Internet of Things heat supply network analysis method and system based on self-attention mechanism

Abstract

The invention provides an Internet of Things heat supply network analysis method and system based on a self-attention mechanism, and the method comprises the steps: 1, calling historical monitoring data of a heat supply network based on Internet of Things equipment, and carrying out the data coding of the historical monitoring data, and obtaining feature sequence data; 2, learning the feature sequence data, and constructing a target model based on a self-attention mechanism; 3, real-time data of operation of the heat supply pipe network are collected in real time based on the Internet of Things equipment, the real-time data are input into the target model to be predicted and analyzed, and a prediction result of the heat consumption behavior of the user is output; 4, dynamically adjusting a heat supply strategy for the user based on the prediction result of the heat consumption behavior of the user, and updating the heat supply state of the user in the heat supply pipe network in real time according to the Internet of Things based on the dynamically adjusted heat supply strategy; and a new solution is provided for intelligent management of the heat supply system, so that the energy utilization efficiency and the user satisfaction of the heat supply system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to an Internet of Things heating pipe network analysis method and system based on a self-attention mechanism. Background Art

[0002] At present, with the rapid development of Internet of Things technology, the Internet of Things has been more and more widely used in various fields. In the field of heating pipe networks, the Internet of Things technology can effectively realize real-time monitoring of operating data in the heating pipe network, thereby effectively improving the operating efficiency of the heating pipe network; However, traditional heating network systems rely on empirical rules and simple statistical models in user demand analysis and energy allocation, lacking in-depth understanding and prediction of user behavior patterns, making it impossible to accurately predict user behavior and provide personalized services in the heating system, which is not conducive to the intelligent development of heating networks, that is, it is impossible to improve energy efficiency and user satisfaction. Therefore, the present invention provides an Internet of Things heating network analysis method and system based on a self-attention mechanism. Summary of the invention

[0003] The present invention provides an Internet of Things heating pipe network analysis method and system based on a self-attention mechanism, which are used to accurately retrieve historical monitoring data of the heating pipe network through an Internet of Things device, and perform data encoding on the historical monitoring data to obtain feature sequence data suitable for target model processing. By learning the feature sequence data, a target model based on the self-attention mechanism is accurately constructed, thereby effectively realizing the prediction and analysis of the user's heating behavior, and then realizing the dynamic adjustment of the user's heating strategy and the real-time update of the heating status, thereby realizing accurate prediction and personalized service of the user behavior of the heating system, providing a new solution for the intelligent management of the heating system, thereby effectively improving the energy utilization efficiency and user satisfaction of the heating system, and having important practical application value.

[0004] A method for analyzing an Internet of Things heating network based on a self-attention mechanism, comprising: Step 1: Retrieve historical monitoring data of the heating network based on the IoT device, encode the historical monitoring data, and obtain characteristic sequence data; Step 2: Learn the feature sequence data and build a target model based on the self-attention mechanism; Step 3: Based on the IoT devices, real-time data of the heating network operation is collected in real time, and the real-time data is input into the target model for prediction analysis, and the prediction results of the user's heating behavior are output; Step 4: Dynamically adjust the heating strategy for users based on the predicted results of users' heat usage behavior, and update the heating status of users in the heating network in real time according to the Internet of Things based on the dynamically adjusted heating strategy.

[0005] Preferably, an IoT heating network analysis method based on a self-attention mechanism, in step 1, historical monitoring data of the heating network is retrieved based on IoT devices, including: Obtain data retrieval tasks, read the data retrieval tasks according to the IoT devices, and determine the central execution point of the data retrieval tasks; Determine the first execution target of the data retrieval task according to the central execution point of the data retrieval task; Obtain a data repository in an IoT device and obtain a data retrieval format of the data repository; Convert the first execution target into a format according to the data retrieval format to obtain a second execution target; The second execution target is input into the data repository, and is located in the data repository, and historical monitoring data is retrieved according to the positioning result, wherein the historical monitoring data includes: operation data of the heating network and user usage data.

[0006] Preferably, a method for analyzing an IoT heating network based on a self-attention mechanism, after retrieving historical monitoring data, further includes: Obtain the data type of the historical monitoring data, extract the data features of each data type corresponding to the historical monitoring data, and construct a feature matrix corresponding to each data type based on the data features; A data recognition model is constructed according to the feature matrix corresponding to each type, and the historical monitoring data is screened based on the data recognition model to determine the data to be cleaned in the historical monitoring data; Clean the data to be cleaned and obtain the target historical monitoring data; The target historical monitoring data is normalized to obtain the final historical monitoring data.

