Coal mine accident situation prediction method, device, equipment, medium and product

Abstract

The invention relates to the technical field of coal mine accident prevention, in particular to a coal mine accident scenario prediction method, device, equipment, medium and product, and the method comprises the steps: collecting coal mine accident cases and classifying coal mine accident scenarios, building a coal mine accident scenario library, determining the accident type and accident feature information in the coal mine accident scenario library, and obtaining a coal mine accident scenario prediction result. After the accidents are classified, the characteristic parameter indexes of different accident scenes are analyzed, and the accident scenes possibly occurring in the coal mine are predicted through the change of the characteristic parameter indexes based on the multi-model fusion neural network, so that the efficiency and the accuracy of the accident scene prediction model are effectively improved; therefore, coal mine accidents are prevented, and the accident handling capacity is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of coal mine accident prevention, and particularly to a method, device, equipment, medium and product for predicting coal mine accident scenarios. Background Art

[0002] During the coal mining process, accidents such as gas outburst, gas explosion, water inrush, and roof disasters are likely to occur. Coal mine accident disasters will not only cause property losses and production suspension, but also seriously endanger the lives of coal miners. Making judgments on the upcoming accidents through various information and predicting in advance before the coal mine accidents occur is of great significance for avoiding accidents or efficiently carrying out emergency rescue work. The production conditions in coal mines are complex, and the factors causing accidents are diverse. Therefore, there is an urgent need for a method that can efficiently and accurately predict coal mine accident scenarios to provide scientific means and effective ways for accident prevention in advance and timely response. Summary of the Invention

[0003] The present invention provides a method, device, equipment, medium and product for predicting coal mine accident scenarios, which solves the technical problem of difficult to accurately predict the occurrence scenarios of coal mine accidents.

[0004] In the first aspect, the present invention provides a method for predicting coal mine accident scenarios, the method comprising: Screening coal mine accident cases to construct a coal mine accident scenario library; Classifying and analyzing the coal mine accident scenario library to obtain accident characteristic information of various accident scenarios, wherein the accident characteristic information includes accident causes, accident processes, and accident consequences; Performing correlation analysis based on the accident characteristic information to obtain characteristic parameter indicators leading to the occurrence of accidents; Training a neural network with multi-model fusion using the characteristic parameter indicators and the labeled coal mine accident scenarios as samples to obtain a coal mine accident scenario prediction model; Inputting the target characteristic parameter indicators into the coal mine accident scenario prediction model to obtain the predicted accident scenarios.

[0005] In the second aspect, the present invention provides a device for predicting coal mine accident scenarios, the device comprising: A screening module for screening coal mine accident cases to construct a coal mine accident scenario library; A characteristic module for classifying and analyzing the coal mine accident scenario library to obtain accident characteristic information of various accident scenarios, wherein the accident characteristic information includes accident causes, accident processes, and accident consequences; A correlation module for performing correlation analysis based on the accident characteristic information to obtain characteristic parameter indicators leading to the occurrence of accidents; A modeling module, which is used to train a neural network with multi-model fusion by using the feature parameter indicators and the labeled coal mine accident scenarios as samples, so as to obtain a coal mine accident scenario prediction model; A prediction module, which is used to input the target feature parameter indicators into the coal mine accident scenario prediction model to obtain the predicted accident scenario.

[0006] In a third aspect, the present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps of the coal mine accident scenario prediction method described in the first aspect.

[0007] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the coal mine accident scenario prediction method described in the first aspect are implemented.

[0008] In a fifth aspect, the present invention provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the coal mine accident scenario prediction method described in the first aspect are implemented.

[0009] The present invention provides a coal mine accident scenario prediction method, device, equipment, medium and product. The method includes: screening coal mine accident cases to construct a coal mine accident scenario library; classifying and analyzing the coal mine accident scenario library to obtain accident feature information of various accident scenarios, where the accident feature information includes accident causes, accident processes, and accident consequences; performing correlation analysis based on the accident feature information to obtain feature parameter indicators that lead to accidents; training a neural network with multi-model fusion by using the feature parameter indicators and the labeled coal mine accident scenarios as samples to obtain a coal mine accident scenario prediction model; inputting the target feature parameter indicators into the coal mine accident scenario prediction model to obtain the predicted accident scenario, so as to be able to predict the occurrence scenario of coal mine accidents. Description of the Drawings

[0010] The present invention will be described in more detail below based on embodiments and with reference to the drawings: Figure 1 It is a schematic flowchart of a coal mine accident scenario prediction method provided by an embodiment of the present invention; Figure 2 It is a schematic structural diagram of a coal mine accident scenario prediction device provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of a coal mine accident scenario prediction method provided by an application example of the present invention; Figure 4 It is a schematic diagram of a method for constructing a neural network with multi-model fusion provided by an application example of the present invention.

[0011] In the accompanying drawings, like reference numerals are used for like components, and the drawings are not drawn to actual scale. Detailed Description of the Invention

[0012] In order to enable those skilled in the art of the present technology to better understand the solution of the present invention, and to fully understand how the present invention uses technical means to solve technical problems and the implementation process of achieving corresponding technical effects and to implement accordingly, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to 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. The embodiments of the present invention and each feature in the embodiments can be combined with each other without conflict, and the formed technical solutions are all within the protection scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

[0013] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned accompanying drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0014] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0015] During the coal mining process, accidents such as gas outburst, gas explosion, water inrush, roof disaster, etc. are likely to occur, endangering the lives of coal miners. With the development of science and technology and the standardized mining in strict accordance with safety regulations, the incidence rate of coal mine accidents has been decreasing year by year, but various coal mine accident disasters still occur continuously. There is a technical problem in this field that it is difficult to predict the occurrence scenario of coal mine accidents.

