A method and system for predicting the risk of severe mining pressure

By constructing a deep learning model based on convolutional neural networks and combining coal mine reserves and environmental factors data, the high mine pressure hazard in coal mines can be monitored in real time. This solves the problem that existing technologies fail to consider the differences between coal mine factors and environmental factors, and achieves accurate prediction of high mine pressure and improves coal mine safety.

CN118708959BActive Publication Date: 2026-11-17ZHUNGEERQIYUNFEI MINING IND CO LTD +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410634684.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-22
Publication Date
2026-11-17
Estimated Expiration
2044-05-22

AI Technical Summary

Technical Problem

Existing technologies fail to effectively consider the differences between coal mine internal factors and environmental factors when predicting the risk of strong coal mine pressure, resulting in inaccurate predictions.

Method used

By acquiring coal mine reserves and environmental factor data related to historical severe mine pressure accidents, a risk index is calibrated, a deep learning model based on convolutional neural networks is constructed, and predictions are made by combining real-time mine environmental data to generate hazard prediction coefficients and hazard levels.

Benefits of technology

It enables accurate prediction of the current high risk level of coal mine pressure, improves the safety and stability of coal mines, and allows for timely preventive measures.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118708959B_ABST
    Figure CN118708959B_ABST
Patent Text Reader

Abstract

This invention provides a method and system for predicting severe mine pressure hazards, relating to the field of mining engineering technology. This invention collects historical data on coal mine reserves and environmental factors where severe mine pressure accidents have occurred, and combines this with a risk index determined by the severity of the accident type. A deep learning model is then trained, enabling the model to obtain a predicted risk index by inputting the reserve and environmental data of the coal mine to be predicted, thus determining the level of potential severe mine pressure hazards. This invention considers real-time environmental data of the coal mine to be predicted and combines it with the predicted risk index to generate a hazard prediction model. This model can monitor the hazard level of the coal mine in real time and take corresponding preventative measures based on the prediction results, thereby improving the safety and stability of the coal mine.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of mining engineering technology, specifically to a method and system for predicting the risk of strong mine pressure. Background Technology

[0002] Severe mine pressure is a common hazardous situation in coal mining, referring to the pressure concentration and instability caused by the interaction between coal seams and rock strata under geomechanical conditions. Severe mine pressure accidents can lead to serious consequences such as mine collapse and gas explosions, threatening the lives of miners and the normal operation of the coal mine. Therefore, predicting and preventing severe mine pressure accidents in advance has become an important task in coal mine safety management.

[0003] In the prior art, CN117668650A discloses a method and apparatus for predicting strong mine pressure hazard. This method predicts the strong mine pressure hazard level of the working face section by acquiring multiple sets of historical strong mine pressure hazard data and using a strong mine pressure hazard level prediction model. The prediction is primarily based on historical data from the working face section. However, the strong mine pressure hazard of a coal mine is not only related to the historical working face conditions but also to factors specific to the mine itself. For example, there is a correlation between strong mine pressure accidents and the mine's own mining time and reserves. Long-term mining activities may lead to coal seam stripping and changes in the coal seam stress state, thereby increasing the instability of the mine structure and raising the probability of strong mine pressure accidents. The scale and duration of mining activities are generally related to the coal mine's reserves. Generally, coal mines with abundant reserves may engage in long-term, large-scale mining activities, which may increase the probability of strong mine pressure accidents. Furthermore, the gradual decrease in coal reserves may also lead to changes in mining methods, thus affecting the stability of the mine structure.