[0007] Preferably, a method for analyzing an IoT heating network based on a self-attention mechanism, in step 1, data encoding is performed on historical monitoring data to obtain feature sequence data, including: Retrieve the time series when obtaining historical monitoring data, and obtain the sequence interval of the time series; Assigning sequence positions to historical monitoring data according to the sequence interval, and determining the sub-historical monitoring data corresponding to each sequence position according to the assignment result; Assign a coding vector to the sub-historical monitoring data corresponding to each sequence position, and perform data encoding on the sub-monitoring data corresponding to each sequence position according to the coding vector; wherein the coding vector for each sequence position is different; The encoded sub-monitoring data are sorted according to the time series, and feature sequence data are generated based on the sorting results.

[0008] Preferably, a method for analyzing an IoT heating network based on a self-attention mechanism, in step 2, learning the feature sequence data and constructing a target model based on the self-attention mechanism, including: Read the feature sequence data to determine the feature types contained in the feature sequence data, wherein the feature types are: feature types of the operation data of the heating network changing with the time series and feature types of the user usage data changing with the time series; Dividing the feature sequence data into first feature sequence data and second feature sequence data according to the feature type contained in the feature sequence data; Divide the first feature sequence data and the second feature sequence data into a corresponding first training set and a first validation set according to a preset ratio; The time in the first training set is used as an independent variable, and the operation data of the heating network or the user usage data is used as a dependent variable to construct a target change curve; The target change curve is read to determine the amplitude point, inflection point and curve fluctuation range in the target change curve, and at the same time, the amplitude point, inflection point and amplitude range in the target change curve are used as the feature center point of the first training set; The self-attention mechanism is learned based on the feature center point to obtain a first model element corresponding to the first feature sequence and a second model element corresponding to the second feature sequence; The first model element and the second model element are encapsulated according to the time series, and based on the encapsulation results, the multi-head self-attention layer, the position encoding layer and the feedforward neural network layer are adaptively trained to obtain a target model based on the self-attention mechanism; The target model is verified according to the first verification set to obtain the model loss value of the target model, and when the model loss value of the target model is within a preset range, the construction of the target model based on the self-attention mechanism is completed.

[0009] Preferably, an IoT heating network analysis method based on a self-attention mechanism, in step 3, real-time data of the heating network operation is collected in real time based on IoT devices, including: Read the communication node locations of IoT devices and the communication node locations of the heating network; Match the target communication connection mode in the preset communication library according to the communication node location of the IoT device and the communication node location of the heating network; Obtain preset communication configuration parameters; Establish a communication link between the IoT device and the heating network based on the target communication connection mode, and configure the communication link based on preset communication configuration parameters to obtain a final communication link; Based on the final communication link, the real-time data of the heating network operation is collected in real time based on the Internet of Things devices.

[0010] Preferably, a method for analyzing an IoT heating network based on a self-attention mechanism, in step 3, inputs real-time data into a target model for predictive analysis, and outputs a prediction result of a user's heating behavior, including: Input the real-time data into the target model, add position information to the real-time data according to the position encoding layer in the target model, and obtain the target sequence of the real-time data according to the adding result; The multi-head self-attention layer in the target model is used to capture key information of the target sequence, and obtain the operation characteristic data of the heating network and the user usage characteristic data in the current real-time data; The operating characteristic data of the heating network and the user usage characteristic data in the current real-time data are input into the feedforward neural network layer in the target model for processing to determine the prediction results of the user's heat usage behavior under the current operating characteristics of the heating network.

[0011] Preferably, in a method for analyzing an IoT heating network based on a self-attention mechanism, in step 4, dynamically adjusting the heating strategy for the user based on the prediction result of the user's heating behavior includes: Read the prediction results of the user's heat consumption behavior, determine the change trend of the heating data provided to the user based on the heating network, and at the same time, perform data simulation on the change trend of the heating data to obtain the heating prediction parameters of the heating network for heating the user; Retrieve the current heating strategy for providing heating to users based on the heating network, and read the heating parameters in the current heating strategy; Matching the heat supply prediction parameters with the heat supply parameters, determining the difference parameters between the heat supply prediction parameters and the heat supply parameters, and using the difference parameters as target adjustment values; The target adjustment value is corrected in real time in the current heating strategy to complete the dynamic adjustment of the user's heating strategy.

