A gas outburst monitoring and early warning method and system based on an improved ConvLSTM model

Abstract

The application discloses a gas outburst monitoring and early warning method and system based on an improved ConvLSTM model, relates to the technical field of gas monitoring, and has the technical scheme as follows: a mine gas data set is constructed; the gas data set is cleaned, unstructured data is converted into structured data, and the structured data is subjected to data cleaning, denoising and standardization; high-order derived feature extraction is performed on the pretreated gas data set; an improved ConvLSTM model is constructed; the improved ConvLSTM model is trained using a historical gas data set; potential high-risk gas areas are identified; based on the analysis result of the improved ConvLSTM model, potential gas outburst and fire risk areas are identified; and a monitoring platform is constructed. Through the improved ConvLSTM model, the application realizes efficient collection and processing of gas data, improves the accuracy and real-time performance of monitoring, and provides strong support for mine safety management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of gas monitoring, more particularly, it relates to a gas outburst monitoring and early warning method and system based on an improved ConvLSTM model. BACKGROUND

[0002] Gas outburst monitoring and early warning has become an important part of mine safety management. The existing gas outburst monitoring and early warning methods mainly rely on traditional sensor data collection and simple data analysis techniques, which have many limitations and cannot meet the high requirements of real-time and accuracy in modern mining.

[0003] Firstly, the existing gas outburst monitoring and early warning methods lack accuracy in data processing and risk assessment. Traditional methods usually use basic data analysis and statistical methods, lack effective intelligent algorithm support, and cannot fully capture the data characteristics in complex mine environments, resulting in inaccurate monitoring results. These methods often rely on a single sensor or limited sensor networks, and the breadth and depth of data collection are insufficient, which cannot fully reflect the safety situation of the mine. At the same time, the data processing efficiency is low, and accurate risk assessment results cannot be obtained quickly under large-scale data, affecting the real-time and effectiveness of monitoring.

[0004] Secondly, the existing monitoring and early warning system also has problems in data collection frequency and layout design. The sensor arrangement of traditional systems is often not reasonable, resulting in insufficient comprehensive data collection, affecting the accuracy of subsequent analysis and risk assessment. Due to the low data collection frequency, the system cannot obtain the latest mine safety data in time, resulting in monitoring lag. In addition, there is a lot of noise and inconsistency in the original data, which has not been effectively preprocessed, resulting in low data quality, further affecting the reliability of the monitoring results.

[0005] To solve these problems, the improved ConvLSTM model is applied to gas outburst detection to provide a new solution. Although the existing particle swarm algorithm performs well in optimization problems, its efficiency and accuracy still need to be improved when dealing with complex gas outburst monitoring and early warning data. The traditional particle swarm algorithm faces the challenge of high-dimensional data calculation when dealing with complex gas outburst early warning data, and is prone to local optimum, affecting the overall monitoring effect.

[0006] Therefore, the present application aims to provide a gas outburst monitoring and early warning method and system based on an improved ConvLSTM model to solve the above problems. SUMMARY

[0007] The purpose of the present application is to provide a gas outburst monitoring and early warning method and system based on an improved ConvLSTM model, which realizes efficient data collection and processing and improves the accuracy and real-time performance of monitoring by improving the ConvLSTM model, thereby providing strong support for mine safety production.

[0008] The above technical purpose of the present application is realized by the following technical scheme: a gas outburst monitoring and early warning method based on an improved ConvLSTM model, comprising the following steps:

[0009] S1, arranging various types of sensors in key areas of the mine to collect gas concentration, rock vibration signals, temperature data, humidity data and airflow pressure, and constructing a gas data set;

[0010] S2, preprocessing the gas data set by converting unstructured data into structured data, and then performing data cleaning, denoising and standardization on the structured data, and retaining the node position information of each data in the gas data set;

[0011] S3, performing high-order derived feature extraction on the preprocessed gas data set to extract time domain features, frequency domain features and time-frequency domain features of the gas data set;

[0012] S4, constructing an improved ConvLSTM model according to the time domain features, frequency domain features and time-frequency domain features, and combining the mine geographical position and adjacent relationship of the sensors, the improved ConvLSTM model aims to monitor five levels of gas risk;

[0013] S5, training the improved ConvLSTM model using historical gas data sets;

[0014] S6, inputting the preprocessed gas data set into the trained improved ConvLSTM model for analysis, the improved ConvLSTM model monitors the gas outburst risk according to the input data and generates an analysis result;

[0015] S7, based on the analysis result of the improved ConvLSTM model, identifying potential gas risk areas, evaluating the risk level of the identified risk areas, and generating corresponding early warning information, when a high-risk area is detected, timely issuing an alarm to notify relevant personnel and systems to take preventive measures, and when a data trend in a preset range is monitored, taking corresponding monitoring feedback measures;

[0016] S8, based on the monitoring feedback information and the monitoring and early warning system, constructing a monitoring platform and providing a visual interface to enable management personnel to view the current monitoring state, historical data trends and early warning information.

