Method for identifying, regulating and controlling abnormal CO emission causes based on drilling gas extraction data

Abstract

The invention relates to the technical field of coal mine gas extraction, in particular to a CO abnormal emission cause identification and regulation method based on drill hole gas extraction data. Comprising the following steps: S1, collecting bedding drilling gas extraction data of a working surface similar to an existing project, and setting a label for a carbon monoxide emission condition; s2, performing data processing, establishing a data set, and dividing the data set into a test set, a verification set and a training set; s3, constructing a classification model based on the neural network, and carrying out training and verification; s4, collecting bedding drill hole gas extraction data of an existing project, and analyzing bedding drill hole gas extraction conditions of the existing project by utilizing the trained classification model; and S5, for the bedding drill hole with the abnormal carbon monoxide emission condition, performing bedding drill hole gas extraction regulation and control based on the corresponding abnormal carbon monoxide emission type. According to the method, the disaster cause of abnormal emission of carbon monoxide can be quickly identified through the existing bedding drilling gas extraction data, so that adjustment is carried out in time, and the safe and efficient extraction effect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of coal mine gas extraction, and in particular to a method for identifying and regulating the cause of abnormal CO emission based on borehole gas extraction data. Background Art

[0002] With the continuous deepening of coal mining activities, the prevention and control of coal mine gas disasters has become an important issue in coal mine safety production. Gas disasters not only threaten the lives of miners, but also seriously affect the economic benefits and sustainable development of coal mines. As an important means of preventing gas disasters, the effectiveness of gas extraction technology is directly related to the safety of coal mine production.

[0003] In the field of coal mine gas, gas extraction by drilling holes in the coal seam is a common method of gas control. This method uses negative pressure to extract gas from deep in the coal seam by drilling holes in the coal seam. However, in the actual extraction process, abnormal carbon monoxide (CO) may sometimes be encountered, which not only increases the difficulty of gas control, but may also indicate deeper safety risks. Accurate identification of the cause of abnormal carbon monoxide emission is crucial to the formulation of effective gas control measures. Traditional identification methods rely on expert analysis and manual experience judgment, which are inefficient and have limited accuracy.

[0004] In recent years, with the development of artificial intelligence technology, new ideas have been provided for the identification of the causes of abnormal carbon monoxide emission. By establishing a classification model based on a neural network, the causes of abnormal carbon monoxide emission in the layered boreholes can be quickly and accurately identified, thereby guiding the gas extraction and control to achieve safe and efficient gas management. In order to address the deficiencies in the prior art, the present invention aims to provide a method for identifying and controlling the causes of abnormal CO emission based on borehole gas extraction data through data processing and neural network technology. Summary of the invention

[0005] Purpose of the invention: In view of the deficiencies in the prior art, the present invention proposes a method for identifying and regulating the cause of abnormal CO emission based on borehole gas extraction data, thereby achieving the purpose of rapid identification of the cause of abnormal carbon monoxide emission during in-layer borehole gas extraction, gas extraction regulation and safe and efficient extraction.

[0006] Technical solution: To achieve the above purpose, the present invention establishes a classification model for abnormal carbon monoxide emission during gas extraction from bed boreholes based on neural networks through basic data collection, so as to classify and identify the carbon monoxide emission during gas extraction from bed boreholes on the working face, and then realize safe extraction through negative pressure regulation and process adjustment. Specifically, the present invention includes the following steps:

[0007] S1. Collect the gas extraction data of the working face with the same or similar coal mining conditions as the existing project, the coal mining conditions include mining parameters and coal quality parameters, the gas extraction data of the layer includes several items, each of which includes extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, extracted gas flow rate, and corresponding carbon monoxide emission condition labels, the carbon monoxide emission conditions include no abnormal emission of CO, primary carbon monoxide, slow oxidation of coal body, oxidation and aggregation of coal residues in adjacent goafs, friction oxidation of drilling tools, and oxidation of air leakage in the layer boreholes;

[0008] S2. Fill in the missing parameters in the in-layer gas extraction data or directly delete the in-layer gas extraction data, and correct the erroneous parameters; detect outliers and remove or correct them; normalize each parameter; obtain the final in-layer gas extraction data set, and divide the data set into a test set, a validation set, and a training set;

[0009] Preferably, in step S2, the normalization processing method includes Z-score normalization and minimum-maximum normalization.

