An intelligent adjustment system and method for mine state based on the Internet of Things

Abstract

The present invention discloses an intelligent adjustment system and method for mine state based on the Internet of Things, including: a state extraction module: obtaining preliminary state factors from operation parameters, and then performing deep clustering optimization on the preliminary state factors to extract state factors; a deep application module: globally modeling the state factors based on the Transformer architecture to obtain the global representation of the state factors, and analyzing the importance sequence representation of the state factors at different time scales based on the global representation; a decision control module: according to the importance sequence representation, selecting the best time interval for data collection for the state parameters corresponding to the state factors in each sample, and obtaining the state parameters collected at each time interval for observation. When an abnormal situation occurs, an alarm is immediately sent and the device operation parameters are adjusted, without being affected by the differences in the change rates of different device states on the importance of the collected data.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control, and particularly to an intelligent adjustment system and method for mine status based on the Internet of Things. Background Art

[0002] The underground operation environment is complex, with various potential safety hazards such as gas explosion, water inrush, and roof fall. Traditional safety monitoring means have problems such as limited monitoring range and untimely information transmission, making it difficult to detect and warn of potential dangers in a timely manner. For example, in coal mines, the accident rate of trackless rubber-tired vehicles has been on the rise in recent years, which has a great impact on underground safety production, resulting in personal injuries and economic losses. The Internet of Things technology can realize real-time and comprehensive monitoring of mine equipment and environment, detect abnormalities in a timely manner and give warnings, providing a strong guarantee for safety production.

[0003] At present, in the traditional mine status adjustment system, the device status data is usually sampled and input at fixed time intervals. This fixed-time-interval method does not take into account the differences in the change rates of different device statuses. For example, the current parameter of mine equipment will fluctuate greatly instantaneously when the equipment starts or the load changes, while the temperature parameter of the equipment changes slowly over a long period of time. If a unified fixed time interval is used, for parameters that change rapidly like current, key change information may be missed due to too large a time interval, and for parameters that change slowly like temperature, the long-term change trend cannot be accurately captured because the sampling time interval is short, and the true operating state of the equipment cannot be accurately obtained, affecting the control decision-making of mine equipment. Therefore, an intelligent adjustment system and method for mine status based on the Internet of Things are proposed here to solve the above problems. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art and to achieve the above object, the present invention proposes the following technical solutions:

[0005] An intelligent adjustment system for mine status based on the Internet of Things, comprising:

[0006] A data acquisition module: collecting mine equipment data based on the Internet of Things technology and extracting operation parameters from the mine equipment data;

[0007] A status extraction module: obtaining preliminary status factors from the operation parameters, and then performing deep clustering optimization on the preliminary status factors to extract status factors;

[0008] A deep application module: globally modeling the status factors based on the Transformer architecture, obtaining the global representation of the status factors, and analyzing the importance sequence representation of the status factors at different time scales based on the global representation;

[0009] Decision control module: According to the importance sequence representation, the optimal time interval is selected for data collection for the state parameters corresponding to the state factors in each sample, and the state parameters collected at each time interval are obtained for observation. When an abnormal situation occurs, an alarm is immediately sent and the equipment operation parameters are adjusted.

[0010] The operating parameters include fast-changing operating parameters and slow-changing operating parameters;

[0011] The fast-changing operating parameters include vibration parameters and voltage parameters;

[0012] The slowly varying operating parameters include temperature parameters and pressure parameters.

[0013] The preliminary state factor is collected in the following manner:

[0014] Based on the operation parameters, feature data is extracted through a convolutional neural network to construct a conditional generative adversarial network. , to operate on the category labels of the parameter data For conditions;

[0015] An integrated model consisting of multiple autoencoders with different structures is established. Each autoencoder extracts and compresses features of the original data or generated data from different angles. The encoding results of multiple autoencoders are fused, and the fused features are represented as preliminary state factors. .

[0016] The integrated model consists of a sparse autoencoder and a variational autoencoder;

[0017] The category label By setting normal threshold ranges for various operating parameters according to the design standards of mine equipment, different operating states of different operating parameters are obtained to obtain category labels. .

