Coal mine goaf spontaneous combustion three-zone dynamic monitoring and liquid carbon dioxide injection regulation and control system

Abstract

The invention discloses a coal mine goaf spontaneous combustion three-zone dynamic monitoring and liquid carbon dioxide injection regulation and control system, and relates to the technical field of mine field flame prevention. Deploying a sensor to dynamically adjust the acquisition layout and frequency; the data processing and analyzing module is used for dividing three spontaneous combustion zones by using deep preprocessing and an improved neural network and judging risks; the liquid carbon dioxide injection control module regulates and controls liquid injection according to needs through a regulating valve and a fuzzy-neural network algorithm; and the remote monitoring and early warning module is used for realizing monitoring, multi-stage early warning and virtual simulation based on cloud and big data, and all the modules are coordinated and can be fused with the existing system. The coal mine goaf spontaneous combustion three-zone accurate monitoring and liquid carbon dioxide efficient injection are achieved through accurate multi-source monitoring, intelligent analysis decision and dynamic accurate regulation and control; the system can early warn hidden dangers in advance, improve prevention and control accuracy, reduce resource waste, and can be integrated with an existing system to guarantee safety production of a coal mine and reduce accident risks.

Description

Technical Field

[0001] The present invention relates to the field of mine fire prevention technology, and in particular to a system for dynamically monitoring three spontaneous combustion zones in a coal mine goaf and injecting liquid carbon dioxide. Background Art

[0002] During coal mining, spontaneous combustion of coal in goafs is a major hidden danger that threatens production safety. With the continuous increase in the depth and intensity of coal mining, the scope of goafs has expanded, and the geological conditions have become increasingly complex, the limitations of traditional monitoring and prevention methods have become increasingly prominent. Existing coal mine goaf spontaneous combustion monitoring mostly relies on a single type of sensor, such as obtaining data only through temperature sensors or carbon monoxide sensors. This monitoring method not only does not collect comprehensive information, but also has low sensor accuracy and slow response speed, making it difficult to accurately capture subtle changes in the early stages of spontaneous combustion in real time, resulting in the inability to promptly detect spontaneous combustion hazards. At the same time, the installation layout of the sensors lacks scientific planning and fails to fully consider factors such as the geological structure and ventilation conditions of the goaf. As a result, the collected data cannot truly reflect the actual situation of the three spontaneous combustion zones, which brings great difficulties to subsequent prevention and control work.

[0003] In terms of spontaneous combustion prevention and control, the injection of liquid carbon dioxide is a commonly used fire prevention and extinguishing technology, but its control process has many problems. The current injection system mostly adopts a fixed injection strategy, which cannot be flexibly adjusted according to the dynamic changes of the three spontaneous combustion zones in the goaf and the actual risk level. For example, in areas with a low risk of coal spontaneous combustion, liquid carbon dioxide is still injected according to the conventional flow rate, resulting in a waste of liquid carbon dioxide; while in high-risk areas, the injection volume and injection time may be insufficient, making it difficult to effectively suppress coal oxidation and spontaneous combustion. In addition, the injection flow control lacks precision, and the existing regulating valve has a slow response speed and low adjustment accuracy. It cannot quickly adapt to the complex and changing working conditions of the goaf, resulting in a significant reduction in the injection effect.

[0004] The entire monitoring and control system lacks effective collaboration and intelligent decision-making capabilities across its components. Data processing and analysis methods are outdated, often relying on simple threshold judgments and statistical analysis methods, making it difficult to deeply explore the patterns of spontaneous combustion and risk trends underlying the data. The remote monitoring system's functionality is limited, enabling only simple data display and alarms, failing to provide managers with a scientific basis for decision-making. Furthermore, the system is independent of the coal mine's existing safety monitoring and automation control systems, preventing data from being shared and creating information silos. This severely restricts the overall effectiveness of spontaneous combustion prevention and control in coal mine goafs, necessitating an urgent need for more advanced and intelligent monitoring and control systems to ensure safe production in coal mines. Summary of the Invention