[0004] Furthermore, factors within a coal mine can also influence the severity of severe mine pressure. For example, temperature data within the mine affects the thermal expansion and contraction properties of rocks; increased temperature leads to rock expansion, impacting mine support and rock mass stability. Therefore, the shortcomings of existing technologies lie in their reliance on historical data to predict the severity of working faces, neglecting factors specific to the coal mine itself. Different coal mines may have different ore seam structures and properties. Relying solely on working face data makes it difficult to accurately predict the severe pressure hazard caused by the coal mine's own properties and structure, as well as the hazards caused by environmental factors. Therefore, it cannot effectively predict the current severity of severe mine pressure in a coal mine. The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure and may therefore include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for predicting the danger of strong mining pressure, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for predicting the risk of severe mining pressure includes the following steps: Step 1: Obtain the coal mine reserve data and environmental factor data of coal mines that have experienced severe mine pressure accidents in the historical data, obtain the accident type of the corresponding coal mine that has experienced severe mine pressure accidents, and mark the risk index according to the severity of the accident type. Map the coal mine reserve data, environmental factor data and risk index of each coal mine one by one. Step 2: Construct a deep learning model based on a convolutional neural network. Use coal mine reserve data and environmental factor data as the training set, and use the risk index as the label. Input these into the deep learning model and train it. Step 3: Obtain coal reserve data and environmental factor data of the coal mine to be predicted, as well as real-time environmental data of the mine. Input the coal reserve data and environmental factor data of the coal mine to be predicted into the trained deep learning model to obtain the prediction risk index. Step 4: Based on the obtained predicted risk index, combine it with the real-time environmental data of the mine, perform dimensionless processing to generate a hazard prediction model, and obtain the hazard prediction coefficient. Step 5: Compare the obtained hazard prediction coefficients with the hazard reference thresholds, perceive the degree of hazard in the coal mine based on the comparison results, and generate the current hazard level of the coal mine.

[0007] Furthermore, the coal mine reserve data that has experienced a severe mine collapse includes the coal mine reserve data at the time of exploration, the mining time data up to the time of the severe mine collapse, and the average monthly mining volume in the three months prior to the time of the accident. The environmental factor data includes the depth of the top layer of the coal mine from the ground surface, the thickness of the coal seam, and the dip angle of the coal seam at the time of exploration.

[0008] Furthermore, the accident types in coal mines that have experienced severe mine pressure accidents include coal seam roof collapse, coal pillar compression, coal wall collapse, coal seam water inrush, and gas explosion. The specific logic behind assigning a risk index based on the severity of the accident type is as follows: The risk index for gas explosion accidents is set at 0.9, the risk index for coal seam roof collapse accidents is set at 0.75, the risk index for coal pillar compression accidents is set at 0.6, the risk index for coal wall collapse accidents is set at 0.45, and the risk index for coal wall collapse accidents is set at 0.3.

[0009] Furthermore, the deep learning model is constructed using a deep neural network based on a multilayer perceptron. The deep neural network of the multilayer perceptron includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. The first hidden layer, the second hidden layer, and the third hidden layer each have at least two neurons and all use ReLU as the activation function.

[0010] Furthermore, the coal reserve data for the coal mine to be predicted includes the reserve data at the time of exploration, the time data up to the current mining period, and the average monthly mining volume for the previous three months. The environmental factor data includes the depth of the top layer of the coal mine from the ground, the thickness of the coal seam, and the dip angle of the coal seam at the time of exploration. The real-time environmental data of the coal mine to be predicted includes the average temperature data, average pressure data, average humidity data, and stress equilibrium data within the mine. The specific logic for generating the average temperature data, average pressure data, and average humidity data within the mine is as follows: Equal distance setting Each sampling point collects temperature, pressure, humidity data inside the mine, and stress data on the upper sidewall of the mine at each sampling point. against The average values ​​of temperature, pressure, and humidity inside the mine are calculated to obtain average temperature, average pressure, and average humidity data. The formula used is as follows:

[0011]

[0012]

[0013] in, , and These represent the average temperature, average pressure, and average humidity data of the coal mine to be predicted, respectively. , and They represent the first Temperature, pressure, and humidity data from each collection point. The number of the collection point. It is a positive integer, and .

[0014] Furthermore, the specific logic underlying the stress equilibrium data of the coal mine to be predicted is as follows: according to Stress data of the upper sidewall of the mine shaft, calculation The variance of the stress data is calculated and calibrated to the stress equilibrium data. The specific formula is as follows:

[0015]

[0016] in, Indicates the coal mine to be predicted Average stress data at the sampling points, Indicates the first Stress data from each collection point This represents the stress equilibrium data of the coal mine to be predicted.

