Intelligent control system and method for digital coal yard

Abstract

The embodiment of the invention discloses a digital coal yard intelligent control system and method, and the method comprises the steps: collecting the environment data (temperature value, humidity value and inflammable gas concentration value) of each key position through distributed sensors disposed at a plurality of key positions of a coal yard; carrying out data preprocessing and time sequence implicit association on the environment data of each key position by adopting a data processing technology based on artificial intelligence, and then carrying out generality-guided combustion-supporting factor time sequence collaboration on a key position temperature time sequence implicit association feature and a key position inflammable gas concentration time sequence implicit association feature; therefore, according to the combustion-supporting-combustion suppression confrontation representation between the key position combustion-supporting factor time sequence representation and the key position humidity time sequence implicit association, the explosion risk of the key position is intelligently judged, and a corresponding early warning prompt is generated. In this way, efficient and intelligent management of the coal yard environment is achieved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent control, and more specifically, relates to a digital coal yard intelligent control system and method. Background Art

[0002] Coal yards are important places for coal storage and transportation, but due to the special properties of coal, coal yards face many safety hazards, especially the risk of fire and explosion. Coal may spontaneously combust due to oxidation reactions during storage. At the same time, there may be flammable gases (such as methane, carbon monoxide, etc.) in coal yards. These gases are prone to explosions when exposed to open flames or high temperatures at a certain concentration. Therefore, the safety management of coal yards is crucial, and monitoring and early warning are required to prevent accidents.

[0003] However, traditional coal yard safety management mainly relies on manual inspections and simple monitoring by fixed sensors. Specifically, manual inspections usually require inspectors to regularly inspect the coal yard, but due to the large area of ​​the coal yard, it is difficult for inspectors to cover all key locations and it is easy to miss potential danger points, especially remote or difficult-to-reach locations, such as inside coal piles and high-risk areas. In addition, the frequency of manual inspections is limited by human resources and time arrangements, and it is usually impossible to achieve all-weather, real-time monitoring, resulting in the inability to timely discover and deal with safety hazards during non-inspection periods. Once an emergency occurs, the inspectors may not be able to respond quickly, delaying the best time to deal with it. Although fixed sensors can monitor in real time, their alarm mechanism is relatively simple, and usually requires manual confirmation before taking action. The response speed is slow, and the number of sensors is limited. It cannot fully cover the entire coal yard and is prone to missed or false alarms.

[0004] Therefore, an intelligent control solution for a digital coal yard is desired. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art, and discloses a digital coal yard intelligent control system and method.

[0006] In a first aspect, an embodiment of the present invention provides a digital coal yard intelligent control system, which includes:

[0007] An environmental data acquisition module is used to collect environmental data of multiple key locations through distributed sensors deployed at the coal yard, wherein the environmental data includes temperature, humidity, and flammable gas concentration.

[0008] An environmental data processing and encoding module, used for performing data preprocessing and sequence encoding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature;

[0009] A key position temperature concentration compensation interaction module is used to perform a combustion-supporting factor time series collaborative analysis based on common feature guidance on the key position temperature time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature to obtain a key position combustion-supporting factor time series representation;

[0010] A key position combustion-supporting-flammable confrontation analysis module is used to use the key position humidity time series implicit correlation feature as the key position ignition suppression factor time series representation, and input the key position combustion-supporting factor time series representation and the key position ignition suppression factor time series representation into the combustion-supporting-flammable confrontation analysis network to obtain the combustion-supporting-flammable confrontation representation;

[0011] The explosion risk judgment module is used to obtain an analysis result of whether the risk of explosion exceeds a preset threshold based on the combustion-supporting-suppressing time sequence confrontation representation, and to judge whether to generate an explosion warning prompt based on the analysis result.

[0012] In some possible embodiments, the environmental data processing encoding module includes:

[0013] A key position environmental data preprocessing unit, used for preprocessing the environmental data of each key position according to the time dimension to obtain a time queue of key position temperature values, a time queue of key position humidity values ​​and a time queue of key position flammable gas concentration values;

[0014] The key position environmental data sequence encoding unit is used to input the time queue of the key position temperature value, the time queue of the key position humidity value and the time queue of the key position flammable gas concentration value into a sequence encoder based on a forward LSTM model respectively to obtain a key position temperature time series implicit association feature vector as the key position temperature time series implicit association feature, a key position humidity time series implicit association feature vector as the key position humidity time series implicit association feature and a key position flammable gas concentration time series implicit association feature vector as the key position flammable gas concentration time series implicit association feature.

[0015] In some possible embodiments, the key position temperature concentration compensation interaction module includes:

[0016] A temperature-concentration time series commonality extraction unit is used to input the key position temperature time series implicit association feature vector and the key position flammable gas concentration time series implicit association feature vector into a commonality feature extraction network to obtain a commonality feature representation vector between key position temperature-concentration time series features;

[0017] The temperature-concentration fine-grained compensation interaction unit is used to perform feature fine-grained compensation interaction on the key position temperature time series implicit association feature vector and the key position flammable gas concentration time series implicit association feature vector based on the common feature representation vector between the key position temperature-concentration time series features to obtain the key position combustion-supporting factor time series representation.

[0018] In some possible embodiments, the temperature concentration fine-grained compensation interaction unit includes:

[0019] The temperature concentration time series feature probabilization subunit is used to input the key position temperature time series implicit association feature vector, the key position flammable gas concentration time series implicit association feature vector and the key position temperature-concentration time series feature common feature representation vector into the probabilization unit to obtain the probabilistic key position temperature time series implicit association feature vector, the probabilistic key position flammable gas concentration time series implicit association feature vector and the probabilistic key position temperature-concentration time series feature common feature representation vector;

[0020] A compensation vector calculation subunit, used to respectively calculate the fine-grained compensation vectors of the probabilistic key position temperature time series implicit association feature vector and the probabilistic key position flammable gas concentration time series implicit association feature vector relative to the probabilistic key position temperature-concentration time series feature common feature representation vector to obtain the key position temperature fine-grained compensation vector and the key position flammable gas concentration fine-grained compensation vector;

[0021] The temperature concentration characteristic compensation subunit is used to perform characteristic compensation on the key position temperature time series implicit correlation characteristic vector and the key position flammable gas concentration time series implicit correlation characteristic vector based on the key position temperature fine-grained compensation vector and the key position flammable gas concentration fine-grained compensation vector to obtain the key position temperature fine-grained characteristic compensation vector and the key position flammable gas concentration fine-grained characteristic compensation vector;

[0022] The temperature-concentration feature multidimensional fusion subunit is used to cascade the key position temperature fine-grained feature compensation vector, the key position flammable gas concentration fine-grained feature compensation vector and the key position temperature-concentration time series feature common feature representation vector to obtain the key position combustion-supporting factor time series representation vector as the key position combustion-supporting factor time series representation.