[0018] The state factor extraction process is:

[0019] Initial state factor Conduct further deep clustering optimization;

[0020] Extracting preliminary state factors through deep learning network Based on the high-level feature representation, a series of clustering variables are obtained, and the category labels are As an evaluation indicator of the state factor, through the objective function Implement clustering;

[0021] Assume the state factor is expressed as , state factor The information gain is , get the sum of state factor importance ;

[0022] Construct the objective function through the sum of state factor importances Construct the objective function , expressed as:

[0023]

[0024] Among them, represents the balance coefficient, is the clustering variable;

[0025] Cluster the preliminary state factors by minimizing the objective function to obtain the state factor .

[0026] The clustering variables include:

[0027] The compactness measure of clustering is , which measures the tightness of high-level feature representations within the same cluster;

[0028] The separability measure is , which measures the separation degree of high-level feature representations between different clusters;

[0029] The label consistency measure is , which measures the matching degree between the clustering result and the class label .

[0030] The process of obtaining the global representation of the state factor is as follows:

[0031] Capture the dependency relationships between state factors through the self-attention mechanism and gradually refine the features with the help of a multi-layer structure;

[0032] Map each state factor in the state factor to a vector and add positional encoding to reflect the order information to obtain the input sequence ;

[0033] Input the state factor sequentially into the model constructed by stacking multiple Transformer blocks based on the input sequence ;

[0034] In a single self-attention mechanism stage, for each state factor corresponding to the input sequence , obtain the attention weights ;

[0035] Obtain the query vector, key vector, and value vector through linear transformation, and perform weighted summation on the value vector according to the attention weights to obtain the attention output of a single head ;

[0036] Then, multiple single-headed attention outputs are concatenated and linearly transformed to obtain the output result . The output is input into a feed-forward neural network to transform and refine the features to obtain ;

[0037] By stacking multiple Transformer blocks, feature extraction and integration are performed to obtain the global representation of the state factor .

[0038] The process of obtaining the importance sequence representation is as follows:

[0039] Based on the global representation , a fully connected layer is added;

[0040] Let the weight matrix of the added fully connected layer be , and the bias vector be ;

[0041] An output value is obtained through the formula , and then the output value is input into the activation function . The activation function maps the output value to the optimal range, and this optimal range is the importance sequence representation .

[0042] The process of obtaining the optimal time interval is as follows:

[0043] Based on the importance sequence , it is divided into categories and sorted from high to low;

[0044] For each state factor corresponding to the importance sequence , the change rate is calculated;

[0045] Suppose the importance is divided into 3 categories, , and a change rate threshold is set;

[0046] Based on the change rate and the change rate threshold , the optimal time interval is obtained.

[0047] An intelligent adjustment method for mine status based on the Internet of Things includes:

[0048] S1: Collect mine equipment data based on Internet of Things technology and extract the operation parameters in the mine equipment data;

[0049] S2: Obtain the preliminary state factors from the operation parameters, and then perform deep clustering optimization on the preliminary state factors to extract the state factors;

[0050] S3: Based on the Transformer architecture, globally model the state factors to obtain the global representation of the state factors, and analyze the importance sequence representation of the state factors at different time scales based on the global representation;

[0051] S4: According to the importance sequence representation, select the best time interval for data collection for the state parameters corresponding to the state factors in each sample, and obtain the state parameters collected at each time interval for observation. When an abnormal situation occurs, immediately send an alarm and adjust the device operation parameters.