[0016] To solve the above technical problem of difficult prediction of the occurrence scenarios of coal mine accidents, the present invention proposes a coal mine accident scenario prediction scheme. The following specifically describes the implementation details of the scheme of the present invention. The following content is only the implementation details provided for convenient understanding and is not necessary for implementing the scheme.

[0017] Embodiment 1 Figure 1 It is a schematic flowchart of a coal mine accident scenario prediction method provided by an embodiment of the present application. As Figure 1 shown, the method includes: S1: Screen coal mine accident cases and construct a coal mine accident scenario library; The technical problem to be solved in this embodiment is how to predict the occurrence scenarios of coal mine accidents. In the technical solution of this embodiment, first, coal mine accident cases are screened to construct a coal mine accident scenario library. A word segmentation tool such as the jieba word segmentation tool is used to preliminarily screen the collected accident cases, and the repeatedly occurring coal mine accidents are removed therefrom, thereby constructing a coal mine accident scenario library.

[0018] S2: Conduct classification analysis on the coal mine accident scenario library to obtain accident characteristic information of various accident scenarios, where the accident characteristic information includes accident causes, accident processes, and accident consequences.

[0019] Specifically, conduct classification analysis on the constructed coal mine accident scenario library, classify the accident scenarios in the coal mine accident scenario library according to different accident types, such as gas accident scenarios, dust accident scenarios, etc., and then obtain accident characteristic information including accident causes, accident processes, and accident consequences for various accident scenarios.

[0020] S3: Conduct correlation analysis based on the accident characteristic information to obtain characteristic parameter indicators leading to the occurrence of accidents; Specifically, conduct correlation analysis based on this accident characteristic information to explore characteristic parameter indicators leading to the occurrence of accidents, such as specific indicators such as the gas concentration within a certain range and the oxygen concentration reaching a corresponding value.

[0021] S4: Train a neural network with multi-model fusion using the characteristic parameter indicators and the labeled coal mine accident scenarios as samples to obtain a coal mine accident scenario prediction model.

[0022] Specifically, use these characteristic parameter indicators and the labeled coal mine accident scenarios as samples to train a neural network with multi-model fusion, and obtain a coal mine accident scenario prediction model through a model composed of a neural network algorithm improved by fusing a BP neural network and a long short-term memory network LSTM.

[0023] S5: Input the target characteristic parameter index into the coal mine accident scenario prediction model to obtain the predicted accident scenario.

[0024] The technical solution of this embodiment, through the above series of steps, first widely collects and screens cases to construct a coal mine accident scenario library, then classifies and analyzes to obtain accident characteristic information, correlates characteristic parameter indexes, and uses a suitable neural network to construct a coal mine accident scenario prediction model, which can predict the coal mine accident occurrence scenario more comprehensively and systematically. In actual coal mine production, in the face of numerous complex factors that may cause accidents, the prediction method of this embodiment can estimate different types of accident scenarios such as gas explosion and roof accident in advance, know the possible risks in advance, so that coal mine enterprises can take corresponding measures in advance, such as adjusting the ventilation system in advance to prevent gas accumulation, strengthening roof support to prevent roof accidents, etc., thereby effectively reducing the probability of coal mine accidents and ensuring the safety of coal mine workers and the normal progress of coal mine production.

[0025] Embodiment 2 Based on the above Embodiment 1, in S5, inputting the target characteristic parameter index into the coal mine accident scenario prediction model to obtain the predicted accident scenario includes: S51: Real-time monitor the characteristic parameter index, where the characteristic parameter index includes gas concentration, gas pressure, and / or temperature.

[0026] The technical problem to be solved in this embodiment is how to perform accident scenario prediction based on the coal mine accident scenario prediction model. In the technical solution of this embodiment, the characteristic parameter index is monitored in real time, and the characteristic parameter index of this embodiment covers key indexes such as gas concentration, gas pressure, temperature, oxygen content, and CO concentration. For example, in the process of coal mine production, corresponding monitoring devices are installed at each key position underground. For example, a gas concentration sensor will detect the gas concentration value in different areas underground in real time, and a temperature sensor will always feedback the temperature situation of the location.

[0027] S52: Input the characteristic parameter index monitored in real time as the target characteristic parameter index into the coal mine accident scenario prediction model to predict the possible accident scenario.

[0028] Specifically, input these real-time monitored characteristic parameter indicators into the already constructed coal mine accident scenario prediction model. For example, input the real-time obtained gas concentration change value, temperature fluctuation value, etc. as the target characteristic parameter indicators into the coal mine accident scenario prediction model built by the improved neural network algorithm based on the fusion of BP neural network and long short-term memory network LSTM. The coal mine accident scenario prediction model will predict the possible accident scenarios according to the input data, based on the rules set inside and the logic formed by previous learning and training. For example, it will judge whether accident scenarios such as gas explosion and roof collapse will occur.

[0029] The technical solution of this embodiment can timely capture the potential accident risk changes in the coal mine production environment by monitoring the characteristic parameter indicators related to the coal mine in real time and accurately and inputting them into the corresponding coal mine accident scenario prediction model. Taking the gas explosion accident as an example, once the gas concentration shows an abnormal increase and the temperature also shows abnormal fluctuations and other parameter changes that meet the conditions for the occurrence of a gas explosion accident, the model can quickly predict the possible gas explosion accident scenario. The coal mine enterprise can immediately take emergency measures such as evacuating personnel urgently and cutting off the power supply, greatly improving the timeliness of dealing with accidents, avoiding the expansion of accidents, reducing casualties and property losses, and ensuring the safe and orderly development of coal mine production.

[0030] Embodiment 3 Based on any of the above embodiments, the coal mine accident scenario prediction method further includes: S6: Improve the coal mine accident cases in the coal mine accident scenario library based on the predicted accident scenarios output by the coal mine accident scenario prediction model.