[0023] In some possible embodiments, the temperature-concentration time series feature probabilization subunit is used to: input the key position temperature time series implicit association feature vector, the key position flammable gas concentration time series implicit association feature vector and the key position temperature-concentration time series feature common feature representation vector into the Sigmoid function for probabilistic processing to obtain the probabilistic key position temperature time series implicit association feature vector, the probabilistic key position flammable gas concentration time series implicit association feature vector and the probabilistic key position temperature-concentration time series feature common feature representation vector.

[0024] In some possible embodiments, the temperature concentration characteristic compensation subunit is used to:

[0025] Inputting the key position temperature fine-grained compensation vector and the key position temperature time series implicit correlation feature vector into a fine-grained compensation module to obtain the key position temperature fine-grained feature compensation vector;

[0026] The key position flammable gas concentration fine-grained compensation vector and the key position flammable gas concentration time series implicit correlation feature vector are input into the fine-grained compensation module to obtain the key position flammable gas concentration fine-grained feature compensation vector.

[0027] In some possible embodiments, the key position combustion-supporting-flammable confrontation analysis module is used to: use the key position humidity time series implicitly associated characteristic vector as the key position combustion suppression factor time series representation vector and the transposed vector of the key position humidity time series implicitly associated characteristic vector to obtain the combustion-supporting-flammable confrontation representation matrix as the combustion-supporting-flammable confrontation representation.

[0028] In some possible embodiments, the explosion risk judgment module is used to:

[0029] Inputting the combustion-supporting-suppressing time series confrontation representation matrix into a classifier-based intelligent analysis module to obtain the analysis result, wherein the analysis result is used to indicate whether the risk of explosion exceeds a preset threshold;

[0030] In response to the analysis result that the risk of explosion exceeds a preset threshold, the explosion warning prompt is generated.

[0031] In a second aspect, an embodiment of the present invention provides a digital coal yard intelligent control method, which includes:

[0032] Collect environmental data of each key location through distributed sensors deployed at multiple key locations in the coal yard, wherein the environmental data includes temperature value, humidity value, and flammable gas concentration value;

[0033] Performing data preprocessing and sequence coding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature;

[0034] Performing a time series collaborative analysis of combustion-supporting factors guided by common features on the key position temperature time series implicit correlation features and the key position flammable gas concentration time series implicit correlation features to obtain a time series representation of the key position combustion-supporting factors;

[0035] Taking the key position humidity time series implicit correlation feature as the key position ignition suppression factor time series representation, inputting the key position combustion supporting factor time series representation and the key position ignition suppression factor time series representation into the combustion supporting-flammable confrontation analysis network to obtain the combustion supporting-ignition suppression time series confrontation representation;

[0036] Based on the combustion-supporting-suppressing time sequence confrontation representation, an analysis result of whether the risk of explosion exceeds a preset threshold is obtained, and based on the analysis result, it is determined whether to generate an explosion warning prompt.

[0037] In some possible embodiments, the environmental data of each key position is preprocessed and sequence-encoded to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature, including:

[0038] Preprocessing the environmental data of each key position according to the time dimension to obtain a time queue of temperature values ​​at the key position, a time queue of humidity values ​​at the key position, and a time queue of flammable gas concentration values ​​at the key position;

[0039] The time queue of the key position temperature values, the time queue of the key position humidity values, and the time queue of the key position flammable gas concentration values ​​are respectively input into a sequence encoder based on a forward LSTM model to obtain a key position temperature time series implicit association feature vector as the key position temperature time series implicit association feature, a key position humidity time series implicit association feature vector as the key position humidity time series implicit association feature, and a key position flammable gas concentration time series implicit association feature vector as the key position flammable gas concentration time series implicit association feature.

[0040] Compared with the prior art, the embodiment of the present invention provides a digital coal yard intelligent control system and method, which collects environmental data (temperature value, humidity value, flammable gas concentration value) of each key position through distributed sensors deployed at multiple key positions of the coal yard, and uses data encoding and processing technology based on artificial intelligence to perform data preprocessing and implicit time series association on the environmental data of each key position, and then performs common-guided combustion factor time series coordination on the key position temperature time series implicit association characteristics and the key position flammable gas concentration time series implicit association characteristics, so as to intelligently judge the risk of explosion at the key position according to the combustion factor time series representation of the key position and the combustion-suppression confrontation representation between the key position humidity time series implicit association and generate corresponding early warning prompts. In this way, by deploying distributed sensors at multiple key positions of the coal yard, the entire coal yard can be fully covered to ensure that there is no monitoring blind spot. In addition, these sensors can collect environmental data in real time. Once an abnormal situation is detected, the system can immediately generate an early warning prompt to help managers take measures quickly, thereby realizing efficient and intelligent management of the coal yard environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0042] Figure 1 4 is a block diagram of a digital coal yard intelligent control system according to an embodiment of the present invention.

[0043] Figure 2 The block diagram is a block diagram of an environmental data processing and encoding module in a digital coal yard intelligent control system according to an embodiment of the present invention.

[0044] Figure 3 The block diagram is a key position temperature concentration compensation interaction module in a digital coal yard intelligent control system according to an embodiment of the present invention.

[0045] Figure 4 The figure is a flow chart of a digital coal yard intelligent control method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0047] Unless otherwise specified, the technical terms or scientific terms used in the embodiments of the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. "Including" or "comprising" used in the embodiments of the present invention neither limit the shapes, numbers, steps, actions, operations, components, originals and / or their groups mentioned, nor exclude the appearance or addition of one or more other different shapes, numbers, steps, actions, operations, components, originals and / or their groups, or the addition of these. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number and order of the indicated technical features. Thus, the features defined as "first" and "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0048] Unless otherwise specifically stated, the relative arrangement of the components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention. At the same time, it should be understood that for ease of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship, and the techniques, methods and devices known to ordinary technicians in the relevant fields may not be discussed in detail, but where appropriate, the techniques, methods and devices shown should be considered as part of the authorized specification. In all examples shown and discussed here, any specific other examples may have different values. It should be noted that similar symbols and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0049] In the description of the embodiments of the present invention, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In the embodiments of the present invention, the schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in the embodiments of the present invention and the features of the different embodiments or examples, without contradiction.

[0050] Coal yards are important places for coal storage and transportation, but due to the special properties of coal, coal yards face many safety hazards, especially the risk of fire and explosion. Coal may spontaneously combust due to oxidation reactions during storage. At the same time, there may be flammable gases (such as methane, carbon monoxide, etc.) in coal yards. These gases are prone to explosions when exposed to open flames or high temperatures at a certain concentration. Therefore, the safety management of coal yards is crucial, and monitoring and early warning are required to prevent accidents.

