Mining underground safety early warning system

Abstract

The invention relates to the technical field of mining, and discloses a mining underground safety early warning system, which comprises the following modules: a data acquisition module used for controlling different types of sensing equipment and performing data transmission, a drilling monitoring module used for monitoring the state of drilling equipment, a data processing and analyzing module, and an early warning module used for monitoring the state of the drilling equipment. The data preprocessing module is used for preprocessing data and carrying out deep analysis on the preprocessed data, and the personnel positioning module is used for determining underground position coordinates of personnel in real time. Through the constructed LSTM model, the action intention state of the underground operation personnel can be analyzed, whether the subsequent action process has potential safety hazards or not is pre-judged according to the action intention obtained through analysis, and after the pre-judgment result is obtained, safety warning is performed on the underground operation personnel in advance through the early warning module; therefore, information hysteresis caused by real-time action information transmission is avoided, and the overall safety of the safety early warning process is also improved.

Description

Technical Field

[0001] The invention relates to the technical field of mining, and in particular to an underground safety early warning system for mining. Background Art

[0002] Mining refers to the process of extracting mineral resources from the earth's crust, including metal ores, non-metallic ores and energy minerals. This process usually involves multiple steps, such as exploration, mining, transportation and processing. Mining methods are mainly divided into underground mining and open-pit mining, depending on the location and nature of the mineral deposits.

[0003] In the existing technology, people often use video portrait recognition to analyze the action status of employees in the current area, so as to obtain the current action classification and make long-distance safety warnings based on whether there are safety hazards in the action classification.

[0004] However, in actual use, due to the certain data response time in the process of transmitting underground information to the ground data processing center, risky action information cannot be identified as safety hazards in time. That is, when a safety warning is issued, the safety hazard has already been transformed into the risk of dangerous behavior, which makes it impossible for people to supervise the behavior of underground operators in time. In view of this, we propose an underground safety warning system for mining. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides a mine underground safety warning system, which solves the problem in the prior art that the warning information cannot be notified to the operator in time because the signal transmission requires a certain response time.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A mine mining underground safety early warning system, including the following modules:

[0007] A data acquisition module is used to control different types of sensor equipment and transmit data to a control center through a data transmission unit;

[0008] The drilling monitoring module is used to monitor the status of the drilling equipment through various sensors installed on the drilling equipment and the instrument data of the drilling equipment itself;

[0009] Data processing and analysis module, which is used to apply data analysis algorithms, including statistical analysis and machine learning algorithms, to conduct in-depth analysis of pre-processed data and identify regularities and abnormal patterns in the data;

[0010] The personnel positioning module is used to determine the position coordinates of personnel underground in real time through positioning tags and UWB positioning technology, and to interconnect data with the positioning base stations arranged underground;

[0011] The early warning module sets the early warning thresholds for each monitoring indicator and determines the early warning status based on the data obtained by the data processing and analysis module. At the same time, it warns the underground personnel through sound and light early warning methods;

[0012] The video analysis module uses intelligent video analysis algorithms, including target detection, behavior recognition, and scene understanding algorithms, to automatically analyze underground video surveillance data, and uses algorithms to analyze whether there are safety hazards in personnel's operating behaviors.

[0013] Preferably, the sensors controlled in the data acquisition module include: a gas sensor for monitoring the gas concentration underground; a dust sensor for monitoring the dust content in various areas underground; a temperature sensor for monitoring the underground temperature; a humidity sensor for monitoring the humidity of the underground environment; and a pressure sensor for monitoring the tunnel pressure.

[0014] Preferably, the drilling monitoring module also includes a drilling parameter monitoring module, a formation characteristic monitoring module and a drilling parameter adaptation module. The drilling parameter monitoring module is used to obtain the current drilling equipment parameters including depth, pressure, speed and torque, and transmit the acquired data to the data processing and analysis module.