[0012] Preferably, a method for analyzing a heating network of the Internet of Things based on a self-attention mechanism, based on a dynamically adjusted heating strategy, updates the heating status of users in the heating network in real time according to the Internet of Things, including: Read the dynamically adjusted heating strategy and determine the heating adjustment instructions; Transmit heating adjustment instructions to the heating network based on the Internet of Things; The heating adjustment instructions are read according to the heating network and the heating status of the user is updated.

[0013] An IoT heating network analysis system based on a self-attention mechanism, comprising: A data retrieval module is used to retrieve historical monitoring data of the heating network based on IoT devices, and encode the historical monitoring data to obtain characteristic sequence data; Model building module, used to learn feature sequence data and build a target model based on the self-attention mechanism; The prediction and analysis module is used to collect real-time data of the heating network operation based on IoT devices, input the real-time data into the target model for prediction and analysis, and output the prediction results of the user's heating behavior; The update module is used to dynamically adjust the heating strategy for users based on the predicted results of users' heat usage behavior, and to update the heating status of users in the heating network in real time according to the Internet of Things based on the dynamically adjusted heating strategy.

[0014] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.

[0015] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is a flow chart of a method for analyzing a heating network of the Internet of Things based on a self-attention mechanism in an embodiment of the present invention; Figure 2 This is a flowchart of step 1 in a method for analyzing a heating network of the Internet of Things based on a self-attention mechanism in an embodiment of the present invention; Figure 3 This is a structural diagram of an Internet of Things heating network analysis system based on a self-attention mechanism in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0018] Embodiment 1: This embodiment provides an IoT heating network analysis method based on a self-attention mechanism. Figure 1 As shown, including: Step 1: Retrieve historical monitoring data of the heating network based on the IoT device, encode the historical monitoring data, and obtain characteristic sequence data; Step 2: Learn the feature sequence data and build a target model based on the self-attention mechanism; Step 3: Based on the IoT devices, real-time data of the heating network operation is collected in real time, and the real-time data is input into the target model for prediction analysis, and the prediction results of the user's heating behavior are output; Step 4: Dynamically adjust the heating strategy for users based on the predicted results of users' heat usage behavior, and update the heating status of users in the heating network in real time according to the Internet of Things based on the dynamically adjusted heating strategy.

[0019] In this embodiment, the historical monitoring data includes: operation data of the heating network and user usage data.

[0020] In this embodiment, data encoding of historical monitoring data can be achieved by allocating and encoding the positions of the historical monitoring data through time series, so that the constructed target model can have the characteristics of capturing the long-term dependencies of time series data, wherein the data encoding conforms to the characteristic sequence data processed by the target model, and at the same time, the target model is a Transformer model.

[0021] In this embodiment, the target model includes: a multi-head self-attention layer, a position encoding layer and a feedforward neural network layer, which can effectively process sequence data and extract key features.

[0022] In this embodiment, the prediction result of the user's heat usage behavior may include, for example, features such as heat usage time, heat usage, and temperature preference.

[0023] This embodiment also includes obtaining user satisfaction and energy consumption feedback data after updating the heating status of users in the heating network, so as to continuously iterate and optimize the Transformer model to adapt to changes in user behavior patterns and the actual operation of the heating system.

[0024] In this embodiment, the heating strategy is dynamically adjusted based on customer feature analysis and prediction results to achieve personalized heating and optimized energy distribution.

[0025] In this embodiment, IoT devices refer to physical devices that are connected through a network and can communicate with each other, collect and transmit data. They usually have components such as sensors, processors, and communication modules, which can realize intelligent perception, control, and management.

[0026] The working principle and beneficial effects of the above technical solution are: accurately retrieve the historical monitoring data of the heating pipeline network through the Internet of Things equipment, and encode the historical monitoring data to obtain feature sequence data suitable for target model processing, and accurately construct a target model based on the self-attention mechanism by learning the feature sequence data, so as to effectively realize the prediction and analysis of the user's heating behavior, and then realize the dynamic adjustment of the user's heating strategy and real-time update of the heating status, thereby realizing accurate prediction and personalized service of the user behavior of the heating system, providing a new solution for the intelligent management of the heating system, thereby effectively improving the energy utilization efficiency and user satisfaction of the heating system, and having important practical application value.