[0051] However, traditional coal yard safety management mainly relies on manual inspections and simple monitoring by fixed sensors. Specifically, manual inspections usually require inspectors to regularly inspect the coal yard, but due to the large area of ​​the coal yard, it is difficult for inspectors to cover all key locations and it is easy to miss potential danger points, especially remote or difficult-to-reach locations, such as inside coal piles and high-risk areas. In addition, the frequency of manual inspections is limited by human resources and time arrangements, and it is usually impossible to achieve all-weather, real-time monitoring, resulting in the inability to timely discover and deal with safety hazards during non-inspection periods. Once an emergency occurs, the inspectors may not be able to respond quickly, delaying the best time to deal with it. Although fixed sensors can monitor in real time, their alarm mechanism is relatively simple, and usually requires manual confirmation before taking action. The response speed is slow, and the number of sensors is limited. It cannot fully cover the entire coal yard and is prone to missed or false alarms.

[0052] Therefore, in view of the above technical problems, the technical concept of the present invention is to collect the environmental data (temperature value, humidity value, flammable gas concentration value) of each key position through distributed sensors deployed at multiple key positions of the coal yard, and use data encoding and processing technology based on artificial intelligence to perform data preprocessing and implicit temporal association on the environmental data of each key position, and then perform common-guided combustion factor temporal coordination on the implicit temporal association characteristics of the key position temperature and the implicit temporal association characteristics of the flammable gas concentration at the key position, so as to intelligently judge the risk of explosion at the key position according to the combustion factor temporal representation of the key position and the combustion-suppression confrontation between the key position humidity temporal implicit association and generate corresponding early warning prompts. In this way, by deploying distributed sensors at multiple key positions of the coal yard, the entire coal yard can be fully covered to ensure that there is no monitoring blind spot. In addition, these sensors can collect environmental data in real time. Once an abnormal situation is detected, the system can immediately generate an early warning prompt to help managers take measures quickly, thereby realizing efficient and intelligent management of the coal yard environment.

[0053] Figure 1 FIG. 1 is a block diagram of a digital coal yard intelligent control system according to an embodiment of the present invention. Figure 1As shown, according to an embodiment of the present invention, the digital coal yard intelligent control system 100 includes: an environmental data acquisition module 110, which is used to collect environmental data of each key position through distributed sensors deployed at multiple key positions of the coal yard, and the environmental data includes temperature value, humidity value, and flammable gas concentration value; an environmental data processing and encoding module 120, which is used to perform data preprocessing and sequence encoding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature; a key position temperature concentration compensation interaction module 130, which is used to perform data preprocessing and sequence encoding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature; The implicit correlation characteristics of the flammable gas concentration time series at key positions are used to perform a coordinated analysis of the combustion-supporting factor time series based on the common characteristics to obtain the combustion-supporting factor time series representation at key positions; the key position combustion-supporting-flammable confrontation analysis module 140 is used to use the implicit correlation characteristics of the humidity time series at key positions as the combustion-suppressing factor time series representation at key positions, and input the combustion-supporting factor time series representation at key positions and the combustion-suppressing factor time series representation at key positions into the combustion-supporting-flammable confrontation analysis network to obtain the combustion-supporting-suppressing time series confrontation representation; the explosion risk judgment module 150 is used to obtain an analysis result of whether the risk of explosion exceeds a preset threshold based on the combustion-suppressing time series confrontation representation, and based on the analysis result, judge whether to generate an explosion warning prompt.

[0054] In the embodiment of the present invention, the environmental data acquisition module 110 is used to collect environmental data of each key position through distributed sensors deployed at multiple key positions of the coal yard, and the environmental data includes temperature value, humidity value, and flammable gas concentration value. It should be understood that in the coal yard environment, temperature value, humidity value and flammable gas concentration value are important indicators for measuring safety conditions. Specifically, the temperature value refers to the temperature of the air inside or outside the coal yard, or the temperature of the surface and inside of the coal pile. High temperature can accelerate the oxidation reaction of coal, leading to spontaneous combustion. When the temperature in the coal pile reaches a certain level, it may spontaneously catch fire. Therefore, monitoring the temperature can help to detect overheating areas early, so as to take cooling or other measures to prevent the occurrence of fire. The humidity value refers to the water vapor content in the air, usually expressed as relative humidity (RH), that is, the percentage of the actual amount of water vapor contained in the air at a certain temperature to the saturated water vapor at that temperature. Too high or too low humidity may affect the properties of coal. Too high humidity may cause coal to absorb moisture and agglomerate, increasing the difficulty of transportation and handling; too low humidity may cause coal dust to fly, increasing the risk of fire and explosion. Therefore, appropriate humidity helps to maintain the stability and safety of the coal yard. The flammable gas concentration value refers to the concentration of a series of flammable gases (such as methane, carbon monoxide, etc.) that may exist in the coal yard. These gases may come from the natural decomposition, oxidation process or other chemical reactions of coal. Flammable gases are prone to explosion when exposed to open flames or high temperatures within a certain concentration range, so monitoring their concentration is the key to preventing explosion accidents. Based on this, in order to comprehensively and comprehensively consider the factors of explosion, in the technical solution of the present invention, environmental data of each key location is collected by distributed sensors deployed at multiple key locations of the coal yard, and the environmental data includes temperature values, humidity values, and flammable gas concentration values.

[0055] To achieve this function, advanced sensor technology and network communication technology are used. First of all, the selection and deployment of sensors are the basis of the entire system. Various types of sensors are installed in key locations in the coal yard, such as around the coal pile, inside the coal pile, near the vents, and in high-risk areas. These sensors have high sensitivity and high reliability and can work stably in harsh industrial environments. For example, temperature sensors can accurately measure temperature changes inside and outside the coal pile, humidity sensors are used to monitor the moisture content in the air, and flammable gas sensors can detect the concentration of flammable gases such as methane and carbon monoxide in real time.

[0056] In the process of sensor deployment, it is necessary to comprehensively consider the geographical distribution of the coal yard, the structural characteristics of the coal pile, and potential risk points. For example, inside the coal pile, embedded or inserted temperature sensors can be used. These sensors are connected to external collection equipment through cables and can monitor the temperature changes inside the coal pile in real time. Around the coal pile, multiple temperature and humidity sensors can be arranged. These sensors can not only monitor the temperature and humidity in the air, but also transmit data to the central processing unit through wireless communication technology. In addition, highly sensitive flammable gas sensors are installed near vents and high-risk areas. These sensors can detect gas leaks in time to prevent potential explosion risks.

