Intelligent inversion early warning method and system for dynamic evolution of coal spontaneous combustion in goaf

Abstract

The invention discloses an intelligent inversion early warning method and system for dynamic evolution of coal spontaneous combustion in a goaf, and the method comprises the following steps: 1, collecting a temperature parameter, a multi-component gas parameter and a gas flow parameter in the goaf through a sensor network composed of a temperature sensor, a multi-component gas sensor and a gas flow sensor; wherein the temperature sensors are distributed in a net shape along the trend and tendency of a goaf, the multi-component gas sensors are arranged according to a gas flowing path of the goaf, and the gas flow sensors are arranged in a potential area of an air leakage channel; the invention relates to the technical field of coal mine safety monitoring. According to the intelligent inversion early warning method and system for dynamic evolution of coal spontaneous combustion in the goaf, through cooperative acquisition of temperature, multi-component gas and gas flow parameters, in combination with scientific arrangement of sensors and a strict calibration process, high-precision and omnibearing perception of key parameters in the goaf is realized, and the problems of single parameter and insufficient precision in traditional monitoring are solved.

Description

Technical Field

[0001] This invention relates to the field of coal mine safety monitoring technology, specifically to an intelligent inversion early warning method and system for the dynamic evolution of spontaneous combustion of coal in goaf areas. Background Technology

[0002] Spontaneous combustion of coal in goaf areas is a major safety hazard in coal mine production. It is characterized by its high degree of concealment, complex evolution process, and wide range of hazards, seriously threatening the lives of underground workers and the order of mine production. When residual coal in a goaf comes into contact with air, it undergoes a slow oxidation reaction and continuously releases heat. When the heat accumulates to a certain level, it can easily ignite a spontaneous combustion fire, not only destroying large amounts of coal resources but also potentially causing secondary disasters such as gas explosions and carbon monoxide poisoning, resulting in enormous economic losses and social impact.

[0003] Currently, monitoring and early warning technologies for spontaneous combustion of coal in goaf areas still have many limitations: traditional monitoring methods mostly rely on the collection of temperature and gas parameters at single points or in local areas, and the sensor deployment lacks systematicity, making it difficult to comprehensively reflect the dynamic changes of the entire goaf area; existing early warning models often only consider the evolution law of a single physical field, ignoring the coupling effect between multiple fields such as temperature, gas concentration, and airflow, resulting in insufficient accuracy in inverting the dynamic evolution process of spontaneous combustion of coal; at the same time, the level of intelligence of early warning systems is low, mostly based on fixed thresholds to judge risk, lacking in-depth mining of historical case data and machine learning capabilities, making it difficult to achieve accurate identification and graded early warning of different stages of spontaneous combustion.

[0004] Furthermore, the complex geological conditions of the goaf, uneven distribution of air leakage channels, and strong spatial heterogeneity of coal physical parameters further increase the uncertainty of coal spontaneous combustion evolution. Existing sensor networks suffer from poor data transmission stability, non-standard parameter calibration, and imperfect outlier handling mechanisms, leading to reduced reliability of monitoring data. Meanwhile, the lack of an update and iteration mechanism for early warning models makes it difficult to adapt to the dynamic changes in the actual working conditions of the goaf, and the accuracy and timeliness of early warnings fail to meet the needs of safe mine production.

[0005] Therefore, developing a method and system that can accurately invert the dynamic evolution of spontaneous combustion of coal in goaf areas and provide intelligent early warning is of great practical significance for improving coal mine fire prevention and control capabilities and ensuring safe production in mines. Summary of the Invention

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas, comprising the following steps:

[0007] Step 1: Collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area using a sensor network consisting of temperature sensors, multi-component gas sensors, and airflow sensors.

[0008] Among them, temperature sensors are distributed in a grid pattern along the direction and inclination of the goaf, multi-component gas sensors are arranged according to the gas flow path of the goaf, and airflow sensors are set in the potential area of ​​the air leakage channel.

[0009] Step 2: Based on the parameters, the dynamic evolution process of spontaneous combustion of coal in the goaf is inverted using a multi-field coupled dynamic evolution model that considers the interaction of temperature field, gas concentration field and airflow field. The inversion process includes simulating the coupling effect of coal oxidation heat release, heat transfer, gas diffusion and convection and air leakage.

[0010] Step 3: Based on the inversion results, generate coal spontaneous combustion early warning information using a machine learning early warning model trained with historical coal spontaneous combustion case data. The early warning information includes the spontaneous combustion risk level and corresponding prevention and control recommendations.

[0011] Preferably, the multi-component gas parameters include the concentration parameters of carbon monoxide, carbon dioxide, oxygen and ethylene, wherein the detection accuracy of carbon monoxide concentration is ≤1ppm and the detection accuracy of ethylene concentration is ≤0.1ppm; the airflow parameters include airflow pressure difference and airflow velocity parameters, and the airflow velocity detection range is 0.1-10m / s.