[0005] The present invention proposes a system for dynamically monitoring the three spontaneous combustion zones in coal mine goafs and injecting liquid carbon dioxide to solve the problems mentioned in the above-mentioned prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a system for dynamic monitoring of three zones of spontaneous combustion in coal mine goaf and injection of liquid carbon dioxide, comprising:

[0007] Sensor data acquisition module: Temperature sensors, oxygen sensors, and carbon monoxide sensors are installed in the tunnels, roofs, and floor positions of the coal mine goaf. Semiconductor carbon monoxide sensors with integrated noise reduction algorithms are used to collect data through algorithms. Remove the interference signal, where N filtered is the filtered signal, N raw is the original signal, α is the noise reduction coefficient, N noise (i) is the i-th noise sample, collecting carbon monoxide concentration data in real time; the methane sensor uses multi-band spectral detection technology. The sensors are distributed in the goaf with a dynamic grid spacing. The spacing is adjusted in real time according to the mining progress and geological changes in the goaf. The data collection frequency is dynamically set according to the data change rate through an adaptive adjustment algorithm. The collected data is transmitted to the data processing and analysis module via wireless ad hoc network communication;

[0008] Data processing and analysis module: Receives data transmitted by the sensor data acquisition module, first pre-processes the data, including adaptive filtering, data fusion and feature extraction; for temperature data, an adaptive filtering algorithm based on wavelet transform is used to automatically adjust the filtering parameters to remove noise according to the fluctuation characteristics of the data; for gas concentration data, a data fusion algorithm based on Bayesian theory is used to improve data accuracy by combining prior knowledge and real-time measurement data; then, a deep neural network self-ignition three-zone division model is used to introduce an attention mechanism through the formula Different weights are assigned to different features to highlight the impact of features on the division of the three spontaneous combustion zones. i is the attention weight of the i-th feature, W a is the attention weight matrix, h i is the i-th eigenvector; at the same time, according to the data change trend, through the formula Calculate the weighted change rate of each gas concentration to determine the spontaneous combustion risk level, where ΔC weighted is the weighted gas concentration change rate, β i is the weight of the i-th gas, C t+1(i) and C t(i) are the concentrations of the i-th gas at adjacent moments, and Δt is the time interval.

[0009] Furthermore, it also includes:

[0010] Liquid carbon dioxide injection control module: Determine the location, flow rate and time of liquid carbon dioxide injection based on the three-zone classification results and spontaneous combustion risk level obtained by the data processing and analysis module; use a flow control valve to control the flow of liquid carbon dioxide, and set the control valve to have self-learning and adaptive adjustment functions. Use a fuzzy neural network control algorithm to adjust the flow rate, and dynamically adjust the opening of the control valve based on oxygen concentration, temperature and carbon monoxide concentration factors; the relationship between the injection flow rate and various factors is mapped through the formula Q = f(O2, T, CO), where Q is the injection flow rate, O2 is the oxygen concentration, and T is the temperature , CO is the carbon monoxide concentration, and the function f is obtained through neural network training. In the oxidation zone, the injection flow rate is dynamically adjusted through the control algorithm according to the oxygen concentration and temperature to reduce the oxygen concentration and inhibit coal oxidation and spontaneous combustion. The injection time is determined according to the duration of the spontaneous combustion risk and the injection effect evaluation model. The injection effect evaluation model is evaluated by the formula E=ω1·ΔO2+ω2·ΔT+ω3·ΔCO, where E is the injection effect evaluation value, ω1, ω2, and ω3 are weight coefficients, and ΔO2, ΔT, and ΔCO are the changes in oxygen concentration, temperature, and carbon monoxide concentration before and after injection, respectively.

[0011] Remote monitoring and early warning module: Establish a remote monitoring center based on cloud computing and big data analysis, connect to the data processing and analysis module through the 5G network, and display the temperature, gas concentration, three-zone division of spontaneous combustion, and liquid carbon dioxide injection status of the goaf in real time; set early warning thresholds at different levels. When the data exceeds the early warning thresholds at different levels, the system automatically issues different types of sound and light alarm signals and notifies relevant personnel via SMS, email, and app push;

[0012] Sensor fault diagnosis and compensation submodule: The module monitors the working status of the sensor in real time, and determines whether the sensor is faulty through feature analysis and anomaly detection algorithms. If the sensor is faulty, it uses a data compensation algorithm based on time-space correlation to compensate for the historical data and the data of adjacent normal sensors. The algorithm is calculated by the formula Combined with the spatiotemporal correlation coefficient to determine the weight to achieve data continuity, where C compensated is the compensated data, γ j is the weight of the jth adjacent sensor, C neighbor(j) is the data of the jth adjacent sensor.