[0027] Embodiment 2: Based on Example 1, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism, such as Figure 2 As shown, in step 1, historical monitoring data of the heating network is retrieved based on the IoT device, including: S101: Obtain a data retrieval task, and read the data retrieval task according to the IoT device to determine a central execution point of the data retrieval task; S102: determining a first execution target of the data retrieval task according to a central execution point of the data retrieval task; S103: Acquire a data repository in the IoT device, and acquire a data retrieval format of the data repository; S104: Convert the format of the first execution target according to the data retrieval format to obtain a second execution target; S105: Input the second execution target into the data repository, locate it in the data repository, and retrieve historical monitoring data according to the location result, wherein the historical monitoring data includes: operation data of the heating network and user usage data.

[0028] In this embodiment, the data retrieval task may be specific data information that needs to be retrieved.

[0029] In this embodiment, the central execution point may be key information that needs to be executed in the data retrieval task, that is, specific parameters related to the retrieved data information.

[0030] In this embodiment, the first execution target may be an object that ultimately needs to be executed, that is, data that ultimately needs to be retrieved.

[0031] In this embodiment, the second execution target may be a result obtained by performing format conversion on the first execution target.

[0032] In this embodiment, the historical monitoring data may be data information finally retrieved from the data storage repository according to the data retrieval task.

[0033] The working principle and beneficial effects of the above technical solution are: by reading the data retrieval task, the central execution point of the data retrieval task is effectively obtained, and then the first execution target of the data retrieval task is effectively determined, and by converting the format of the first execution target according to the data retrieval format, it is ensured that the obtained second execution target is effectively and accurately identified in the data repository, thereby ensuring the accuracy and effectiveness of the retrieval of historical monitoring data.

[0034] Embodiment 3: Based on Example 2, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism, which, after retrieving historical monitoring data, also includes: Obtain the data type of the historical monitoring data, extract the data features of each data type corresponding to the historical monitoring data, and construct a feature matrix corresponding to each data type based on the data features; A data recognition model is constructed according to the feature matrix corresponding to each type, and the historical monitoring data is screened based on the data recognition model to determine the data to be cleaned in the historical monitoring data; Clean the data to be cleaned and obtain the target historical monitoring data; The target historical monitoring data is normalized to obtain the final historical monitoring data.

[0035] In this embodiment, the data feature may be information such as the value range and structural components of the historical monitoring data corresponding to each data type.

[0036] In this embodiment, the feature matrix can be constructed based on data features and is a parameter basis for constructing a data recognition model, that is, data of corresponding data types can be accurately and effectively recognized based on the feature matrix.

[0037] In this embodiment, the data to be cleaned may be data in historical monitoring data that needs to be cleaned, and may be data that is identified by a data identification model and is irrelevant to the current data type.

[0038] In this embodiment, the target historical monitoring data may be final data obtained after cleaning the data that needs to be cleaned in the historical monitoring data.

[0039] In this embodiment, the normalization process may be to unify the value range of the historical monitoring data.

[0040] The beneficial effect of the above technical solution is: by constructing a data recognition model according to the data type of historical monitoring data, the data to be cleaned in the historical monitoring data can be effectively identified and cleaned according to the data recognition model, thereby improving the data quality of the historical monitoring data and reducing the deviation of the historical monitoring data, thereby providing protection and reliability for data encoding of historical monitoring data.

[0041] Embodiment 4: On the basis of Example 1, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism. In step 1, historical monitoring data is encoded to obtain feature sequence data, including: Retrieve the time series when obtaining historical monitoring data, and obtain the sequence interval of the time series; Assigning sequence positions to historical monitoring data according to the sequence interval, and determining the sub-historical monitoring data corresponding to each sequence position according to the assignment result; Assign a coding vector to the sub-historical monitoring data corresponding to each sequence position, and perform data encoding on the sub-monitoring data corresponding to each sequence position according to the coding vector; wherein the coding vector for each sequence position is different; The encoded sub-monitoring data are sorted according to the time series, and feature sequence data are generated based on the sorting results.

[0042] In this embodiment, the time series may be specific time information corresponding to the historical monitoring data.

[0043] In this embodiment, the sequence interval may be the time interval that exists when retrieving historical monitoring data.

[0044] In this embodiment, the sub-historical monitoring data may be specific historical monitoring data corresponding to each sequence position obtained after matching the historical monitoring data with the corresponding sequence position.

[0045] In this embodiment, the encoding vector may be an encoding parameter required for encoding the sub-historical monitoring data corresponding to each sequence position.

[0046] In this embodiment, the characteristic sequence data may be a result obtained by sorting the encoded sub-monitoring data according to a time series.