[0012] Preferably, before acquiring parameters, the sensor network is debugged and calibrated, specifically including: calibrating the temperature sensor at four temperature points: -10℃, 50℃, 100℃, and 200℃, with a calibration error ≤0.5℃; calibrating the gas sensor by introducing a standard gas sample containing carbon monoxide, carbon dioxide, oxygen, and ethylene, with a concentration error ≤3% for three consecutive detections; and conducting a 72-hour data transmission stability test on the sensor network, with a data packet loss rate ≤0.1%.

[0013] The coordinated acquisition of temperature, multi-component gas and airflow parameters, combined with the scientific deployment of sensors and a rigorous calibration process, has enabled high-precision and comprehensive perception of key parameters in the goaf area. This provides a reliable data foundation for subsequent inversion and early warning, and solves the problems of single parameters and insufficient accuracy in traditional monitoring.

[0014] Preferably, before inverting using the multi-field coupled dynamic evolution model, the collected parameters are preprocessed. The preprocessing includes: using the 3σ criterion to remove abnormal values ​​of temperature and gas concentration, using linear interpolation to fill in ≤5 consecutive missing data points, and standardizing the temperature parameters (-50~300℃) and gas concentration parameters (0~100%) to the [0,1] interval.

[0015] Preferably, the multi-field coupled dynamic evolution model is constructed based on geological and physical parameters such as coal seam thickness, porosity, permeability, specific heat capacity, and thermal conductivity of the goaf. It uses the finite volume method for spatial discretization, with a time step of 5-15 minutes, and outputs three-dimensional dynamic evolution results once per hour. The machine learning early warning model is a support vector machine or neural network model. During training, the input historical coal spontaneous combustion case data includes historical temperature parameters, multi-component gas parameters, airflow parameters, and corresponding coal spontaneous combustion stage information. The coal spontaneous combustion stage information includes a slow oxidation stage, an accelerated self-heating stage, and a combustion stage.

[0016] The multi-field coupled dynamic evolution model comprehensively considers the interaction of temperature field, gas concentration field, and gas flow field, and combines geological and physical parameters. It achieves real-time inversion of the three-dimensional dynamic evolution process through the finite volume method and reasonable time step. It can accurately simulate key processes such as coal oxidation exothermic and heat transfer, and overcomes the inversion bias caused by the neglect of multi-field coupling effect in traditional models.

[0017] Preferably, during the application of the machine learning early warning model, iterative optimization is performed every 7-30 days based on newly added goaf monitoring data and corresponding coal spontaneous combustion status information. The optimization process includes adjusting the model weight coefficients and threshold parameters. The method also includes an early warning verification and feedback step: comparing the early warning information with the actual on-site detection results, calculating the early warning accuracy rate, and triggering a model parameter recalibration process when the accuracy rate is lower than 90%.

[0018] The machine learning early warning model trained on historical cases can directly generate risk levels and prevention and control suggestions based on the inversion results. Through regular iterative optimization and accuracy feedback calibration, it ensures that the early warning adapts to the dynamic changes in the goaf area. The parallel computing unit and efficient transmission module further ensure the real-time nature of the early warning, buying time for on-site prevention and control.

[0019] An intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas includes:

[0020] The sensing module consists of a temperature sensor, a multi-component gas sensor, and an airflow sensor, used to collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area; the temperature sensor is a distributed fiber optic temperature sensor with a sensing distance ≥10km and spatial resolution ≤1m; the multi-component gas sensor has a built-in gas filtration and dehumidification unit, and can work stably in an environment with a relative humidity of 80%-95%.

[0021] The sensing module uses distributed fiber optic sensors and high-humidity gas sensors to adapt to the complex environment of the goaf; the transmission module supports access from 800+ nodes and uses SM4 encryption to ensure data security; the display module is compatible with multiple terminals and, combined with offline caching, ensures the effective transmission and application of early warning information under complex mining conditions.

[0022] The processing module is used to receive parameters collected by the sensing module, invert the dynamic evolution process of coal spontaneous combustion using a multi-field coupled dynamic evolution model, and generate early warning information through a machine learning early warning model. The processing module is equipped with a parallel computing unit, and the single model solution time is ≤10 minutes. The processing module includes a data processing unit, a model calculation unit, and an early warning generation unit.

[0023] The display module is used to show the inversion results and early warning information, including a 3D visualized goaf model, dynamic evolution curves, and early warning signal pop-ups;

[0024] The transmission module is used to transmit the parameters collected by the sensing module to the processing module. It adopts a hybrid transmission method of 5G and industrial Ethernet, with a data transmission rate of ≥10Mbps, and has data encryption function. It uses the SM4 national cryptographic algorithm to encrypt the transmitted data.

