A method and system for visual monitoring of coal mine filling mining area

Abstract

The present invention discloses a method and system for visual monitoring of coal mine backfilling mining area, including: collecting equipment operation data of backfilling mining area in real time through monitoring equipment; deeply analyzing the collected equipment operation data to obtain the current operation status of equipment in backfilling mining area; combining environmental monitoring data and equipment operation data of backfilling mining area, using artificial neural network intelligent fault diagnosis method to evaluate the equipment operation status of backfilling mining area, and timely warning before equipment failure; feeding back the obtained warning information to designated equipment and designated personnel. The present invention is of great significance to improving production efficiency of mining area and inhibiting surface subsidence.

Description

Technical Field

[0001] The present invention relates to the technical field of visual monitoring, and in particular to a method and system for visual monitoring of a coal mine filling mining area. Background Art

[0002] With the increase of mining depth and the improvement of social awareness of environmental protection, the proportion of backfill mining method applied in mining enterprises is increasing, and there is a trend of further development and expansion. Backfill mining method can adapt to various complex and changeable ore body conditions, such as irregularly distributed coal block sections, three-down coal corner coal, etc., to maximize the coal recovery rate and reduce the mining depletion rate, effectively control the mine pressure activities and ensure the safety of mining production, protect the surface from damage and maintain the existing ecological environment.

[0003] Therefore, in actual operation, visual monitoring of the filling mining area is particularly important during filling mining. However, in the prior art, visual monitoring of the filling mining area in coal mines is rare. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention proposes a method and system for visual monitoring of coal mine filling mining areas, which is of great significance to improving the production efficiency of mining areas and inhibiting surface subsidence.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for visual monitoring of coal mine filling mining area, comprising the following steps:

[0007] Collect equipment operation data in the filling and mining area in real time through monitoring equipment;

[0008] Conduct in-depth analysis on the collected equipment operation data to obtain the current operating status of the equipment in the filling mining area;

[0009] Combined with the environmental monitoring data and equipment operation data of the filling mining area, the artificial neural network intelligent fault diagnosis method is used to evaluate the equipment operation status in the filling mining area, and timely warning is given before the equipment fails;

[0010] The obtained warning information will be fed back to the designated equipment and designated personnel.

[0011] Preferably, the equipment operation data of the filling mining area includes: the operation data of the filling preparation station, the filling pipeline, the filling drilling equipment, the filling dehydration equipment, the conveyor, the mixer, the filling pump and other auxiliary equipment.

[0012] Preferably, the method for in-depth analysis of the collected equipment operation data includes:

[0013] Configure targeted data collection strategies based on the operating characteristics of the equipment, capture, analyze, and locate effective vibration, temperature, current, voltage, and speed data of the equipment, and use intelligent alarm strategies to promptly detect abnormal operating status data of the equipment and transmit it back to the on-site monitoring system server;

[0014] At the same time, on-site equipment maintenance and management personnel can access the intelligent monitoring system in this server through computer terminals to understand the real-time operating status of the monitoring unit;

[0015] Dedicated diagnostic experts can remotely perform data analysis and fault diagnosis on all monitored equipment.

[0016] Preferably, the method of using an artificial neural network intelligent fault diagnosis method to evaluate the equipment operating status in the filling mining area includes:

[0017] Through deep learning technology, especially the improved long short-term memory network LSTM, data mining is performed on the real-time operation data of the equipment to obtain the equipment status operation information;

[0018] Based on the operating status information and corresponding sensor information, deep reinforcement learning (DRL) is constructed to make intelligent decisions to optimize the performance of the equipment and extend its service life.

[0019] Preferably, the improved long short-term memory network LSTM is a CNN-LSTM state analysis network integrating an attention mechanism, specifically comprising:

[0020] The CNN convolution-pooling structure is used to abstract and extract effective data features. The generated feature vector is multiplied by the feature importance vector after the attention mechanism to obtain a time series feature vector with important differences.

[0021] The time correlation of features is found through the LSTM network, and finally the state classification is achieved through the softmax classifier.

[0022] Preferably, based on the operating status information and the corresponding sensor information, a method for constructing a deep reinforcement learning (DRL) to make intelligent decisions to optimize the performance of the equipment and extend its service life includes:

[0023] The performance of the agent is defined as the efficiency and load response capability of the equipment. The environment includes the internal operating status of the equipment and the external working conditions. The actuator is the control system of the equipment, and the sensor measures the real-time operating data of the equipment.