[0010] Preferably, in step S2, if the number of data items of in-layer borehole gas extraction under a certain label is small, the number of data items of in-layer borehole gas extraction under the label is made close to the number of data items of in-layer borehole gas extraction under other labels by copying or resampling using oversampling technology.

[0011] S3. Select a gated recurrent neural network to construct a classification model for abnormal carbon monoxide emission in gas extraction from in-layer boreholes, wherein the classification model uses extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, extracted gas cumulative volume, and extracted gas flow rate as input parameters, and uses a carbon monoxide emission label as an output parameter; train the classification model based on the training set; use an optimization algorithm to perform optimization training, use a validation set to evaluate the performance of the classification model under different hyperparameter settings, and select the optimal hyperparameter configuration to establish a classification model for abnormal carbon monoxide emission in gas extraction from in-layer boreholes;

[0012] Preferably, in step S3, a test set is used to evaluate the performance of the final classification model, thereby verifying the performance of the classification model in practical applications.

[0013] Preferably, in step S3, the Adam gradient descent optimization algorithm is used to optimize the internal parameters of the classification model, and the genetic algorithm is used to adjust and optimize the hyperparameters in the classification model to finally obtain the optimal value of the number of hidden layer units in the classification model.

[0014] S4. Collect gas extraction data from existing projects in-layer boreholes, including extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, and extracted gas flow rate; use the trained classification model to analyze the gas extraction situation of existing projects in-layer boreholes, and determine the in-layer boreholes with abnormal carbon monoxide emission and the corresponding abnormal carbon monoxide emission types;

[0015] S5. For the in-layer boreholes with abnormal carbon monoxide outflow, gas extraction control and process improvement of in-layer boreholes are carried out based on the corresponding abnormal carbon monoxide outflow type, including: For native carbon monoxide: hydraulic cutting and high-pressure water jet hole expansion technology are used to increase coal seam fissures and improve coal seam permeability, and high-extraction extraction is used to reduce the amount of carbon monoxide outflowing from in-layer boreholes; For slow oxidation of coal body: optimize ventilation mode to improve air fluidity and remove heat, and use coal seam grouting technology to reduce the slow oxidation rate of coal body, reduce coal body exposure area and time ; Regarding the oxidation and accumulation of coal residues in adjacent goafs: seal the goafs to reduce air entry, reduce the contact between coal residues and air, reduce the possibility of oxidation reactions, fill the goafs, clean the coal residues, set up ventilation isolation zones, and inject inert gas to replace oxygen; Regarding friction oxidation during drilling with drill tools: select drill tool materials with low friction coefficients to reduce friction heat during drilling, install a cooling system on the drill tools, and control the drilling speed; Regarding air leakage oxidation in bedding boreholes: improve the sealing of bedding boreholes, inject nitrogen into bedding boreholes to drive out oxygen, and adjust the sealing depth.

[0016] The inventive point and beneficial effects of the present invention: The method for identifying and regulating the cause of abnormal CO outburst based on borehole gas extraction data of the present invention can quickly identify the cause of abnormal carbon monoxide outburst through the existing in-layer borehole gas extraction data, so as to make timely adjustments to achieve the effect of safe and efficient extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The present invention is a flow chart of the method for identifying and controlling the cause of abnormal CO emission based on borehole gas extraction data. DETAILED DESCRIPTION

[0018] In the specific implementation section, the Figure 1 The technical solution of the present invention is elaborated in detail.

[0019] Specifically, the method for identifying and controlling the cause of abnormal CO emission based on borehole gas extraction data of the present invention comprises the following steps:

[0020] S1. Collect gas extraction data of the seam boreholes of the working face with the same or similar coal mining conditions as the existing engineering, the coal mining conditions include mining parameters such as coal seam mining height, coal seam mining depth, coal seam inclination, working face size, and coal quality parameters such as coal type and coal seam gas content, the seam borehole gas extraction data includes several items, each of which includes extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, extracted gas flow parameters, and corresponding carbon monoxide emission condition labels, the carbon monoxide emission conditions include but are not limited to abnormal emission without CO, primary carbon monoxide (referring to carbon monoxide naturally generated in coal formation and coalification, that is, during the formation of coal seams), slow oxidation of coal body, oxidation and aggregation of coal residues in adjacent goafs, friction oxidation of drilling tools, and oxidation of air leakage in seam boreholes;

[0021] Note: Generally, carbon monoxide will not appear during gas extraction. When carbon monoxide appears, it is an abnormal phenomenon.