[0013] Liquid CO2 storage and supply optimization module: The module monitors the liquid level, pressure and temperature parameters of the liquid CO2 storage tank in real time, and calculates the state of the liquid CO2 in the storage tank through a thermodynamic model; the thermodynamic model takes into account the material properties of the storage tank, ambient temperature changes and the phase change process of CO2, and calculates the state of the liquid CO2 in the storage tank through the formula Predict the pressure change in the storage tank, where P is pressure, n is the amount of substance, R is the gas constant, T is temperature, V is volume, δ is the ambient temperature influence coefficient, T env The ambient temperature is set; according to the demand forecast of the liquid carbon dioxide injection control module, the optimization strategy based on genetic algorithm is adopted to optimize the storage and supply strategy of liquid carbon dioxide in combination with the supply cost, transportation distance and storage capacity factors to arrange the replenishment and distribution of liquid carbon dioxide; at the same time, a safety valve and a pressure regulating device are set to ensure the safety of the storage tank. The safety valve automatically adjusts the opening pressure according to the pressure change in the storage tank. The formula P open =P base +β·ΔP for dynamic adjustment, where P open is the opening pressure, P base is the basic opening pressure, β is the pressure adjustment coefficient, and ΔP is the pressure change.

[0014] Furthermore, the sensors in the sensor data acquisition module adopt a modular design, which is convenient for installation, maintenance and replacement. The installation position of the sensor is optimized and determined through three-dimensional geological modeling and numerical simulation technology. According to the geological conditions, ventilation conditions and coal spontaneous combustion characteristics of the goaf, the monitoring effects of the sensors at different installation positions are simulated, and the optimal installation plan is selected to accurately collect data reflecting the spontaneous combustion status of the goaf.

[0015] Furthermore, the deep neural network spontaneous combustion three-zone division model in the data processing and analysis module adopts transfer learning and incremental learning technology; at the same time, the model adopts model fusion technology to fuse neural network models of different structures, and obtains the final division result through weighted averaging, thereby improving the accuracy of the spontaneous combustion three-zone division.

[0016] Furthermore, the flow control valve in the liquid carbon dioxide injection control module adopts piezoelectric drive technology, and the control algorithm of the control valve adopts a model-predictive control method. By establishing a dynamic model of the injection process, the injection effect in the future is predicted, and the opening of the control valve is adjusted in real time according to the prediction results; at the same time, the module is also equipped with a flow feedback adjustment function, which monitors the injection flow in real time through a flow sensor installed on the injection pipeline and compares it with the set flow, and performs error correction through a PID control algorithm.

[0017] Furthermore, the remote monitoring and early warning module is equipped with a decision-making assistance function, which integrates and associates the geological information, mining process, and safety specification knowledge of the coal mine goaf through knowledge graph technology. When abnormalities occur in the monitoring data, the system performs reasoning and analysis based on the knowledge graph to provide decision-making suggestions for relevant personnel; at the same time, the module is also equipped with a visual interaction function, and users interact with the virtual simulation model through a touch screen or mouse device to adjust the working parameters in real time.

[0018] Furthermore, the feature analysis and anomaly detection algorithms in the sensor fault diagnosis and compensation module adopt a distributed computing architecture to distribute data processing tasks to computing nodes for parallel processing; at the same time, the module is also equipped with a fault prediction function. By analyzing the historical fault data and current working status data of the sensor, a time series analysis algorithm is used to predict the time and type of sensor failure so that maintenance and replacement can be carried out in advance.

[0019] Furthermore, the genetic algorithm-based optimization strategy in the liquid carbon dioxide storage and supply optimization module adopts an adaptive genetic operator to dynamically adjust the crossover probability and mutation probability according to the fitness value during the optimization process, thereby improving the algorithm's search efficiency and convergence speed; at the same time, the module also takes into account the quality change factor of liquid carbon dioxide, and adjusts the supply strategy according to the test results by regularly testing the purity and impurity content of liquid carbon dioxide.