[0057] In order to ensure the accuracy and reliability of the data, a multi-layer verification and correction mechanism is adopted. First, the sensor itself has a self-calibration function, which can automatically calibrate before each data collection to ensure the accuracy of the data. Secondly, the system adopts a redundant design, that is, multiple sensors of the same type are deployed at key locations, and the data of multiple sensors are comprehensively analyzed through data fusion technology to eliminate the errors that may occur in a single sensor. For example, if the temperature values ​​measured by multiple temperature sensors at the same location differ greatly, the system will automatically select the data closest to the average value as the final result, thereby improving the reliability of the data.

[0058] In terms of data transmission, a variety of communication technologies are used, including wired communication and wireless communication. Wired communication is mainly used for sensors inside coal piles and fixed positions, and transmits data to the central processing unit through cables. Wireless communication is used for sensors around coal piles and mobile positions, which transmit data to the central processing unit through wireless networks (such as Wi-Fi, LoRa, Zigbee, etc.). The advantage of wireless communication technology is its high flexibility and can cover a wider area, especially in large coal yards, which can effectively reduce wiring costs and maintenance difficulties.

[0059] In order to ensure the stability and security of data transmission, a variety of data transmission protocols and encryption technologies are also used. For example, the system can use the MQTT (Message Queuing Telemetry Transport) protocol, which is a lightweight message transmission protocol that is particularly suitable for communication between IoT devices. The MQTT protocol supports the publish / subscribe model, which can achieve efficient data transmission and low-latency response. In addition, the system also uses the TLS (Transport Layer Security) protocol to encrypt data transmission to ensure that the data is not stolen or tampered with during transmission.

[0060] During the data collection process, it also has automatic diagnosis and fault alarm functions. When a sensor fails or the data is abnormal, the system will automatically issue an alarm and record the relevant information in the log, which is convenient for maintenance personnel to carry out timely maintenance. For example, if the data measured by a temperature sensor exceeds the normal range for many consecutive times, the system will automatically mark the sensor and generate a fault report to remind maintenance personnel to check and replace it. This automated fault diagnosis and alarm mechanism greatly improves the reliability and maintenance efficiency of the system.

[0061] In an embodiment of the present invention, the environmental data processing and encoding module 120 is used to perform data preprocessing and sequence encoding on the environmental data of each key position to obtain the key position temperature time series implicit correlation characteristics, the key position humidity time series implicit correlation characteristics and the key position flammable gas concentration time series implicit correlation characteristics. Figure 2 FIG. 1 is a block diagram of an environmental data processing and encoding module in a digital coal yard intelligent control system according to an embodiment of the present invention. Specifically, Figure 2As shown, the environmental data processing and encoding module 120 includes: a key position environmental data preprocessing unit 121, which is used to perform data preprocessing on the environmental data of each key position according to the time dimension to obtain a time queue of key position temperature values, a time queue of key position humidity values, and a time queue of key position flammable gas concentration values; a key position environmental data sequence encoding unit 122, which is used to input the time queue of key position temperature values, the time queue of key position humidity values, and the time queue of key position flammable gas concentration values ​​into a sequence encoder based on a forward LSTM model to obtain a key position temperature time series implicit association feature vector as the key position temperature time series implicit association feature, a key position humidity time series implicit association feature vector as the key position humidity time series implicit association feature, and a key position flammable gas concentration time series implicit association feature vector as the key position flammable gas concentration time series implicit association feature. Accordingly, considering that each environmental parameter in the environmental data of each key position has a trend of change and fluctuation over time, and each environmental parameter has mutual influence and correlation on the time scale, therefore, in order to be able to analyze each environmental parameter independently in a more detailed and centralized manner, in the technical solution of the present invention, the environmental data of each key position are preprocessed according to the time dimension to obtain a time queue of the temperature value at the key position, a time queue of the humidity value at the key position, and a time queue of the flammable gas concentration value at the key position. In this way, the trend of each environmental parameter over time can be observed and analyzed. For example, whether the temperature value continues to rise within a specific time period, whether the flammable gas concentration suddenly increases under certain conditions, etc. Then, considering that in the coal yard environment, some safety risks may require long-term historical data to be accurately predicted. For example, the gradual increase in temperature may take several days or even weeks to reach a dangerous level. That is, the time queues of each environmental parameter have different degrees of time correlation and change trends between different time spans. And considering that the forward LSTM model is used to process the long-term dependency problem in sequence data. That is, the traditional RNN is prone to gradient vanishing or gradient exploding problems when processing long sequences, while the forward LSTM can effectively remember and utilize long-distance information by introducing cell states and gating mechanisms (input gate, output gate, and forget gate), so as to better capture and associate parameter correlations between different time spans. Therefore, in the technical solution of the present invention, a sequence encoder based on a forward LSTM model is used to process the time queue of the key position temperature value, the time queue of the key position humidity value, and the time queue of the flammable gas concentration value to dynamically adjust the output of the current time step according to the previous time step of the model, so as to better mine and refine the dependent correlation relationship between different time steps.

[0062] It is worth mentioning that the forward LSTM model is a special recurrent neural network (RNN) that is particularly suitable for processing sequence data with long-term dependencies. In the digital coal yard intelligent control system, the forward LSTM model is used to encode the time series data of environmental parameters such as temperature, humidity and flammable gas concentration to extract their temporal implicit correlation features. The following is an example of the network structure of a forward LSTM model, which details its various components and their working principles.

[0063] The core of the forward LSTM model lies in its unique cell state and gating mechanism, which enables the model to effectively remember and utilize long-distance information, thereby better capturing and associating parameter correlations between different time spans. Specifically, the forward LSTM model consists of multiple LSTM units, each of which contains three gating mechanisms: input gate, output gate and forget gate, as well as a cell state.

[0064] Here we use the time queue of the temperature value at the key position to illustrate. Assume that there is a time queue of the temperature value at the key position X = [x1, x2, x3, ..., x t ], where t represents the length of the time series. The input for each time step t is the temperature value.

