method, the update is performed. In the update process, the initial position update method is first used for updating. In the initial position update method, the above moth position update method is adaptively updated through the inertia weight:

[0043] Mi=Di·e bθ ·cos(2πθ)+ω·Fj

[0044] Among them, b is a constant used to define the shape of the logarithmic spiral; θ is a random number between [-1,1]; Mi represents the i-th moth, that is, the individual; Fj represents the j-th flame; Di represents the distance between the i-th moth and the j-th flame, and the formula is the constant of the logarithmic spiral shape; the formula is: Di=|Fj-Mi|, ω is the inertia weight, which refers to the inertia weight idea of ​​the particle swarm algorithm. The larger the ω, the greater the search intensity in the global range of the algorithm, and the smaller the ω, the higher the search accuracy in the local range of the algorithm.

[0045]

[0046] Where t is the current iteration number; T is the maximum iteration number; a and d are constants respectively.

[0047] 5) Determine whether the flame data satisfies the requirement of only one flame remaining. If only one flame remains, directly output the final result;

[0048] 6) Determine whether the maximum number of iterations has been reached. If not, repeat steps 2-6 above. If reached, output the optimal moth individual, that is, the location with the highest possible gas concentration that is finally found. The maximum number of iterations is set according to manual experience. The location may be a non-sensor monitoring location. By judging the location, the predicted gas concentration is obtained through the connection model and the prediction model. Whether the predicted gas concentration and air pressure exceed the threshold value, if the threshold value is exceeded, the location may have a risk of gas outburst. If the threshold value is not exceeded, the location as the location with the highest gas concentration also has certain risks, and its location is prompted and warned.

[0049] Through the above technical scheme, the present invention collects monitoring data of coal mine tunnels through a sensor group, monitors the monitoring data in real time, obtains monitoring results, and monitors the coal mine tunnels in real time. At the same time, a prediction model is constructed, and the monitoring data is predicted by the prediction model to obtain prediction data. Based on the prediction data, the possible gas outbursts in the future are monitored; as an extension, the sensor group collects data from limited points, and a connection model can be constructed to extend it to all tunnel point data. Through the optimization algorithm, the connection model and the prediction model, the predicted gas outburst position is obtained to realize the prediction of the possible gas outburst position.

[0050] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A coal mine tunnel gas outburst monitoring and early warning method, characterized in that: include: Collect monitoring data of coal mine tunnels through the sensor group, monitor the monitoring data in real time, obtain detection results, build a prediction model, predict the monitoring data through the prediction model to obtain prediction data, judge the prediction data, and obtain early warning results; Constructing a connection model, the connection model is used to map the correlation between the coal mine tunnel position and the monitoring data, obtain the initial position, optimize the initial position through the optimization algorithm, the connection model and the prediction model, and obtain the predicted gas outburst position; The prediction model adopts a deep learning model, wherein the deep learning model adopts a fully connected neural network model, and the prediction model is trained by historical monitoring data of coal mine tunnels; The connection model adopts a deep learning model, wherein the deep learning model adopts a fully connected neural network model, and the connection model is trained in sequence according to the time sequence using the location information of the sensor group and the monitoring data at a previous fixed time corresponding to the sensor group and the monitoring data at the current time; The process of predicting monitoring data through the prediction model includes: Acquire monitoring data of the current sensor group and several neighboring sensors at the current time and the previous fixed time, encode the monitoring data according to the number and time of the sensor group, generate a coding matrix, input the coding matrix into the prediction model, and obtain the predicted monitoring data at the future interval time; The input of the connection model is the location information of a test point in the coal mine tunnel, and the output is the connection gas concentration and air pressure of the test point; The process of optimizing the initial position includes: The goal is to maximize the predicted monitoring data obtained by sequentially processing the position through the connection model and the prediction model; update the initial position, input the updated position into the connection model to obtain updated connection monitoring data, filter the adjacent monitoring data of several closest sensor groups according to the updated position, encode the updated connection monitoring data and the adjacent monitoring data to obtain an updated encoding matrix, predict the updated encoding matrix through the prediction model to obtain predicted monitoring data, judge the predicted monitoring data, and update the updated position again, repeat the above process until the maximum number of iterations is reached to obtain the predicted gas outburst position; The optimization algorithm adopts the moth-to-flame algorithm.

2. The method according to claim 1, characterized in that: The monitoring data includes gas concentration information, air pressure information and wind speed and direction information.

Images