[0020] Furthermore, the system also includes an interface module that integrates with the coal mine's existing safety monitoring system and automation control system; the interface module adopts standardized data communication protocols and interface specifications to realize data transmission and real-time sharing; through data fusion, this system utilizes the resources and data of the coal mine's existing system, and at the same time feeds back the monitoring data and control information of this system to the existing system, realizing the integrated management of coal mine safety monitoring and control.

[0021] Compared with the existing technology, the beneficial effects of the present invention are:

[0022] In the monitoring process, by deploying a variety of high-precision, customized sensors and adopting dynamic grid layout and adaptive acquisition frequency adjustment, key data such as goaf temperature and gas concentration can be obtained comprehensively, accurately and in real time. Compared with traditional monitoring methods, data accuracy and timeliness are greatly improved, and spontaneous combustion hazards can be discovered in advance.

[0023] The data processing and analysis module utilizes advanced algorithms and models to not only accurately delineate the three spontaneous combustion zones but also more scientifically determine the level of spontaneous combustion risk through weighted calculations of gas concentration change rates, providing a reliable basis for prevention and control efforts. The liquid carbon dioxide injection control module's intelligent flow control valve and advanced control algorithms dynamically and precisely adjust the injection location, flow rate, and timing based on risk levels and real-time monitoring data, effectively suppressing coal spontaneous combustion while avoiding waste of liquid carbon dioxide.

[0024] The remote monitoring and early warning module, based on cloud computing, big data, and virtual simulation technologies, enables efficient data display, multi-level early warning, and intelligent decision-making support. Through a visual interactive interface, managers can intuitively understand the status of goaf areas and obtain decision-making recommendations, improving management efficiency and the scientific nature of decision-making. Furthermore, the coordinated cooperation between the system's submodules and its deep integration with existing coal mine systems have broken down information silos, enabling integrated management of safety monitoring and control. This has comprehensively improved the overall effectiveness of spontaneous combustion prevention and control in coal mine goaf areas, effectively ensuring safe production in coal mines and reducing economic losses and the occurrence of safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic block diagram of the system for dynamic monitoring of three zones of spontaneous combustion in coal mine goaf and injection of liquid carbon dioxide proposed by the present invention;

[0026] Figure 2 This is a schematic diagram comparing the accuracy of sensor data acquisition for the three-zone dynamic monitoring of spontaneous combustion in coal mine goafs and liquid carbon dioxide injection in different systems proposed by the present invention. DETAILED DESCRIPTION

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0028] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0029] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below with reference to the accompanying drawings.

[0030] Reference Figure 1 and Figure 2 :A specific implementation method of a system for dynamic monitoring of three zones of spontaneous combustion in coal mine goaf and injection of liquid carbon dioxide

[0031] During coal mining operations, the various modules of this application system work closely together to comprehensively ensure the safety of coal mine goafs, from data collection to control execution. The specific implementation details are as follows:

[0032] Multi-sensor data acquisition module: In terms of sensor selection, the temperature sensor uses a distributed fiber Bragg grating temperature sensor that is resistant to high temperature and electromagnetic interference. Its temperature measurement accuracy is -40℃-60℃, with a range of ±0.05℃, and an accuracy of ±0.1℃ in the range of 60℃-120℃; the oxygen sensor uses an electrochemical sensor based on micro-nano structure, which has fast response characteristics, a response time of less than 5 seconds, a measurement range of 0-25%VOL, and an accuracy of ±0.08%VOL; the carbon monoxide sensor is a semiconductor sensor with an integrated intelligent noise reduction algorithm, which uses the noise reduction formula (where α is preset to 0.3 based on the on-site environment, and n takes the most recent 10 noise samples), improving the measurement accuracy to ±0.8ppm; the methane sensor uses multi-band spectral detection technology, with a measurement range of 0-100%LEL and an accuracy of ±0.8%LEL.