[0017] Furthermore, the specific logic underlying the generation of the hazard prediction model to derive the hazard prediction coefficients is as follows:

[0018] in, Indicates the risk prediction coefficient. This represents the predicted risk index obtained through a deep learning model. This indicates the standard temperature inside a coal mine.

[0019] Furthermore, based on the comparison between the obtained hazard prediction coefficient and the hazard reference threshold, the specific logic for generating the current hazard level of the coal mine is as follows: when At that time, the level of danger generated is low risk; when At that time, the level of danger generated was medium risk; when At that time, the level of danger generated is high risk; in, This indicates the danger reference threshold.

[0020] The present invention also provides a strong mine pressure hazard prediction system, the prediction system being used to execute the above-described strong mine pressure hazard prediction method, comprising: The training set acquisition module is used to acquire coal mine reserve data and environmental factor data that have experienced severe mine pressure accidents in historical data, and to acquire the accident type of the corresponding coal mine that has experienced severe mine pressure accidents. It also calibrates the risk index according to the severity of the accident type, and maps the coal mine reserve data, environmental factor data and risk index of each coal mine one by one. Model building module: Used to build a deep learning model based on convolutional neural networks. Coal mine reserve data and environmental factor data are used as training sets, and risk index is used as a label. The data are input into the deep learning model to train it. The data generation module is used to acquire coal reserve data and environmental factor data of the coal mine to be predicted, as well as real-time environmental data of the coal mine to be predicted. The coal reserve data and environmental factor data of the coal mine to be predicted are input into the trained deep learning model to obtain the prediction risk index. The data processing module is used to combine the obtained predicted risk index with real-time mine environmental data, perform dimensionless processing to generate a hazard prediction model, and obtain hazard prediction coefficients. The hazard prediction module compares the acquired hazard prediction coefficients with hazard reference thresholds, perceives the hazard level of the coal mine based on the comparison results, and generates the current hazard level of the coal mine. Compared with the prior art, the beneficial effects of the present invention are: This invention collects historical data on coal mine reserves and environmental factors that have experienced severe mine collapse accidents. It then uses this data to establish a risk index based on the severity of the accident type, trains a deep learning model, and obtains a predicted risk index by inputting the reserve and environmental data of the coal mine to be predicted. This allows the model to determine the level of potential severe mine collapse hazards. Furthermore, this invention considers real-time environmental data of the coal mine to be predicted and combines it with the predicted risk index to generate a hazard prediction model. This model can monitor the hazard level of the coal mine in real time and take corresponding preventative measures based on the prediction results, thereby improving the safety and stability of the coal mine.

[0021] This invention uses coal mine data and mining conditions, combined with internal environmental data, to provide early warning and real-time monitoring. By using historical data as a reference, it determines the risk level of a potential strong mine pressure accident in the current coal mine, and assesses the impact of the current environment on the strong mine pressure accident based on environmental data. This comprehensive assessment reflects the degree of danger of a strong mine pressure accident in the current coal mine, facilitating timely preventive measures. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a schematic diagram of the overall system structure of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0024] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0025] Example:

[0026] Please see Figure 1 The present invention provides a technical solution: A method for predicting the risk of severe mining pressure includes the following steps: Step 1: Obtain the coal mine reserve data and environmental factor data of coal mines that have experienced severe mine collapse accidents in the historical data, obtain the accident type of the corresponding coal mine that has experienced severe mine collapse accidents, and label the risk index according to the severity of the accident type. Map the coal mine reserve data, environmental factor data and risk index of each coal mine one by one.