[0052] The present invention has the following beneficial effects:

[0053] In the present invention, first, the operation parameters are extracted from a large amount of collected data, making data processing more targeted. A large amount of data will be generated during the operation of mine equipment. By accurately extracting the operation parameters, the attention can be focused on the factors that directly affect the operation state of the equipment, reducing unnecessary data interference, and improving the efficiency and effectiveness of data analysis;

[0054] Secondly, the preliminary state factors are obtained from the operation parameters, completing the preliminary screening and integration of the original data, converting the operation parameters into state factors with greater analysis value, and laying a foundation for subsequent in-depth analysis. This helps to more accurately describe the operation state of the equipment and improve the ability to identify abnormal states of the equipment;

[0055] Then, based on the Transformer architecture, globally model the state factors, which can fully capture the mutual relationships and dependency information of the state factors at different time scales and spatial dimensions, and obtain the global representation of the state factors. This enables the model to understand the operation state of the equipment from a macroscopic perspective and avoid misjudgment caused by local information analysis. Analyzing the importance sequence representation of the state factors at different time scales based on the global representation can clarify the importance degree and role of each state factor in the operation of the equipment. This helps to focus resources and attention on key state parameters during equipment monitoring and maintenance, and preferentially process factors that have a greater impact on equipment operation, improving the accuracy and effectiveness of equipment management;

[0056] Finally, select the best collection time interval according to the importance of the state factors to achieve intelligent management of data collection. That is, high-frequency collection of key important factors and reduction of the collection frequency of non-key factors can better make targeted control decisions for mine equipment;

[0057] The importance of the collected data will not be affected by the differences in the change rates of different device states. When an abnormal situation occurs, the device operating parameters can be immediately adjusted, realizing the real-time monitoring and automatic control of the mine equipment. Description of the Drawings

[0058] Figure 1 It is a system block diagram of an intelligent mine state adjustment system and method based on the Internet of Things proposed by the present invention.

[0059] Figure 2 It is a method step diagram of an intelligent mine state adjustment system and method based on the Internet of Things proposed by the present invention. Detailed Embodiments

[0060] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0061] Embodiment 1: As Figure 1 shown, an intelligent mine state adjustment system based on the Internet of Things proposed by the present invention includes:

[0062] Data acquisition module: Collect mine equipment data based on Internet of Things technology and extract the operating parameters in the mine equipment data;

[0063] The operating parameters include fast-changing operating parameters and slow-changing operating parameters;

[0064] The fast-changing operating parameters include vibration parameters and voltage parameters, and the slow-changing operating parameters include temperature parameters and pressure parameters;

[0065] Specifically, the acquisition method for slow-changing parameters:

[0066] Under normal conditions, the temperature sensor collects data every 10 seconds. When the detected temperature change rate exceeds 0.2 °C per minute, the acquisition frequency is automatically increased to once every 2 seconds;

[0067] The pressure sensor normally collects data every 15 seconds. When the pressure change rate exceeds 0.2 MPa per minute, the acquisition frequency is adjusted to once every 3 seconds;

[0068] The collected data includes temperature values, pressure values, and corresponding timestamps;

[0069] The acquisition method for fast-changing parameters:

[0070] The initial acquisition frequency of the vibration sensor is set to , when the device is in startup, braking, overload or other abnormal working conditions, the change rate of characteristic parameters such as the peak value of the vibration signal is monitored in real time through the built-in intelligent algorithm. When the change rate exceeds the preset threshold, the acquisition frequency can be increased to a maximum of ;

[0071] The initial acquisition frequency of the current and voltage sensors is , when the operating state of the device changes, such as when the current fluctuation exceeds of the rated current or the voltage fluctuation exceeds of the rated voltage, the acquisition frequency is increased to ;

[0072] The collected data includes vibration data, current values, voltage values and corresponding timestamps.

[0073] State extraction module: Obtain the preliminary state factors from the operating parameters, and then perform deep clustering optimization on the preliminary state factors to extract the state factors;

[0074] The process of extracting the preliminary state factors is as follows:

[0075] Based on the operating parameters, extract the feature data through a convolutional neural network and construct a conditional generative adversarial network , with the class label of the operating parameter data (the label corresponding to different operating states of the device) as the condition;

[0076] Among them, the generator of the conditional generative adversarial network aims to learn to generate data samples that are similar in characteristics to the real operating parameter data but have diversity , is a random noise vector;

[0077] The discriminator of the conditional generative adversarial network is used to distinguish between real data and generated data ;

[0078] Specifically, the method for obtaining the class label is as follows:

[0079] According to the design standards of the mine equipment, set the normal threshold range for each operating parameter. For example:

[0080] For the current parameter of the underground motor, the current fluctuates within a certain range during normal operation. When the current exceeds the normal threshold, it is marked as "overload operation state";

[0081] When it is lower than the normal threshold, it is marked as "light load operation state";

[0082] When it is within the normal range, it is marked as "normal operation state";

[0083] For the temperature parameter, if the temperature of the key components of the device exceeds the safety threshold, it is marked as "high temperature abnormal state", and if it is within the safe range, it is marked as "normal temperature state".