[0025] Preferably, the data processing unit is used to preprocess the collected parameters, with a processing capacity of ≥1000 data points / second; the model operation unit is used to run a multi-field coupled dynamic evolution model, supporting online adjustment of model parameters; the early warning generation unit is used to run a machine learning early warning model and generate early warning signals of different levels, including level one, level two and level three early warnings.

[0026] Preferably, the early warning generation unit includes an emergency linkage module. When a level-three early warning signal is generated, the emergency linkage module automatically triggers the start-up instructions of prevention and control equipment such as nitrogen injection in the goaf and sealing of air leakage channels, and sends an emergency alarm to the coal mine safety monitoring center.

[0027] Preferably, the transmission module supports access from more than 800 sensor nodes, with an end-to-end data transmission latency of ≤200ms; the display module is compatible with the mine dispatch room's large screen and mobile terminals, and supports offline caching and synchronous updates of early warning information.

[0028] This invention provides an intelligent inversion and early warning method and system for the dynamic evolution of spontaneous combustion of coal in goaf areas. It has the following beneficial effects:

[0029] (I) The intelligent inversion and early warning method and system for the dynamic evolution of coal spontaneous combustion in the goaf, through the coordinated acquisition of temperature, multi-component gas and airflow parameters, combined with the scientific deployment of sensors and strict calibration process, has achieved high-precision and all-round perception of key parameters in the goaf, providing a reliable data foundation for subsequent inversion and early warning, and solving the problems of single parameters and insufficient accuracy in traditional monitoring.

[0030] (II) The intelligent inversion early warning method and system for the dynamic evolution of coal spontaneous combustion in the goaf is based on a multi-field coupled dynamic evolution model. It comprehensively considers the interaction of temperature field, gas concentration field and air flow field, and combines geological and physical parameters. It realizes the real-time inversion of the three-dimensional dynamic evolution process through the finite volume method and reasonable time step. It can accurately simulate key processes such as coal oxidation heat release and heat transfer, and overcomes the inversion deviation caused by the neglect of multi-field coupling effect in traditional models.

[0031] (III) The intelligent inversion early warning method and system for the dynamic evolution of coal spontaneous combustion in the goaf can directly generate risk levels and prevention and control suggestions based on the inversion results through a machine learning early warning model trained on historical cases. Furthermore, through regular iterative optimization and accuracy feedback calibration, the early warning is ensured to adapt to the dynamic changes in the goaf. The parallel computing unit and efficient transmission module further ensure the real-time nature of the early warning, buying time for on-site prevention and control.

[0032] (iv) The intelligent inversion early warning method and system for the dynamic evolution of coal spontaneous combustion in the goaf adopts distributed optical fiber sensors and high humidity resistant gas sensors in the sensing module to adapt to the complex environment of the goaf; the transmission module supports access of 800+ nodes and adopts SM4 encryption to ensure data security; the display module is compatible with multiple terminals and combined with offline caching function to ensure the effective transmission and application of early warning information under complex mining conditions. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating the steps of the intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas according to the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] Example 1: The present invention provides a technical solution:

[0036] A smart inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas includes the following steps:

[0037] Step 1: Collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area using a sensor network consisting of temperature sensors, multi-component gas sensors, and airflow sensors.

[0038] The multi-component gas parameters include the concentration parameters of carbon monoxide, carbon dioxide, oxygen and ethylene, wherein the detection accuracy of carbon monoxide concentration is ≤1ppm and the detection accuracy of ethylene concentration is ≤0.1ppm; the airflow parameters include airflow pressure difference and airflow velocity parameters, and the airflow velocity is detected at 0.1m / s.

[0039] Before collecting parameters, the sensor network is debugged and calibrated, specifically including: calibrating the temperature sensor at four temperature points: -10℃, 50℃, 100℃, and 200℃, with a calibration error ≤0.5℃; calibrating the gas sensor by introducing standard gas samples containing carbon monoxide, carbon dioxide, oxygen, and ethylene, with an error ≤3% for three consecutive detections; and conducting a 72-hour data transmission stability test on the sensor network, with a data packet loss rate ≤0.1%.

[0040] The coordinated acquisition of temperature, multi-component gas and airflow parameters, combined with the scientific deployment of sensors and a rigorous calibration process, has enabled high-precision and comprehensive perception of key parameters in the goaf area. This provides a reliable data foundation for subsequent inversion and early warning, and solves the problems of single parameters and insufficient accuracy in traditional monitoring.

[0041] Among them, temperature sensors are distributed in a grid pattern along the direction and inclination of the goaf, multi-component gas sensors are arranged according to the gas flow path of the goaf, and airflow sensors are set in the potential area of ​​the air leakage channel.

[0042] Step 2: Based on the parameters, the dynamic evolution process of spontaneous combustion of coal in the goaf is inverted using a multi-field coupled dynamic evolution model that considers the interaction of temperature field, gas concentration field and airflow field. The inversion process includes simulating the coupling effect of coal oxidation heat release, heat transfer, gas diffusion and convection and air leakage.