[0024] The selection of control strategy is based on the current state of the device and is performed by the DRL algorithm at each decision moment; the state S is defined as the set of operating parameters of the device, the action a is the control command, and the reward function r is based on the operating efficiency and stability of the device;

[0025] The design of the DQN network will include an input layer, multiple hidden layers and an output layer. The input layer receives the current state S, the hidden layer extracts features through a deep neural network, and the output layer provides a Q value for each action, indicating the expected return of taking the action;

[0026] Through continuous iterative training, the intelligent agent will learn to choose the optimal action under a given state to achieve efficient and stable operation of the equipment. The ultimate goal is that the equipment can maintain optimal performance under various working conditions while preventing possible failures.

[0027] Preferably, the optimization goal is to maximize the action value function Q π* (S, a), the cumulative reward for taking action a in state S, following the optimal policy π*:

[0028]

[0029] In the formula, r t represents the discount factor after time t, and R(t) represents the reward for taking action a when the operating state of the electromechanical equipment is S.

[0030] The present invention also provides a visual monitoring system for coal mine filling mining area, including: a collection module, an analysis module, an evaluation module and a feedback module;

[0031] The acquisition module is used to collect equipment operation data of the filling and mining area in real time through monitoring equipment;

[0032] The analysis module is used to perform in-depth analysis on the collected equipment operation data to obtain the current operation status of the equipment in the filling mining area;

[0033] The evaluation module is used to combine the environmental monitoring data and equipment operation data of the filling mining area, and use the artificial neural network intelligent fault diagnosis method to evaluate the equipment operation status in the filling mining area, and timely warn before the equipment fails;

[0034] The feedback module is used to feed back the obtained warning information to the designated equipment and designated personnel.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The present invention discloses a method and system for visual monitoring of the filling mining area of ​​a coal mine, including real-time collection of equipment operation data of the filling mining area by monitoring equipment; in-depth analysis of the collected equipment operation data to obtain the current operation status of the equipment in the filling mining area; combining the environmental monitoring data and equipment operation data of the filling mining area, using an artificial neural network intelligent fault diagnosis method to evaluate the equipment operation status of the filling mining area, and timely warning before the equipment fails; and feeding back the obtained warning information to the designated equipment and designated personnel. The present invention is of great significance to improving the production efficiency of the mining area and suppressing the surface subsidence. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0038] Figure 1 The present invention is a flow chart of a method for visually monitoring a coal mine backfill mining area according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0041] Embodiment 1

[0042] like Figure 1 As shown, the present invention provides a method for visual monitoring of coal mine filling mining area, comprising the following steps:

[0043] Collect equipment operation data in the filling and mining area in real time through monitoring equipment;

[0044] Conduct in-depth analysis on the collected equipment operation data to obtain the current operating status of the equipment in the filling mining area;

[0045] Combined with the environmental monitoring data and equipment operation data of the filling mining area, the artificial neural network intelligent fault diagnosis method is used to evaluate the equipment operation status in the filling mining area, and timely warning is given before the equipment fails;

[0046] The obtained warning information will be fed back to the designated equipment and designated personnel.

[0047] In this embodiment, the equipment operation data of the filling mining area includes: the operation data of the filling preparation station, the filling pipeline, the filling drilling equipment, the filling dehydration equipment, the conveyor, the mixer, the filling pump and other auxiliary equipment.

[0048] In this embodiment, the method for performing in-depth analysis on the collected device operation data includes:

[0049] Configure targeted data collection strategies based on the operating characteristics of the equipment, capture, analyze, and locate effective vibration, temperature, current, voltage, and speed data of the equipment, and use intelligent alarm strategies to promptly detect abnormal operating status data of the equipment and transmit it back to the on-site monitoring system server;

[0050] At the same time, on-site equipment maintenance and management personnel can access the intelligent monitoring system in this server through computer terminals to understand the real-time operating status of the monitoring unit;

[0051] Dedicated diagnostic experts can remotely perform data analysis and fault diagnosis on all monitored equipment.

[0052] Specifically, it is implemented through an online monitoring system, and the system technical path is composed of three levels.

[0053] (1) Data collection layer. Install integrated sensors on the key equipment to be monitored to collect raw data. Configure targeted data collection strategies based on the operating characteristics of the equipment to capture, analyze, and locate the effective vibration, temperature, current, voltage, and speed data of the equipment. With the help of intelligent alarm strategies, timely discover the abnormal operating status data of the equipment and transmit it to the data collection station via wired or wireless means. The data collection station transmits the data to the online monitoring and remote diagnosis system server deployed in the filling and mining area via wired or wireless AP networking.

[0054] (2) On-site monitoring system. A server is provided with the monitoring system. The server is installed in the server room of the control room and is used to store the equipment status data sent back from the data acquisition station. At the same time, on-site equipment maintenance and management personnel can access the intelligent monitoring system in the server through a computer terminal to understand the real-time operating status of the monitoring unit.

[0055] (3) Remote diagnosis center system. Professional diagnosis experts can remotely analyze data and diagnose faults on all monitored equipment.

