A heat supply station heat supply area heat network data visualization monitoring method and system

Abstract

This invention discloses a data visualization monitoring method and system for heating networks in heating station supply areas, specifically relating to the management field. The method includes synchronously collecting multi-dimensional data, establishing a heating network spatial coordinate system to locate the data collection nodes, and storing the coordinate data in a data storage center. It employs a K-means clustering algorithm to extract three typical heating patterns and simultaneously evaluates the matching degree of user heating patterns. An optimization cycle is set to extract optimized samples, and the gradient descent method is used to minimize the sum of squared errors between model predictions and actual values, updating relevant model parameters. Based on data analysis results, this invention generates three targeted decision-making suggestions: resource allocation, inspection priority, and heating guidance. Combined with historical data tracing and operational pattern mining, it provides comprehensive support for short-term operational optimization and long-term planning of the heating network, effectively improving the efficiency of heating resource utilization, reducing operational risks, and minimizing resource waste, thus exhibiting significant economic and management benefits.

Description

Technical Field

[0001] This invention relates to the field of management technology, and more specifically, to a method and system for visual monitoring of heating network data in a heating station's heating area. Background Technology

[0002] In the operation and management of heating station systems, heat network data monitoring is a core element in ensuring the quality of heating services and optimizing resource allocation. Existing heat network monitoring technologies mainly collect basic operational data such as supply and return water temperature, pressure, and flow rate through sensors, and then store and display this data using a data management platform. This provides basic data query services for operation and management personnel, supporting the daily management of the heating system.

[0003] While existing technologies can collect and present basic data to meet minimum monitoring needs, they still have significant shortcomings from the perspective of digital management and decision support: First, data collection is limited to a single dimension, focusing only on the operating parameters of the heating network itself, without integrating key related data such as environmental characteristics of the heating area and user-side heating behavior, resulting in a one-sided data system that cannot fully support operational decision analysis; Second, data processing is insufficient, limited to basic operations such as deduplication and statistics, lacking in-depth processing capabilities such as multi-dimensional data correlation mining and user behavior pattern analysis, making it difficult to uncover the underlying operational patterns and potential risks; Third, visualization formats are rigid, mostly static tables and single curves, failing to intuitively display the relationships between data, spatial distribution characteristics, and trend evolution patterns, making it difficult for managers to quickly extract effective information; Fourth, risk warning and decision support are weak, triggering warnings only through simple threshold judgments, without constructing a multi-factor comprehensive risk assessment system, and lacking a decision suggestion generation mechanism based on data analysis, failing to provide scientific guidance for operational optimization and resource allocation.

[0004] To address the problems of limited data dimensions, insufficient processing depth, poor visualization effects, and weak decision support in existing technologies, this invention proposes a data visualization monitoring method and system for heating networks in heating station supply areas. It focuses on the entire process of data collection, processing, analysis, visualization, and decision support. Through multi-dimensional data integration, in-depth correlation analysis, innovative visualization presentation, and generation of scientific decision suggestions, it achieves intelligent management and efficient utilization of heating network operation data. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for visual monitoring of heating network data in the heating area of ​​a heating station, which solves the problems mentioned in the background art through the following solutions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for visually monitoring heat network data in a heating station's heating area, comprising: S1. Synchronously collect multi-dimensional data, establish a spatial coordinate system for the heating network to complete the positioning of the collection nodes, and store the coordinate data in the data storage center; S2. Outliers are removed using the Grubbs criterion, and missing values ​​are filled in using the weighted average of data from the same node in adjacent collection periods. The cleaned data is then standardized. S3. Construct a multi-dimensional correlation model for heating network data and obtain the operational coordination index; S4. Use the K-means clustering algorithm to extract three typical heating patterns, and evaluate the matching degree of user heating patterns. S5. Integrate multi-dimensional data to construct a fault risk assessment model and obtain a fault risk index; S6. Generate a 3D topology map of the heating network to visualize spatial relationships; generate a heat map of heating network risks to visualize risk distribution; draw dynamic time series curves of key parameters to visualize trend evolution; and use radar charts to visualize pattern matching. S7. Based on the fault risk index, three levels of early warning are divided, and corresponding level early warning prompts are triggered and pushed to mobile devices. Based on the operation coordination index, user heat usage pattern matching degree, fault risk index, and historical operation data, three types of decision suggestions are generated. S8. Stores full-process data, provides multi-condition combined query and data export functions, generates standardized data reports, and uses an LSTM neural network model to construct a heating network operation trend prediction model and calculate the predicted value of the coordination index for the next three days. S9. Set an optimization period, extract optimization samples, use gradient descent to minimize the sum of squared errors between the model predictions and actual values, and update the relevant model parameters.