[0043] Before inversion using the multi-field coupled dynamic evolution model, the collected parameters are preprocessed. The preprocessing includes: using the 3σ criterion to remove abnormal values ​​of temperature and gas concentration, using linear interpolation to fill in ≤5 consecutive missing data points, and standardizing the temperature parameters and gas concentration parameters to the [0,1] interval.

[0044] The multi-field coupled dynamic evolution model is constructed based on geological and physical parameters such as coal seam thickness, porosity, permeability, specific heat capacity, and thermal conductivity in the goaf. It uses the finite volume method for spatial discretization, with a time step of 5 minutes, and outputs three-dimensional dynamic evolution results once per hour. The machine learning early warning model is a support vector machine or neural network model. During training, the input historical coal spontaneous combustion case data includes historical temperature parameters, multi-component gas parameters, airflow parameters, and corresponding coal spontaneous combustion stage information. The coal spontaneous combustion stage information includes a slow oxidation stage, an accelerated self-heating stage, and a combustion stage.

[0045] During the application of the machine learning early warning model, iterative optimization is performed every 7-30 days based on newly added monitoring data of goaf areas and corresponding coal spontaneous combustion status information. The optimization process includes adjusting the model weight coefficients and threshold parameters. The method also includes an early warning verification and feedback step: comparing the early warning information with the actual on-site detection results, calculating the early warning accuracy, and triggering the model parameter recalibration process when the accuracy is lower than 90%.

[0046] The machine learning early warning model trained on historical cases can directly generate risk levels and prevention and control suggestions based on the inversion results. Through regular iterative optimization and accuracy feedback calibration, it ensures that the early warning adapts to the dynamic changes in the goaf area. The parallel computing unit and efficient transmission module further ensure the real-time nature of the early warning, buying time for on-site prevention and control.

[0047] The multi-field coupled dynamic evolution model comprehensively considers the interaction of temperature field, gas concentration field, and airflow field, and combines geological and physical parameters. It realizes the real-time inversion of the three-dimensional dynamic evolution process through the finite volume method and reasonable time step. It can accurately simulate key processes such as coal oxidation exothermic and heat transfer, and overcomes the inversion deviation caused by the neglect of multi-field coupling effect in traditional models.

[0048] Step 3: Based on the inversion results, generate coal spontaneous combustion early warning information using a machine learning early warning model trained with historical coal spontaneous combustion case data. The early warning information includes the spontaneous combustion risk level and corresponding prevention and control recommendations.

[0049] An intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas includes:

[0050] The sensing module consists of a temperature sensor, a multi-component gas sensor, and an airflow sensor, used to collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area. The temperature sensor is a distributed fiber optic temperature sensor with a sensing distance of ≥10km and a spatial resolution of ≤1m. The multi-component gas sensor has a built-in gas filtration and dehumidification unit, enabling it to operate stably in an environment with a relative humidity of 80%.

[0051] The sensing module uses distributed fiber optic sensors and high-humidity gas sensors to adapt to the complex environment of the goaf; the transmission module supports access from 800+ nodes and uses SM4 encryption to ensure data security; the display module is compatible with multiple terminals and, combined with offline caching, ensures the effective transmission and application of early warning information under complex mining conditions.

[0052] The processing module is used to receive parameters collected by the sensing module, invert the dynamic evolution process of coal spontaneous combustion using a multi-field coupled dynamic evolution model, and generate early warning information through a machine learning early warning model. The processing module is equipped with a parallel computing unit, and the single model solution time is ≤10 minutes. The processing module includes a data processing unit, a model calculation unit, and an early warning generation unit.

[0053] The data processing unit is used to preprocess the collected parameters, with a processing capacity of ≥1000 data points / second; the model operation unit is used to run a multi-field coupled dynamic evolution model, supporting online adjustment of model parameters; the early warning generation unit is used to run a machine learning early warning model and generate early warning signals of different levels, including level one, level two, and level three early warnings.

[0054] The early warning generation unit includes an emergency linkage module. When a level 3 early warning signal is generated, the emergency linkage module automatically triggers the start instructions for prevention and control equipment such as nitrogen injection in the goaf and sealing of air leakage channels, and sends an emergency alarm to the coal mine safety monitoring center.

[0055] The display module is used to show the inversion results and early warning information, including a 3D visualized goaf model, dynamic evolution curves, and early warning signal pop-ups;

[0056] The transmission module is used to transmit the parameters collected by the sensing module to the processing module. It adopts a hybrid transmission method of 5G and industrial Ethernet, with a data transmission rate of ≥10Mbps, and has data encryption function. It uses the SM4 national cryptographic algorithm to encrypt the transmitted data.