[0015] Preferably, the formation characteristic monitoring module is used to perform multi-source data fusion monitoring of the current formation characteristics through geological radar, sonic logging and electromagnetic induction, and to perform in situ real-time analysis of the chemical composition and mineral composition of the rock core at the drilling site through a spectrometer during the drilling process. The drilling parameter adaptive module is used to automatically adjust the depth, pressure, speed and torque of the drilling equipment according to the formation characteristics and drilling parameter changes monitored in real time during the drilling process, and use a machine learning algorithm to optimize its own parameter adjustment strategy by learning the drilling data.

[0016] Preferably, the data processing and analysis module uses a data preprocessing algorithm to filter, denoise, and calibrate the collected raw data to remove abnormal data points caused by sensor failure or interference from complex underground environment.

[0017] Preferably, the data processing and analysis module synchronously establishes an algorithm model based on the current mine data, and based on the historical gas concentration, temperature, ventilation volume data series, predicts the changing trend of gas concentration in the future and generates a corresponding risk distribution map, and issues potential danger warning signals in advance.

[0018] Preferably, the personnel positioning module calculates and generates the current three-dimensional model map of the underground according to the current mining status before performing calculations, and combines UWB technology and BLE technology with the existing positioning infrastructure underground, and displays each person underground in the generated three-dimensional model map by highlighting the points of light.

[0019] Preferably, before building the personnel positioning module, it is necessary to divide the data communication partitions according to the established underground three-dimensional modeling, and build distributed computing nodes in each group of divided partitions. At the same time, the data is preliminarily calculated by the distributed computing nodes to obtain the positioning data of the personnel in the area, and the distributed calculation results are transmitted to the data processing and analysis module for summary display.

[0020] Preferably, the early warning module customizes personalized early warning push methods according to the responsibilities and needs of different personnel. For drilling operators, when the early warning is triggered, the early warning information is directly transmitted to the operator through the smart helmets and smart watch devices they wear, including vibration, voice prompts, and simple text display methods. For ground management personnel, in addition to receiving detailed early warning reports through mobile phones and tablet computers, targeted decision-making suggestions are pushed according to their management scope and responsibilities, including dispatching rescue resources and adjusting drilling plan information.

[0021] Preferably, the video analysis module uses image analysis technology to identify people and their behavior information, and further predicts intentions based on the person behavior identification information, and identifies safety hazards in the scene.

[0022] The present invention provides a safety early warning system for underground mining. It has the following beneficial effects:

[0023] 1. The LSTM model constructed by the present invention can analyze the action intention state of the underground workers, and predict whether there are safety hazards in their subsequent action process based on the action intention obtained by the analysis, and issue a safety warning in advance through the early warning module after the prediction result is obtained, thereby avoiding the information lag caused by the instant action information transmission, and also improving the overall safety of the safety warning process.

[0024] 2. The present invention uses UWB technology and BLE technology to perform three-dimensional positioning of employees working underground, and feeds back the obtained three-dimensional positioning information to the generated three-dimensional modeling diagram of the underground and displays it in a highlighted state, so that managers can intuitively understand the distribution of employees in various areas underground.

[0025] 3. The present invention estimates the gas concentration attribute value of unknown points by utilizing the information of known gas concentration monitoring points, and considers the spatial correlation of the data by calculating the spatial variation function, so as to generate the gas concentration distribution of the entire underground area, and conduct risk assessment and early warning based on this, so as to further ensure the safety of the underground environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is the flow chart of the underground safety early warning system for this mine;

[0027] Figure 2 It is a schematic diagram of the drilling monitoring module of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the specification 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.