[0084] Obtain the different operating states, i.e., class labels, of these different operating parameters.

[0085] Build an integrated model composed of multiple (5) autoencoders with different structures. Each autoencoder extracts and compresses features from the original data or generated data from different perspectives, and fuses the encoding results of multiple autoencoders. The fused features are represented as preliminary state factors , where i is the index of the number of preliminary state factors;

[0086] The process of extracting state factors is as follows:

[0087] Specifically, the integrated model consists of a sparse autoencoder and a variational autoencoder;

[0088] Perform further deep clustering optimization on the preliminary state factors ;

[0089] Let the state factor be represented as ;

[0090] Extract the high-level feature representation of the preliminary state factors through a deep learning network, obtain a series of clustering variables based on the high-level feature representation, and use the class label as the evaluation index of the state factor. Through the objective function achieve clustering;

[0091] Specifically, a series of clustering variables include:

[0092] The compactness measure of clustering is , which is used to measure the tightness of the high-level feature representations within the same cluster;

[0093] The separability measure is , which measures the separation degree of the high-level feature representations between different clusters;

[0094] The label consistency measure is , which measures the matching degree between the clustering result and the class label ;

[0095] Obtain the importance of the state factor, which is obtained by calculating the information gain index. Let the information gain of the state factor be , then the total importance of the state factors is: , since the information gain of the state factor is obtained based on the preliminary state factor obtained, so the total importance of the state factors contains elements, which have the same quantity as the preliminary state factor Same quantity;

[0096] Then, based on the total importance of the state factors Construct the objective function as follows: , where represents the balance coefficient;

[0097] By minimizing the objective function, cluster the preliminary state factors to obtain the final state factors , where is the quantity index of the state factors, which is the same as the quantity of the preliminary state factors. One preliminary state factor corresponds to one state factor optimized through clustering;

[0098] Specifically, during the clustering optimization process, the objective function will continuously adjust the clustering partition and the screening of the state factors. When the objective function converges to a certain extent or meets the preset stopping conditions, the obtained clustering results and the corresponding parameters are the optimized state factors.

[0099] Deep application module: Based on the Transformer architecture, globally model the state factors to obtain the global representation of the state factors, and analyze the importance sequence representation of the state factors at different time scales based on the global representation;

[0100] The process of globally modeling the state factors based on the Transformer architecture to obtain the global representation of the state factors is as follows:

[0101] Capture the dependency relationships between state factors through the self-attention mechanism and gradually refine the features with the help of a multi-layer structure;

[0102] Map each state factor in the state factors to a vector, and add position encoding to reflect the order information to obtain the input sequence , where the number of input sequences is the same as the number of state factors, both being ;

[0103] Input the state factors sequentially into the model constructed by stacking multiple (12) Transformer blocks based on the input sequence ;

[0104] The core components of each Transformer block consist of a multi-head self-attention mechanism and a feed-forward neural network, that is, one Transformer corresponds to one head of the multi-head self-attention mechanism;

[0105] In a single self-attention mechanism stage, for the state factor corresponding to each input sequence , obtain the attention weights ;

[0106] Obtain query vectors, key vectors, and value vectors through linear transformation, and perform weighted summation on the value vectors according to the attention weights to obtain the attention output of a single head ;

[0107] Then concatenate the attention outputs of multiple (12) single heads and perform linear transformation on the concatenated result to obtain the output result , and input the output into a feed-forward neural network to further transform and refine the features, which is expressed by the formula:

[0108]