[0027] In this embodiment, the coal mine reserve data for which a severe mine pressure accident occurred includes the coal mine reserve data at the time of exploration, the mining time data up to the time of the severe mine pressure accident, and the average monthly mining volume for the three months prior to the accident. The mining time data refers to the time span from the time of the severe pressure accident to the time of mining. The environmental factor data includes the depth of the top layer of the coal mine from the ground surface, the coal seam thickness data, and the coal seam dip angle data at the time of exploration. The coal seam dip angle data refers to the data measured and recorded during the exploration process, representing the degree of inclination of the coal seam relative to the horizontal plane. This data can be used to assess the stability of the coal seam and the mining plan. The coal seam thickness data refers to the maximum vertical distance of a continuous coal deposit in the coal mine.

[0028] The coal mines that have experienced severe mine pressure accidents refer to those that have experienced accidents such as coal seam roof collapse, coal pillar compression, coal wall collapse, coal seam water inrush, and gas explosion. Coal seam roof collapse, coal pillar compression, coal wall collapse, coal seam water inrush, and gas explosion are the main severe mine pressure accidents that occur in coal mines. Coal mining is often accompanied by complex geological and engineering conditions, such as the coal reserves, mining time cycle, and the dip angle of the coal seam. Coal seam mining will lead to stress redistribution. If the support design is inappropriate or the countermeasures are insufficient, it is very easy to cause problems such as roof collapse, coal pillar compression, and coal wall collapse.

[0029] In coal mining, the working face is the most active area and also the place where stress is concentrated. Strong mine pressure mainly results from the pressure on the coal seam and its roof and floor exceeding the rock's own bearing capacity during mining. Abnormal release of this pressure can lead to accidents such as roof collapse and coal pillar compression. Coal mines are high-risk areas for gas accumulation. Under certain conditions, such as when the gas concentration reaches the explosive limit or an ignition source is present, a gas explosion may occur. Poor gas management, including inadequate detection, ventilation, and gas drainage, all increase the risk of explosion.

[0030] In this embodiment, during coal mining, there is a certain correlation between mining time and monthly mining volume, the depth of the top layer of the coal mine from the ground during exploration, coal seam thickness, and coal seam dip angle, and coal mine accidents. This is because these factors directly or indirectly affect the stability and safety of the mine, potentially leading to coal mine accidents. Mining time and monthly mining volume are related to the workload and pressure within the mine. Excessive mining time or large mining volume may result in excessive pressure within the mine, increasing the risk of roof collapse and coal pillar compression. These accidents can lead to mine subsidence and collapse. The depth of the top layer of the coal mine from the ground during exploration, coal seam thickness, and coal seam dip angle are important references for assessing mine stability and safety. If the top layer of the coal mine is shallow, the coal seam is thin, and the coal seam dip angle is large, the mine structure and stability may be poor, making it prone to accidents such as coal wall collapse, coal seam water inrush, and gas explosions.

[0031] In this embodiment, the accident types in coal mines that have experienced severe mine pressure accidents include coal seam roof collapse, coal pillar compression, coal wall collapse, coal seam water inrush, and gas explosion. The specific logic for determining the risk index based on the severity of the accident type is as follows: The risk index for gas explosion accidents is set at 0.9, the risk index for coal seam roof collapse accidents is set at 0.75, the risk index for coal pillar compression accidents is set at 0.6, the risk index for coal wall collapse accidents is set at 0.45, and the risk index for coal wall collapse accidents is set at 0.3.

[0032] In this embodiment, risk indices are assigned according to severity from highest to lowest: gas explosions are the most serious type of accident, causing serious casualties and damage to mine facilities, and therefore have the highest risk index. Coal seam roof collapses may cause roadway or working face collapses, affecting mine ventilation and escape, and have the second highest risk index. Coal pillar compression may cause coal pillar deformation, damage, or collapse, affecting mine support stability and safe production. Coal wall collapses may cause roadway closure or blockage, affecting normal mine production and safety. Coal seam water inrush may cause roadway flooding, affecting mine ventilation and escape, but compared to other types of accidents, the losses are usually smaller, so the risk index is the lowest.