[0029] Example:

[0030] Please see attached Figure 1 - Attachment Figure 2 The embodiment of the present invention provides a mine mining underground safety early warning system, including the following modules:

[0031] A data acquisition module is used to control different types of sensor equipment and transmit data to a control center through a data transmission unit;

[0032] The drilling monitoring module is used to monitor the status of the drilling equipment through various sensors installed on the drilling equipment and the instrument data of the drilling equipment itself;

[0033] Data processing and analysis module, which is used to apply data analysis algorithms, including statistical analysis and machine learning algorithms, to conduct in-depth analysis of pre-processed data and identify regularities and abnormal patterns in the data;

[0034] The personnel positioning module is used to determine the position coordinates of personnel underground in real time through positioning tags and UWB positioning technology, and to interconnect data with the positioning base stations arranged underground;

[0035] The early warning module sets the early warning thresholds for each monitoring indicator and determines the early warning status based on the data obtained by the data processing and analysis module. At the same time, it warns the underground personnel through sound and light early warning methods;

[0036] The video analysis module uses intelligent video analysis algorithms, including target detection, behavior recognition, and scene understanding algorithms, to automatically analyze underground video surveillance data, and uses algorithms to analyze whether there are safety hazards in personnel's operating behaviors.

[0037] The sensors controlled in the data acquisition module include: a gas sensor for monitoring the gas concentration underground; a dust sensor for monitoring the dust content in each area underground; a temperature sensor for monitoring the underground temperature; a humidity sensor for monitoring the humidity of the underground environment; and a pressure sensor for monitoring the tunnel pressure.

[0038] The drilling monitoring module also includes a drilling parameter monitoring module, a formation characteristic monitoring module and a drilling parameter adaptive module. The drilling parameter monitoring module is used to obtain the current drilling equipment parameters including depth, pressure, speed and torque, and transmit the acquired data to the data processing and analysis module.

[0039] The formation characteristic monitoring module is used to perform multi-source data fusion monitoring of the current formation characteristics through geological radar, sonic logging and electromagnetic induction, and to perform in-situ real-time analysis of the chemical composition and mineral composition of the rock core at the drilling site through a spectrometer during the drilling process. The drilling parameter adaptive module is used to automatically adjust the depth, pressure, speed and torque of the drilling equipment according to the formation characteristics and drilling parameter changes monitored in real time during the drilling process, and use a machine learning algorithm to optimize its own parameter adjustment strategy by learning the drilling data.

[0040] The data processing and analysis module uses a data preprocessing algorithm to filter, remove noise, and calibrate the collected raw data to remove abnormal data points caused by sensor failure or interference from the complex underground environment.

[0041] The data processing and analysis module synchronously establishes an algorithm model based on the current mine data, and predicts the changing trend of gas concentration in the future and generates a corresponding risk distribution map based on the historical gas concentration, temperature, and ventilation data series, and issues a potential danger warning signal in advance. The algorithm principle is to use the information of known points (known gas concentration monitoring points) to estimate the attribute value (gas concentration) of unknown points, and consider the spatial correlation of data by calculating the spatial variation function, so as to generate the gas concentration distribution of the entire area, and then generate a risk distribution map;

[0042] Algorithm formula:

[0043] Assume that Z(x) is the gas concentration at spatial position x. It is known that there are n monitoring points x1, x2, ...x n , and the corresponding gas concentration values ​​are Z(x1), Z(x2), ...Z(xn );

[0044] First, the spatial variation function γ(h) is calculated. It is a function that describes the degree of variation of gas concentration when the distance between two points is h. It is calculated by the following formula:

[0045]

[0046] Where N(h) is the number of point pairs with distance h;

[0047] Then, the system of equations is constructed based on the spatial variogram:

[0048]

[0049] Where x0 is the position of the point to be estimated, λ j is the weight coefficient, μ is the Lagrange multiplier;

[0050] Solve this system of equations to get the weight coefficient λ j , then the gas concentration value Z(x0) of the estimated point x0 can be estimated by the following formula;

[0051]

[0052] By repeating the above calculation process for each point to be estimated in the entire area, the gas concentration distribution of the entire area can be obtained, and then a risk distribution map can be generated according to the set risk threshold and other conditions.