[0109] where, represents the feed-forward neural network, denotes the activation function, is the weight matrix of the first layer of the feed-forward neural network, with a dimension of , is the weight matrix of the second layer of the feed-forward neural network, with a dimension of , is the bias vector of the first layer of the feed-forward neural network, with a dimension of , is the bias vector of the second layer of the feed-forward neural network, with a dimension of ;

[0110] Furthermore, represents the model dimension, that is, the dimension of the feature vector of the data (the output of the multi-head self-attention mechanism) input to the feed-forward neural network, represents the intermediate layer dimension of the feed-forward neural network, and ;

[0111] Then, by stacking multiple Transformer blocks, continuously perform feature extraction and integration on to finally obtain the global representation of the state factor . Similarly, one state factor corresponds to one global representation of the state factor, so the quantities are both ;

[0112] Specifically, by taking the output of the previous Transformer block as the input of the next Transformer block, the features of the state factors are further extracted and integrated after passing through each Transformer block. For example, after layer normalization and residual connection of the output of the first Transformer block, it is used as the input of the second Transformer block, and the calculation processes of the multi-head self-attention mechanism and the feed-forward neural network are repeated to obtain , and so on. As the Transformer blocks are stacked, the model can continuously optimize the relationships between the state factors and finally obtain the global representation of the state factors ;

[0113] Based on the global representation The process of analyzing the importance sequence representation of the state factors at different time scales is as follows:

[0114] After obtaining the global representation , a fully connected layer is added to analyze the importance of the state factors;

[0115] Specifically, since the global representation integrates the comprehensive information of all state factors at different time scales, analyzing the importance of the state factors is to analyze from this global representation the importance of each state factor at different time scales, presented in the form of a sequence, which is convenient for subsequent targeted processing according to the importance degree of the state factors, such as control strategy formulation, fault warning, etc.;

[0116] Let the weight matrix of the added fully connected layer be , and the bias vector be ;

[0117] Among them, the dimension of the weight matrix of the fully connected layer is , where is the output dimension of the fully connected layer, and (the number of state factors, in order to obtain the importance score corresponding to each state factor), The dimension of is

[0118] Through the formula an output value is obtained, and then the output value is input into the activation function . The activation function maps the output value to the optimal range, and this optimal range is the importance sequence representation . One global representation corresponds to one importance sequence representation, so the quantity is ;

[0119] Specifically, the sum of all the importance sequences output through the above steps is 1. It is the proportion of the th state factor in the importance of all state factors. The higher the proportion, the more important the state factor.

[0120] Decision control module: According to the representation of the importance sequence, select the best time interval for data acquisition for the state parameters corresponding to the state factors in each sample, and obtain the state parameters collected at each time interval for observation. When an abnormal situation occurs, immediately send an alarm and adjust the device operation parameters.

[0121] The process of obtaining the best time interval is as follows:

[0122] Based on the importance sequence , it is divided into categories, and sorted from high to low, denoted as , where the importance ranking is , each category contains several values of the importance sequence, and the total number of all importance sequences is ;

[0123] For each state factor corresponding to the importance sequence , calculate its change rate over a period of time. The calculation process is as follows:

[0124]

[0125] Among them, is the value of the state factor at time step , is the value of the state factor at time step , and the change rate of the state factor is obtained;

[0126] For the change rates of state factors in different categories corresponding to the devices and their change rate ranges, establish a mapping rule for the time interval. Suppose the importance is divided into 3 categories, namely , and at the same time set a change rate threshold . Based on the change rate and the change rate threshold , obtain the best time interval. Specifically:

[0127] For the state factors corresponding to the devices whose importance sequence belongs to (high) and the change rate is greater than the change rate threshold , select the first time interval Data collection is carried out at the shortest time interval of 5 seconds to 1 minute (high importance and high change rate).

[0128] For the devices corresponding to the status factors with importance belonging to (high) and the change rate less than or equal to the change rate threshold select the second time interval Data collection is carried out at a shorter time interval of 1 to 5 minutes (high importance and low change rate).

[0129] For the devices corresponding to the status factors with importance belonging to (medium) and the change rate greater than the change rate threshold select the third time interval Data collection is carried out at a shorter time interval of 5 to 10 minutes (medium importance and high change rate).