[0033] By using risk indices to categorize high-risk mine accidents, the likelihood of different accident types can be quantitatively assessed. By categorizing accident risk indices, safety risks can be abstracted into specific values, which helps to conduct a more comprehensive and objective assessment of coal mine safety risks. The risk index reflects the severity of the accident.

[0034] Step 2: Construct a deep learning model based on a convolutional neural network. Use coal mine reserve data and environmental factor data as the training set, and use the risk index as the label. Input these data into the deep learning model and train it.

[0035] In this embodiment, the deep learning model is constructed using a deep neural network based on a multilayer perceptron. The deep neural network of the multilayer perceptron includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. The first hidden layer, the second hidden layer, and the third hidden layer each have at least two neurons and all use ReLU as the activation function.

[0036] The deep neural network structure of multilayer perceptrons can better handle nonlinear relationships. In coal mine data, there are often complex nonlinear relationships between coal reserves and environmental factors and risk indices. Deep neural networks can better capture complex patterns and features in the data through multilayer nonlinear transformations. Multilayer perceptron deep neural networks possess strong expressive power. By increasing the number of hidden layers and neurons, deep neural networks can learn richer and more abstract feature representations, thereby improving the model's predictive performance. The ReLU function has shown good performance in practice, featuring fast computation and reduced gradient vanishing. The deep learning model in this embodiment can effectively handle large-scale training data and complex nonlinear relationships, contributing to improved deep neural network performance.

[0037] In this embodiment, when training the model, the coal reserves data and environmental factor data of the corresponding coal mine are used as training data inputs, and the risk index of the accident type of the corresponding coal mine is used as a label to train the model.

[0038] Step 3: Obtain coal reserve data and environmental factor data of the coal mine to be predicted, as well as real-time environmental data of the mine. Input the coal reserve data and environmental factor data of the coal mine to be predicted into the trained deep learning model to obtain the prediction risk index.

[0039] In this embodiment, the coal reserve data of the coal mine to be predicted includes the coal reserve data at the time of exploration, the time data up to the current mining period, and the average monthly mining quantity for the previous three months. The time data up to the current mining period is the time span data of mining that has been carried out up to the current time. When calculating the average monthly mining quantity for the previous three months, if the mining period is less than three months, the average monthly ore quantity mined is calculated for all mining time periods.

[0040] The environmental factor data includes the depth of the top layer of the coal mine from the ground, the thickness of the coal seam, and the dip angle of the coal seam during exploration. The real-time environmental data of the coal mine to be predicted includes the average temperature, average pressure, average humidity, and stress equilibrium data within the mine. The specific logic for generating the average temperature, average pressure, and average humidity data within the mine is as follows: Equal distance setting Each sampling point collects temperature, pressure, humidity data inside the mine, and stress data on the upper sidewall of the mine at each sampling point. against The average values ​​of temperature, pressure, and humidity inside the mine are calculated to obtain average temperature, average pressure, and average humidity data. The formula used is as follows:

[0041]

[0042]

[0043] in, , and These represent the average temperature, average pressure, and average humidity data of the coal mine to be predicted, respectively. , and They represent the first Temperature, pressure, and humidity data from each collection point. The number of the collection point. It is a positive integer, and .

[0044] In this embodiment, the average temperature data inside the mine affects the thermal expansion and contraction properties of the rock. Increased temperature will cause the rock to expand, affecting mine support and rock mass stability. The average pressure data reflects the stress state of the rock inside the mine. Excessive pressure will cause rock damage and deformation, increasing the risk of strong mine pressure. The average humidity data has a certain impact on the strength and cohesion of the rock. Excessive humidity will cause the rock to soften and become unstable.

[0045] Step 4: Based on the obtained predicted risk index, combine it with the real-time environmental data of the mine, perform dimensionless processing to generate a hazard prediction model, and obtain the hazard prediction coefficient.