[0031] The technical problem to be solved in this embodiment is how to improve the coal mine accident cases in the coal mine accident scenario library. In the technical solution of this embodiment, the coal mine accident cases in the coal mine accident scenario library are improved based on the coal mine accident scenario prediction model that has been constructed. For example, in the process of using the coal mine accident scenario prediction model to predict the accident scenario, for those previously collected coal mine accident cases that may have missing data or inaccurate parts, the various data and logical relationships used in the model analysis are used to supplement and correct them. For example, in some accident cases, the ventilation conditions at the time of the accident are not recorded in detail. Through the analysis of the correlation between ventilation conditions and accidents in the model, based on the gas concentration, other related gas concentrations, and the location of the accident, etc., it is possible to reasonably infer and improve the specific ventilation status at the time, such as the size of the ventilation volume, whether the ventilation is smooth, and other information, and then add these improved information to the corresponding coal mine accident cases, so as to achieve the improvement of the coal mine accident cases in the coal mine accident scenario library. For example, some coal mine accident scenarios have not yet occurred, but the model of this embodiment predicts that an accident may also occur in this scenario, and outputs the predicted accident scenario. In this regard, the predicted accident scenario is added to the accident scenario library, and the coal mine accident cases in the coal mine accident scenario library are improved.

[0032] The technical solution of this embodiment, by utilizing the analysis and correlation capabilities of the coal mine accident scenario prediction model, checks for omissions and improves and updates the original accident cases in the coal mine accident scenario library. In the long-term coal mine production and accident analysis process, the coal mine accident scenario library continues to accumulate cases, but some cases collected in the early stage may have data defects due to the limitations of the conditions at the time. Through the improvement method of this embodiment, the case information in the coal mine accident scenario library is more accurate and comprehensive, and when these cases are subsequently used for analysis, model building and other operations, more accurate results can be obtained, such as more accurately analyzing the causes of different accidents, more reasonably dividing accident types, etc., further improving the accuracy and reliability of the entire coal mine accident scenario prediction system, and helping coal mine enterprises to better prevent accidents.

[0033] Embodiment 4 On the basis of any of the above embodiments, S1, screening coal mine accident cases and building a coal mine accident scenario library includes: S11: Collect coal mine accident cases from accident reports.

[0034] The technical problem to be solved in this embodiment is how to screen coal mine accident cases and build a coal mine accident scenario library. In the technical solution of this embodiment, the first step is to collect coal mine accident cases from accident reports. These reports contain information on coal mine accidents in many different regions and of different types, which are an important source of materials for building a scenario library.

[0035] S12: Segment and screen coal mine accident cases, and eliminate the repeatedly occurring coal mine accidents in the coal mine accident cases.

[0036] Specifically, for segmenting and screening coal mine accident cases, in this embodiment, a specific word segmentation tool is adopted, such as the jieba word segmentation tool. Use it to process the collected coal mine accident case texts, identify and eliminate the repeatedly occurring coal mine accidents therein, and avoid the impact of duplicate data on subsequent analysis and the construction of the coal mine accident scenario library.

[0037] S13: Construct a coal mine accident scenario library based on the screened coal mine accident cases.

[0038] Specifically, construct a coal mine accident scenario library based on the screened coal mine accident cases. Organize and classify the remaining valid and valuable coal mine accident cases after screening according to preset rules and logic, so as to form a coal mine accident scenario library that is convenient for subsequent query, analysis, and for constructing a prediction model. For example, classify and store them according to the time sequence of accident occurrence, accident type, etc.

[0039] The technical solution of this embodiment collects cases from professional accident reports, then uses a suitable word segmentation tool for precise screening, and finally constructs the scenario library in an orderly manner, so that the constructed coal mine accident scenario library has high quality. With such a screened and well-organized coal mine accident scenario library, when analyzing different types of accidents such as gas accidents and roof accidents, relevant cases can be extracted more quickly and accurately for reference, which helps to more efficiently discover the characteristics and laws of different accidents, provides strong support for further constructing an accurate coal mine accident scenario prediction model, and ensures coal mine safety production.

[0040] Embodiment Five Based on any of the above embodiments, S3, perform correlation analysis based on accident characteristic information to obtain characteristic parameter indicators leading to the accident, which may include: S31: Construct a hierarchical model based on accident characteristic information, and the hierarchical model includes an element layer and an attribute layer.

[0041] The technical problem to be solved in this embodiment is how to perform correlation analysis based on the accident characteristic information to obtain the characteristic parameter indicators that cause the accident. In the technical solution of this embodiment, first, a hierarchical model is constructed based on the accident characteristic information. According to various objects such as personnel, equipment, and environment involved in the accident scenario and the attribute status of the objects, it is divided into four dimensions: accident characteristics, causes of occurrence, affected objects, and emergency response. Each dimension contains two layers: the element layer and the attribute layer. The element layer contains factors such as objects related to the accident scenario, and the attribute layer refers to the quantitative indicators of each factor. This hierarchical model contains the element layer and the attribute layer. The element layer covers many factors related to the accident scenario, such as ventilation conditions, gas concentration, personnel status, etc. The attribute layer is a description of the quantitative indicators corresponding to these factors, such as the specific value of the ventilation volume, the specific percentage of the gas concentration, etc.

[0042] S32: Perform correlation analysis on the element layer and the attribute layer to obtain the characteristic parameter indicators that cause the accident.

[0043] Specifically, correlation analysis can be performed on the element layer and the attribute layer to explore the correlation relationships between these different factors and their corresponding quantitative indicators, so as to obtain the characteristic parameter indicators that cause the accident. For example, it is analyzed that when conditions such as the gas concentration is in a certain range, the oxygen concentration reaches a certain value, and the ventilation volume is within a specific range are simultaneously satisfied, it is easy to trigger a gas explosion accident.