[0053] The personnel positioning module generates a three-dimensional model of the current underground according to the current mining status before calculation, and combines UWB technology and BLE technology with the existing positioning infrastructure underground, and displays each person underground in the generated three-dimensional model by highlighting the points.

[0054] Before the personnel positioning module is built, it is necessary to divide the data communication partitions according to the established underground three-dimensional modeling, and build distributed computing nodes in each group of partitions. At the same time, the distributed computing nodes perform preliminary calculations on the data to obtain the positioning data of the personnel in the area, and transmit the distributed calculation results to the data processing and analysis module for summary display. The specific calculation process is as follows:

[0055] First, multiple UWB positioning base stations are deployed underground, and personnel carry positioning tags. The positioning tags send UWB signals, and the base stations receive the signals and record the reception time;

[0056] Assume that the time when the bit tag sends the signal is t0, and the time when the base station i receives the signal is t i , then the distance d between the base station and the positioning tag iIt can be calculated based on the signal propagation speed c (approximately the speed of light in air) and the time difference:

[0057] d i =c×(t i -t0);

[0058] The position coordinates of the positioning tag (i.e., person) are calculated using the trilateration method or the multilateration method through the distance measurement information of multiple base stations (at least 3) and the positioning tag. Taking the trilateration method as an example, assuming that the coordinates of the three base stations are (x1, y1, z1), (x2, y2, z2), and (x3, y3, z3), and the corresponding distances to the positioning tag are d1, d2, and d3, respectively, the coordinates (x, y, z) of the positioning tag can be solved by the following set of equations;

[0059]

[0060] By solving the equations, the position coordinates of the personnel in the well can be obtained;

[0061] The early warning module customizes personalized early warning push methods according to the responsibilities and needs of different personnel. For drilling operators, when the early warning is triggered, the early warning information is directly transmitted to the operator through the smart helmets and smart watches they wear, including vibration, voice prompts, and simple text display methods. For ground management personnel, in addition to receiving detailed early warning reports through mobile phones and tablet computers, targeted decision-making suggestions are pushed according to their management scope and responsibilities, including dispatching rescue resources and adjusting drilling plan information.

[0062] The video analysis module uses image analysis technology to identify people and their behavior information, and further predicts intentions based on the person's behavior identification information, and identifies potential safety hazards in the scene. The deep learning algorithm is used in the process. The specific measurement method is as follows:

[0063] We also collected a large amount of video data on underground personnel behavior, and annotated the personnel behavior in each video clip in detail, not only annotating the specific behavior, but also clarifying its corresponding intention. For example, the behavior of "slowly approaching the drilling equipment and picking up the tool" is marked as "preparing for equipment maintenance";

[0064] Build a long short-term memory network (LSTM) model. LSTM consists of input gate, output gate, forget gate and cell state, which can effectively handle long-term dependencies in sequence data. The action feature sequence of people extracted from the video is used as input data and input into the LSTM model;

[0065] The labeled data is divided into training set, validation set and test set. The LSTM model is trained using the training set. During the training process, the weight parameters of the model are continuously adjusted through the back propagation algorithm combined with the optimizer (such as Adam optimizer) to minimize the difference between the predicted intent and the actual intent.

[0066] For the video to be analyzed, extract the action feature sequence of the person and input it into the trained LSTM model. The LSTM model will calculate based on the input sequence data and output the predicted intention of the person. For example, after inputting the action feature sequence of the person "quickly walks towards the drilling equipment and adjusts the equipment parameters", the model outputs the predicted intention as "eager to start drilling operations";

[0067] Assume that the input action feature sequence of personnel is x = {x1, x2, ..., x T} (where T is the sequence length), the input gate, output gate, and forget gate of the LSTM unit at time t are denoted by i t , o t 、f t , the cell state is denoted as c t , the hidden state is recorded as h t

[0068] Input gate update formula:

[0069] i t =σ(W xi x t +W hi h t-1 +b i )

[0070] Where σ is the sigmoid function, W xi , W hi is the weight matrix, b i is the bias vector;

[0071] Forget gate update formula:

[0072] f t =σ(W xf x t +W hf h t-1 +b f )

[0073] Cell state update formula:

[0074] c t =f t c t-1 +i t tanh(W xc x t +Whc h t-1 +b c )

[0075] Output gate update formula:

[0076] o t =σ(W xo x t +W ho h t-1 +b o )

[0077] Hidden state update formula:

[0078] h t =o t tanh(c t )

[0079] Finally, after being processed by the LSTM network, the output layer is based on the hidden state h t Output the predicted personnel intention. The specific output method depends on the specific classification method (such as classification output through the fully connected layer and softmax function, etc.);

[0080] Assume that there is a new video clip to be analyzed, which comes from the real-time monitoring of the underground drilling site. The video clip is processed according to the above feature extraction method to extract the behavior feature vector sequence of the personnel;

[0081] These extracted features are combined into a feature vector sequence and fed into the LSTM model as input data. For example, for a video clip, its feature vector sequence is as follows:

[0082] Frame 1: [personnel position coordinate change 1, body posture feature 1, action duration 1]

[0083] Frame 2: [personnel position coordinate change 2, body posture feature 2, action duration 2]

[0084] …

[0085] Frame n: [person position coordinate change n, body posture feature n, action duration n]

[0086] The extracted feature vector sequence is input into the trained LSTM model. The model will process the input sequence data through the LSTM layer and the fully connected layer in turn, and finally output the predicted personnel intention in the output layer;

[0087] The model structure is as follows:

[0088] Input layer: receives feature vector sequences with dimensions (batch_size, sequence_length, input_dim), where batch_size is the number of samples for each training, sequence_length is the length of the feature vector sequence (i.e., the number of video frames), and input_dim is the dimension of the input data (here preset to 128);

[0089] LSTM layer: contains 64 hidden units, processes the input feature vector sequence and captures the long-term dependencies;

[0090] Fully connected layer: Connect the output of the LSTM layer to a fully connected layer to further integrate and transform the data;

[0091] Output layer: The softmax function is used to convert the output of the fully connected layer into a probability distribution. Each element represents the predicted probability of the corresponding intent category. The dimension is (batch_size, num_clusses), where num_clusses is the number of intent categories (5 here).

[0092] Assume that the probability distribution of the output is as follows:

[0093] Prepare for equipment maintenance: 0.1

[0094] Eager to adjust drilling parameters: 0.6

[0095] Check the device running status: 0.2

[0096] Other meanings Figure 1 :0.05

[0097] Other meanings Figure 2 :0.05

[0098] According to the output probability distribution, the category with the highest probability is selected as the predicted person's intention. In this example, the probability of "anxious to adjust drilling parameters" is the highest, so it is predicted that the person's intention in this video clip is to adjust drilling parameters in a hurry.

[0099] Through such an LSTM-based deep learning algorithm example, we can use the video data of underground personnel behavior to predict personnel behavior intentions through steps such as feature extraction, model construction and training, thereby providing more valuable information for underground mining safety management and discovering potential safety hazards in advance.

[0100] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A mine mining underground safety early warning system, characterized in that: Includes the following modules: A data acquisition module is used to control different types of sensor equipment and transmit data to a control center through a data transmission unit; The drilling monitoring module is used to monitor the status of the drilling equipment through various sensors installed on the drilling equipment and the instrument data of the drilling equipment itself; Data processing and analysis module, which is used to apply data analysis algorithms, including statistical analysis and machine learning algorithms, to conduct in-depth analysis of pre-processed data and identify regularities and abnormal patterns in the data; The personnel positioning module is used to determine the position coordinates of personnel underground in real time through positioning tags and UWB positioning technology, and to interconnect data with the positioning base stations arranged underground; The early warning module sets the early warning thresholds for each monitoring indicator and determines the early warning status based on the data obtained by the data processing and analysis module. At the same time, it warns the underground personnel through sound and light early warning methods; The video analysis module uses intelligent video analysis algorithms, including target detection, behavior recognition, and scene understanding algorithms, to automatically analyze underground video surveillance data, and uses algorithms to analyze whether there are safety hazards in personnel's operating behaviors.