[0007] Preferably, the multidimensional data includes heating network operation data, heating area environmental data, and user-side heat consumption data; the heating network operation data includes the water supply pipe temperature at each data collection node. Water supply pipe pressure , return water pipe temperature , return water pipe pressure Pipeline flow Thermal conductivity of pipe material Pipe wall thickness The environmental data for the heating area includes the ambient temperature corresponding to each data collection node. Ambient humidity Wind speed Sunshine duration The user-side heating data includes the indoor temperature of the users covered by each data collection node. Heating duration Set temperature Number of users The spatial coordinate system of the heating network has its origin at the geographical center of the heating station, with the extension direction of the main pipeline of the heating network as the X-axis, the horizontal direction perpendicular to the main pipeline as the Y-axis, and the vertical direction perpendicular to the XY plane as the Z-axis. The location of the data acquisition nodes involves assigning the spatial coordinates of each data acquisition node to the corresponding data acquisition node number. .

[0008] Preferably, the Grubbs criterion for removing outliers is based on the assumption that the measured value of a certain data indicator is... Calculate the mean and standard deviation Calculate the Grubbs statistic for each measurement. ,like The measured value is then determined to be an outlier and removed. This is the Grubbs critical value. The weighted mean of data from the same node in adjacent acquisition periods to fill in missing values ​​refers to the process of using this method to fill in missing values ​​in the dataset after removing outliers. Complete the task, including Data from the previous period. This is for the data in the next period; the standardization process involves converting all cleaned data into a unified dimension, through... Obtain standardized data, among which This is the historical average of the indicator. is the historical standard deviation of this indicator; the standardized data follows a standard normal distribution.

[0009] Preferably, the multi-dimensional correlation model of the heating network data is a model constructed by integrating heating network operation data, heating area environmental data, and user-side heating data to quantify the degree of coordination and matching among various parameters; the operation coordination index An indicator used to characterize the level of coordination and adaptation between the environmental conditions of the heating network operation at each node and the user's heating demand.

[0010] Preferably, the K-means clustering algorithm extracts three typical heating patterns for clustering historical user heating data, setting the number of clusters to three corresponding high, medium, and low heating demand patterns, thus obtaining feature vectors for the three typical heating patterns; the feature vectors of the typical heating patterns include a set temperature. Heating duration Number of users Specifically, it is: Mode 1 The user's heating mode matching degree The index is used to characterize the degree to which actual heat consumption behavior matches typical patterns; the failure risk index ,in The average of the coordination indexes across all nodes. This represents the historical average pressure of the water supply pipe. This represents the historical average pressure of the return water pipe.

[0011] Preferably, the spatial visualization of the three-dimensional topology map of the heating network is achieved by generating the three-dimensional topology map of the heating network based on a spatial coordinate system and labeling the coordinates of each node. The operation coordination index is represented by a color gradient; the heat network risk heat map visualizes the risk distribution by generating a heat network risk heat map, radiating outwards from the core of each node, with the radiation range positively correlated with the fault risk index, and the color intensity indicating the level of risk, overlaying to display the specific values ​​of the fault risk index; the key parameter time series dynamic curve visualizes the trend evolution by plotting the water supply pipe temperature. , return water pipe temperature Pipeline flow Operational Coordination Index Failure Risk Index The time series dynamic curve supports switching between day, month, and week for viewing. The curve inflection points are automatically marked and associated with the corresponding environmental data changes for the time period. The radar chart realizes pattern matching visualization by using a radar chart to show the matching degree between the user's heat usage pattern at each node and three typical patterns. Each dimension of the radar chart corresponds to the feature parameters.