[0047] The working principle and beneficial effects of the above technical solution are: by retrieving the time series when obtaining historical monitoring data and obtaining the sequence interval of the time series, the sequence position is effectively assigned to the historical monitoring data according to the sequence interval, and then the sub-historical monitoring data corresponding to each sequence position is determined according to the assignment result, and by assigning a coding vector to the sub-historical monitoring data corresponding to each sequence position, data encoding of the sub-monitoring data corresponding to each sequence position is achieved, and by sorting the data-encoded sub-monitoring data according to the time series, the feature sequence data is effectively obtained, which lays the foundation for the target model to recognize the feature sequence data.

[0048] Embodiment 5: On the basis of Example 1, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism. In step 2, the feature sequence data is learned to construct a target model based on a self-attention mechanism, including: Read the feature sequence data to determine the feature types contained in the feature sequence data, wherein the feature types are: feature types of the operation data of the heating network changing with the time series and feature types of the user usage data changing with the time series; Dividing the feature sequence data into first feature sequence data and second feature sequence data according to the feature type contained in the feature sequence data; Divide the first feature sequence data and the second feature sequence data into a corresponding first training set and a first validation set according to a preset ratio; The time in the first training set is used as an independent variable, and the operation data of the heating network or the user usage data is used as a dependent variable to construct a target change curve; The target change curve is read to determine the amplitude point, inflection point and curve fluctuation range in the target change curve, and at the same time, the amplitude point, inflection point and amplitude range in the target change curve are used as the feature center point of the first training set; The self-attention mechanism is learned based on the feature center point to obtain a first model element corresponding to the first feature sequence and a second model element corresponding to the second feature sequence; The first model element and the second model element are encapsulated according to the time series, and based on the encapsulation results, the multi-head self-attention layer, the position encoding layer and the feedforward neural network layer are adaptively trained to obtain a target model based on the self-attention mechanism; The target model is verified according to the first verification set to obtain the model loss value of the target model, and when the model loss value of the target model is within a preset range, the construction of the target model based on the self-attention mechanism is completed.

[0049] In this embodiment, the first feature sequence data and the second feature sequence data may be specific results obtained after dividing the feature sequence data according to feature types.

[0050] In this embodiment, the preset ratio is set in advance.

[0051] In this embodiment, the first training set and the first validation set may be results obtained by dividing the first feature sequence data and the second feature sequence data respectively.

[0052] In this embodiment, the target change curve can be a curve constructed by taking the time in the first training set and the operation data of the heating network or the user usage data as dependent variables, and is used to characterize the changes in the operation data of the heating network or the user usage data over time.

[0053] In this embodiment, the amplitude point may be a specific value situation corresponding to different amplitudes in the target change curve.

[0054] In this embodiment, the feature center point can be the amplitude point, inflection point and amplitude range in the target change curve as key parameters for learning the self-attention mechanism.

[0055] In this embodiment, the first model element and the second model element may be model structure parameters obtained by learning the self-attention mechanism through the feature center point.

[0056] In this embodiment, the multi-head self-attention layer, the position encoding layer and the feedforward neural network layer are the data layers in the first model element and the second model element, and are the structural parameters that constitute the target model.

[0057] In this embodiment, the model loss value may be a parameter used to characterize the performance loss of the target model. A larger value indicates a worse performance of the target model.

[0058] In this embodiment, the preset range is known in advance.

[0059] The working principle and beneficial effects of the above technical solution are: by reading the feature sequence data, the feature type contained in the feature sequence data is effectively determined, and at the same time, the feature sequence data is divided into first feature sequence data and second feature sequence data according to the feature type contained in the feature sequence data, and the first feature sequence data and the second feature sequence data are divided into corresponding first training sets and first verification sets according to a preset ratio, respectively, to provide reliable data support for building a target model, and secondly, the self-attention mechanism is learned through the training set to achieve effective acquisition of the first model element and the second model element, and the target model is effectively constructed through the first model element and the second model element, and finally, the target model is verified according to the first verification set, which effectively guarantees the prediction and analysis of real-time data by the target model based on the self-attention mechanism, thereby improving the intelligence of the heating network.

[0060] Embodiment 6: On the basis of Example 1, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism. In step 3, real-time data of the heating network operation is collected in real time based on IoT devices, including: Read the communication node locations of IoT devices and the communication node locations of the heating network; Match the target communication connection mode in the preset communication library according to the communication node location of the IoT device and the communication node location of the heating network; Obtain preset communication configuration parameters; Establish a communication link between the IoT device and the heating network based on the target communication connection mode, and configure the communication link based on preset communication configuration parameters to obtain a final communication link; Based on the final communication link, the real-time data of the heating network operation is collected in real time based on the Internet of Things devices.