[0056] In this embodiment, the method of using the artificial neural network intelligent fault diagnosis method to evaluate the equipment operation status in the filling mining area includes:

[0057] Through deep learning technology, especially the improved long short-term memory network LSTM, data mining is performed on the real-time operation data of the equipment to obtain the equipment status operation information;

[0058] Based on the operating status information and corresponding sensor information, deep reinforcement learning (DRL) is constructed to make intelligent decisions to optimize the performance of the equipment and extend its service life.

[0059] Among them, the improved long short-term memory network LSTM is a CNN-LSTM state analysis network that integrates the attention mechanism, specifically including:

[0060] The CNN convolution-pooling structure is used to abstract and extract effective data features. The generated feature vector is multiplied by the feature importance vector after the attention mechanism to obtain a time series feature vector with important differences.

[0061] The time correlation of features is found through the LSTM network, and finally the state classification is achieved through the softmax classifier.

[0062] Specifically, the original data set is divided into multiple data subsets at equal time intervals, and each data subset enters the CNN-based feature extraction module and attention branch module respectively. Then, the results of the two parallel branches are multiplied to obtain multiple time series feature sequences. The fused feature sequence is input into the LSTM hidden unit structure, and the LSTM's ability to grasp the time series is used to find the correlation of the data features before and after the time. The hidden unit selectively forgets or retains the information, extracts the features of the sequence in the time dimension, and finally realizes the multi-label classification of the equipment operation status through the softmax fully connected network.

[0063] Among them, the convolution layer of the CNN structure of the feature extraction network is composed of k layers of two-dimensional data graphs, which are stacked by m pieces of equipment operation status data with a length of n. The downsampling layer of the CNN structure still uses the MaxPooling operation to achieve feature map compression and dimensionality reduction. After multiple convolution layers and pooling, feature sequences with high-level abstract characteristics are gradually extracted. At the same time, the attention branch realizes feature aggregation and scale recovery of the input data, extracts the feature importance sequence through a convolution kernel of size (m*1*1), and then constrains the value between 0 and 1 through the tanh function.

[0064] The entire architecture fully considers the complex characteristics of data in time and space dimensions, effectively prevents problems such as loss of key features and time gradient diffusion, and comprehensively captures and learns the features of equipment operation status data.

[0065] In this embodiment, based on the operating status information and the corresponding sensor information, a method for constructing a deep reinforcement learning (DRL) to make intelligent decisions to optimize the performance of the equipment and extend its service life includes:

[0066] First, the performance of the intelligent agent is defined as the efficiency and load response capability of the equipment. The environment includes the internal operating status of the equipment, such as temperature, current, voltage, etc., as well as external working conditions, such as load demand and underground environment. Actuators are the control systems of the equipment, which are used to adjust the operating parameters of the equipment, such as the settings of the frequency converter. Sensors measure the real-time operating data of the motor, such as current, voltage, speed and temperature, as well as the real-time analysis of the equipment operating status.

[0067] The selection of control strategy is based on the current state of the device, and is performed by the DRL algorithm at each decision moment; the state S is defined as the set of operating parameters of the device, the action a is the control command, and the reward function r is based on the operating efficiency and stability of the device. The state vector can be expressed as:

[0068] S t =[s voltage (t), s cunent (t), s temperaure (t), s load (t)]

[0069] The motion space is defined by the adjustable range of the motor control parameters:

[0070] A t =[a1(t), a2(t)...a3(t)]

[0071] The reward function aims to optimize the efficiency and responsiveness of the motor while minimizing maintenance and energy consumption:

[0072] r t =f(e, r, cos t main , cost consumption )

[0073] In the formula, e represents the work efficiency, r represents the response signal, and cost represents the loss value in a certain aspect.

[0074] The design of the DQN network will include an input layer, multiple hidden layers and an output layer. The input layer receives the current state S t , the hidden layer extracts features through a deep neural network, and the output layer provides a Q value for each action, which represents the expected reward of taking that action;

[0075] The objective function will calculate an expected long-term return for the current state and the selected action. DQN training updates parameters through the following loss function:

[0076]

[0077] Through continuous iterative training, the agent will learn to choose the best action under a given state, thereby achieving efficient and stable operation of the equipment. The ultimate goal is for the equipment to maintain optimal performance under various working conditions, while preventing possible failures and ensuring the continuity and safety of coal mine production. This intelligent control strategy will significantly improve the operating efficiency and reliability of equipment in the coal mine filling mining area, reduce maintenance costs, and provide strong support for the efficient operation of equipment in the coal mine filling mining area.