[0057] The transmission module supports the access of more than 800 sensor nodes, with an end-to-end data transmission latency of ≤200ms; the display module is compatible with the mine dispatch room's large screen and mobile terminals, and supports offline caching and synchronous updates of early warning information.

[0058] Example 2: The present invention provides a technical solution:

[0059] A smart inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas includes the following steps:

[0060] Step 1: Collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area using a sensor network consisting of temperature sensors, multi-component gas sensors, and airflow sensors.

[0061] The multi-component gas parameters include the concentration parameters of carbon monoxide, carbon dioxide, oxygen and ethylene, wherein the detection accuracy of carbon monoxide concentration is ≤1ppm and the detection accuracy of ethylene concentration is ≤0.1ppm; the airflow parameters include airflow pressure difference and airflow velocity parameters, and the airflow velocity is detected at 10m / s;

[0062] Before collecting parameters, the sensor network is debugged and calibrated, specifically including: calibrating the temperature sensor at four temperature points: -10℃, 50℃, 100℃, and 200℃, with a calibration error ≤0.5℃; calibrating the gas sensor by introducing standard gas samples containing carbon monoxide, carbon dioxide, oxygen, and ethylene, with an error ≤3% for three consecutive detections; and conducting a 72-hour data transmission stability test on the sensor network, with a data packet loss rate ≤0.1%.

[0063] The coordinated acquisition of temperature, multi-component gas and airflow parameters, combined with the scientific deployment of sensors and a rigorous calibration process, has enabled high-precision and comprehensive perception of key parameters in the goaf area. This provides a reliable data foundation for subsequent inversion and early warning, and solves the problems of single parameters and insufficient accuracy in traditional monitoring.

[0064] Among them, temperature sensors are distributed in a mesh pattern along the direction and inclination of the goaf, multi-component gas sensors are arranged according to the gas flow path of the goaf, and airflow sensors are set in the potential area of ​​the air leakage channel.

[0065] Step 2: Based on the parameters, the dynamic evolution process of spontaneous combustion of coal in the goaf is inverted using a multi-field coupled dynamic evolution model that considers the interaction of temperature field, gas concentration field and airflow field. The inversion process includes simulating the coupling effect of coal oxidation heat release, heat transfer, gas diffusion and convection and air leakage.

[0066] Before inversion using the multi-field coupled dynamic evolution model, the collected parameters are preprocessed. The preprocessing includes: using the 3σ criterion to remove outliers in temperature and gas concentration, using linear interpolation to fill in ≤5 consecutive missing data points, and standardizing the temperature and gas concentration parameters.

[0067] The multi-field coupled dynamic evolution model is constructed based on geological and physical parameters such as coal seam thickness, porosity, permeability, specific heat capacity, and thermal conductivity in the goaf. It uses the finite volume method for spatial discretization, with a time step of 15 minutes, and outputs three-dimensional dynamic evolution results once per hour. The machine learning early warning model is a support vector machine or neural network model. During training, the input historical coal spontaneous combustion case data includes historical temperature parameters, multi-component gas parameters, airflow parameters, and corresponding coal spontaneous combustion stage information. The coal spontaneous combustion stage information includes a slow oxidation stage, an accelerated self-heating stage, and a combustion stage.

[0068] During the application of the machine learning early warning model, iterative optimization is performed every 30 days based on newly added monitoring data of goaf areas and corresponding coal spontaneous combustion status information. The optimization process includes adjusting the model weight coefficients and threshold parameters. The method also includes an early warning verification and feedback step: comparing the early warning information with the actual on-site detection results, calculating the early warning accuracy rate, and triggering the model parameter recalibration process when the accuracy rate is lower than 90%.

[0069] The machine learning early warning model trained on historical cases can directly generate risk levels and prevention and control suggestions based on the inversion results. Through regular iterative optimization and accuracy feedback calibration, it ensures that the early warning adapts to the dynamic changes in the goaf area. The parallel computing unit and efficient transmission module further ensure the real-time nature of the early warning, buying time for on-site prevention and control.

[0070] The multi-field coupled dynamic evolution model comprehensively considers the interaction of temperature field, gas concentration field, and airflow field, and combines geological and physical parameters. It realizes the real-time inversion of the three-dimensional dynamic evolution process through the finite volume method and reasonable time step. It can accurately simulate key processes such as coal oxidation exothermic and heat transfer, and overcomes the inversion deviation caused by the neglect of multi-field coupling effect in traditional models.

[0071] Step 3: Based on the inversion results, generate coal spontaneous combustion early warning information using a machine learning early warning model trained with historical coal spontaneous combustion case data. The early warning information includes the spontaneous combustion risk level and corresponding prevention and control recommendations.