[0061] In this embodiment, the communication node location may be the specific location of the IoT device and each communication node in the heating network.

[0062] In this embodiment, the preset communication library is known in advance and is used to store a plurality of different types of communication connection modes.

[0063] In this embodiment, the target communication connection method may be a method suitable for communication connection between the current Internet of Things device and the heating network.

[0064] In this embodiment, the preset communication configuration parameters are known in advance.

[0065] The beneficial effect of the above technical solution is: effectively ensuring the efficiency and accuracy of communication, thereby providing strong guarantee and reliability for obtaining real-time data.

[0066] Embodiment 7: Based on Example 1, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism. In step 3, real-time data is input into a target model for predictive analysis, and the prediction results of user heat use behavior are output, including: Input the real-time data into the target model, add position information to the real-time data according to the position encoding layer in the target model, and obtain the target sequence of the real-time data according to the adding result; The multi-head self-attention layer in the target model is used to capture key information of the target sequence, and obtain the operation characteristic data of the heating network and the user usage characteristic data in the current real-time data; The operating characteristic data of the heating network and the user usage characteristic data in the current real-time data are input into the feedforward neural network layer in the target model for processing to determine the prediction results of the user's heat usage behavior under the current operating characteristics of the heating network.

[0067] In this embodiment, the target sequence may be a result obtained by adding position information to the real-time data, including the arrangement of the real-time data.

[0068] In this embodiment, the key information may be core data in the target sequence.

[0069] The beneficial effect of the above technical solution is that by inputting real-time data into the target model for processing, accurate prediction of the user behavior of the heating system and personalized service are achieved.

[0070] Embodiment 8: Based on Example 1, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism. In step 4, the heating strategy for the user is dynamically adjusted based on the prediction result of the user's heating behavior, including: Read the prediction results of the user's heat consumption behavior, determine the change trend of the heating data provided to the user based on the heating network, and at the same time, perform data simulation on the change trend of the heating data to obtain the heating prediction parameters of the heating network for heating the user; Retrieve the current heating strategy for providing heating to users based on the heating network, and read the heating parameters in the current heating strategy; Matching the heat supply prediction parameters with the heat supply parameters, determining the difference parameters between the heat supply prediction parameters and the heat supply parameters, and using the difference parameters as target adjustment values; The target adjustment value is corrected in real time in the current heating strategy to complete the dynamic adjustment of the user's heating strategy.

[0071] In this embodiment, the heating data change trend may be a representation of the heating change situation of the user.

[0072] In this embodiment, the heat supply prediction parameter may be obtained by simulating the change trend of the user's heat supply data, and represents the prediction result of the heat supply to the user.

[0073] In this embodiment, the heating parameters may be specific parameter values ​​defined in the current heating strategy for providing heating to the user.

[0074] In this embodiment, the difference parameter may be a different parameter between the heating prediction parameter and the heating parameter.

[0075] The working principle and beneficial effects of the above technical solution are: by reading the predicted results of the user's heat usage behavior, the user's heating data change trend is determined. At the same time, the heating data change trend is simulated to obtain the heating prediction parameters for the heating network to provide heating to the user. Secondly, the heating prediction parameters are matched with the heating parameters to effectively lock the difference parameters between the heating prediction parameters and the heating parameters, and real-time correction is made to the current heating strategy according to the difference parameters. Finally, the dynamic adjustment of the user's heating strategy is completed, which effectively provides a new solution for the intelligent management of the heating system, thereby effectively improving the energy utilization efficiency and user satisfaction of the heating system.

[0076] Embodiment 9: Based on Example 1, this embodiment provides an IoT heating network analysis method based on a self-attention mechanism, which updates the heating status of users in the heating network in real time according to the IoT based on a dynamically adjusted heating strategy, including: Read the dynamically adjusted heating strategy and determine the heating adjustment instructions; Transmit heating adjustment instructions to the heating network based on the Internet of Things; The heating adjustment instructions are read according to the heating network and the heating status of the user is updated.

[0077] In this embodiment, the heating adjustment instruction may be determined after reading the dynamically adjusted heating strategy, and is used to control the heating supply of the heating network.

[0078] The beneficial effect of the above technical solution is: by determining the heating adjustment instructions, intelligent management of the heating status of the heating network to the user is achieved.