[0033] During the installation and deployment phase, sensor nodes were installed every 20 meters in the tunnels and every 50 square meters in the roof and floor, based on the 3D geological model of the goaf and the mining progress. A professional construction team embedded fiber Bragg grating temperature sensors in 3-5 cm deep grooves in the coal lining and secured with a high-temperature adhesive. Other gas sensors were installed in specially designed explosion-proof protective housings and fixed to the tunnel sidewalls at a height of 1.5-2 meters. The sensors initially collected data at a rate of once per minute. The system monitors the rate of change in data in real time. If the temperature change rate in a particular area exceeds 1°C per 10 minutes, an adaptive adjustment algorithm is used to increase the frequency of temperature sensor data collection to once every 2 minutes. Data transmission utilizes the ZigBee wireless ad hoc networking protocol. Each sensor node is equipped with a low-power processor. After preliminary encoding, the collected data is transmitted via a multi-hop relay to a hub node at the edge of the goaf. From there, it is sent via the fiber optic network to the data processing and analysis module.

[0034] Data Processing and Analysis Module: After data transmitted from the aggregation node enters the preprocessing unit, the temperature data is processed using an adaptive filtering algorithm based on wavelet transform. The sym4 wavelet basis is used for three-layer decomposition, and the threshold is automatically adjusted based on the energy distribution of high-frequency coefficients to effectively filter out noise. Gas concentration data fusion is based on Bayesian theory, using historical monitoring data to determine the prior probability distribution of each sensor's measurement value. The posterior probability is calculated based on real-time data to obtain the fusion result.

[0035] The improved deep neural network self-ignition three-band division model is built based on the TensorFlow framework and contains 5 convolutional layers and 3 fully connected layers. The attention mechanism is introduced through the formula Calculate feature weights, where W a The matrix parameters were optimized using a back-propagation algorithm. The training dataset, comprising 200,000 samples from field monitoring data and laboratory simulations from 10 typical coal mines nationwide, was used. Stochastic gradient descent was used with a learning rate of 0.001, and the model was trained over 50 iterations.

[0036] When calculating the weighted gas concentration change rate, set the oxygen concentration weight β1 = 0.4, the carbon monoxide concentration weight β2 = 0.3, and the methane concentration weight β3 = 0.3. According to the formula The weighted gas concentration change rate of each area is calculated every 10 minutes to assess the spontaneous combustion risk level.

[0037] Liquid CO2 injection control module: The intelligent flow control valve utilizes piezoelectric drive technology, achieving a response time of less than 0.5 seconds. It is installed at a branch node in the goaf's injection pipeline. A fuzzy-neural network control algorithm is implemented using MATLAB. The fuzzy controller inputs are oxygen concentration deviation, temperature deviation, and CO concentration deviation, and the output is a control signal for the valve's opening. The neural network utilizes a three-layer structure, with three nodes in the input layer, ten nodes in the hidden layer, and one node in the output layer. Weights are determined through offline training.

[0038] During injection, the algorithm calculates the injection flow rate Q using the formula Q = f(O2, T, CO) by inputting real-time data on oxygen concentration (O2), temperature (T), and carbon monoxide concentration (CO). When an area is in the oxidation zone and the oxygen concentration exceeds 18%, the injection flow rate in that area is prioritized. In the injection effect evaluation model, ω1 = 0.5, ω2 = 0.3, and ω3 = 0.2 are set. Every 30 minutes of injection, the effect is evaluated using the formula E = ω1·ΔO2 + ω2·ΔT + ω3·ΔCO. If the standard is not met, the injection strategy is adjusted.

[0039] Remote Monitoring and Early Warning Module: The remote monitoring center is located in the mine's ground-level dispatching room. The server uses a Huawei FusionServer Pro series (128GB of memory, a 24-core CPU), running the Windows Server 2019 operating system and proprietary monitoring software. Data is received over a 5G network, and a 3D visualization interface displays the goaf's temperature field, gas concentration distribution, the three spontaneous combustion zones, and the status of the injection pipeline in real time.