[0072] An intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas includes:

[0073] The sensing module consists of a temperature sensor, a multi-component gas sensor, and an airflow sensor, used to collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area; the temperature sensor is a distributed fiber optic temperature sensor with a sensing distance ≥10km and spatial resolution ≤1m; the multi-component gas sensor has a built-in gas filtration and dehumidification unit, and can work stably in an environment with a relative humidity of 95%.

[0074] The sensing module uses distributed fiber optic sensors and high-humidity gas sensors to adapt to the complex environment of the goaf; the transmission module supports access from 800+ nodes and uses SM4 encryption to ensure data security; the display module is compatible with multiple terminals and, combined with offline caching, ensures the effective transmission and application of early warning information under complex mining conditions.

[0075] The processing module is used to receive parameters collected by the sensing module, invert the dynamic evolution process of coal spontaneous combustion using a multi-field coupled dynamic evolution model, and generate early warning information through a machine learning early warning model. The processing module is equipped with a parallel computing unit, and the single model solution time is ≤10 minutes. The processing module includes a data processing unit, a model calculation unit, and an early warning generation unit.

[0076] The data processing unit is used to preprocess the collected parameters, with a processing capacity of ≥1000 data points / second; the model operation unit is used to run a multi-field coupled dynamic evolution model, supporting online adjustment of model parameters; the early warning generation unit is used to run a machine learning early warning model and generate early warning signals of different levels, including level one, level two, and level three early warnings.

[0077] The early warning generation unit includes an emergency linkage module. When a level 3 early warning signal is generated, the emergency linkage module automatically triggers the start instructions for prevention and control equipment such as nitrogen injection in the goaf and sealing of air leakage channels, and sends an emergency alarm to the coal mine safety monitoring center.

[0078] The display module is used to show the inversion results and early warning information, including a 3D visualized goaf model, dynamic evolution curves, and early warning signal pop-ups;

[0079] The transmission module is used to transmit the parameters collected by the sensing module to the processing module. It adopts a hybrid transmission method of 5G and industrial Ethernet, with a data transmission rate of ≥10Mbps, and has data encryption function. It uses the SM4 national cryptographic algorithm to encrypt the transmitted data.

[0080] The transmission module supports the access of more than 800 sensor nodes, with an end-to-end data transmission latency of ≤200ms; the display module is compatible with the mine dispatch room's large screen and mobile terminals, and supports offline caching and synchronous updates of early warning information.

[0081] Example 3: The present invention provides a technical solution:

[0082] A smart inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas includes the following steps:

[0083] Step 1: Collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area using a sensor network consisting of temperature sensors, multi-component gas sensors, and airflow sensors.

[0084] The multi-component gas parameters include the concentration parameters of carbon monoxide, carbon dioxide, oxygen and ethylene, wherein the detection accuracy of carbon monoxide concentration is ≤1ppm and the detection accuracy of ethylene concentration is ≤0.1ppm; the airflow parameters include airflow pressure difference and airflow velocity parameters, and the airflow velocity detection range is 0.1-10m / s;

[0085] Before collecting parameters, the sensor network is debugged and calibrated, specifically including: calibrating the temperature sensor at four temperature points: -10℃, 50℃, 100℃, and 200℃, with a calibration error ≤0.5℃; calibrating the gas sensor by introducing standard gas samples containing carbon monoxide, carbon dioxide, oxygen, and ethylene, with an error ≤3% for three consecutive detections; and conducting a 72-hour data transmission stability test on the sensor network, with a data packet loss rate ≤0.1%.

[0086] The coordinated acquisition of temperature, multi-component gas and airflow parameters, combined with the scientific deployment of sensors and a rigorous calibration process, has enabled high-precision and comprehensive perception of key parameters in the goaf area. This provides a reliable data foundation for subsequent inversion and early warning, and solves the problems of single parameters and insufficient accuracy in traditional monitoring.

[0087] Among them, temperature sensors are distributed in a grid pattern along the direction and inclination of the goaf, multi-component gas sensors are arranged according to the gas flow path of the goaf, and airflow sensors are set in the potential area of ​​the air leakage channel.

[0088] Step 2: Based on the parameters, the dynamic evolution process of spontaneous combustion of coal in the goaf is inverted using a multi-field coupled dynamic evolution model that considers the interaction of temperature field, gas concentration field and airflow field. The inversion process includes simulating the coupling effect of coal oxidation heat release, heat transfer, gas diffusion and convection and air leakage.

[0089] Before inversion using the multi-field coupled dynamic evolution model, the collected parameters are preprocessed. The preprocessing includes: using the 3σ criterion to remove outliers in temperature and gas concentration, using linear interpolation to fill in ≤5 consecutive missing data points, and standardizing the temperature and gas concentration parameters.