[0079] Embodiment 10: This embodiment provides an IoT heating network analysis system based on a self-attention mechanism. Figure 3 As shown, including: A data retrieval module is used to retrieve historical monitoring data of the heating network based on IoT devices, and encode the historical monitoring data to obtain characteristic sequence data; Model building module, used to learn feature sequence data and build a target model based on the self-attention mechanism; The prediction and analysis module is used to collect real-time data of the heating network operation based on IoT devices, input the real-time data into the target model for prediction and analysis, and output the prediction results of the user's heating behavior; The update module is used to dynamically adjust the heating strategy for users based on the predicted results of users' heat usage behavior, and to update the heating status of users in the heating network in real time according to the Internet of Things based on the dynamically adjusted heating strategy.

[0080] The working principle and beneficial effects of the above technical solution are: accurately retrieve the historical monitoring data of the heating pipeline network through the Internet of Things equipment, and encode the historical monitoring data to obtain feature sequence data suitable for target model processing, and accurately construct a target model based on the self-attention mechanism by learning the feature sequence data, so as to effectively realize the prediction and analysis of the user's heating behavior, and then realize the dynamic adjustment of the user's heating strategy and real-time update of the heating status, thereby realizing accurate prediction and personalized service of the user behavior of the heating system, providing a new solution for the intelligent management of the heating system, thereby effectively improving the energy utilization efficiency and user satisfaction of the heating system, and having important practical application value.

[0081] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for analyzing heating pipe networks in the Internet of Things based on a self-attention mechanism, characterized in that: include: Step 1: Retrieve historical monitoring data of the heating network based on the IoT device, encode the historical monitoring data, and obtain characteristic sequence data; Step 2: Learn the feature sequence data and build a target model based on the self-attention mechanism; Step 3: Based on the IoT devices, real-time data of the heating network operation is collected in real time, and the real-time data is input into the target model for prediction analysis, and the prediction results of the user's heating behavior are output; Step 4: Dynamically adjust the heating strategy for users based on the predicted results of users' heat usage behavior, and update the heating status of users in the heating network in real time according to the Internet of Things based on the dynamically adjusted heating strategy.

2. According to claim 1, a method for analyzing an Internet of Things heating network based on a self-attention mechanism is characterized in that: In step 1, historical monitoring data of the heating network is retrieved based on the IoT device, including: Obtain data retrieval tasks, read the data retrieval tasks according to the IoT devices, and determine the central execution point of the data retrieval tasks; Determine the first execution target of the data retrieval task according to the central execution point of the data retrieval task; Obtain a data repository in an IoT device and obtain a data retrieval format of the data repository; Convert the first execution target into a format according to the data retrieval format to obtain a second execution target; The second execution target is input into the data repository, and is located in the data repository, and historical monitoring data is retrieved according to the positioning result, wherein the historical monitoring data includes: operation data of the heating network and user usage data.

3. According to claim 2, a method for analyzing an Internet of Things heating network based on a self-attention mechanism is characterized in that: After retrieving historical monitoring data, it also includes: Obtain the data type of the historical monitoring data, extract the data features of each data type corresponding to the historical monitoring data, and construct a feature matrix corresponding to each data type based on the data features; A data recognition model is constructed according to the feature matrix corresponding to each type, and the historical monitoring data is screened based on the data recognition model to determine the data to be cleaned in the historical monitoring data; Clean the data to be cleaned and obtain the target historical monitoring data; The target historical monitoring data is normalized to obtain the final historical monitoring data.

4. According to the method of analyzing the Internet of Things heating network based on the self-attention mechanism of claim 1, it is characterized in that: In step 1, the historical monitoring data is encoded to obtain characteristic sequence data, including: Retrieve the time series when obtaining historical monitoring data, and obtain the sequence interval of the time series; Assigning sequence positions to historical monitoring data according to the sequence interval, and determining the sub-historical monitoring data corresponding to each sequence position according to the assignment result; Assign a coding vector to the sub-historical monitoring data corresponding to each sequence position, and perform data encoding on the sub-monitoring data corresponding to each sequence position according to the coding vector; wherein the coding vector for each sequence position is different; The encoded sub-monitoring data are sorted according to the time series, and feature sequence data are generated based on the sorting results.