[0065] In particular, in a specific embodiment of the present invention, the processing process of the forward LSTM model is as follows:

[0066] 1. The input gate determines how much new information is added to the cell state at the current time step t. The input gate combines the input and the hidden state of the previous time step through a sigmoid activation function, and outputs a value between 0 and 1, indicating the degree of information retention. Specifically, the output of the input gate is obtained by concatenating the input and the hidden state of the previous time step, multiplying it by a weight matrix, adding a bias term, and finally passing it through the sigmoid activation function. 2. The forget gate determines how much old cell state information is discarded. The forget gate works similarly to the input gate, and also combines the input and the hidden state of the previous time step through a sigmoid activation function, and outputs a value between 0 and 1, indicating the degree of information retention. Specifically, the output of the forget gate is obtained by concatenating the input and the hidden state of the previous time step, multiplying it by a weight matrix, adding a bias term, and finally passing it through the sigmoid activation function. 3. The cell state is updated by combining the outputs of the forget gate and the input gate. First, the candidate cell state is calculated by concatenating the input and the hidden state of the previous time step, multiplying it by a weight matrix, adding a bias term, and finally passing it through the hyperbolic tangent activation function. Then, the cell state is updated by element-wise multiplying the cell state of the previous time step with the output of the forget gate, plus the element-wise multiplication of the candidate cell state with the output of the input gate. 4. The output gate determines how much of the cell state will be output as the hidden state of the current time step. The output gate works similarly to the input gate and the forget gate. It also combines the input and the hidden state of the previous time step through a sigmoid activation function and outputs a value between 0 and 1, indicating the degree of information retention. Specifically, the output of the output gate is obtained by concatenating the input and the hidden state of the previous time step, multiplying it by a weight matrix, adding a bias term, and finally passing it through the sigmoid activation function. 5. The hidden state of the current time step is calculated by element-wise multiplying the output of the output gate with the activation value of the cell state. The activation value of the cell state is obtained through the hyperbolic tangent activation function.

[0067] Through the above process, the forward LSTM model can effectively capture and associate the changes in environmental parameters between different time spans, providing strong data support for the safety management of coal yards. This model can not only process complex sequence data, but also maintain the validity of information over a long time span, thereby improving the ability to identify and warn of potential safety risks.

[0068] In an embodiment of the present invention, the key position temperature concentration compensation interaction module 130 is used to perform a time series collaborative analysis of the combustion-supporting factors based on the common feature guidance on the key position temperature time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature to obtain the key position combustion-supporting factor time series representation. Accordingly, considering that the key position temperature time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature are two different environmental parameters, they each contain rich information. Therefore, in order to capture the time series interaction relationship between the two modal data, so as to more comprehensively understand the environmental conditions of the coal yard, it is necessary to integrate these two features. However, when integrating information from multiple data sources, each source (temperature and concentration) may provide a unique data perspective, but may also contain redundant content. Traditional integration techniques sometimes fail to fully consider the data points of these unique features, which may result in incomplete information expression. Moreover, if the features of different data sources are not properly coordinated, this inconsistency may weaken the overall effect of feature integration. Based on this, a time series collaborative analysis of combustion factors at key locations is performed based on the common feature guidance on the implicit correlation features of the temperature time series at key locations and the implicit correlation features of the flammable gas concentration time series at key locations to capture and refine the common features between the two, while supplementing and retaining the information in each original feature to obtain the time series representation of combustion factors at key locations.

[0069] Figure 3 FIG. 1 is a block diagram of a temperature concentration compensation interaction module for key positions in a digital coal yard intelligent control system according to an embodiment of the present invention. Specifically, Figure 3 As shown, the key position temperature concentration compensation interaction module 130 includes: a temperature concentration time series commonality extraction unit 131, which is used to input the key position temperature time series implicit association feature vector and the key position flammable gas concentration time series implicit association feature vector into a commonality feature extraction network to obtain a commonality feature representation vector between key position temperature-concentration time series features; a temperature concentration fine-grained compensation interaction unit 132, which is used to perform feature fine-grained compensation interaction on the key position temperature time series implicit association feature vector and the key position flammable gas concentration time series implicit association feature vector based on the commonality feature representation vector between the key position temperature-concentration time series features to obtain the key position combustion-supporting factor time series representation.

[0070] The processing process of the temperature concentration time series commonality extraction unit 131 can be expressed as follows:

[0071]

[0072] Wherein, V1 and V2 are the implicit correlation feature vector of the key position temperature time series and the implicit correlation feature vector of the key position flammable gas concentration time series, respectively. To add by position, W VT and b VT are the weight matrix and bias vector respectively, tanh(·) is the tanh function, V J It is the common feature representation vector between the key position temperature-concentration time series features. It should be understood that common features refer to the features that exist in both the temperature and flammable gas concentration time series. Through the common feature extraction network, the common change trends and patterns between the two modal data can be extracted and reflected.

[0073] Specifically, in the embodiment of the present invention, the temperature concentration fine-grained compensation interaction unit 132 includes: a temperature concentration time series feature probabilization subunit, which is used to input the key position temperature time series implicit association feature vector, the key position flammable gas concentration time series implicit association feature vector and the key position temperature-concentration time series feature common feature representation vector into the probabilization unit to obtain the probabilistic key position temperature time series implicit association feature vector, the probabilistic key position flammable gas concentration time series implicit association feature vector and the probabilistic key position temperature-concentration time series feature common feature representation vector; a compensation vector calculation subunit, which is used to respectively calculate the fine-grained compensation of the probabilistic key position temperature time series implicit association feature vector and the probabilistic key position flammable gas concentration time series implicit association feature vector relative to the probabilistic key position temperature-concentration time series feature common feature representation vector. A compensation vector is used to obtain a fine-grained compensation vector for the temperature at a key position and a fine-grained compensation vector for the flammable gas concentration at a key position; a temperature-concentration feature compensation subunit is used to perform feature compensation on the implicitly associated feature vector of the temperature time series at a key position and the implicitly associated feature vector of the flammable gas concentration at a key position based on the fine-grained compensation vector for the temperature at a key position and the fine-grained compensation vector for the flammable gas concentration at a key position respectively to obtain a fine-grained feature compensation vector for the temperature at a key position and a fine-grained feature compensation vector for the flammable gas concentration at a key position; a temperature-concentration feature multidimensional fusion subunit is used to cascade the common feature representation vector between the fine-grained feature compensation vector for the temperature at a key position, the fine-grained feature compensation vector for the flammable gas concentration at a key position and the temperature-concentration time series features at a key position to obtain a time series representation vector of a combustion-supporting factor at a key position as the time series representation of the combustion-supporting factor at a key position.

[0074] More specifically, in an embodiment of the present invention, the temperature-concentration time series feature probabilization subunit is used to: input the key position temperature time series implicit association feature vector, the key position flammable gas concentration time series implicit association feature vector and the key position temperature-concentration time series feature common feature representation vector into the Sigmoid function for probabilistic processing to obtain the probabilistic key position temperature time series implicit association feature vector, the probabilistic key position flammable gas concentration time series implicit association feature vector and the probabilistic key position temperature-concentration time series feature common feature representation vector.