[0090] The multi-field coupled dynamic evolution model is constructed based on geological and physical parameters such as coal seam thickness, porosity, permeability, specific heat capacity, and thermal conductivity in the goaf. It uses the finite volume method for spatial discretization, with a time step of 10 minutes, and outputs three-dimensional dynamic evolution results once per hour. The machine learning early warning model is a support vector machine or neural network model. During training, the input historical coal spontaneous combustion case data includes historical temperature parameters, multi-component gas parameters, airflow parameters, and corresponding coal spontaneous combustion stage information. The coal spontaneous combustion stage information includes a slow oxidation stage, an accelerated self-heating stage, and a combustion stage.

[0091] During the application of the machine learning early warning model, iterative optimization is performed every 7-30 days based on newly added monitoring data of goaf areas and corresponding coal spontaneous combustion status information. The optimization process includes adjusting the model weight coefficients and threshold parameters. The method also includes an early warning verification and feedback step: comparing the early warning information with the actual on-site detection results, calculating the early warning accuracy, and triggering the model parameter recalibration process when the accuracy is lower than 90%.

[0092] The machine learning early warning model trained on historical cases can directly generate risk levels and prevention and control suggestions based on the inversion results. Through regular iterative optimization and accuracy feedback calibration, it ensures that the early warning adapts to the dynamic changes in the goaf area. The parallel computing unit and efficient transmission module further ensure the real-time nature of the early warning, buying time for on-site prevention and control.

[0093] The multi-field coupled dynamic evolution model comprehensively considers the interaction of temperature field, gas concentration field, and airflow field, and combines geological and physical parameters. It realizes the real-time inversion of the three-dimensional dynamic evolution process through the finite volume method and reasonable time step. It can accurately simulate key processes such as coal oxidation exothermic and heat transfer, and overcomes the inversion deviation caused by the neglect of multi-field coupling effect in traditional models.

[0094] Step 3: Based on the inversion results, generate coal spontaneous combustion early warning information using a machine learning early warning model trained with historical coal spontaneous combustion case data. The early warning information includes the spontaneous combustion risk level and corresponding prevention and control recommendations.

[0095] An intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas includes:

[0096] The sensing module consists of a temperature sensor, a multi-component gas sensor, and an airflow sensor, and is used to collect temperature parameters, multi-component gas parameters, and airflow parameters in the goaf area. The temperature sensor is a distributed fiber optic temperature sensor with a sensing distance of ≥10km and a spatial resolution of ≤1m. The multi-component gas sensor has a built-in gas filtration and dehumidification unit and can work stably in an environment with a relative humidity of 90%.

[0097] The sensing module uses distributed fiber optic sensors and high-humidity gas sensors to adapt to the complex environment of the goaf; the transmission module supports access from 800+ nodes and uses SM4 encryption to ensure data security; the display module is compatible with multiple terminals and, combined with offline caching, ensures the effective transmission and application of early warning information under complex mining conditions.

[0098] The processing module is used to receive parameters collected by the sensing module, invert the dynamic evolution process of coal spontaneous combustion using a multi-field coupled dynamic evolution model, and generate early warning information through a machine learning early warning model. The processing module is equipped with a parallel computing unit, and the single model solution time is ≤10 minutes. The processing module includes a data processing unit, a model calculation unit, and an early warning generation unit.

[0099] The data processing unit is used to preprocess the collected parameters, with a processing capacity of ≥1000 data points / second; the model operation unit is used to run a multi-field coupled dynamic evolution model, supporting online adjustment of model parameters; the early warning generation unit is used to run a machine learning early warning model and generate early warning signals of different levels, including level one, level two, and level three early warnings.

[0100] The early warning generation unit includes an emergency linkage module. When a level 3 early warning signal is generated, the emergency linkage module automatically triggers the start instructions for prevention and control equipment such as nitrogen injection in the goaf and sealing of air leakage channels, and sends an emergency alarm to the coal mine safety monitoring center.

[0101] The display module is used to show the inversion results and early warning information, including a 3D visualized goaf model, dynamic evolution curves, and early warning signal pop-ups;

[0102] The transmission module is used to transmit the parameters collected by the sensing module to the processing module. It adopts a hybrid transmission method of 5G and industrial Ethernet, with a data transmission rate of ≥10Mbps, and has data encryption function. It uses the SM4 national cryptographic algorithm to encrypt the transmitted data.

[0103] The transmission module supports the access of more than 800 sensor nodes, with an end-to-end data transmission latency of ≤200ms; the display module is compatible with the mine dispatch room's large screen and mobile terminals, and supports offline caching and synchronous updates of early warning information.

[0104] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0105] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas, characterized in that, Includes the following steps: Step 1: Collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area using a sensor network consisting of temperature sensors, multi-component gas sensors, and airflow sensors. Among them, temperature sensors are distributed in a mesh pattern along the direction and inclination of the goaf, multi-component gas sensors are arranged according to the gas flow path of the goaf, and airflow sensors are set in the potential area of ​​the air leakage channel. Step 2: Based on the parameters, the dynamic evolution process of spontaneous combustion of coal in the goaf is inverted using a multi-field coupled dynamic evolution model that considers the interaction of temperature field, gas concentration field and airflow field. The inversion process includes simulating the coupling effect of coal oxidation heat release, heat transfer, gas diffusion and convection and air leakage. Step 3: Based on the inversion results, generate coal spontaneous combustion early warning information using a machine learning early warning model trained with historical coal spontaneous combustion case data. The early warning information includes the spontaneous combustion risk level and corresponding prevention and control recommendations.