[0017] The application is further provided with: the method for constructing the gas data set in step S1 comprises the following steps:

[0018] S11, arranging a gas concentration sensor, a rock mass vibration signal sensor, a temperature sensor, a humidity sensor and an airflow pressure sensor in a key area of the mine, respectively, for collecting gas concentration G(t), rock mass vibration signal S(t), temperature data T(t), humidity data H(t) and airflow pressure A(t);

[0019] S12, collecting and summarizing the data to construct a gas data set D:

[0020] D={(t, G(t), S(t), T(t), H(t), A(t), L)};

[0021] Wherein, t represents the collection time, and L represents the geographical position information of the sensor.

[0022] The application is further provided with: the method for preprocessing in step S2 comprises the following steps:

[0023] S21, performing preliminary cleaning on the gas data set D to convert unstructured data into structured data;

[0024] S22, performing data cleaning on the structured gas data set to remove abnormal values and incomplete data points:

[0025] If |d i -μ D |>σ D , d final is an abnormal value and needs to be removed;

[0026] Wherein, d norm is a single data point in the gas data set, μ norm is the mean of the gas data set D, and σ norm is the standard deviation of the gas data set D;

[0027] S23, performing denoising processing on the cleaned gas data set, removing noise components in the data by using a filtering algorithm to obtain smooth mine environment data;

[0028] S24, performing standardization processing on the denoised gas data set to convert data of different dimensions into the same dimension range;

[0029] S25, retaining the node position information L of each data point in the gas data set, so that the consistency and traceability of the data in the geographical space are ensured:

[0030] D norm ={(t, P norm(t), S norm (t), T norm (t), H norm (t), A norm (t), L)};

[0031] wherein, D final is the final gas data set after retaining node position information, P norm (t), S norm (t), T norm (t), H norm (t), A norm (t) are the normalized gas concentration, rock vibration signal, temperature data, humidity data and air flow pressure respectively.

[0032] The application further provides that: the method for combining the mine geographical position and the adjacent relationship of the sensors according to the time domain features, the frequency domain features and the time-frequency domain features in the step S4 comprises the following steps.

[0033] S41, according to the time domain features, the frequency domain features and the time-frequency domain features of the gas data set D final , the corresponding feature matrix F is extracted:

[0034] F={F t , F f , F tf};

[0035] wherein, F t represents the time domain feature matrix, F f represents the frequency domain feature matrix, and F tf represents the time-frequency domain feature matrix;

[0036] S42, the geographical position matrix G is constructed by combining the mine geographical position L and the adjacent relationship of the sensors, and is used for describing the spatial relationship between the sensors:

[0037] G={g ij};

[0038] wherein, g ij represents the geographical adjacent relationship between the sensors i and j, if the sensors i and j are adjacent, g ij =1, otherwise g ij =0;

[0039] S43, the improved ConvLSTM model is constructed based on the feature matrix F and the geographical position matrix G, and the model comprises a convolution layer, a graph convolution network layer, a long short-term memory network layer and a self-attention mechanism layer.

[0040] The application further provides that: the method for constructing the improved ConvLSTM model in the step S43 comprises the following steps:

[0041] S431、In the convolutional layer, a multi-scale convolutional kernel structure is adopted to capture the features of gas concentration G(t), rock vibration signal S(t), temperature data T(t), humidity data H(t) and air flow pressure A(t) at different spatial and temporal scales:

[0042]

[0043] wherein, represents the output feature value of the lth layer convolutional layer at position (x, y), represents the (m, n)th weight of the 1st layer convolutional kernel, b (l) represents the bias of the lth layer, σ represents the activation function, D final (x+m, y+n) represents the value of the normalized gas data set at position (x+m, y+n), φ(m, n) represents a newly introduced multi-scale convolutional kernel adjustment function for adjusting feature extraction at different scales to capture multi-scale features related to gas outburst;

[0044] S432, introduce an improved graph convolutional network to fuse the geographical position matrix G with the feature matrix, capture the spatial relationship and correlation between sensors by introducing a new node relationship modeling method, and the graph convolutional operation is represented as:

[0045]

[0046] wherein, represents the output feature of the kth layer graph convolutional network at node v, represents the neighbor node set of node v, c vu represents the normalization constant, G vu represents the weight between nodes v and u in the geographical position matrix, represents the output feature of the k-1th layer convolutional network at node u, represents the weight matrix of the kth layer graph convolutional network, represents the bias of the kth layer graph convolutional network, ψ(v, u) represents a new node relationship modeling function for modeling the spatial relationship between sensor nodes and optimizing the spatial correlation analysis of gas outburst monitoring;

[0047] S433, after the graph convolutional network layer, an adaptive time-frequency analysis mechanism is introduced to dynamically extract and weight the features at different time periods and frequency ranges to generate comprehensive time-frequency features:

[0048]