[0075] The processing process of the above temperature concentration time series feature probabilization subunit can be expressed as follows:

[0076] V a =sigmoid(V1)

[0077] V b =sigmoid(V2)

[0078] V c =sigmoid(V J )

[0079] Among them, V1 and V2 are the implicit correlation feature vector of the key position temperature time series and the implicit correlation feature vector of the key position flammable gas concentration time series, respectively. J is the common feature representation vector between the key position temperature-concentration time series features, sigmoid(·) is the sigmoid function, V a is the implicit correlation feature vector of the temperature time series at the key position in the probability, V b is the implicit correlation feature vector of the flammable gas concentration time series at the probabilistic key position, V c It is a common feature representation vector between the temperature-concentration time series features at the probabilistic key position. In this way, the input features are mapped to the (0,1) interval to achieve normalization. The feature vector processed by the Sigmoid function can be regarded as a certain probability value, indicating the probability of the occurrence of temperature, flammable gas concentration or temperature-concentration common features in a certain period of time. This helps to convert the feature vector into a more intuitive and easier to understand form to reflect which features are significant and which are secondary.

[0080] The processing process of the above compensation vector calculation subunit can be expressed by the formula:

[0081]

[0082] in, is the eigenvalue of each position in the implicitly associated eigenvector of the temperature time series of the probabilistic key position, is the eigenvalue of each position in the implicit correlation eigenvector of the probabilistic flammable gas concentration time series at the key position, is the characteristic value of each position in the common feature representation vector between the temperature-concentration time series features of the probabilistic key position, log represents the logarithmic function value with base 2, exp(·) represents the exponential function value with the natural constant e as the base, softmax(·) is the softmax function, v ac is the fine-grained compensation vector of temperature at key locations, v bc is the fine-grained compensation vector of the flammable gas concentration at the key position. In this way, the unique characteristic information in temperature and flammable gas concentration can be emphasized, which may not be obvious or omitted in the common features, but is crucial for the analysis and processing of the model. This means that through fine feature extraction, the details that are extremely important for in-depth understanding and accurate prediction of explosion behavior can be captured, thereby improving the performance and accuracy of the model.

[0083] More specifically, in an embodiment of the present invention, the temperature-concentration characteristic compensation subunit is used to: input the key position temperature fine-grained compensation vector and the key position temperature time series implicit association characteristic vector into a fine-grained compensation module to obtain the key position temperature fine-grained characteristic compensation vector; input the key position flammable gas concentration fine-grained compensation vector and the key position flammable gas concentration time series implicit association characteristic vector into the fine-grained compensation module to obtain the key position flammable gas concentration fine-grained characteristic compensation vector.

[0084] The processing process of the temperature concentration characteristic compensation subunit can be expressed as follows:

[0085] v 1a =V1⊙v ac

[0086] v 2b =V2⊙v bc

[0087] Among them, V1 and V2 are the implicit correlation feature vector of the key position temperature time series and the implicit correlation feature vector of the key position flammable gas concentration time series, respectively. ac is the fine-grained compensation vector of temperature at key locations, v bc is the fine-grained compensation vector of the flammable gas concentration at the key position, ⊙ is the point multiplication by position, v 1a is the fine-grained feature compensation vector of key position temperature, v 2bis the fine-grained feature compensation vector of flammable gas concentration at key locations. In this way, by integrating the compensation vector into the initial feature vector, the key elements of the original temperature and flammable gas concentration characteristics can be enhanced, ensuring that these elements are more prominent and easy to identify when performing explosion judgment and classification tasks, which helps to improve the model's effectiveness in accurate classification.

[0088] The processing process of the temperature concentration feature multidimensional fusion subunit can be expressed as follows:

[0089] v f =concat[V J ;v 1a ;v 2b ]

[0090] Among them, V J is the common feature representation vector between the temperature-concentration time series features at the key position, v 1a is the fine-grained feature compensation vector of key position temperature, v 2b is the fine-grained feature compensation vector of the flammable gas concentration at the key position, concat[·;·;·] is a cascade operation on the vector, v f is the time series representation vector of the combustion factors at the key positions. In this way, a representation that combines the two features is generated, which contains the common features of their temperature and concentration, and also retains their unique features.

[0091] In an embodiment of the present invention, the key position combustion-supporting-flammable confrontation analysis module 140 is used to use the key position humidity time series implicit correlation feature as the key position ignition suppression factor time series representation, and input the key position combustion-supporting factor time series representation and the key position ignition suppression factor time series representation into the combustion-supporting-flammable confrontation analysis network to obtain the combustion-supporting-flammable confrontation representation. Specifically, the key position combustion-supporting-flammable confrontation analysis module is used to: use the key position humidity time series implicit correlation feature vector as the key position ignition suppression factor time series representation, multiply the key position combustion-supporting factor time series representation vector and the key position humidity time series implicit correlation feature vector to obtain the combustion-supporting-flammable confrontation representation matrix as the combustion-supporting-flammable confrontation representation. Accordingly, considering that the key position combustion-supporting factor time series representation vector reflects the impact of temperature and flammable gas concentration on coal yard safety, these factors usually increase the risk of fire and explosion. The key position combustion suppression factor time series representation (key position humidity time series implicit correlation feature vector) reflects the impact of humidity on the safety of the coal yard, and high humidity usually helps to suppress the occurrence of fire and explosion. Therefore, in order to comprehensively evaluate the interaction between the combustion-supporting and combustion-suppressing factors and to more comprehensively understand the safety status of the coal yard, in the technical solution of the present invention, the key position humidity time series implicit correlation feature is used as the key position combustion suppression factor time series representation, and the key position combustion-supporting factor time series representation and the key position combustion suppression factor time series representation are input into the combustion-supporting-flammable confrontation analysis network to obtain the combustion-supporting-suppressing time series confrontation representation. In particular, in a specific embodiment of the present invention, the combustion-supporting-suppressing time series confrontation representation matrix obtained by multiplying the transposed vector of the key position combustion-supporting factor time series representation vector and the key position humidity time series implicit correlation feature vector is used as the combustion-supporting-suppressing time series confrontation representation. It should be understood that the combustion-supporting and combustion-suppressing factors will show different time series dynamic changes and interactions at different times and different positions. By calculating the transfer matrix of the two vectors, these time series dynamic changes and trends can be captured, which helps to understand which factors dominate in different time periods and thus more accurately assess the safety status of the coal yard.