2. The underground mine safety warning system according to claim 1 is characterized in that: The sensors controlled in the data acquisition module include: a gas sensor for monitoring the gas concentration underground; a dust sensor for monitoring the dust content in each area underground; a temperature sensor for monitoring the underground temperature; a humidity sensor for monitoring the humidity of the underground environment; and a pressure sensor for monitoring the tunnel pressure.

3. The underground mine safety warning system according to claim 1 is characterized in that: The drilling monitoring module also includes a drilling parameter monitoring module, a formation characteristic monitoring module and a drilling parameter adaptive module. The drilling parameter monitoring module is used to obtain the current drilling equipment parameters including depth, pressure, speed and torque, and transmit the acquired data to the data processing and analysis module.

4. The underground mine safety warning system according to claim 3 is characterized in that: The formation characteristic monitoring module is used to perform multi-source data fusion monitoring of the current formation characteristics through geological radar, sonic logging and electromagnetic induction, and to perform in-situ real-time analysis of the chemical composition and mineral composition of the rock core at the drilling site through a spectrometer during the drilling process. The drilling parameter adaptive module is used to automatically adjust the depth, pressure, speed and torque of the drilling equipment according to the formation characteristics and drilling parameter changes monitored in real time during the drilling process, and use a machine learning algorithm to optimize its own parameter adjustment strategy by learning the drilling data.

5. The underground safety early warning system for mining according to claim 1 is characterized in that: The data processing and analysis module uses a data preprocessing algorithm to filter, remove noise, and calibrate the collected raw data to remove abnormal data points caused by sensor failure or interference from the complex underground environment.

6. The underground safety warning system for mining according to claim 1 is characterized in that: The data processing and analysis module synchronously establishes an algorithm model based on the current mine data, and predicts the changing trend of gas concentration in the future and generates a corresponding risk distribution map based on the historical gas concentration, temperature, and ventilation volume data series, and issues potential danger warning signals in advance.

7. The underground safety warning system for mining according to claim 1 is characterized in that: The personnel positioning module generates a three-dimensional model of the current underground according to the current mining status before calculation, and combines UWB technology and BLE technology with the existing positioning infrastructure underground, and displays each person underground in the generated three-dimensional model by highlighting the points.

8. The underground safety warning system for mining according to claim 7 is characterized in that: Before building the personnel positioning module, it is necessary to divide the data communication partitions according to the established underground three-dimensional modeling, and build distributed computing nodes in each group of divided partitions. At the same time, the distributed computing nodes perform preliminary calculations on the data to obtain the positioning data of the personnel in the area, and transmit the distributed calculation results to the data processing and analysis module for summary display.

9. The underground mine safety warning system according to claim 1, characterized in that: The early warning module customizes personalized early warning push methods according to the responsibilities and needs of different personnel. For drilling operators, when the early warning is triggered, the early warning information is directly transmitted to the operator through the smart helmets and smart watches they wear, including vibration, voice prompts, and simple text display methods. For ground management personnel, in addition to receiving detailed early warning reports through mobile phones and tablet computers, targeted decision-making suggestions are pushed according to their management scope and responsibilities, including dispatching rescue resources and adjusting drilling plan information.

10. The underground mine safety warning system according to claim 1, characterized in that: The video analysis module uses image analysis technology to identify people and their behavior information, and further predicts intentions based on the person behavior identification information, and identifies safety hazards in the scene.

Images