[0049] wherein, F tf_adaptive (t, f) represents the adaptive time-frequency feature at time t and frequency f, ατv (t, f) denotes the dynamic weight on time t and frequency f, F gcn (τ, v) represents the output feature of the graph convolution network at time τ and frequency v, γ(τ, v) represents a new time-frequency weighting function, which dynamically adjusts the weight of different time and frequency features, and optimizes the sensitivity of the improved ConvLSTM model to the gas outburst features at different time-frequency distributions;

[0050] S434, in the long short-term memory network layer, the feature matrix processed by the graph convolution network is combined to model the spatial relationship between the sensors:

[0051]

[0052] wherein, represents the output feature of the hth long short-term memory network at time t, represents the weight matrix input to the hidden layer, represents the weight matrix from the hidden layer to the hidden layer, H( h - 1 ) represents the value of the (h-1)th hidden state at time t-1, b( h ) represents the bias of the hth layer, and λ(t) represents a time adjustment function.

[0053] S435, the self-attention mechanism is introduced, the features at different time steps are calculated by weighting, the time features are focused, and the prediction ability of the improved ConvLSTM model to the key moment of gas outburst ignition is optimized:

[0054]

[0055] wherein, F attention (t) represents the attention weighted feature at time t, β i (t) represents the attention weight at time step i, F lstm (i) represents the output feature of the long short-term memory network at time step i, and θ(i) represents an attention weighting function for focusing on the features at key time steps.

[0056] S436, according to the output of the position-aware long short-term memory network layer and the self-attention mechanism layer, a multilayer perceptron is used to predict the gas outburst risk level, and five levels of gas outburst risk levels R are defined, which are no risk r1, low risk r2, medium risk r3, medium-high risk r4 and high risk r5.

[0057] The application further provides that: the method for identifying the potential gas risk area based on the analysis result of the improved ConvLSTM model in step S7 comprises the following steps:

[0058] S71. Based on the analysis results of the improved ConvLSTM model, potential gas outburst risk areas are identified, and the model outputs the gas outburst risk score matrix R score :

[0059] R score ={r ij};

[0060] Among them, r ij represents the gas outburst risk score at location (i, j);

[0061] S72. Evaluate the identified risk areas and determine their risk level R level The risk levels are classified as follows:

[0062]

[0063] Among them, θ1, θ2, θ3, and θ4 are thresholds for different risk levels;

[0064] S73, according to the risk level R level Generate corresponding warning information W, which includes risk level, specific location and recommended measures;

[0065] S74. When a high-risk area r5 is detected, the system promptly issues an alarm A, notifies relevant personnel and systems, and takes preventive measures;

[0066] S75, when the data trend T within the preset range is monitored trend When the system takes corresponding monitoring feedback measures, the data trend monitoring formula is:

[0067] T trend ={t|f(t)>φ};

[0068] Where f(t) represents the data trend function, φ represents the preset trend threshold, and when the trend function f(t) exceeds the threshold φ, corresponding monitoring feedback measures are taken;

[0069] S76. Record the warning information W and alarm A in the system log.

[0070] The present invention also provides a gas outburst monitoring and early warning system based on an improved ConvLSTM model, which is characterized by comprising a data acquisition module, a data preprocessing module, a feature extraction module, an improved ConvLSTM model module, a model training module, a risk assessment and early warning module, and a monitoring platform module;

[0071] The data acquisition module is used to collect gas concentration, rock mass vibration signals, temperature data, humidity data and air flow pressure to generate a gas data set;

[0072] The data preprocessing module is used for cleaning the collected gas data set, converting unstructured data into structured data, and then performing data cleaning, denoising and standardization on the structured data, and retaining the node position information of each data of the gas data set;

[0073] The feature extraction module is used for high-order derived feature extraction on the preprocessed gas data set, extracts time domain features, frequency domain features and time-frequency domain features of the gas data set, and constructs a feature matrix and a geographical position matrix in combination with the mine geographical position and adjacent relationship of the sensor;

[0074] The improved ConvLSTM model module is used for constructing an improved ConvLSTM model based on the feature matrix and the geographical position matrix, wherein the model includes a convolutional layer, a graph convolutional network layer, a long short-term memory network layer and a self-attention mechanism layer, and is used for monitoring five levels of gas outburst risk;

[0075] The model training module is used for training the improved ConvLSTM model by using historical gas data set, optimizing the model parameters, so that the model can accurately predict the risk level of gas outburst;

[0076] The risk assessment and early warning module is used for identifying potential gas outburst risk areas based on the analysis result of the improved ConvLSTM model, evaluating the risk level of the identified risk area, and generating corresponding early warning information, when a high-risk area is detected, timely issuing an alarm to notify relevant personnel and system to take preventive measures, and when a data trend in a preset range is monitored, taking corresponding monitoring feedback measures;

[0077] The monitoring platform module provides a visual interface for managers to view the current monitoring state, historical data trend and early warning information in real time, and records the early warning information and alarm information.