[0092] In an embodiment of the present invention, the explosion risk judgment module 150 is used to obtain an analysis result of whether the risk of explosion exceeds a preset threshold based on the combustion-suppression time series confrontation representation, and judge whether to generate an explosion warning prompt based on the analysis result. Specifically, in an embodiment of the present invention, the explosion risk judgment module is used to: input the combustion-suppression time series confrontation representation matrix into the intelligent analysis module based on the classifier to obtain the analysis result, and the analysis result is used to indicate whether the risk of explosion exceeds the preset threshold; in response to the analysis result that the risk of explosion exceeds the preset threshold, the explosion warning prompt is generated. That is, the combustion-suppression time series confrontation representation obtained by associating the key position humidity time series implicit association feature vector and the key position suppression factor time series representation vector is used to perform classification processing, so as to intelligently judge the risk of explosion at the key position and generate corresponding warning prompts. In this way, by deploying distributed sensors at multiple key positions in the coal yard, the entire coal yard can be fully covered to ensure that there is no monitoring blind spot. In addition, these sensors can collect environmental data in real time. Once an abnormality is detected at one of the key points, the system can immediately generate an early warning prompt to help managers take quick measures, thereby achieving efficient and intelligent management of the coal yard environment.

[0093] In particular, the processing steps of the above process are as follows: 1. The intelligent analysis module extracts and preprocesses the combustion-supporting-suppressing time series confrontation representation matrix. This step includes standardizing and normalizing the data to ensure the consistency and comparability of the data. Feature extraction may involve extracting key features in the matrix, such as maximum value, minimum value, mean value, variance, etc., which can help the classifier to perform analysis more effectively. Through these preprocessing steps, the information in the matrix is ​​converted into a format suitable for classifier processing. 2. The classifier-based intelligent analysis module uses a pre-trained classifier to evaluate the extracted features. The classifier can be a support vector machine (SVM), a random forest, a neural network, etc., which learns and identifies high-risk patterns through historical data and known explosion cases. During the evaluation process, the classifier calculates the probability of an explosion under the current environmental state, and this probability value reflects the current safety risk level of the coal yard. 3. The evaluation result is a numerical value indicating the risk of an explosion. The intelligent analysis module compares this numerical value with a preset threshold. The preset threshold is determined based on historical data and expert experience to distinguish between normal and high-risk states. If the assessment results show that the risk of explosion exceeds the preset threshold, the intelligent analysis module will generate an explosion warning. In particular, the warning can be conveyed to managers in a variety of ways, such as alarm sounds, text messages, emails, etc. The warning usually contains specific high-risk locations and times, as well as recommended response measures, to help managers take timely actions to avoid potential safety accidents. In this way, not only the safety management level of the coal yard is improved, but also potential safety risks are reduced, providing strong support for the efficient and safe operation of the coal yard.

[0094] In summary, the digital coal yard intelligent control system 100 based on the embodiment of the present invention is explained, which collects the environmental data (temperature value, humidity value, flammable gas concentration value) of each key position through distributed sensors deployed at multiple key positions of the coal yard, and uses data encoding and processing technology based on artificial intelligence to perform data preprocessing and implicit timing association on the environmental data of each key position, and then performs common-guided combustion factor timing coordination on the key position temperature timing implicit association characteristics and the key position flammable gas concentration timing implicit association characteristics, so as to intelligently judge the risk of explosion at the key position according to the combustion factor timing representation of the key position and the combustion-suppression confrontation representation between the key position humidity timing implicit association and generate corresponding early warning prompts. In this way, by deploying distributed sensors at multiple key positions of the coal yard, the entire coal yard can be fully covered to ensure that there is no monitoring blind spot. In addition, these sensors can collect environmental data in real time. Once an abnormal situation is detected, the system can immediately generate an early warning prompt to help managers take measures quickly, thereby realizing efficient and intelligent management of the coal yard environment.

[0095] As described above, the digital coal yard intelligent control system 100 according to the embodiment of the present invention can be implemented in various wireless terminals, such as a server with a digital coal yard intelligent control algorithm. In a possible implementation, the digital coal yard intelligent control system 100 according to the embodiment of the present invention can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the digital coal yard intelligent control system 100 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the digital coal yard intelligent control system 100 can also be one of the many hardware modules of the wireless terminal.

[0096] Alternatively, in another example, the digital coal yard intelligent control system 100 and the wireless terminal may also be separate devices, and the digital coal yard intelligent control system 100 may be connected to the wireless terminal via a wired and / or wireless network, and transmit interactive information in an agreed data format.

[0097] Figure 4 Flow chart of the digital coal yard intelligent control method according to an embodiment of the present invention. Figure 4 As shown, the digital coal yard intelligent control method according to the embodiment of the present invention includes: S110, collecting environmental data of each key position through distributed sensors deployed at multiple key positions of the coal yard, the environmental data including temperature value, humidity value, and flammable gas concentration value; S120, performing data preprocessing and sequence encoding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature; S130, performing data preprocessing and sequence encoding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature; The implicit correlation characteristics of the body concentration time series are used to perform a coordinated analysis of the combustion-supporting factor time series based on the common characteristics to obtain the combustion-supporting factor time series representation at the key position; S140, using the implicit correlation characteristics of the humidity time series at the key position as the flame retardant factor time series representation at the key position, and inputting the combustion-supporting factor time series representation at the key position and the flame retardant factor time series representation at the key position into the combustion-supporting-flammable confrontation analysis network to obtain the combustion-supporting-flammability confrontation representation; S150, based on the combustion-supporting-flammability confrontation representation, obtaining an analysis result of whether the risk of explosion exceeds a preset threshold, and based on the analysis result, determining whether to generate an explosion warning prompt.

[0098] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned digital coal yard intelligent control method have been referred to above. Figures 1 to 3 The digital coal yard intelligent control system has been introduced in detail, and therefore, its repeated description will be omitted.

[0099] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A digital coal yard intelligent control system, characterized in that: include: An environmental data acquisition module is used to collect environmental data of multiple key locations through distributed sensors deployed at the coal yard, wherein the environmental data includes temperature, humidity, and flammable gas concentration. An environmental data processing and encoding module, used for performing data preprocessing and sequence encoding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature; A key position temperature concentration compensation interaction module is used to perform a combustion-supporting factor time series collaborative analysis based on common feature guidance on the key position temperature time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature to obtain a key position combustion-supporting factor time series representation; A key position combustion-supporting-flammable confrontation analysis module is used to use the key position humidity time series implicit correlation feature as the key position ignition suppression factor time series representation, and input the key position combustion-supporting factor time series representation and the key position ignition suppression factor time series representation into the combustion-supporting-flammable confrontation analysis network to obtain the combustion-supporting-flammable confrontation representation; The explosion risk judgment module is used to obtain an analysis result of whether the risk of explosion exceeds a preset threshold based on the combustion-supporting-suppressing time sequence confrontation representation, and to judge whether to generate an explosion warning prompt based on the analysis result.