[0046] In this embodiment, the specific logic upon which the stress equilibrium data of the coal mine to be predicted is based is as follows: according to Stress data of the upper sidewall of the mine shaft, calculation The variance of the stress data is calculated and calibrated to the stress equilibrium data. The specific formula is as follows:

[0047]

[0048] in, Indicates the coal mine to be predicted Average stress data at the sampling points, Indicates the first Stress data from each collection point This represents the stress equilibrium data of the coal mine to be predicted.

[0049] The stress state at the top of the mine, i.e. the upper sidewall, directly affects the stability of the mine. If the stress state at the top is abnormal, such as stress concentration or uneven stress, it may trigger a strong mine pressure accident. Stress equilibrium data reflects whether the stress distribution at the top of the coal mine is uniform. If it is uneven, it will result in a large variance, i.e. a large stress equilibrium data, which is more likely to trigger a strong mine pressure accident.

[0050] In this embodiment, the specific logic upon which the hazard prediction model is derived to obtain the hazard prediction coefficient is as follows:

[0051] in, Indicates the risk prediction coefficient. This represents the predicted risk index obtained through a deep learning model. This indicates the standard temperature inside a coal mine.

[0052] Furthermore, the hazard prediction coefficient indicates the degree to which a mine is predicted to experience a severe mine pressure hazard. The larger the hazard prediction coefficient, the higher the predicted severe mine pressure hazard level, and the more likely a severe mine pressure accident will occur. The standard temperature inside the coal mine represents the standard temperature suitable for mining. If the average temperature data deviates significantly from the standard temperature, the impact on the thermal expansion and contraction properties of the rock is greater, which in turn has a greater impact on mine support and rock mass stability, resulting in a higher hazard prediction coefficient. The higher the average humidity, the more severe the softening and instability of the rock, resulting in a higher hazard prediction coefficient. When the average pressure is higher, excessive pressure can lead to rock damage and deformation, resulting in a higher hazard prediction coefficient. As the mine depth increases, the underground pressure increases, and the rock's supporting force may weaken. Under high pressure, the rock is more likely to undergo plastic deformation or fracture, thereby reducing its supporting capacity.

[0053] In this embodiment, the predicted risk index is obtained through a deep learning model. This reflects the degree of danger caused by a strong mine collapse accident in a coal mine. If the predicted risk index is higher, the damage caused by the strong mine collapse accident will be greater. Therefore, the higher the predicted risk index, the greater the corresponding risk prediction coefficient.

[0054] Step 5: Compare the obtained hazard prediction coefficients with the hazard reference thresholds, perceive the degree of hazard in the coal mine based on the comparison results, and generate the current hazard level of the coal mine.

[0055] In this embodiment, the specific logic for generating the current risk level of the coal mine by comparing the obtained risk prediction coefficient with the risk reference threshold is as follows: when At this time, the risk level is low, indicating that the mine is stable and the risk of a strong mine pressure accident is low. when At this time, the risk level is medium, indicating that the stability of the mine is slightly above the normal range and there is a small risk of strong mine pressure. when At this time, the danger level is high, indicating that the stability of the mine is severely compromised and there is an extremely high risk of strong mine pressure. in, This indicates the danger reference threshold.

[0056] In this embodiment, by collecting historical data on coal mine reserves and environmental factors where severe mine collapse accidents have occurred, and combining this with a risk index based on the severity of the accident type, a deep learning model is established and trained. This model can obtain a predicted risk index by inputting the reserve and environmental data of the coal mine to be predicted, thereby determining the level of potential severe mine collapse hazards. This invention considers real-time environmental data of the coal mine to be predicted and combines it with the predicted risk index to generate a hazard prediction model. This model can monitor the hazard level of the coal mine in real time and take corresponding preventative measures based on the prediction results, thereby improving the safety and stability of the coal mine.