[0078] In summary, the present application has the following advantages:

[0079] 1、The improved ConvLSTM model of the application, the convolutional layer adopts a multi-scale convolution kernel structure, which can capture the characteristics of gas concentration, rock vibration signal, temperature data, humidity data and air flow pressure at different spatial and temporal scales, and improve the feature extraction capability of complex gas outburst monitoring and early warning data. The improved graph convolution network is introduced to fuse the geographical position matrix and the feature matrix, and the spatial relationship and correlation between sensors are captured through a new node relationship modeling method, which optimizes the analysis capability of the model for the spatial relationship between sensor nodes. After the graph convolution network layer, an adaptive time-frequency analysis mechanism is introduced to dynamically extract and weight the features of different time periods and frequency ranges, generate comprehensive time-frequency features, and improve the sensitivity and prediction accuracy of the model for gas outburst features at different time-frequency distributions. The self-attention mechanism is introduced, the features of different time steps are calculated by weighting, and important time features are focused on, which optimizes the gas outburst prediction ability of the model at key moments. This mechanism significantly improves the attention and recognition ability of the model to key features, enhances the timeliness and accuracy of the warning, and by combining convolutional neural network, graph convolution network, long short-term memory network and self-attention mechanism, a multi-level, multi-scale and multi-dimensional comprehensive gas outburst monitoring model is constructed. Not only improves the processing ability and analysis accuracy of complex gas outburst monitoring and early warning data, but also enhances the robustness and stability of the model, which can realize efficient and accurate risk assessment under large-scale data;

[0080] 2、Based on the analysis results of the improved ConvLSTM model, the system can identify potential gas outburst risk areas in real time, evaluate the risk level of the identified risk areas, and generate corresponding warning information. When a high-risk area is detected, the system timely issues an alarm to inform relevant personnel and systems to take preventive measures to effectively prevent disasters. The system builds an integrated monitoring platform to provide a visual interface, allowing management personnel to view the current monitoring status, historical data trends and warning information in real time. By recording and analyzing warning information and alarm information, management personnel can make effective decisions and interventions to improve the overall level of mine safety management. BRIEF DESCRIPTION OF DRAWINGS

[0081] Fig. 1 is a flowchart of a gas outburst monitoring and early warning method based on an improved ConvLSTM model in embodiment 1 of the application;

[0082] Fig. 2 is a structure diagram of the improved ConvLSTM model of the gas outburst monitoring and early warning method based on the improved ConvLSTM model in embodiment 1 of the application. DETAILED DESCRIPTION

[0083] The following will be described in conjunction with the accompanying Figs. 1-2Further details of the present application are provided below.

[0084] Embodiment 1: A gas outburst monitoring and early warning method based on an improved ConvLSTM model, comprising the following steps:

[0085] S1, arranging multiple types of sensors in key areas of the mine to collect gas concentration, rock vibration signal, temperature data, humidity data and airflow pressure, and constructing a gas data set.

[0086] The method for constructing the gas data set in step S1 of this embodiment comprises the following steps:

[0087] S11, arranging gas concentration sensors, rock vibration signal sensors, temperature sensors, humidity sensors and airflow pressure sensors in key areas of the mine, respectively for collecting gas concentration G(t), rock vibration signal S(t), temperature data T(t), humidity data H(t) and airflow pressure A(t);

[0088] S12, aggregating the collected data to construct a gas data set D:

[0089] D={(t,G(t),S(t),T(t),H(t),A(t),L)};

[0090] Wherein, t represents the collection time, and L represents the geographical position information of the sensor.

[0091] S2, preprocessing the gas data set, converting unstructured data into structured data, and then performing data cleaning, denoising and standardization on the structured data, and retaining the node position information of each data of the gas data set.

[0092] The preprocessing method in step S2 of this embodiment comprises the following steps:

[0093] S21, preliminarily cleaning the gas data set D to convert unstructured data into structured data;

[0094] S22, performing data cleaning on the structured data gas data set to remove abnormal values and incomplete data points:

[0095] If |d i -μ D |>σ D , d i is an abnormal value and needs to be removed;

[0096] Wherein, d i is a single data point in the gas data set, μ D is the mean of the gas data set D, and σ D is the standard deviation of the gas data set D.

[0097] S23, the cleaned gas data set is denoised, and a filtering algorithm is used to remove noise components in the data to obtain smooth mine environment data;

[0098] S24, the denoised gas data set is standardized, and data of different dimensions are converted to the same dimension range;

[0099] S25, the node position information L of each data point in the gas data set is retained, so that the consistency and traceability of the data in geographical space:

[0100] D final = {(t, P norm (t), S norm (t), T norm (t), H norm (t), A norm (t), L)};

[0101] Wherein, D final is the final gas data set after retaining node position information, P norm (t), S norm (t), T norm (t), H norm (t), A norm (t) are the standardized gas concentration, rock vibration signal, temperature data, humidity data and air flow pressure respectively.

[0102] S3, the preprocessed gas data set is subjected to high-order derived feature extraction, and time domain features, frequency domain features and time-frequency domain features of the gas data set are extracted.

[0103] S4, according to the time domain features, frequency domain features and time-frequency domain features, combined with the mine geographical position and adjacent relationship of the sensor, an improved ConvLSTM model is constructed, and the improved ConvLSTM model is used to monitor five levels of gas risk.