2. The digital coal yard intelligent control system according to claim 1 is characterized in that: The environmental data processing and encoding module comprises: A key position environmental data preprocessing unit, used for preprocessing the environmental data of each key position according to the time dimension to obtain a time queue of key position temperature values, a time queue of key position humidity values ​​and a time queue of key position flammable gas concentration values; The key position environmental data sequence encoding unit is used to input the time queue of the key position temperature value, the time queue of the key position humidity value and the time queue of the key position flammable gas concentration value into a sequence encoder based on a forward LSTM model respectively to obtain a key position temperature time series implicit association feature vector as the key position temperature time series implicit association feature, a key position humidity time series implicit association feature vector as the key position humidity time series implicit association feature and a key position flammable gas concentration time series implicit association feature vector as the key position flammable gas concentration time series implicit association feature.

3. The digital coal yard intelligent control system according to claim 2 is characterized in that: The key position temperature concentration compensation interactive module includes: A temperature-concentration time series commonality extraction unit is used to input the key position temperature time series implicit association feature vector and the key position flammable gas concentration time series implicit association feature vector into a commonality feature extraction network to obtain a commonality feature representation vector between key position temperature-concentration time series features; The temperature-concentration fine-grained compensation interaction unit is used to perform feature fine-grained compensation interaction on the key position temperature time series implicit association feature vector and the key position flammable gas concentration time series implicit association feature vector based on the common feature representation vector between the key position temperature-concentration time series features to obtain the key position combustion-supporting factor time series representation.

4. The digital coal yard intelligent control system according to claim 3 is characterized in that: The temperature concentration fine-grained compensation interaction unit includes: The temperature concentration time series feature probabilization subunit is used to input the key position temperature time series implicit association feature vector, the key position flammable gas concentration time series implicit association feature vector and the key position temperature-concentration time series feature common feature representation vector into the probabilization unit to obtain the probabilistic key position temperature time series implicit association feature vector, the probabilistic key position flammable gas concentration time series implicit association feature vector and the probabilistic key position temperature-concentration time series feature common feature representation vector; A compensation vector calculation subunit, used to respectively calculate the fine-grained compensation vectors of the probabilistic key position temperature time series implicit association feature vector and the probabilistic key position flammable gas concentration time series implicit association feature vector relative to the probabilistic key position temperature-concentration time series feature common feature representation vector to obtain the key position temperature fine-grained compensation vector and the key position flammable gas concentration fine-grained compensation vector; The temperature concentration characteristic compensation subunit is used to perform characteristic compensation on the key position temperature time series implicit correlation characteristic vector and the key position flammable gas concentration time series implicit correlation characteristic vector based on the key position temperature fine-grained compensation vector and the key position flammable gas concentration fine-grained compensation vector to obtain the key position temperature fine-grained characteristic compensation vector and the key position flammable gas concentration fine-grained characteristic compensation vector; The temperature-concentration feature multidimensional fusion subunit is used to cascade the key position temperature fine-grained feature compensation vector, the key position flammable gas concentration fine-grained feature compensation vector and the key position temperature-concentration time series feature common feature representation vector to obtain the key position combustion-supporting factor time series representation vector as the key position combustion-supporting factor time series representation.

5. The digital coal yard intelligent control system according to claim 4 is characterized in that: The temperature-concentration time series feature probabilization subunit is used to: input the key position temperature time series implicit association feature vector, the key position flammable gas concentration time series implicit association feature vector and the key position temperature-concentration time series feature common feature representation vector into the Sigmoid function for probabilistic processing to obtain the probabilistic key position temperature time series implicit association feature vector, the probabilistic key position flammable gas concentration time series implicit association feature vector and the probabilistic key position temperature-concentration time series feature common feature representation vector.

6. The digital coal yard intelligent control system according to claim 5 is characterized in that: The temperature concentration characteristic compensation subunit is used for: Inputting the key position temperature fine-grained compensation vector and the key position temperature time series implicit correlation feature vector into a fine-grained compensation module to obtain the key position temperature fine-grained feature compensation vector; The key position flammable gas concentration fine-grained compensation vector and the key position flammable gas concentration time series implicit correlation feature vector are input into the fine-grained compensation module to obtain the key position flammable gas concentration fine-grained feature compensation vector.

7. The digital coal yard intelligent control system according to claim 6 is characterized in that: The key position combustion-supporting-flammable confrontation analysis module is used to: use the key position humidity time series implicitly associated characteristic vector as the key position combustion suppression factor time series representation vector and the transposed vector of the key position humidity time series implicitly associated characteristic vector to obtain the combustion-supporting-flammable timing confrontation representation matrix as the combustion-supporting-flammable timing confrontation representation.

8. The digital coal yard intelligent control system according to claim 7 is characterized in that: The explosion risk judgment module is used to: Inputting the combustion-supporting-suppressing time series confrontation representation matrix into a classifier-based intelligent analysis module to obtain the analysis result, wherein the analysis result is used to indicate whether the risk of explosion exceeds a preset threshold; In response to the analysis result that the risk of explosion exceeds a preset threshold, the explosion warning prompt is generated.

9. A digital coal yard intelligent control method, characterized in that: include: Collect environmental data of each key location through distributed sensors deployed at multiple key locations in the coal yard, wherein the environmental data includes temperature value, humidity value, and flammable gas concentration value; Performing data preprocessing and sequence coding on the environmental data of each key position to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature; Performing a time series collaborative analysis of combustion-supporting factors guided by common features on the key position temperature time series implicit correlation features and the key position flammable gas concentration time series implicit correlation features to obtain a time series representation of the key position combustion-supporting factors; Taking the key position humidity time series implicit correlation feature as the key position ignition suppression factor time series representation, inputting the key position combustion supporting factor time series representation and the key position ignition suppression factor time series representation into the combustion supporting-flammable confrontation analysis network to obtain the combustion supporting-ignition suppression time series confrontation representation; Based on the combustion-supporting-suppressing time sequence confrontation representation, an analysis result of whether the risk of explosion exceeds a preset threshold is obtained, and based on the analysis result, it is determined whether to generate an explosion warning prompt.

10. The digital coal yard intelligent control method according to claim 9 is characterized in that: The environmental data of each key position is preprocessed and sequenced to obtain the key position temperature time series implicit correlation feature, the key position humidity time series implicit correlation feature and the key position flammable gas concentration time series implicit correlation feature, including: Preprocessing the environmental data of each key position according to the time dimension to obtain a time queue of temperature values ​​at the key position, a time queue of humidity values ​​at the key position, and a time queue of flammable gas concentration values ​​at the key position; The time queue of the key position temperature values, the time queue of the key position humidity values, and the time queue of the key position flammable gas concentration values ​​are respectively input into a sequence encoder based on a forward LSTM model to obtain a key position temperature time series implicit association feature vector as the key position temperature time series implicit association feature, a key position humidity time series implicit association feature vector as the key position humidity time series implicit association feature, and a key position flammable gas concentration time series implicit association feature vector as the key position flammable gas concentration time series implicit association feature.

Images