[0044] S33: If there is a lack of characteristic parameter indicators, perform numerical simulation on the missing characteristic parameter indicators to supplement the missing characteristic parameter indicators.

[0045] Specifically, if there is a lack of characteristic parameter indicators, for example, in a gas explosion accident, the specific value of the gas emission time is difficult to directly observe, then numerical simulation is performed on the missing characteristic parameter indicators. For example, numerical simulation technologies such as FLUENT, COMSOL, and FLAC3d are used to simulate the accident development process, construct a dynamic scenario of the accident occurrence, and then supplement and improve the missing parameter indicators, so that the obtained characteristic parameter indicators that cause the accident are more complete and accurate.

[0046] The technical solution of this embodiment can accurately and comprehensively obtain the characteristic parameter indicators that lead to accidents by constructing a hierarchical model, conducting correlation analysis, and numerically simulating and supplementing missing indicators. In actual coal mine production, clarifying the parameter indicators plays a key role in preventing accidents in advance. Taking a gas explosion accident as an example, after knowing the accurate ranges of parameter indicators such as gas concentration and oxygen concentration, coal mine enterprises can specifically set monitoring and early warning values. Once the monitoring data approaches these dangerous indicator ranges, measures such as ventilation adjustment and gas drainage can be taken in a timely manner to avoid accidents. At the same time, it also provides a reliable data basis for constructing an accurate accident scenario prediction model later, improving the scientificity and effectiveness of the entire coal mine accident prediction and prevention system.

[0047] Embodiment Six Based on any of the above embodiments, the multi-model fusion neural network may include: an LSTM network, and a BP neural network, a random forest, and / or a decision tree, and a fusion network; wherein, S4, using the characteristic parameter indicators and the labeled coal mine accident scenarios as samples to train the multi-model fusion neural network, obtaining a coal mine accident scenario prediction model, including: S41: Input the characteristic parameter indicators (such as parameters like gas concentration, gas emission volume, CO concentration, oxygen content, ventilation volume, etc. in a gas explosion accident) into the LSTM network for processing to obtain the time series characteristics of the coal mine gas explosion parameter indicators.

[0048] The technical problem to be solved in this embodiment is how to input the characteristic parameter indicators into the multi-model fusion neural network to train and obtain a coal mine accident scenario prediction model. In the technical solution of this embodiment, first, the characteristic parameter indicators are processed by LSTM. LSTM, that is, the long short-term memory network, can well process time series data. For example, for data such as gas concentration and temperature that change over time, through LSTM processing, the time series characteristics of the characteristic parameter indicators can be obtained. For example, the rising and falling trends and other change situations of the gas concentration within a period of time can be captured.

[0049] S42: Input the time series characteristics into the attention layer to calculate the attention weights for weighting different time series characteristics.

[0050] S43: Input the weighted time series characteristics of the characteristic parameter indicators into the BP neural network, the random forest, and / or the decision tree respectively to obtain the coal mine accident scenarios predicted under at least one single prediction network.

[0051] Specifically, input the time series features of the characteristic parameter indicators into different models such as BP neural network, random forest, and / or decision tree, allowing different models to analyze and learn these data features from their respective perspectives, mining the hidden laws behind the data, and respectively predicting the coal mine accident scenarios predicted under a single prediction network.

[0052] S44: Input the coal mine accident scenarios predicted under the at least one single prediction network and the corresponding attention weights into a fusion network for feature fusion to obtain the coal mine accident scenario predicted under the composite network.

[0053] Specifically, input the coal mine accident scenarios output by the BP neural network, random forest, and / or decision tree into the fusion network. For example, a BP network or a logistic regression model can be selected as the fusion network to predict and integrate the outputs of the above multiple models, synthesize the advantages of each model, and finally obtain a coal mine accident scenario prediction model, making it have higher accuracy and stronger prediction ability.

[0054] S45: Adopt a loss function constructed based on the labeled coal mine accident scenarios and the predicted coal mine accident scenarios, and combine the regularization term of the attention weights to jointly train the BP neural network, random forest, decision tree, and attention layer to obtain a coal mine accident scenario prediction model.

[0055] Among them, the loss function is as follows:

[0056] Among them, L is the loss value, n is the number of samples, , and are weight coefficients with a sum value of 1, , are weight coefficients with a sum value of 1, , , , are successively the characteristic values corresponding to the single prediction network output by the BP neural network, random forest, and / or decision tree and the i th predicted coal mine accident scenario output under the composite network, and the is the characteristic value corresponding to the i th labeled coal mine accident scenario.

[0057] The technical solution of this embodiment constructs a coal mine accident scenario prediction model by using LSTM to process time series features and combining the method of multi-model fusion, giving full play to the advantages of different models. In the face of the complex and changeable coal mine production environment and numerous factors affecting the occurrence of accidents, the model constructed by the technical solution of this embodiment can more accurately predict accident scenarios based on the changes in characteristic parameter indicators. For example, when predicting accidents such as gas explosions and roof collapses, compared with a single model, the technical solution of this embodiment can more sensitively capture the risk signals contained in the subtle changes of various parameter indicators, more accurately judge the possibility of accidents, enabling coal mine enterprises to make full preparations in advance, reasonably allocate resources for accident prevention and response, and effectively ensuring coal mine production safety and the lives of miners.

[0058] Embodiment 7 Figure 2 is a structural schematic diagram of a coal mine accident scenario prediction device provided by an embodiment of the present application. The device includes: A screening module, configured to screen coal mine accident cases to construct a coal mine accident scenario library; A feature module, configured to perform classification analysis on the coal mine accident scenario library to obtain accident feature information for various accident scenarios, where the accident feature information includes accident causes, accident processes, and accident consequences; A correlation module, configured to perform correlation analysis based on the accident feature information to obtain characteristic parameter indicators that cause accidents; A modeling module, configured to train a neural network with multi-model fusion using the characteristic parameter indicators and labeled coal mine accident scenarios as samples to obtain a coal mine accident scenario prediction model; A prediction module, configured to input target characteristic parameter indicators into the coal mine accident scenario prediction model to obtain a predicted accident scenario.