[0104] The method of the present embodiment in step S4 according to the time domain features, frequency domain features and time-frequency domain features, combined with the mine geographical position and adjacent relationship of the sensor, includes the following steps:

[0105] S41, according to the time domain features, frequency domain features and time-frequency domain features of the gas data set D final , the corresponding feature matrix F is extracted:

[0106] F = {F t , F f , F tf};

[0107] Wherein, F tdenotes a time domain feature matrix, F f denotes a frequency domain feature matrix, F tf denotes a time-frequency domain feature matrix;

[0108] S42, in combination with the geographical position L and the adjacent relationship of the sensor, a geographical position matrix G is constructed to describe the spatial relationship between the sensors:

[0109] G={gi j};

[0110] wherein, g ij denotes the geographical adjacent relationship between sensors i and j, if sensors i and j are adjacent, g ij =1, otherwise g ij =0;

[0111] S43, based on the feature matrix F and the geographical position matrix G, an improved ConvLSTM model is constructed, which includes a convolutional layer, a graph convolutional network layer, a long short-term memory network layer and a self-attention mechanism layer.

[0112] The method for constructing the improved ConvLSTM model in step S43 includes the following steps:

[0113] S431, in the convolutional layer, a multi-scale convolution kernel structure is adopted to capture the features of gas concentration G(t), rock vibration signal S(t), temperature data T(t), humidity data H(t) and air flow pressure A(t) at different spatial and temporal scales:

[0114]

[0115] wherein, denotes the output feature value of the lth layer convolutional layer at position (x, y), denotes the (m, n)th weight of the lth layer convolutional kernel, b (l) denotes the bias of the lth layer, σ denotes the activation function, D final (x+m, y+n) denotes the value of the normalized gas data set at position (x+m, y+n), φ(m, n) denotes a newly introduced multi-scale convolution kernel adjustment function for adjusting feature extraction at different scales to capture multi-scale features related to gas outburst;

[0116] S432, an improved graph convolutional network is introduced to fuse the geographical position matrix G and the feature matrix, and a new node relationship modeling method is introduced to capture the spatial relationship and correlation between the sensors, and the graph convolutional operation is represented as:

[0117]

[0118] wherein, denotes the output feature of the k-th layer graph convolutional network at node v, denotes the set of neighbor nodes of node v, c vu denotes the normalization constant, G vu denotes the weight between nodes v and u in the geographical position matrix, denotes the output feature of the k-1-th layer convolutional network at node u, denotes the weight matrix of the k-th layer graph convolutional network, denotes the bias of the k-th layer graph convolutional network, and ψ(v, u) denotes a new node relationship modeling function for modeling the spatial relationship between sensor nodes, optimizing the spatial correlation analysis of gas outburst monitoring;

[0119] S433, after the graph convolutional network layer, an adaptive time-frequency analysis mechanism is introduced to dynamically extract and weight the features of different time periods and frequency ranges, and generate comprehensive time-frequency features:

[0120]

[0121] where F tf_adaptive (t, f) denotes the adaptive time-frequency feature at time t and frequency f, α τv (t, f) denotes the dynamic weight at time τ and frequency v, F gcn (τ, v) denotes the output feature of the graph convolutional network at time τ and frequency v, and γ(τ, v) denotes a new time-frequency weighting function for dynamically adjusting the weight of different time and frequency features, optimizing the sensitivity of the improved ConvLSTM model to the distribution of gas outburst features at different time and frequency;

[0122] S434, in the long short-term memory network layer, the feature matrix processed by the graph convolutional network is combined to model the spatial relationship between sensors:

[0123]

[0124] where, denotes the output feature of the h-th layer long short-term memory network at time t, denotes the weight matrix input to the hidden layer, denotes the weight matrix from the hidden layer to the hidden layer, H (h-1) denotes the value of the h-1-th layer hidden state at time t-1, b (h) denotes the bias of the h-th layer, and λ(t) denotes a time adjustment function;

[0125] S435, a self-attention mechanism is introduced to focus on the time features by weighted calculation of the features at different time steps, optimizing the prediction ability of the improved ConvLSTM model for gas outburst ignition at key moments:

[0126]

[0127] where F attention (t) represents the attention weighted feature at time t, β i (t) represents the attention weight at time step i, F lstm (i) represents the output feature of the long short-term memory network at time step i, and θ(i) represents the attention weighting function for focusing on the features of the key time steps.

[0128] S436, according to the outputs of the position-aware long short-term memory network layer and the self-attention mechanism layer, combining the multi-layer perceptron to predict the gas outburst risk level, defining five levels of gas outburst risk level R, respectively, no risk r1, low risk r2, medium risk r3, medium-high risk r4 and high risk r5.

[0129] S5, using the historical gas data set to train the improved ConvLSTM model.

[0130] S6, input the preprocessed gas data set into the trained improved ConvLSTM model for analysis, and the improved ConvLSTM model monitors the gas outburst risk according to the input data to generate an analysis result.