[0012] Preferably, the three-level early warning level classification is based on a fault risk index, dividing risks into three levels: low risk, medium risk, and high risk, where low risk is... Medium risk High risk is The corresponding risk level warning prompts are triggered and pushed to mobile devices by a visual display terminal that pops up a warning prompt corresponding to the risk level: low risk is indicated by a blue prompt box, medium risk by a yellow prompt box, and high risk by a red prompt box. Simultaneously, the prompts are pushed to the mobile devices of operations management personnel. The generation of the three types of decision suggestions is based on the operational coordination index. User hot mode matching degree and failure risk index Based on historical operational data, resource allocation suggestions, inspection priority suggestions, and heat usage guidance suggestions are generated; the resource allocation suggestions are based on the operational synergy index. And user hot mode matching degree The node simultaneously obtains the resource allocation adjustment coefficient. Used to guide the optimization of heating plans, among which The average of the coordination index for all nodes; the inspection priority suggestion: obtain the inspection priority. The inspection plan recommendation list is generated by sorting the data from highest to lowest priority; the heating guidance suggestions are based on the matching degree of the user's heating mode. The nodes, combined with typical pattern characteristics, generate user heating habit optimization guidance schemes.

[0013] Preferably, the full-process data storage involves a data storage center storing collected data, preprocessing results, calculated indicators, early warning records, and decision-making suggestions; the multi-condition combined query and data export function generates standardized data reports by supporting multi-condition combined queries based on time intervals, node numbers, parameter types, and risk levels, exporting corresponding raw data and analysis results, and generating standardized data reports; the LSTM neural network model used to construct the heating network operation trend prediction model is a model built based on historical data using an LSTM neural network model to predict the heating network operation trend; the calculation of the predicted value of the coordination index for the next three days involves inputting the operation coordination index for the past seven days. Environmental data and user heat consumption data, through The predicted values ​​of the synergy index for the next three days are obtained, among which... , The weight matrix, For the bias vector, for Activation function The hyperbolic tangent activation function is used. This is the hidden layer state. For the input vector, It is in the cellular state.

[0014] Preferably, the optimization period is set to 30 days; the optimization sample extraction is to extract all data within each optimization period as optimization samples; the gradient descent method to minimize the sum of squared errors is to use the gradient descent method to minimize the sum of squared errors between the model's predicted value and the actual value; the relevant model parameter update is to update the feature weights in the calculation of the collaborative index, the parameter coefficients of the fault assessment model, and the weight matrix and bias vector of the LSTM model, thereby improving the accuracy of model analysis and prediction.

[0015] Preferably, a data visualization monitoring system for a heating network in a heating station's heating area includes: Data acquisition and positioning module: synchronously acquires multi-dimensional data, establishes a spatial coordinate system for the heating network to complete the positioning of acquisition nodes, and stores the coordinate data in the data storage center; Data cleaning and preprocessing module: Uses Grubbs' criterion to remove outliers, uses the weighted average of data from the same node in adjacent collection periods to fill in missing values, and performs standardization processing on the cleaned data; Association Model Calculation Module: Constructs a multi-dimensional association model for heating network data and obtains the operational synergy index; Heating pattern matching module: Uses K-means clustering algorithm to extract three typical heating patterns, and evaluates the matching degree of user heating patterns; Fault Risk Assessment Module: Integrates multi-dimensional data to construct a fault risk assessment model and obtain a fault risk index; The data intelligent visualization module generates a 3D topology map of the heating network to visualize spatial relationships, generates a heat map of heating network risks to visualize risk distribution, plots dynamic time series curves of key parameters to visualize trend evolution, and uses radar charts to visualize pattern matching. Early warning decision generation module: Based on the fault risk index, it divides the early warning level into three levels, triggers the corresponding level of early warning prompts and pushes them to the mobile terminal, and generates three types of decision suggestions based on the operation coordination index, user heating mode matching degree and fault risk index combined with historical operation data; Data storage and prediction module: Stores full-process data, provides multi-condition combined query and data export functions, generates standardized data reports, and uses an LSTM neural network model to build a heating network operation trend prediction model and calculate the predicted value of the coordination index for the next three days; Model parameter optimization module: Set the optimization period, extract optimization samples, use gradient descent to minimize the sum of squared errors between the model predictions and actual values, and update the relevant model parameters.