[0059] Optionally, the prediction module may be specifically configured to monitor characteristic parameter indicators in real time, where the characteristic parameter indicators include gas concentration, gas pressure, and / or temperature; input the real-time monitored characteristic parameter indicators as the target characteristic parameter indicators into the coal mine accident scenario prediction model to predict possible accident scenarios.

[0060] Optionally, the above coal mine accident scenario prediction device may further include: A perfection module, configured to perfect the coal mine accident cases in the coal mine accident scenario library based on the predicted accident scenarios output by the coal mine accident scenario prediction model.

[0061] Optionally, a screening module, specifically used to collect coal mine accident cases from accident reports; perform word segmentation and screening on the coal mine accident cases, and eliminate the repeatedly occurring coal mine accidents in the coal mine accident cases; construct a coal mine accident scenario library based on the screened coal mine accident cases.

[0062] Optionally, a correlation module, specifically used to construct a hierarchical model based on the accident feature information, where the hierarchical model includes an element layer and an attribute layer; perform correlation analysis on the element layer and the attribute layer to obtain characteristic parameter indicators that cause accidents; if there are missing characteristic parameter indicators, perform numerical simulation on the missing characteristic parameter indicators to supplement the missing characteristic parameter indicators.

[0063] Optionally, the multi-model fusion neural network includes: an LSTM network, and a BP neural network, a random forest, and / or a decision tree, and a fusion network; Correspondingly, the modeling module is specifically used for: Input the characteristic parameter indicators into the LSTM network for processing to obtain the time series characteristics of the characteristic parameter indicators; Input the time series characteristics into the attention layer to calculate the attention weights for weighting different time series characteristics; Input the weighted time series characteristics of the characteristic parameter indicators into the BP neural network, the random forest, and / or the decision tree respectively to obtain the coal mine accident scenarios predicted under at least one single prediction network; Input the coal mine accident scenarios predicted under the at least one single prediction network and the corresponding attention weights into the fusion network for feature fusion to obtain the predicted coal mine accident scenarios under the composite network; Adopt a loss function constructed based on the labeled coal mine accident scenarios and the predicted coal mine accident scenarios, and combine the regularization term of the attention weights to jointly train the BP neural network, the random forest, the decision tree, and the attention layer to obtain a coal mine accident scenario prediction model; Among them, the loss function is as follows:

[0064] Among them, L is the loss value, n is the number of samples, , and are weight coefficients with a sum value of 1, , are weight coefficients with a sum value of 1, , , , The single prediction network output by the BP neural network, random forest and decision tree and the output by the composite network are respectively i The characteristic value corresponding to the predicted coal mine accident scenario is Marked i The characteristic value corresponding to each coal mine accident scenario.

[0065] The functions of each module in the embodiment of the present device correspond to the steps in the aforementioned method embodiment, and the relevant technical problems to be solved and the technical effects achieved can also be found in the description in the aforementioned method embodiment.

[0066] Embodiment 8 In the technical solution of this embodiment, a computer device is provided, including a memory, a processor and a computer program stored in the memory, and the processor executes the computer program to implement the steps of any coal mine accident scenario prediction method of the above embodiments.

[0067] In the technical solution of this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of any coal mine accident scenario prediction method of the above-mentioned embodiments are implemented.

[0068] In the technical solution of this embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps of any of the coal mine accident scenario prediction methods of the above embodiments.

[0069] The processor may include, but is not limited to, for example, one or more processors or microprocessors, etc. Each processor may be implemented by an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components, and is used to execute the methods in the above embodiments. The computer-readable storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof. The computer-readable storage medium may include, but is not limited to, for example, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (such as hard disks, floppy disks, solid state drives, removable disks, CD-ROMs, DVD-ROMs, Blu-ray discs, etc.).

[0070] The computer-readable storage medium may also store at least one computer-executable program / instructions, and the computer-executable program / instructions are, for example, computer-readable instructions. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may, for example, include random access memory (RAM) and / or cache memory, etc. The computer-readable storage medium may, for example, include read-only memory (ROM), hard disks, flash memory, etc. For example, the non-transitory computer-readable storage medium may be connected to a computing device such as a computer. Then, when the computing device runs the computer-readable instructions stored on the computer-readable storage medium, the various methods described above may be performed.

[0071] In addition, the computer device may also include (but is not limited to) a data bus, an input / output (I / O) bus, a display, and input / output devices (such as a keyboard, a mouse, a speaker, etc.). The processor may communicate with external devices via the I / O bus through a wired or wireless network. In one embodiment, the at least one computer-executable instruction may also be compiled into or form a software product / computer program product, and when one or more computer-executable instructions are run by the processor, the steps of the various functions and / or methods in the embodiments described in the present technology are performed.

[0072] Embodiment Nine Based on the above embodiments, this embodiment provides an application example.

[0073] This application example provides a coal mine accident scenario prediction scheme. By collecting coal mine accident cases and classifying coal mine accident scenarios, a historical scenario database / coal mine accident scenario database is established. The accident types, accident occurrence times, accident occurrence locations, and accident impact scopes in the coal mine accident scenario database are clarified. After classifying the accidents, the characteristic parameter indicators of different accident scenarios are analyzed, and based on a neural network with multi-model fusion, the possible accident scenarios in the coal mine are predicted through the changes in coal mine index parameters. Among them, while predicting the coal mine accident scenarios, the scenario data in the coal mine accident scenario database can also be improved, improving the efficiency and accuracy of the accident scenario prediction model, reducing the probability of accidents occurring, and enhancing the accident response and emergency rescue capabilities.