5. The method for analyzing the Internet of Things heating network based on the self-attention mechanism according to claim 1 is characterized in that: In step 2, the feature sequence data is learned to build a target model based on the self-attention mechanism, including: Read the feature sequence data to determine the feature types contained in the feature sequence data, wherein the feature types are: feature types of the operation data of the heating network changing with the time series and feature types of the user usage data changing with the time series; Dividing the feature sequence data into first feature sequence data and second feature sequence data according to the feature type contained in the feature sequence data; Divide the first feature sequence data and the second feature sequence data into a corresponding first training set and a first validation set according to a preset ratio; The time in the first training set is used as an independent variable, and the operation data of the heating network or the user usage data is used as a dependent variable to construct a target change curve; The target change curve is read to determine the amplitude point, inflection point and curve fluctuation range in the target change curve, and at the same time, the amplitude point, inflection point and amplitude range in the target change curve are used as the feature center point of the first training set; The self-attention mechanism is learned based on the feature center point to obtain a first model element corresponding to the first feature sequence and a second model element corresponding to the second feature sequence; The first model element and the second model element are encapsulated according to the time series, and based on the encapsulation results, the multi-head self-attention layer, the position encoding layer and the feedforward neural network layer are adaptively trained to obtain a target model based on the self-attention mechanism; The target model is verified according to the first verification set to obtain the model loss value of the target model, and when the model loss value of the target model is within a preset range, the construction of the target model based on the self-attention mechanism is completed.

6. The method for analyzing the Internet of Things heating network based on the self-attention mechanism according to claim 1 is characterized in that: In step 3, real-time data of the heating network operation is collected based on the IoT devices, including: Read the communication node locations of IoT devices and the communication node locations of the heating network; Match the target communication connection mode in the preset communication library according to the communication node location of the IoT device and the communication node location of the heating network; Obtain preset communication configuration parameters; Establish a communication link between the IoT device and the heating network based on the target communication connection mode, and configure the communication link based on preset communication configuration parameters to obtain a final communication link; Based on the final communication link, the real-time data of the heating network operation is collected in real time based on the Internet of Things devices.

7. The method for analyzing the Internet of Things heating network based on the self-attention mechanism according to claim 1 is characterized in that: In step 3, the real-time data is input into the target model for prediction analysis, and the prediction results of the user's thermal behavior are output, including: Input the real-time data into the target model, add position information to the real-time data according to the position encoding layer in the target model, and obtain the target sequence of the real-time data according to the adding result; The multi-head self-attention layer in the target model is used to capture key information of the target sequence, and obtain the operation characteristic data of the heating network and the user usage characteristic data in the current real-time data; The operating characteristic data of the heating network and the user usage characteristic data in the current real-time data are input into the feedforward neural network layer in the target model for processing to determine the prediction results of the user's heat usage behavior under the current operating characteristics of the heating network.

8. The method for analyzing the Internet of Things heating network based on the self-attention mechanism according to claim 1 is characterized in that: In step 4, the heating strategy for the user is dynamically adjusted based on the prediction result of the user's heat usage behavior, including: Read the prediction results of the user's heat consumption behavior, determine the change trend of the heating data provided to the user based on the heating network, and at the same time, perform data simulation on the change trend of the heating data to obtain the heating prediction parameters of the heating network for heating the user; Retrieve the current heating strategy for providing heating to users based on the heating network, and read the heating parameters in the current heating strategy; Matching the heat supply prediction parameters with the heat supply parameters, determining the difference parameters between the heat supply prediction parameters and the heat supply parameters, and using the difference parameters as target adjustment values; The target adjustment value is corrected in real time in the current heating strategy to complete the dynamic adjustment of the user's heating strategy.

9. The method for analyzing the Internet of Things heating network based on the self-attention mechanism according to claim 1, characterized in that: Based on the dynamically adjusted heating strategy, the heating status of users in the heating network is updated in real time according to the Internet of Things, including: Read the dynamically adjusted heating strategy and determine the heating adjustment instructions; Transmit heating adjustment instructions to the heating network based on the Internet of Things; The heating adjustment instructions are read according to the heating network and the heating status of the user is updated.

10. An IoT heating network analysis system based on self-attention mechanism, characterized in that: include: A data retrieval module is used to retrieve historical monitoring data of the heating network based on IoT devices, and encode the historical monitoring data to obtain characteristic sequence data; Model building module, used to learn feature sequence data and build a target model based on the self-attention mechanism; The prediction and analysis module is used to collect real-time data of the heating network operation based on IoT devices, input the real-time data into the target model for prediction and analysis, and output the prediction results of the user's heating behavior; The update module is used to dynamically adjust the heating strategy for users based on the predicted results of users' heat usage behavior, and to update the heating status of users in the heating network in real time according to the Internet of Things based on the dynamically adjusted heating strategy.

Images