[0131] S7, based on the analysis result of the improved ConvLSTM model, identify the potential gas risk area, evaluate the risk level of the identified risk area, and generate corresponding warning information, when a high-risk area is detected, timely alarm, notify relevant personnel and system, take preventive measures, and when the data trend in the preset range is monitored, take corresponding monitoring feedback measures.

[0132] The method for identifying the potential gas risk area in step S7 based on the analysis result of the improved ConvLSTM model includes the following steps:

[0133] S71, based on the analysis result of the improved ConvLSTM model, identify the potential gas outburst risk area, and the model outputs a gas outburst risk score matrix R score :

[0134] R score ={r ij};

[0135] where r ij represents the gas outburst risk score at position (i, j);

[0136] S72, evaluate the identified risk area to determine its risk level R level , the risk level is divided according to the following:

[0137]

[0138] Wherein θ1, θ2, θ3, θ4 are the threshold values of different risk levels respectively;

[0139] S73, according to the risk level R level Generate corresponding warning information W, which contains risk level, specific location and suggested measures;

[0140] S74, when detecting a high-risk area r5, the system timely issues an alarm A, notifies relevant personnel and systems, and takes preventive measures;

[0141] S75, when monitoring the data trend T trend of the preset range, the system takes corresponding monitoring feedback measures, and the data trend monitoring formula is:

[0142] T trend ={t|f(t)>φ};

[0143] Wherein f(t) represents the data trend function, and φ represents the preset trend threshold value. When the trend function f(t) exceeds the threshold value φ, corresponding monitoring feedback measures are taken;

[0144] S76, record the warning information W and the alarm A in the system log.

[0145] S8, based on the monitoring feedback information and the monitoring and warning system, a monitoring platform is constructed, and a visual interface is provided to enable management personnel to view the current monitoring state, historical data trend and warning information.

[0146] Embodiment 2: A gas outburst monitoring and warning system based on an improved ConvLSTM model, comprising a data acquisition module, a data preprocessing module, a feature extraction module, an improved ConvLSTM model module, a model training module, a risk assessment and warning module and a monitoring platform module.

[0147] The data acquisition module in this embodiment is used to acquire gas concentration, rock vibration signal, temperature data, humidity data and air flow pressure, and generate a gas data set.

[0148] The data preprocessing module is used to clean the collected gas data set, convert unstructured data into structured data, and then clean, denoise and standardize the structured data, and retain the node position information of each data in the gas data set.

[0149] The feature extraction module is used for high-order derived feature extraction on the preprocessed gas data set, extracts the time domain features, frequency domain features and time-frequency domain features of the gas data set, and combines the mine geographical position and adjacent relationship of the sensor to construct a feature matrix and a geographical position matrix.

[0150] The improved ConvLSTM model module is used for constructing an improved ConvLSTM model based on the feature matrix and the geographical position matrix, wherein the model comprises a convolution layer, a graph convolution network layer, a long short-term memory network layer and a self-attention mechanism layer, and is used for monitoring the gas outburst risk of five levels.

[0151] The model training module is used for training the improved ConvLSTM model by using the historical gas data set, and optimizing the model parameters so that the model can accurately predict the risk level of gas outburst.

[0152] The risk assessment and early warning module is used for identifying potential gas outburst risk areas based on the analysis result of the improved ConvLSTM model, evaluating the risk level of the identified risk areas, and generating corresponding early warning information, and when a high-risk area is detected, timely issuing an alarm to inform relevant personnel and systems to take preventive measures, and when a preset range of data trends is monitored, taking corresponding monitoring feedback measures.

[0153] The monitoring platform module is used for real-time viewing of the current monitoring state, historical data trends and early warning information by the management personnel through the provision of a visual interface, and records the early warning information and alarm information.

[0154] The specific embodiment is only an explanation of the present application, and is not a limitation of the present application. Those skilled in the art can make modifications to the embodiment without creative contribution according to the needs after reading the specification, but as long as the modifications are within the scope of the claims of the present application, they are protected by the patent law.