[0074] The object of the present invention is to provide a coal mine accident scenario prediction method. By classifying coal mine accident scenarios, a historical scenario database / coal mine accident scenario database is established and the parameter indicators of different accident scenarios are analyzed. The missing index parameters are supplemented through methods such as numerical simulation, and the possible accident scenarios in the coal mine are risk-predicted through the changes in index parameters by a neural network with multi-model fusion.

[0075] To achieve the above object, the present invention provides the following application example: A word segmentation tool is used to preliminarily screen the accident cases, eliminate the repeatedly occurring coal mine accidents, and construct a coal mine accident scenario database; The accident scenario parameters such as the accident type, accident occurrence time, accident occurrence location, and accident impact scope in the coal mine accident scenario database are clarified. At the same time, coal mine data such as the accident coal mine address, equipment operation status, and historical accident records are collected, and based on this, the causes of the accident (including direct and indirect causes), the accident development process, and the consequences of the accident are analyzed; The accident scenarios in the coal mine accident scenario database are classified. According to different accident types, coal mine accidents are divided into: gas accident scenarios, dust accident scenarios, roof accident scenarios, mine water disaster accident scenarios, and mechanical and electrical accident scenarios; The information of each type of accident scenario after classification is analyzed respectively to generate an accident scenario characteristic library corresponding to the information of each type of accident scenario. The accident scenario characteristic library includes accident characteristic information such as the direct and indirect causes of the corresponding accident, the accident development process, and the accident impact scope; Construct a hierarchical model for the accident characteristic information, and divide it into four dimensions: accident characteristics, causes of occurrence, affected objects, and emergency response according to various objects such as personnel, equipment, and environment involved in the accident scenario and the attribute status of the objects. Each dimension contains two layers: the element layer and the attribute layer. The element layer includes factors such as objects related to the accident scenario, and the attribute layer refers to the quantitative indicators of each factor; According to the attribute status of the accident characteristic information, conduct correlation analysis on the element layer and the attribute layer of the historical accident scenario to generate specific characteristic parameter indicators that lead to the corresponding accident scenario; conduct comparative observations on the characteristic parameter indicators that lead to the accident scenario, and use numerical simulation methods such as FLUENT, COMSOL, and FLAC for accident scenarios with some difficult-to-observe or missing parameter indicators 3d to simulate the accident development process, construct a dynamic scenario of the accident occurrence, and supplement and improve the missing parameter indicators; Conduct feature combination on the information of the characteristic parameter indicators of the same type of accident scenario to form comprehensive parameter indicator elements for the accident scenario occurrence. Based on the interaction and coupling relationship of carrier substances, energy, and information, combined with the requirements of standards and specifications and the disaster-causing, disaster-bearing, disaster-bearing and disaster-causing, and response mechanism models of the carrier, adopt quantitative and semi-quantitative methods, and through threshold analysis and acceptability analysis, construct a calculation and evaluation method for entity accident parameter indicators and an evolution rule for entity interaction to form a structured accident scenario parameter indicator library for such coal mine accidents; Furthermore, according to the accident scenario parameter indicator library of coal mine accidents, use a multi-model fusion neural network to generate a coal mine accident scenario prediction model. The coal mine accident scenario prediction model is a neural network algorithm improved by the fusion of the BP neural network and the long short-term memory network LSTM. The BP neural network includes parts such as the input layer, output layer, hidden layer, transfer function selection, and network structure diagram design. Input measurement index parameters such as gas concentration, pressure, temperature, and oxygen content into the input layer of the BP neural network, extract non-linear characteristic relationships in the BP neural network, and generate a preliminary risk assessment result. Subsequently, process the dynamic change trend through the time series processing module (LSTM) to obtain the time series change characteristics of the characteristic parameter indicators. Input the time series characteristics into the attention layer to calculate the attention weights for weighting different time series characteristics Conduct real-time dynamic monitoring on the characteristic parameter indicators such as geological conditions, gas extraction, and ventilation system during the coal mine production process, and connect the monitored characteristic parameter indicators into the coal mine accident scenario prediction model. When the indicators show abnormal fluctuations, predict the possible accident scenario through this accident scenario prediction model, evaluate the occurrence risk of coal mine accidents, and reduce the occurrence probability of coal mine accidents.

[0076] The technical effects corresponding to this embodiment are as follows: 1. Establish an accident scenario prediction model based on historical coal mine accident scenario cases, predict the coal mine accident scenario in advance through the changes of characteristic parameter indicators, reduce the probability of accidents, and improve the accident response and emergency rescue capabilities.

[0077] 2. Use methods such as numerical simulation to improve the partially missing accident scenario data, and be able to improve the scenario data in the database while predicting the coal mine accident scenario.

[0078] 3. Adopt a neural network with multi-model fusion, which improves the efficiency and accuracy of the accident scenario prediction model.

[0079] The following further illustrates the present invention with examples. As Figure 3 shown, the present invention provides a method for predicting coal mine accident scenarios. Specifically, taking the scenario prediction of coal mine gas explosion accidents as an example, it includes the following steps: (1) Use the jieba word segmentation tool to preliminarily screen coal mine accident cases, eliminate the repeatedly occurring coal mine accidents, and construct a coal mine accident scenario library.

[0080] (2) Define accident scenario parameters such as accident types, accident occurrence times, accident occurrence locations, and accident influence scopes in the coal mine accident scenario library. At the same time, collect coal mine data such as the address data of the accident coal mine, equipment operation status, and historical accident records, and analyze accident characteristic information such as the causes of accidents, the development process of accidents, and the consequences caused by accidents.

[0081] (3) In (1), for example, classify the accident scenarios in the coal mine accident scenario library, and uniformly screen and extract the accident scenarios involving gas explosion in the coal mine accident scenario library and classify them into the gas explosion accident scenarios.