[0040] There are three levels of warning thresholds: Level 1: Temperature exceeding 50°C, O2 concentration below 18%, and CO concentration exceeding 100 ppm; Level 2: Temperature exceeding 60°C, O2 concentration below 16%, and CO concentration exceeding 200 ppm; and Level 3: Temperature exceeding 70°C, O2 concentration below 14%, and CO concentration exceeding 300 ppm. When different warning levels are triggered, the system issues yellow, orange, and red audible and visual alarms, respectively, and sends messages via SMS to ten relevant personnel, including the mine manager and safety director. The virtual simulation function, developed on the Unity3D platform, imports a geological model of the goaf and real-time data, allowing users to simulate and adjust injection parameters, ventilation conditions, and other working conditions to assist in decision-making.

[0041] Sensor Fault Diagnosis and Compensation Submodule: The sensor fault diagnosis program is written in Python and deployed on edge computing nodes. Multi-scale feature analysis utilizes the PyWavelets library to perform 5-scale wavelet decomposition and extract energy and entropy features. Anomaly detection is based on the support vector machine algorithm from the Scikit-learn library, with a training set consisting of 100,000 sets of normal data and 20,000 sets of fault simulation data. After a fault occurs, a data compensation algorithm based on spatiotemporal correlation is initiated, determining the weight γ by calculating the spatiotemporal correlation coefficient of adjacent sensors. j , according to the formula Complete data compensation to ensure monitoring continuity.

[0042] Liquid CO2 storage and supply optimization submodule: The liquid CO2 storage tank is equipped with high-precision liquid level, pressure and temperature sensors, collecting data 5 times per minute. The improved thermodynamic model is implemented through C# programming and deployed in the storage tank control system to calculate the pressure in real time. When the pressure exceeds 1.8MPa, the intelligent safety valve is activated according to the formula P open =P base +β·ΔP(P base =1.8MPa, β = 0.2) automatically adjusts the opening pressure. This optimization strategy, based on a genetic algorithm, was developed on a Java platform. The population size was 50, with an initial crossover probability of 0.8 and a mutation probability of 0.1, and adaptive adjustments were made. Liquid carbon dioxide purity was tested weekly. If the purity fell below 98%, the tank of liquid was prioritized for use and replenished promptly.

[0043] Data representation and interpretation

[0044] Comparison items Traditional systems This application system Average time to discover spontaneous combustion hazards (days) 5 1 Average amount of liquid carbon dioxide saved (tons / month) 0 30 Occurrence rate of spontaneous combustion accidents in goaf (times / year) 2 0.3 Average decision response time (minutes) 60 10

[0045] Due to the limitations of sensor performance and backward data processing methods, traditional systems have delayed the discovery of spontaneous combustion hazards, taking an average of 5 days, missing the best opportunity for prevention and control. However, the system of this application, with its high-precision sensors and advanced algorithms, has significantly shortened the time to discover hidden dangers to 1 day, providing sufficient time for early intervention. In the use of liquid carbon dioxide, the traditional system has no intelligent regulation, resulting in a waste of resources. The system of this application saves 30 tons per month through precise control. In terms of safety, the traditional system has an average of 2 spontaneous combustion accidents per year. The system of this application has reduced the accident rate to 0.3 times through real-time monitoring and intelligent regulation. In terms of decision-making and response, the traditional system relies on manual analysis, which takes an average of 60 minutes. The system of this application uses remote intelligent monitoring and virtual simulation to complete decision-making and response within 10 minutes, which greatly improves the ability of coal mines to deal with spontaneous combustion risks and ensures safe production.