Claims

1. A gas outburst monitoring and early warning method based on an improved ConvLSTM model, characterized by: The following steps are involved: S1. Deploy various sensors in key areas of the mine to collect gas concentration, rock vibration signals, temperature data, humidity data, and airflow pressure to construct a gas data set. S2. Clean and preprocess the gas dataset, converting unstructured data into structured data. Then, clean, denoise, and standardize the structured data, preserving the node location information of each data point in the gas dataset. S3. performing high-order derived feature extraction on the preprocessed gas data set to extract time domain features, frequency domain features, and time-frequency domain features of the gas data set; S4. Based on the time domain features, frequency domain features, and time-frequency domain features, combined with the mine geographical location and adjacent relationships of the sensors, an improved ConvLSTM model is constructed. The improved ConvLSTM model aims to monitor five levels of gas risk. S5. Use the historical gas dataset to train the improved ConvLSTM model; S6. Input the preprocessed gas data set into the trained improved ConvLSTM model for analysis. The improved ConvLSTM model monitors the gas outburst risk based on the input data and generates analysis results. S7. Based on the analysis results of the improved ConvLSTM model, potential gas risk areas are identified, the risk levels of the identified risk areas are assessed, and corresponding early warning information is generated. When high-risk areas are detected, an alarm is promptly issued to notify relevant personnel and systems, and preventive measures are taken. When data trends within a preset range are monitored, corresponding monitoring feedback measures are taken; S8. Based on monitoring feedback information and monitoring early warning system, build a monitoring platform and provide a visual interface to enable managers to view current monitoring status, historical data trends and early warning information; The method of combining the mine geographical location and the neighboring relationship of the sensor in step S4 according to the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics includes the following steps: S41. According to the gas data set D final The time domain features, frequency domain features and time-frequency domain features are extracted to extract the corresponding feature matrix F: F={F t ,F f ,F tf }; Among them, F t represents the time domain feature matrix, F f Represents the frequency domain feature matrix, F tf Represents the time-frequency domain feature matrix; S42. Combine the mine geographical location L and the adjacent relationship of the sensors to construct a geographical location matrix G to describe the spatial relationship between the sensors: G={g ij }; Among them, g ij represents the geographical proximity relationship between sensors i and j. If sensors i and j are adjacent, then g ij =1, otherwise g ij =0; S43. Based on the feature matrix F and the geographic location matrix G, an improved ConvLSTM model is constructed. The model includes a convolutional layer, a graph convolutional network layer, a long short-term memory network layer, and a self-attention mechanism layer.

2. The gas outburst monitoring and early warning method based on the improved ConvLSTM model according to claim 1 is characterized by: The method for constructing the gas dataset in step S1 includes the following steps: S11. Deploy gas concentration sensors, rock mass vibration signals, temperature sensors, humidity sensors, and airflow pressure sensors in key areas of the mine to collect gas concentration G(t), rock mass vibration signals S(t), temperature data T(t), humidity data H(t), and airflow pressure A(t), respectively. S12. Summarize the collected data to construct a gas dataset D: D={(t,G(t),S(t),T(t),H(t),A(t),L)}; Where t represents the acquisition time, and L represents the geographic location information of the sensor.

3. The gas outburst monitoring and early warning method based on the improved ConvLSTM model according to claim 2 is characterized by: The method for pre-processing in step S2 comprises the following steps: S21. Perform preliminary cleaning on the gas dataset D to convert unstructured data into structured data; S22. Clean the structured gas data set to remove outliers and incomplete data points: If |d i -μ D |>σ D , then d i It is an outlier and needs to be removed; Among them, d i is a single data point in the gas dataset, μ D is the mean of the gas data set D, σ D is the standard deviation of the gas data set D; S23, performing denoising processing on the cleaned gas data set, using a filtering algorithm to remove noise components in the data, and obtaining smooth mine environment data; S24, performing standardization processing on the denoised gas data set to convert data of different dimensions into the same dimension range; S25. Retain the node location information L of each data point in the gas data set to ensure consistency and traceability of the data in geographic space: D final ={(t,P norm (t),S norm (t),T norm (t),H norm (t),A norm (t),L)}; Among them, D final To preserve the final gas data set after node location information, P norm (t),S norm (t),T norm (t),H norm (t),A norm (t) are the standardized gas concentration, rock vibration signal, temperature data, humidity data and airflow pressure, respectively.