[0130] For the devices corresponding to the status factors with importance belonging to (medium) and the change rate less than or equal to the change rate threshold select the fourth time interval Data collection is carried out at a longer time interval of 10 to 15 minutes (medium importance and low change rate).

[0131] For the devices corresponding to the status factors with importance belonging to (low), regardless of the change rate select the fifth time interval Data collection is carried out at a long time interval of 20 minutes (low importance and any change rate).

[0132] Specifically, the importance of the status factors is divided into 3 categories ( high, medium, low). This classification method helps to distinguish the influence degree of different status factors on the operation of the device or system. For example, in the power system, parameters such as voltage and current of key devices can be classified as category, the operation status parameters of some auxiliary devices can be classified as category, and the parameters with less influence on the system operation such as the surface temperature of the device can be classified as ;

[0133] According to the above mapping rules, for each device corresponding to each state factor in each sample, considering its importance category and change rate, determine the corresponding acquisition time interval. For devices corresponding to state factors with high importance and fast changes, select a shorter time interval for high-frequency acquisition. For devices corresponding to state factors with low importance and slow changes, select a longer time interval. Monitor the parameter data for anomalies based on the time interval and control the device according to the monitoring results of the data;

[0134] Based on the anomalies of different state factors and their influence degrees on the device operation, formulate corresponding device control strategies;

[0135] For example, when parameters such as voltage and current of key devices ( type state factors) abnormally increase and exceed the safety threshold, first give an early warning and reduce the device operation load, and at the same time conduct a safety inspection; when a certain parameter of an auxiliary device ( type state factors) appears abnormal, first give an early warning and record it, and then take corresponding measures; when the surface temperature of the device ( type state factors) shows an upward trend but does not affect the overall operation, first give an early warning and record it, and continuously monitor its change trend.

[0136] Embodiment 2: As Figure 2 shown, an intelligent adjustment method for mine status based on the Internet of Things includes:

[0137] S1: Collect mine device data based on Internet of Things technology and extract operation parameters from the mine device data;

[0138] S2: Obtain preliminary state factors from the operation parameters, and then perform deep clustering optimization on the preliminary state factors to extract state factors;

[0139] S3: Perform global modeling on the state factors based on the Transformer architecture to obtain the global representation of the state factors, and analyze the importance sequence representation of the state factors at different time scales based on the global representation;

[0140] S4: According to the importance sequence representation, select the best time interval for data acquisition for the state parameters corresponding to the state factors in each sample, and obtain the state parameters collected at each time interval for observation. When an abnormal situation occurs, immediately send an alarm and adjust the device operation parameters.

[0141] In the application, several formulas involved are calculated by taking the numerical values after dimensionless. The establishment of the formulas is obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. Some coefficients or weights in the formulas are set by those skilled in the art according to the actual situation, so no more details will be described here.

[0142] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those of ordinary skill in the art will appreciate that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution.

[0143] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art will understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent mine status adjustment system based on the Internet of Things, characterized in that: include: Data acquisition module: collects mine equipment data based on the Internet of Things technology and extracts operating parameters from the mine equipment data; State extraction module: obtains preliminary state factors from operating parameters, and then performs deep clustering optimization on the preliminary state factors to extract state factors; Deep Application Module: Globally model the state factors based on the Transformer architecture, obtain the global representation of the state factors, and analyze the importance sequence representation of the state factors at different time scales based on the global representation; The process of obtaining the global representation of the state factor is as follows: The self-attention mechanism is used to capture the dependencies between state factors, and the multi-layer structure is used to gradually refine the features. Each state factor is mapped to a vector, and the position code is added to reflect the order information to obtain the input sequence ; State Factor Based on the input sequence The input sequence is sequentially fed into a model constructed by stacking multiple Transformer blocks in sequence. In a single self-attention mechanism stage, the state factor corresponding to each input sequence is , get the attention weight ; The query vector, key vector and value vector are obtained through linear transformation, and the attention weights are calculated. Perform weighted summation on the value vector to get the attention output of a single head ; Then output multiple single-head attention Splice and linearly transform to get the output result , the output Input into the feedforward neural network to transform and refine the features ; By stacking multiple Transformer blocks, Perform feature extraction and integration to obtain a global representation of the state factor ; The acquisition process of the importance sequence representation is: Based on global representation , add a fully connected layer; Suppose the weight matrix of the added fully connected layer is , the bias vector is ; By formula Get an output value, and then input the output value into the activation function In the activation function Map the output value to the optimal range, which is the importance sequence representation ; Decision control module: According to the importance sequence representation, the optimal time interval is selected for data collection for the state parameters corresponding to the state factors in each sample, and the state parameters collected at each time interval are obtained for observation. When an abnormal situation occurs, an alarm is immediately sent and the equipment operation parameters are adjusted.