[0078] Embodiment 2

[0079] The present invention also provides a visual monitoring system for coal mine filling mining area, including: a collection module, an analysis module, an evaluation module and a feedback module;

[0080] The collection module is used to collect equipment operation data of the filling and mining area in real time through monitoring equipment;

[0081] The analysis module is used to conduct in-depth analysis on the collected equipment operation data to obtain the current operating status of the equipment in the filling mining area;

[0082] The evaluation module is used to combine the environmental monitoring data and equipment operation data of the filling mining area, and use the artificial neural network intelligent fault diagnosis method to evaluate the equipment operation status in the filling mining area, and give timely warnings before equipment failure occurs;

[0083] The feedback module is used to feed back the obtained warning information to the designated equipment and designated personnel.

[0084] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for visual monitoring of coal mine filling mining area, characterized in that: The following steps are involved: Collect equipment operation data in the filling and mining area in real time through monitoring equipment; Conduct in-depth analysis on the collected equipment operation data to obtain the current operating status of the equipment in the filling mining area; Combined with the environmental monitoring data and equipment operation data of the filling mining area, the artificial neural network intelligent fault diagnosis method is used to evaluate the equipment operation status in the filling mining area, and timely warning is given before the equipment fails; Feedback the obtained warning information to designated equipment and designated personnel; The method of using artificial neural network intelligent fault diagnosis method to evaluate the equipment operation status in the filling mining area includes: Through deep learning technology, especially the improved long short-term memory network LSTM, data mining is performed on the real-time operation data of the equipment to obtain the equipment operation status information; Based on the operating status information and corresponding sensor information, deep reinforcement learning (DRL) is constructed to make intelligent decisions to optimize the performance of the equipment and extend its service life. The improved long short-term memory network LSTM is a CNN-LSTM state analysis network that integrates the attention mechanism, specifically including: The CNN convolution-pooling structure is used to abstract and extract effective data features. The generated feature vector is multiplied by the feature importance vector after the attention mechanism to obtain a time series feature vector with important differences. The time correlation of features is found through the LSTM network, and finally the state classification is realized through the softmax classifier; Based on the operating status information and the corresponding sensor information, a method for constructing a deep reinforcement learning (DRL) to make intelligent decisions to optimize the performance of the equipment and extend its service life includes: The performance of the agent is defined as the efficiency and load response capability of the equipment. The environment includes the internal operating status of the equipment and the external working conditions. The actuator is the control system of the equipment, and the sensor measures the real-time operating data of the equipment. The selection of control strategy is based on the current state of the device and is performed by the DRL algorithm at each decision moment; the state S is defined as the set of operating parameters of the device, the action a is the control command, and the reward function r is based on the operating efficiency and stability of the device; The design of the DQN network will include an input layer, multiple hidden layers and an output layer. The input layer receives the current state S, the hidden layer extracts features through a deep neural network, and the output layer provides a Q value for each action, indicating the expected return of taking the action; Through continuous iterative training, the agent will learn to choose the best action under a given state to achieve efficient and stable operation of the equipment. The ultimate goal is to ensure that the equipment can maintain optimal performance under various working conditions while preventing possible failures. The optimization goal is to maximize the action value function The cumulative reward for taking action a in state S, following the optimal strategy π*: In the formula, r t represents the discount factor after time t, and R(t) represents the reward for taking action a when the operating state of the electromechanical equipment is S.

2. The method for visual monitoring of coal mine filling mining area according to claim 1, characterized in that: The equipment operation data of the filling mining area includes: filling preparation station, filling pipeline, filling drilling equipment, filling dehydration equipment, conveyor, mixer, filling pump and other auxiliary equipment operation data.

3. The method for visual monitoring of coal mine filling mining area according to claim 1, characterized in that: Methods for in-depth analysis of collected equipment operation data include: Configure targeted data collection strategies based on the operating characteristics of the equipment, capture, analyze, and locate effective vibration, temperature, current, voltage, and speed data of the equipment, and use intelligent alarm strategies to promptly detect abnormal operating status data of the equipment and transmit it back to the on-site monitoring system server; At the same time, on-site equipment maintenance and management personnel can access the intelligent monitoring system in this server through computer terminals to understand the real-time operating status of the monitoring unit; Dedicated diagnostic experts can remotely perform data analysis and fault diagnosis on all monitored equipment.

4. A visual monitoring system for coal mine filling mining area, characterized in that: The system is used to implement the method described in any one of claims 1 to 3, comprising: a collection module, an analysis module, an evaluation module and a feedback module; The acquisition module is used to collect equipment operation data of the filling and mining area in real time through monitoring equipment; The analysis module is used to perform in-depth analysis on the collected equipment operation data to obtain the current operation status of the equipment in the filling mining area; The evaluation module is used to combine the environmental monitoring data and equipment operation data of the filling mining area, and use the artificial neural network intelligent fault diagnosis method to evaluate the equipment operation status in the filling mining area, and timely warn before the equipment fails; The feedback module is used to feed back the obtained warning information to the designated equipment and designated personnel.

Images