[0046] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A system for dynamic monitoring of three zones of spontaneous combustion in coal mine goaf and injection of liquid carbon dioxide, characterized in that: include: Sensor data acquisition module: Temperature sensors, oxygen sensors, and carbon monoxide sensors are installed in the tunnels, roofs, and floor positions of the coal mine goaf. Semiconductor carbon monoxide sensors with integrated noise reduction algorithms are used to collect data through algorithms. Remove the interference signal, where N filtered is the filtered signal, N raw is the original signal, α is the noise reduction coefficient, N noise (i) is the i-th noise sample, which collects carbon monoxide concentration data in real time; The methane sensor uses multi-band spectral detection technology. The sensors are distributed in the goaf with a dynamic grid spacing. The spacing is adjusted in real time according to the mining progress and geological changes in the goaf. The data collection frequency is dynamically set according to the data change rate using an adaptive adjustment algorithm. The collected data is transmitted to the data processing and analysis module via wireless ad hoc network communication. Data processing and analysis module: Receives data transmitted by the sensor data acquisition module, first pre-processes the data, including adaptive filtering, data fusion and feature extraction; for temperature data, an adaptive filtering algorithm based on wavelet transform is used to automatically adjust the filtering parameters to remove noise according to the fluctuation characteristics of the data; for gas concentration data, a data fusion algorithm based on Bayesian theory is used to improve data accuracy by combining prior knowledge and real-time measurement data; then, a deep neural network self-ignition three-zone division model is used to introduce an attention mechanism through the formula Different weights are assigned to different features to highlight the impact of features on the division of the three spontaneous combustion zones. i is the attention weight of the i-th feature, W a is the attention weight matrix, h i is the i-th eigenvector; According to the data change trend, the formula Calculate the weighted change rate of each gas concentration to determine the spontaneous combustion risk level, where ΔC weighted is the weighted gas concentration change rate, β i is the weight of the i-th gas, C t+1(i) and C t(i) are the concentrations of the i-th gas at adjacent moments, and Δt is the time interval.

2. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 1 is characterized in that: Also includes: Liquid carbon dioxide injection control module: determines the location, flow rate and time of liquid carbon dioxide injection based on the three-zone classification results and spontaneous combustion risk level obtained by the data processing and analysis module; A flow control valve is used to control the flow of liquid carbon dioxide. The control valve is set with self-learning and adaptive adjustment functions, and the flow is adjusted by a fuzzy neural network control algorithm to achieve dynamic adjustment of the valve opening according to the oxygen concentration, temperature and carbon monoxide concentration factors. The relationship between the injection flow rate and each factor is mapped by the formula Q = f(O2, T, CO), where Q is the injection flow rate, O2 is the oxygen concentration, T is the temperature, CO is the carbon monoxide concentration, and the function f is obtained through neural network training. In the oxidation zone, the injection flow rate is dynamically adjusted according to the oxygen concentration and temperature through the control algorithm to reduce the oxygen concentration and inhibit coal oxidation and spontaneous combustion. The injection time is determined according to the duration of the spontaneous combustion risk and the injection effect evaluation model. The injection effect evaluation model is evaluated by the formula E = ω1·ΔO2+ω2·ΔT+ω3·ΔCO, where E is the injection effect evaluation value, ω1, ω2, ω3 are weight coefficients, and ΔO2, ΔT, and ΔCO are the changes in oxygen concentration, temperature, and carbon monoxide concentration before and after injection, respectively. Remote monitoring and early warning module: Establish a remote monitoring center based on cloud computing and big data analysis, connect to the data processing and analysis module through the 5G network, and display the temperature, gas concentration, three-zone division of spontaneous combustion, and liquid carbon dioxide injection status of the goaf in real time; set early warning thresholds at different levels. When the data exceeds the early warning thresholds at different levels, the system automatically issues different types of sound and light alarm signals and notifies relevant personnel via SMS, email, and app push; Sensor fault diagnosis and compensation submodule: The module monitors the working status of the sensor in real time, and determines whether the sensor is faulty through feature analysis and anomaly detection algorithms. If the sensor is faulty, it uses a data compensation algorithm based on time-space correlation to compensate for the historical data and the data of adjacent normal sensors. The algorithm is calculated by the formula Combined with the spatiotemporal correlation coefficient to determine the weight to achieve data continuity, where C compensated is the compensated data, γ j is the weight of the jth adjacent sensor, C neighbor(j) is the data of the jth adjacent sensor.

3. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 1 is characterized in that: Also includes: Liquid CO2 storage and supply optimization module: The module monitors the liquid level, pressure and temperature parameters of the liquid CO2 storage tank in real time, and calculates the state of the liquid CO2 in the storage tank through a thermodynamic model; the thermodynamic model takes into account the material properties of the storage tank, ambient temperature changes and the phase change process of CO2, and calculates the state of the liquid CO2 in the storage tank through the formula Predict the pressure change in the storage tank, where P is pressure, n is the amount of substance, R is the gas constant, T is temperature, V is volume, δ is the ambient temperature influence coefficient, T env The ambient temperature is set; according to the demand forecast of the liquid carbon dioxide injection control module, the optimization strategy based on genetic algorithm is adopted to optimize the storage and supply strategy of liquid carbon dioxide in combination with the supply cost, transportation distance and storage capacity factors to arrange the replenishment and distribution of liquid carbon dioxide; at the same time, a safety valve and a pressure regulating device are set to ensure the safety of the storage tank. The safety valve automatically adjusts the opening pressure according to the pressure change in the storage tank. The formula P open =P base +β·ΔP for dynamic adjustment, where P open is the opening pressure, P base is the basic opening pressure, β is the pressure adjustment coefficient, and ΔP is the pressure change.

4. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 1 is characterized in that: The sensors in the sensor data acquisition module adopt a modular design, which is easy to install, maintain and replace. The installation position of the sensor is optimized and determined through three-dimensional geological modeling and numerical simulation technology. According to the geological conditions, ventilation conditions and coal spontaneous combustion characteristics of the goaf, the monitoring effects of the sensors at different installation positions are simulated, and the optimal installation plan is selected to accurately collect data reflecting the spontaneous combustion status of the goaf.

5. The system for dynamic monitoring of three zones of spontaneous combustion in coal mine goaf and injection of liquid carbon dioxide according to claim 1 is characterized in that: The deep neural network spontaneous combustion three-zone division model in the data processing and analysis module adopts transfer learning and incremental learning technology; at the same time, the model adopts model fusion technology to fuse neural network models of different structures, and obtains the final division result through weighted averaging, thereby improving the accuracy of the spontaneous combustion three-zone division.

6. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 2 is characterized in that: The flow control valve in the liquid carbon dioxide injection control module adopts piezoelectric drive technology, and the control algorithm of the control valve adopts a model-predictive control method. By establishing a dynamic model of the injection process, the injection effect in the future is predicted, and the opening of the control valve is adjusted in real time according to the prediction results. At the same time, the module is also equipped with a flow feedback adjustment function. The injection flow is monitored in real time by a flow sensor installed on the injection pipeline and compared with the set flow, and error correction is performed through the PID control algorithm.

7. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 2 is characterized in that: The remote monitoring and early warning module is equipped with a decision-making assistance function. It integrates and associates the geological information, mining process, and safety specification knowledge of the coal mine goaf through knowledge graph technology. When abnormalities occur in the monitoring data, the system performs reasoning and analysis based on the knowledge graph to provide decision-making suggestions for relevant personnel. At the same time, the module is also equipped with a visual interaction function. Users can interact with the virtual simulation model through a touch screen or mouse device to adjust the working parameters in real time.

8. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 2 is characterized in that: The feature analysis and anomaly detection algorithms in the sensor fault diagnosis and compensation module adopt a distributed computing architecture to distribute data processing tasks to computing nodes for parallel processing. At the same time, the module is also equipped with a fault prediction function. By analyzing the historical fault data and current working status data of the sensor, a time series analysis algorithm is used to predict the time and type of sensor failure so that maintenance and replacement can be carried out in advance.

9. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 3 is characterized in that: The genetic algorithm-based optimization strategy in the liquid carbon dioxide storage and supply optimization module uses an adaptive genetic operator to dynamically adjust the crossover probability and mutation probability based on the fitness value during the optimization process, thereby improving the algorithm's search efficiency and convergence speed. The module also takes into account the quality changes of liquid carbon dioxide, regularly testing the purity and impurity content of liquid carbon dioxide and adjusting the supply strategy based on the test results.

10. The system for dynamic monitoring of three zones of spontaneous combustion and injection of liquid carbon dioxide in coal mine goaf according to claim 1, characterized in that: It also includes an interface module for integration with the existing safety monitoring system and automation control system of the coal mine; the interface module adopts standardized data communication protocols and interface specifications to realize data transmission and real-time sharing; through data fusion, this system utilizes the resources and data of the existing system of the coal mine, and at the same time feeds back the monitoring data and control information of this system to the existing system, realizing the integrated management of coal mine safety monitoring and control.

Images