2. The intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 1, characterized in that: The multi-component gas parameters include the concentration parameters of carbon monoxide, carbon dioxide, oxygen and ethylene, wherein the detection accuracy of carbon monoxide concentration is ≤1ppm and the detection accuracy of ethylene concentration is ≤0.1ppm; the airflow parameters include airflow pressure difference and airflow velocity parameters, and the airflow velocity detection range is 0.1-10m / s.

3. The intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 1, characterized in that: Before collecting parameters, the sensor network is debugged and calibrated, specifically including: calibrating the temperature sensor at four temperature points: -10℃, 50℃, 100℃, and 200℃; calibrating the gas sensor by introducing a standard gas sample containing carbon monoxide, carbon dioxide, oxygen, and ethylene; and conducting a 72-hour data transmission stability test on the sensor network.

4. The intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 1, characterized in that: Before inversion using the multi-field coupled dynamic evolution model, the collected parameters are preprocessed. The preprocessing includes: using the 3σ criterion to remove outliers in temperature and gas concentration, using linear interpolation to fill in ≤5 consecutive missing data points, and standardizing the temperature and gas concentration parameters.

5. The intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 1, characterized in that: The multi-field coupled dynamic evolution model is constructed based on geological and physical parameters such as coal seam thickness, porosity, permeability, specific heat capacity, and thermal conductivity in the goaf. It uses the finite volume method for spatial discretization and outputs three-dimensional dynamic evolution results once per hour. The machine learning early warning model is a support vector machine or neural network model. During training, the input historical coal spontaneous combustion case data includes historical temperature parameters, multi-component gas parameters, airflow parameters, and corresponding coal spontaneous combustion stage information. The coal spontaneous combustion stage information includes a slow oxidation stage, an accelerated self-heating stage, and a combustion stage.

6. The intelligent inversion and early warning method for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 5, characterized in that: During the application of the machine learning early warning model, iterative optimization is performed every 7-30 days based on newly added goaf monitoring data and corresponding coal spontaneous combustion status information. The optimization process includes adjusting the model weight coefficients and threshold parameters. The method also includes an early warning verification and feedback step: comparing the early warning information with the actual on-site detection results, calculating the early warning accuracy rate, and triggering the model parameter recalibration process when the accuracy rate is lower than 90%.

7. An intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas, characterized in that, include: The sensing module consists of a temperature sensor, a multi-component gas sensor, and an airflow sensor, and is used to collect temperature parameters, multi-component gas parameters, and airflow parameters within the goaf area; the temperature sensor is a distributed fiber optic temperature sensor, and the multi-component gas sensor has a built-in gas filtration and dehumidification unit. The processing module receives parameters collected by the sensing module, uses a multi-field coupled dynamic evolution model to invert the dynamic evolution process of coal spontaneous combustion, and generates early warning information through a machine learning early warning model. The processing module is configured with a parallel computing unit, and the processing module includes a data processing unit, a model calculation unit, and an early warning generation unit. The display module is used to show the inversion results and early warning information, including a 3D visualized goaf model, dynamic evolution curves, and early warning signal pop-ups; The transmission module is used to transmit the parameters collected by the sensing module to the processing module. It adopts a hybrid transmission method of 5G and industrial Ethernet, with a data transmission rate of ≥10Mbps, and has data encryption function. It uses the SM4 national cryptographic algorithm to encrypt the transmitted data.

8. The intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 7, characterized in that: The data processing unit is used to preprocess the collected parameters, with a processing capacity of ≥1000 data points / second; the model operation unit is used to run a multi-field coupled dynamic evolution model, supporting online adjustment of model parameters; the early warning generation unit is used to run a machine learning early warning model and generate early warning signals of different levels, including level one, level two and level three early warnings.

9. The intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 8, characterized in that: The early warning generation unit includes an emergency linkage module. When a level 3 early warning signal is generated, the emergency linkage module automatically triggers the start instructions for prevention and control equipment such as nitrogen injection in the goaf and sealing of air leakage channels, and sends an emergency alarm to the coal mine safety monitoring center.

10. The intelligent inversion and early warning system for the dynamic evolution of spontaneous combustion of coal in goaf areas according to claim 9, characterized in that: The transmission module supports the access of multiple sensor nodes, and the display module is compatible with the mine dispatch room's large screen and mobile terminals, supporting offline caching and synchronous updates of early warning information.

Images