[0082] (4) Analyze the classified gas explosion accident scenario information. Exclude 110 accident cases with unclear accident occurrence locations. Among the remaining 615 accident cases, gas explosions will occur in coal mining faces, tunneling faces, return air roads, transportation roads, goafs, shaft bottom yards, underground chambers, connecting roads, roadway intersections, sealed areas, and other locations. During the statistics process, since there are too many locations where gas explosions have occurred, the locations with extremely few gas explosion occurrences are included in others. After statistics, gas explosion accidents mainly occur in tunneling faces and coal mining faces. There are 174 accidents occurring in tunneling faces, accounting for 28.29% of the total number of accidents; there are 133 accidents occurring in coal mining faces, accounting for 21.63% of the total number of accidents.

[0083] There are many reasons for gas accumulation in coal mines, which can be roughly divided into four categories: poor ventilation, abnormal gas outburst, improper treatment of accumulated gas, and gas accumulation in goafs. Poor ventilation is the main source of gas accumulation. The irrational ventilation system includes improper system design, airflow short circuit, series ventilation, and circulating air from local ventilators. Poor air duct management includes air duct extrusion and air leakage. Abnormal outburst includes geological structure, blasting, roof collapse, old ponds, old cellars, etc.

[0084] After eliminating 48 accident cases with unknown ignition sources, the fire sources that caused gas explosion accidents in the remaining 677 cases included electrical sparks, blasting flames, friction sparks, smoking, spontaneous combustion of coal, and other factors (material combustion, gas cutting, etc.). Among them, electrical sparks and blasting flames are the main fire sources. Electrical sparks occurred 355 times, accounting for 48.97% of the total; blasting flames occurred 165 times, accounting for 22.76% of the total. Electrical sparks include the use of mining lamps, blasting busbar short circuit, electrical explosion failure, wire and cable short circuit, electric welding, live maintenance, etc.; blasting flames mainly include explosives, unqualified blasting mud length, blasting, short-interval blasting, etc. The above accident feature information is summarized and generated into the corresponding accident scenario feature library.

[0085] (5) A hierarchical model is constructed for the characteristic information of coal mine gas explosion accidents. According to the various objects involved in the accident scenario, such as personnel, equipment, environment, and the attribute states of the objects, it is divided into four dimensions: accident characteristics, causes, affected objects, and emergency response. Each dimension contains two layers: element layer and attribute layer. The element layer contains factors related to the accident scenario, namely ventilation, temperature, gas concentration, gas outflow, oxygen concentration, air stoppage, power outage, ventilator, wind tube, outflow, gas inspector, personnel, collapse, toxic gas, etc. The attribute layer refers to the quantitative index of each factor. The attribute of coal mine gas explosion accident scenario elements can be mainly divided into three types. ① Parameter type, this type of attribute is mainly manifested as the parameters of the indicator, such as oxygen content, coal seam inclination, etc. The parameter type uses specific parameters to represent a factor or indicator. ② Feature description type, this type of attribute summarizes the characteristics of the factor through text description. For example, the mine types are divided into: high gas mines, low gas mines, coal and gas outburst mines. ③ State type, this type of attribute refers to variables with only states but no clear interval boundaries of the numerical range. For example, the status of the ventilation system can be divided into: normal, partially damaged, severely damaged, and completely interrupted; the status of casualties can be divided into: minor injury, light injury, serious injury, and death.

[0086] (6)Based on the attribute status of the characteristic information of coal mine gas explosion accidents, conduct a correlation analysis on the element layer and attribute layer of the historical accident scenarios of coal mine gas explosions, and generate specific characteristic parameter indicators that lead to the occurrence of gas explosion accident scenarios, such as gas concentration between 5% and 16%, oxygen concentration not less than 12%, ignition source temperature higher than 650 °C, ignition source energy higher than 0.28 mJ, and related temperature, humidity, and ventilation volume of the coal seam. Conduct comparative observations on the parameter indicators of the accident scenarios. For indicators that are not easily observable in accidents such as gas emission time, explosion spread speed, explosion temperature, and harmful gas diffusion speed, as well as fuzzy parameter indicators such as large air volume and local damage, use numerical simulation methods such as FLUENT, COMSOL, and FLAC 3d to simulate the development process of gas explosion accidents, construct a dynamic scene of the accident occurrence, and supplement and improve the missing parameter indicators.

[0087] (7)According to the accident scenario parameter index library of coal mine accidents, use a neural network with multi-model fusion to generate a coal mine accident scenario prediction model. As Figure 4 shown, input the characteristic parameter indicators in the coal mine gas explosion accident scenario, such as gas concentration, gas pressure, gas content, oxygen concentration, CO content, and temperature, into the input layer of the BP neural network. Use LSTM to process time series data and capture the dynamic change trends of parameters such as gas concentration. Subsequently, perform preprocessing on the data, such as data cleaning, missing value filling, and data standardization. Then divide the data into a training set, a validation set, and a test set according to 7:2:1. Use the preprocessed time series features as the input of the BP network to further generate a risk assessment result. Use the outputs of multiple models such as BP neural network, random forest, and decision tree as features, and then use an ensemble model (such as a BP network or a logistic regression model) to perform a final prediction on the outputs of the above multiple models to improve the accuracy of the model. Introduce a weighted cross-entropy loss function or the loss function in Example 6 above to make the model more sensitive to minority class data. Finally, through an incremental training method, enable the model to be quickly updated when receiving new data without having to retrain the entire model.

[0088] (8)Conduct real-time dynamic monitoring on the characteristic parameter indicators such as geological conditions, gas drainage, and ventilation system during the coal mine production process, and connect the detected characteristic parameter indicators into the coal mine accident scenario prediction model. When characteristic parameter indicators such as gas concentration, oxygen content, related temperature of the coal seam, and humidity show abnormal fluctuations and reach or are about to reach the minimum conditions for the occurrence of coal mine gas explosion accidents, the accident scenario prediction model predicts the possible gas explosion accident scenario, indicating the existence of coal mine accident risks, thereby reducing the occurrence probability of coal mine gas explosion accidents and improving the response and emergency rescue capabilities for coal mine gas explosion accidents.