[0057] In this embodiment, early warning and real-time monitoring are conducted by combining the coal mine's own data and mining conditions with environmental data within the mine. Historical data is used as a reference to determine the risk level of a potential strong mine collapse accident. Furthermore, environmental data is used to assess the impact of the current environment on the strong mine collapse accident, comprehensively reflecting the degree of danger of a strong mine collapse accident and facilitating timely preventative measures. Please see Figure 2 The present invention also provides a strong mine pressure hazard prediction system, the prediction system being used to execute the above-described strong mine pressure hazard prediction method, comprising: The training set acquisition module is used to acquire coal mine reserve data and environmental factor data that have experienced severe mine pressure accidents in historical data, and to acquire the accident type of the corresponding coal mine that has experienced severe mine pressure accidents. It also calibrates the risk index according to the severity of the accident type, and maps the coal mine reserve data, environmental factor data and risk index of each coal mine one by one. Model building module: Used to build a deep learning model based on convolutional neural networks. Coal mine reserve data and environmental factor data are used as training sets, and risk index is used as a label. The data are input into the deep learning model to train it. The data generation module is used to acquire coal reserve data and environmental factor data of the coal mine to be predicted, as well as real-time environmental data of the coal mine to be predicted. The coal reserve data and environmental factor data of the coal mine to be predicted are input into the trained deep learning model to obtain the prediction risk index. The data processing module is used to combine the obtained predicted risk index with real-time mine environmental data, perform dimensionless processing to generate a hazard prediction model, and obtain hazard prediction coefficients. The hazard prediction module is used to compare the acquired hazard prediction coefficients with the hazard reference thresholds, perceive the degree of hazard in the coal mine based on the comparison results, and generate the current hazard level of the coal mine. The above formulas are all dimensionless numerical calculations. The formulas are derived from software simulation based on a large amount of collected data to obtain the most recent real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0058] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0059] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0060] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for predicting the risk of severe mine pressure, characterized in that, The specific steps include: Step 1: Obtain the coal mine reserve data and environmental factor data of coal mines that have experienced severe mine pressure accidents in the historical data, obtain the accident type of the corresponding coal mine that has experienced severe mine pressure accidents, and mark the risk index according to the severity of the accident type. Map the coal mine reserve data, environmental factor data and risk index of each coal mine one by one. Step 2: Construct a deep learning model based on a convolutional neural network. Use coal mine reserve data and environmental factor data as the training set, and use the risk index as the label. Input these into the deep learning model and train it. Step 3: Obtain coal reserve data and environmental factor data of the coal mine to be predicted, as well as real-time environmental data of the mine. Input the coal reserve data and environmental factor data of the coal mine to be predicted into the trained deep learning model to obtain the prediction risk index. Step 4: Based on the obtained predicted risk index, combine it with the real-time environmental data of the mine, perform dimensionless processing to generate a hazard prediction model, and obtain the hazard prediction coefficient. Step 5: Compare the obtained hazard prediction coefficients with the hazard reference thresholds, perceive the degree of hazard in the coal mine based on the comparison results, and generate the current hazard level of the coal mine; The specific logic underlying the generation of hazard prediction models and the resulting hazard prediction coefficients is as follows: in, Indicates the risk prediction coefficient. This represents the predicted risk index obtained through a deep learning model. This indicates the standard temperature inside a coal mine. , and These represent the average temperature, average pressure, and average humidity data of the coal mine to be predicted, respectively. This represents the stress equilibrium data of the coal mine to be predicted.

2. The method for predicting the risk of strong mine pressure according to claim 1, characterized in that: The coal mine reserve data that has experienced a severe mine collapse includes the coal mine reserve data at the time of exploration, the mining time data up to the time of the severe mine collapse, and the average monthly mining volume in the three months prior to the time of the accident. The environmental factor data includes the depth of the top layer of the coal mine from the ground surface, the thickness of the coal seam, and the dip angle of the coal seam at the time of exploration.

3. The method for predicting the risk of severe mine pressure according to claim 2, characterized in that: The types of accidents that have occurred in coal mines with severe mine pressure include roof collapse, coal pillar compression, coal wall collapse, coal seam water inrush, and gas explosion. The specific logic behind assigning a risk index based on the severity of the accident type is as follows: The risk index for gas explosion accidents is set at 0.9, the risk index for coal seam roof collapse accidents is set at 0.75, the risk index for coal pillar compression accidents is set at 0.6, the risk index for coal wall collapse accidents is set at 0.45, and the risk index for coal wall collapse accidents is set at 0.