4. The gas outburst monitoring and early warning method based on the improved ConvLSTM model according to claim 3 is characterized by: The method for constructing the improved ConvLSTM model in step S43 includes the following steps: S431. In the convolution layer, a multi-scale convolution kernel structure is used to capture the characteristics of gas concentration G(t), rock vibration signal S(t), temperature data T(t), humidity data H(t), and airflow pressure A(t) at different spatial and temporal scales: in, Represents the output feature value of the lth convolutional layer at position (x, y), represents the (m,n)th weight of the lth convolution kernel, b (l) represents the bias of the lth layer, σ represents the activation function, D final (x+m, y+n) represents the value of the standardized gas dataset at position (x+m, y+n), and φ(m,n) represents the newly introduced multi-scale convolution kernel adjustment function, which is used to adjust feature extraction at different scales and capture multi-scale features related to gas prominence. S432. Introduce an improved graph convolutional network to fuse the geographic location matrix G with the feature matrix. By introducing a new node relationship modeling method, the spatial relationship and correlation between sensors are captured. The graph convolution operation is expressed as: in, represents the output feature of the k-th layer graph convolutional network at node v, represents the set of neighbor nodes of node v, c vu represents the normalization constant, G vu represents the weight between nodes v and u in the geographic location matrix, represents the output feature of the k-1th layer convolutional network at node u, represents the weight matrix of the k-th layer graph convolutional network, represents the bias of the k-th layer graph convolutional network, and ψ(v,u) represents a new node relationship modeling function, which is used to model the spatial relationship between sensor nodes and optimize the spatial correlation analysis of gas outburst monitoring; S433. After the graph convolutional network layer, an adaptive time-frequency analysis mechanism is introduced to dynamically extract and weight features of different time periods and frequency ranges to generate comprehensive time-frequency features: Among them, F tf_adaptive (t,f) represents the adaptive time-frequency feature at time t and frequency f, α τν (t,f) represents the dynamic weight at time τ and frequency ν, F gcn (τ,ν) represents the output features of the graph convolutional network at time τ and frequency ν. γ(τ,ν) represents a new time-frequency weighting function, which is used to dynamically adjust the weights of different time and frequency features to optimize and improve the sensitivity of the ConvLSTM model to gas salient features in different time and frequency distributions. S434. In the long short-term memory network layer, the spatial relationship between sensors is modeled by combining the feature matrix processed by the graph convolutional network: in, represents the output feature of the h-th layer long short-term memory network at time t, represents the weight matrix input to the hidden layer, Represents the weight matrix from hidden layer to hidden layer, H (h-1) represents the value of the hidden state of the h-1th layer at time t-1, b (h) represents the bias of the hth layer, λ(t) represents the time adjustment function; S435. Introducing the self-attention mechanism, by weighted calculation of features at different time steps, focusing on time features, and optimizing and improving the ConvLSTM model's ability to predict gas outburst ignition at critical moments: Among them, F attention (t) represents the attention weighted feature at time t, β i (t) represents the attention weight at time step i, F lstm (i) represents the output features of the long short-term memory network at time step i, and θ(i) represents the attention weighting function, which is used to focus on the features of the key time steps; S436. Based on the outputs of the position-aware long-short-term memory network layer and the self-attention mechanism layer, the multi-layer perceptron is combined to predict the gas outburst risk level, and five levels of gas outburst risk level R are defined, namely no risk r1, low risk r2, medium risk r3, medium-high risk r4 and high risk r5.

5. The gas outburst monitoring and early warning method based on the improved ConvLSTM model according to claim 4 is characterized by: The method for identifying potential gas risk areas based on the analysis results of the improved ConvLSTM model in step S7 includes the following steps: S71. Based on the analysis results of the improved ConvLSTM model, potential gas outburst risk areas are identified, and the model outputs the gas outburst risk score matrix R score : R score ＝{r ij}； Among them, r ij represents the gas outburst risk score at location (i, j); S72. Evaluate the identified risk areas and determine their risk level R level The risk levels are classified as follows: Among them, θ1, θ2, θ3, and θ4 are thresholds for different risk levels; S73, according to the risk level R level Generate corresponding warning information W, which includes risk level, specific location and recommended measures; S74. When a high-risk area r5 is detected, the system promptly issues an alarm A, notifies relevant personnel and systems, and takes preventive measures; S75, when the data trend T within the preset range is monitored trend When the system takes corresponding monitoring feedback measures, the data trend monitoring formula is: T trend ={t∣f(t)>φ}; Where f(t) represents the data trend function, φ represents the preset trend threshold, and when the trend function f(t) exceeds the threshold φ, corresponding monitoring feedback measures are taken; S76. Record the warning information W and alarm A in the system log.

6. A gas outburst monitoring and early warning system based on an improved ConvLSTM model, applied to a gas outburst monitoring and early warning method based on an improved ConvLSTM model as described in any one of claims 1 to 5, characterized in that: It includes data acquisition module, data preprocessing module, feature extraction module, improved ConvLSTM model module, model training module, risk assessment and early warning module and monitoring platform module; The data acquisition module is used to collect gas concentration, rock mass vibration signals, temperature data, humidity data and air flow pressure to generate a gas data set; The data preprocessing module is used to clean the collected gas data set, convert unstructured data into structured data, and then clean, denoise and standardize the structured data to retain the node position information of each data in the gas data set; The feature extraction module is used to extract high-order derived features from the preprocessed gas data set, extract the time domain features, frequency domain features, and time-frequency domain features of the gas data set, and construct a feature matrix and a geographic location matrix based on the mine geographic location and adjacent relationship of the sensor; The improved ConvLSTM model module is used to construct an improved ConvLSTM model based on the feature matrix and the geographic location matrix. The model includes a convolutional layer, a graph convolutional network layer, a long short-term memory network layer, and a self-attention mechanism layer, and is used to monitor five levels of gas outburst risks. The model training module uses the historical gas data set to improve the ConvLSTM model for training and optimizes the model parameters so that it can accurately predict the risk level of gas outburst; The risk assessment and early warning module, based on the analysis results of the improved ConvLSTM model, is used to identify potential gas outburst risk areas, assess the risk level of the identified risk areas, and generate corresponding early warning information. When high-risk areas are detected, an alarm is promptly issued to notify relevant personnel and systems, and preventive measures are taken. When data trends within a preset range are monitored, corresponding monitoring feedback measures are taken; The monitoring platform module provides a visual interface for management personnel to view the current monitoring status, historical data trends and warning information in real time, and record warning information and alarm information.

Images