[0089] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the above-mentioned module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0090] It should be noted that in the present invention, the term "comprising", "including" or any other variant thereof is intended to cover a non-exclusive inclusion, such that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element limited by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0091] Although the disclosed embodiments of the present invention are as above, the above content is only an embodiment adopted for the convenience of understanding the present invention and is not intended to limit the present invention. Any person skilled in the art within the technical field to which the present invention pertains may make any modifications and changes in the form of implementation and details without departing from the spirit and scope disclosed by the present invention. However, the scope of patent protection of the present invention shall still be subject to the scope defined by the appended claims.

Claims

1. A method for predicting coal mine accident scenarios, characterized in that, The method includes: Screening coal mine accident cases to construct a coal mine accident scenario database; Conducting classification analysis on the coal mine accident scenario database to obtain accident characteristic information for various accident scenarios, where the accident characteristic information includes accident causes, accident processes, and accident consequences; Performing correlation analysis based on the accident characteristic information to obtain characteristic parameter indicators leading to accidents; Training a neural network with multi-model fusion using the characteristic parameter indicators and labeled coal mine accident scenarios as samples to obtain a coal mine accident scenario prediction model; Inputting the target characteristic parameter indicators into the coal mine accident scenario prediction model to obtain the predicted accident scenarios.

2. The coal mine accident scenario prediction method according to claim 1, characterized in that The step of inputting the target characteristic parameter indicators into the coal mine accident scenario prediction model to obtain the predicted accident scenarios includes: Real-time monitoring of characteristic parameter indicators, where the characteristic parameter indicators include gas concentration, gas pressure, and / or temperature; Taking the real-time monitored characteristic parameter indicators as the target characteristic parameter indicators and inputting them into the coal mine accident scenario prediction model to predict possible accident scenarios.

3. The coal mine accident scenario prediction method according to claim 1 or 2, characterized in that The method further includes: Improving the coal mine accident cases in the coal mine accident scenario database based on the predicted accident scenarios output by the coal mine accident scenario prediction model.

4. The coal mine accident scenario prediction method according to claim 1, wherein The step of screening coal mine accident cases to construct a coal mine accident scenario database includes: Collecting coal mine accident cases from accident reports; Performing word segmentation screening on the coal mine accident cases to remove duplicate coal mine accidents in the coal mine accident cases; Constructing the coal mine accident scenario database based on the screened coal mine accident cases.

5. The coal mine accident scenario prediction method according to claim 1, wherein Performing correlation analysis based on the accident characteristic information to obtain characteristic parameter indicators leading to accidents, including: Constructing a hierarchical model based on the accident characteristic information, where the hierarchical model includes an element layer and an attribute layer; Performing correlation analysis on the element layer and the attribute layer to obtain characteristic parameter indicators leading to accidents; If there are missing characteristic parameter indicators, numerical simulation is performed on the missing characteristic parameter indicators to supplement the missing characteristic parameter indicators.

6. The coal mine accident scenario prediction method according to claim 1, characterized in that, The neural network with multi-model fusion includes: an LSTM network, a BP neural network, a random forest, and / or a decision tree, and a fusion network; training a neural network with multi-model fusion using the characteristic parameter indicators and labeled coal mine accident scenarios as samples to obtain a coal mine accident scenario prediction model includes: Inputting the characteristic parameter indicators into the LSTM network for processing to obtain the time series characteristics of the characteristic parameter indicators; Inputting the time series characteristics into the attention layer to calculate the attention weights for weighting different time series characteristics; Inputting the weighted time series characteristics of the characteristic parameter indicators into the BP neural network, random forest, and / or decision tree respectively to obtain the predicted coal mine accident scenarios under at least one single prediction network; Inputting the predicted coal mine accident scenarios under the at least one single prediction network and the corresponding attention weights into the fusion network for feature fusion to obtain the predicted coal mine accident scenarios under the composite network; Adopt a loss function constructed based on the labeled coal mine accident scenarios and the predicted coal mine accident scenarios, and combine the regularization term of the attention weight to jointly train the BP neural network, random forest, decision tree, and attention layer to obtain a coal mine accident scenario prediction model; Among them, the loss function is as follows: Among them, L is the loss value, n is the number of samples, , and are the weight coefficients with a sum value of 1, , are the weight coefficients with a sum value of 1, , , , are successively the eigenvalues corresponding to the single prediction network output by the BP neural network, random forest, and decision tree, and the i th predicted coal mine accident scenario under the composite network, and the is the eigenvalue corresponding to the i th labeled coal mine accident scenario.

7. A device for predicting coal mine accident scenarios, characterized in that, The device includes: A screening module for screening coal mine accident cases to construct a coal mine accident scenario library; A feature module for classifying and analyzing the coal mine accident scenario library to obtain accident feature information for various accident scenarios, where the accident feature information includes accident causes, accident processes, and accident consequences; An association module for performing association analysis based on the accident feature information to obtain characteristic parameter indicators leading to the occurrence of accidents; A modeling module for training a neural network with multi-model fusion using the characteristic parameter indicators and the labeled coal mine accident scenarios as samples to obtain a coal mine accident scenario prediction model; A prediction module for inputting target characteristic parameter indicators into the coal mine accident scenario prediction model to obtain predicted accident scenarios.

8. A computer device, comprising a memory, a processor, and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the steps of the coal mine accident scenario prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the coal mine accident scenario prediction method according to any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the coal mine accident scenario prediction method according to any one of claims 1 to 6.

Images