[0022] S2. Fill in the missing parameters in the layer borehole gas extraction data or directly delete the layer borehole gas extraction data, and correct the erroneous parameters; use quartiles to define the normal range of each parameter, so as to detect outliers, and perform elimination or correction operations; normalize each parameter, and the normalization method includes Z score normalization and minimum-maximum normalization; in the case where the number of layer borehole gas extraction data under a certain label is small, the layer borehole gas extraction data of this category can be randomly copied until its number is close to the number of layer borehole gas extraction data under other labels, or a layer borehole gas extraction data and a layer borehole gas extraction data close to it can be selected, and new layer borehole gas extraction data can be generated according to the difference between the two layer borehole gas extraction data and a random number, or resampling can be performed using oversampling technology (SMOTE); then the final layer borehole gas extraction data set is obtained, and the data set is divided into a test set, a validation set, and a training set;

[0023] S3. Select a gated recurrent neural network (GRU) to construct a classification model for abnormal carbon monoxide emission in gas extraction from in-layer boreholes. The classification model uses extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, extracted gas cumulative volume, and extracted gas flow rate as input parameters, and uses the carbon monoxide emission condition (type) label as output parameter. The number of hidden layers of the classification model is 4, the number of neurons in each hidden layer is set as a random number, and the cross entropy loss is the loss function. The classification model is trained based on the training set. The optimization algorithm is used to calculate the error. The optimization training is performed by using the validation set to evaluate the performance of the classification model under different hyperparameter settings, and the optimal hyperparameter configuration is selected to establish a classification model for abnormal carbon monoxide emission from gas extraction in bedding boreholes. The performance of the final classification model is evaluated using the test set, thereby verifying the performance of the classification model in practical applications. In this example, the Adam gradient descent optimization algorithm is used to optimize the internal parameters of the classification model, and the genetic algorithm is used to adjust and optimize the hyperparameters in the classification model. Finally, the optimal value of the number of hidden layer units of the final classification model is obtained, and the classification model at this time is the target classification model.

[0024] S4. Collect gas extraction data of existing engineering in-layer boreholes through sensors and other data acquisition devices, including extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, and extracted gas flow parameters; use the trained classification model to analyze the gas extraction situation of existing engineering in-layer boreholes, and determine the in-layer boreholes with abnormal carbon monoxide emission and the corresponding abnormal carbon monoxide emission types;

[0025] S5. For the in-layer boreholes with abnormal carbon monoxide outflow, gas extraction control and process improvement of in-layer boreholes are carried out based on the corresponding abnormal carbon monoxide outflow type, including: For native carbon monoxide: hydraulic cutting and high-pressure water jet hole expansion technology are used to increase coal seam fissures and improve coal seam permeability, and high-extraction extraction is used to reduce the amount of carbon monoxide outflowing from in-layer boreholes; For slow oxidation of coal body: optimize ventilation mode to improve air fluidity and remove heat, and use coal seam grouting technology to reduce the slow oxidation rate of coal body, reduce coal body exposure area and time ; Regarding the oxidation and accumulation of coal residues in adjacent goafs: seal the goafs to reduce air entry, reduce the contact between coal residues and air, reduce the possibility of oxidation reactions, fill the goafs, clean the coal residues, set up ventilation isolation zones, and inject inert gas to replace oxygen; Regarding friction oxidation during drilling with drill tools: select drill tool materials with low friction coefficients to reduce friction heat during drilling, install a cooling system on the drill tools, and control the drilling speed; Regarding air leakage oxidation in bedding boreholes: improve the sealing of bedding boreholes, inject nitrogen into bedding boreholes to drive out oxygen, and adjust the sealing depth.