2. According to the Internet of Things-based intelligent mine status adjustment system of claim 1, it is characterized in that: The operating parameters include fast-changing operating parameters and slow-changing operating parameters; The fast-changing operating parameters include vibration parameters and voltage parameters; The slowly varying operating parameters include temperature parameters and pressure parameters.

3. The intelligent mine status adjustment system based on the Internet of Things according to claim 1 is characterized in that: The collection method of the preliminary state factor is: Based on the operation parameters, feature data is extracted through a convolutional neural network to construct a conditional generative adversarial network. , to operate on the category labels of the parameter data For conditions; An integrated model consisting of multiple autoencoders with different structures is established. Each autoencoder extracts and compresses features of the original data or generated data from different angles. The encoding results of multiple autoencoders are fused, and the fused features are represented as preliminary state factors. .

4. The intelligent mine state adjustment system based on the Internet of Things according to claim 3 is characterized in that: The integrated model consists of a sparse autoencoder and a variational autoencoder; The category label By setting normal threshold ranges for various operating parameters according to the design standards of mine equipment, different operating states of different operating parameters are obtained to obtain category labels. .

5. The intelligent mine state adjustment system based on the Internet of Things according to claim 1 is characterized in that: The state factor extraction process is: Initial state factor Conduct further deep clustering optimization; Extracting preliminary state factors through deep learning network Based on the high-level feature representation, a series of clustering variables are obtained, and the category labels are As an evaluation indicator of the state factor, through the objective function Implement clustering; Assume the state factor is expressed as , state factor The information gain is , get the sum of state factor importance ; By sum of state factor importance Constructing the objective function , expressed as: ； in, represents the balance coefficient, is the clustering variable; By minimizing the objective function, the initial state factor Perform clustering to obtain the state factor .

6. The intelligent mine state adjustment system based on the Internet of Things according to claim 5 is characterized in that: The clustering variables include: The compactness measure of clustering is , measures the closeness of high-level feature representation within the same cluster; The separation measure is , measures the degree of separation of high-level feature representations between different clusters; The label consistency metric is , measure the clustering results and class labels degree of match.

7. The intelligent mine status adjustment system based on the Internet of Things according to claim 1 is characterized in that: The process of obtaining the optimal time interval is as follows: Based on importance sequence , divided into categories and sort them from high to low; For each importance sequence Corresponding state factor, calculate the rate of change ; Assume that importance is divided into three categories: , set a change rate threshold ; Based on the rate of change and change rate threshold Get the best time interval.

8. A method for intelligently adjusting the state of a mine based on the Internet of Things, using any system described in claims 1 to 7, characterized in that: include: S1: Collect mine equipment data based on the Internet of Things technology and extract operating parameters from the mine equipment data; S2: Obtain preliminary state factors from the operating parameters, and then perform deep clustering optimization on the preliminary state factors to extract state factors; S3: Globally model the state factor based on the Transformer architecture, obtain the global representation of the state factor, and analyze the importance sequence representation of the state factor at different time scales based on the global representation; S4: According to the importance sequence representation, select the best time interval for data collection for the state parameters corresponding to the state factors in each sample, and obtain the state parameters collected at each time interval for observation. When an abnormal situation occurs, send an alarm immediately and adjust the equipment operation parameters.

Images