3.

4. The method for predicting the risk of strong mine pressure according to claim 1, characterized in that: The deep learning model is constructed using a deep neural network based on a multilayer perceptron. The deep neural network of the multilayer perceptron includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. The first hidden layer, the second hidden layer, and the third hidden layer each have at least two neurons and all use ReLU as the activation function.

5. The method for predicting the risk of strong mine pressure according to claim 1, characterized in that: The coal reserve data for the coal mine to be predicted includes the reserve data at the time of exploration, the time data up to the current mining period, and the average monthly mining volume for the previous three months. The environmental factor data includes the depth of the top layer of the coal mine from the ground, the thickness of the coal seam, and the dip angle of the coal seam at the time of exploration. The real-time environmental data of the coal mine to be predicted includes the average temperature data, average pressure data, average humidity data, and stress equilibrium data within the mine. The specific logic for generating the average temperature data, average pressure data, and average humidity data within the mine is as follows: Equal distance setting Each sampling point collects temperature, pressure, humidity data inside the mine, and stress data on the upper sidewall of the mine at each sampling point. against The average values ​​of temperature, pressure, and humidity inside the mine are calculated to obtain average temperature, average pressure, and average humidity data. The formula used is as follows: in, , and These represent the average temperature, average pressure, and average humidity data of the coal mine to be predicted, respectively. , and They represent the first Temperature, pressure, and humidity data from each collection point. The number of the collection point. It is a positive integer, and .

6. The method for predicting the risk of strong mine pressure according to claim 5, characterized in that: The specific logic underlying the stress equilibrium data of the coal mine to be predicted is as follows: according to Stress data of the upper sidewall of the mine shaft, calculation The variance of the stress data is calculated and calibrated to the stress equilibrium data. The specific formula is as follows: in, Indicates the coal mine to be predicted Average stress data at the sampling points, Indicates the first Stress data from each collection point This represents the stress equilibrium data of the coal mine to be predicted.

7. The method for predicting the risk of strong mine pressure according to claim 1, characterized in that: The specific logic for generating the current risk level of the coal mine by comparing the obtained risk prediction coefficient with the risk reference threshold is as follows: when At that time, the level of danger generated is low risk; when At that time, the level of danger generated was medium risk; when At that time, the level of danger generated is high risk; in, This indicates the danger reference threshold.

8. A system for predicting the risk of severe mine pressure, characterized in that: The prediction system is used to execute the strong mine pressure hazard prediction method according to any one of claims 1-7, including: The training set acquisition module is used to acquire coal mine reserve data and environmental factor data that have experienced severe mine pressure accidents in historical data, and to acquire the accident type of the corresponding coal mine that has experienced severe mine pressure accidents. It also calibrates the risk index according to the severity of the accident type, and maps the coal mine reserve data, environmental factor data and risk index of each coal mine one by one. Model building module: Used to build a deep learning model based on convolutional neural networks. Coal mine reserve data and environmental factor data are used as training sets, and risk index is used as a label. The data are input into the deep learning model to train it. The data generation module is used to acquire coal reserve data and environmental factor data of the coal mine to be predicted, as well as real-time environmental data of the coal mine to be predicted. The coal reserve data and environmental factor data of the coal mine to be predicted are input into the trained deep learning model to obtain the prediction risk index. The data processing module is used to combine the obtained predicted risk index with real-time mine environmental data, perform dimensionless processing to generate a hazard prediction model, and obtain hazard prediction coefficients. The hazard prediction module is used to compare the acquired hazard prediction coefficients with hazard reference thresholds, perceive the degree of hazard in the coal mine based on the comparison results, and generate the current hazard level of the coal mine.

Citation Information

Patent Citations

  • Assessment method for predicting underground rock burst danger of coal mine

    CN103244179A

  • Strong mine pressure risk prediction method and device

    CN117668650A