[0026] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for identifying and controlling the cause of abnormal CO emission based on borehole gas extraction data, characterized in that: The following steps are involved: S1. Collect gas extraction data of the in-layer boreholes of the working face with the same or similar coal mining conditions as the existing engineering, the coal mining conditions include mining parameters and coal quality parameters, the in-layer borehole gas extraction data includes several items, each of which includes extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, extracted gas flow rate, and corresponding carbon monoxide emission condition labels, the carbon monoxide emission conditions include no abnormal emission of CO, primary carbon monoxide, slow oxidation of coal body, oxidation and aggregation of coal residues in adjacent goafs, friction oxidation of drilling tools, and oxidation of air leakage in in-layer boreholes; S2. Fill in the missing parameters in the gas extraction data of the layer borehole or directly delete the gas extraction data of the layer borehole, and correct the wrong parameters; detect abnormal values ​​and perform elimination or correction operations; Normalize each parameter; Obtain the final in-seam gas extraction data set and divide the data set into a test set, a validation set, and a training set; S3. Select a gated recurrent neural network to construct a classification model for abnormal carbon monoxide emission during gas extraction from a layered borehole, wherein the classification model uses extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, and extracted gas flow rate as input parameters, and uses a carbon monoxide emission label as an output parameter; and trains the classification model based on the training set; The optimization algorithm is used for optimization training, and the validation set is used to evaluate the performance of the classification model under different hyperparameter settings. The optimal hyperparameter configuration is selected to establish a classification model for abnormal carbon monoxide emission in gas extraction from in-layer boreholes. S4. Collect gas extraction data of existing project in-layer boreholes, including extraction negative pressure, extracted gas concentration, extracted carbon monoxide concentration, extracted gas temperature, extracted gas cumulative volume and extracted gas flow rate parameters; use the trained classification model to analyze the gas extraction situation of existing project in-layer boreholes, determine the in-layer boreholes with abnormal carbon monoxide emission and the corresponding abnormal carbon monoxide emission types; S5. For the in-layer boreholes with abnormal carbon monoxide outburst, gas extraction control and process improvement of in-layer boreholes are carried out based on the corresponding abnormal carbon monoxide outburst type, including: For native carbon monoxide: hydraulic cutting and high-pressure water jet hole expansion technology are used to increase coal seam fissures and improve coal seam permeability, and high-extraction extraction is used to reduce the amount of carbon monoxide outburst from in-layer boreholes; For slow oxidation of coal body: ventilation mode is optimized to increase air fluidity and remove heat, and coal seam grouting technology is used to reduce the slow oxidation rate of coal body and reduce the exposed area and time of coal body; For oxidation and accumulation of coal residue in adjacent goafs: the goafs are sealed to reduce air entry, reduce the contact between coal residue and air, reduce the possibility of oxidation reaction, fill the goafs, clean the coal residue, set up ventilation isolation zones, and inject inert gas to replace oxygen; Regarding the friction oxidation during drilling of drilling tools: select drilling tool materials with low friction coefficient to reduce friction heat during drilling, install cooling system on the drilling tools and control drilling speed; Regarding the air leakage oxidation during bedding boreholes: improve the sealing of bedding boreholes, inject nitrogen into bedding boreholes to drive out oxygen, and adjust the sealing depth.

2. The method for identifying and controlling the cause of abnormal CO outburst according to claim 1, characterized in that: In step S2, the normalization processing method includes Z-score normalization and minimum-maximum normalization.

3. The method for identifying and controlling the cause of abnormal CO outburst according to claim 1 or 2, characterized in that: When the number of borehole gas extraction data under a certain label is small, the number of borehole gas extraction data under the label is made close to the number of gas extraction data under other labels by copying or resampling with oversampling technology.

4. The method for identifying and controlling the cause of abnormal CO outburst according to claim 1, characterized in that: In step S3, the test set is used to evaluate the performance of the final classification model, thereby verifying the performance of the classification model in practical applications.

5. The method for identifying and controlling the cause of abnormal CO outburst according to claim 1 or 4, characterized in that: In step S3, the Adam gradient descent optimization algorithm is used to optimize the internal parameters of the classification model, and the genetic algorithm is used to adjust and optimize the hyperparameters in the classification model to finally obtain the optimal value of the number of hidden layer units in the classification model.

Images