[0016] Technical effects and advantages of the present invention: 1. This invention focuses on the core aspects of data collection, processing, analysis, and decision support. By integrating multi-dimensional data on heating network operation, environment, and user-side data, it constructs a standardized preprocessing workflow, solving the problems of limited data dimensions and inconsistent data quality in existing technologies. Furthermore, by leveraging a synergistic index to quantify parameter correlation strength, it achieves in-depth data mining, fully realizing the value of data assets and providing a scientific basis for heating network operation decisions. 2. This invention innovatively designs a multi-level intelligent visualization presentation method, intuitively displaying the spatial correlation, operational status, and risk distribution of the heating network through 3D topology maps, risk heat maps, and dynamic trend curves. This solves the pain points of poor visualization effects and inefficient information transmission in existing technologies. It constructs a multi-factor fault risk assessment model to achieve accurate prediction and graded early warning of potential risks, overcoming the limitations of simple threshold-based early warning in existing technologies. This facilitates operation and management personnel in quickly focusing on key issues, improving the refinement and efficiency of heating network operation and management, and promoting the transformation of heating network monitoring towards data-driven intelligent decision-making. 3. This invention generates three types of targeted decision-making suggestions based on data analysis results: resource allocation, inspection priority, and heat usage guidance. Combined with historical data tracing and operational pattern mining, it provides comprehensive support for short-term operation optimization and long-term planning of the heating network, solving the problem of weak decision support in existing technologies. Through iterative optimization of model parameters, it continuously improves the accuracy of analysis and prediction, ensuring the adaptability and reliability of the method, effectively improving the efficiency of heating resource utilization, reducing operational risks, and reducing resource waste, thus possessing significant economic and management benefits. Attached Figure Description

[0017] Figure 1 This is a flowchart of the method steps of the present invention; Figure 2 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] As attached Figure 1 The method for visually monitoring heating network data in a heating station's heating area, as shown, includes the following steps: S1. Synchronously collect multi-dimensional data, establish a spatial coordinate system for the heating network to complete the positioning of the collection nodes, and store the coordinate data in the data storage center; In this embodiment, it should be specifically noted that: the multidimensional data includes heating network operation data, heating area environmental data, and user-side heat consumption data; the heating network operation data includes the water supply pipe temperature of each data collection node. Water supply pipe pressure , return water pipe temperature , return water pipe pressure Pipeline flow Thermal conductivity of pipe material Pipe wall thickness The environmental data for the heating area includes the ambient temperature corresponding to each data collection node. Ambient humidity Wind speed Sunshine duration The user-side heating data includes the indoor temperature of the users covered by each data collection node. Heating duration Set temperature Number of users The spatial coordinate system of the heating network has its origin at the geographical center of the heating station, with the extension direction of the main pipeline of the heating network as the X-axis, the horizontal direction perpendicular to the main pipeline as the Y-axis, and the vertical direction perpendicular to the XY plane as the Z-axis. The location of the data acquisition nodes involves assigning the spatial coordinates of each data acquisition node to the corresponding data acquisition node number. The operation of the heating network is influenced by a combination of factors, including its own condition, regional environment, and user needs. Comprehensive collection of three types of core data can provide complete data support for subsequent analysis. Spatial positioning can clearly identify the physical nodes corresponding to the data, ensuring that the analysis results accurately correspond to the actual heating network nodes, which meets the spatial correlation requirements of heating network monitoring. The collected data are all key indicators affecting the operation of the heating network. The coordinate system is established in accordance with the general logic of spatial positioning, which can achieve accurate identification of the collected nodes and lay the foundation for subsequent multi-level analysis.

[0020] S2. Outliers are removed using the Grubbs criterion, and missing values ​​are filled in using the weighted average of data from the same node in adjacent collection periods. The cleaned data is then standardized. In this embodiment, it should be specifically noted that: the Grubbs criterion for removing outliers is based on the assumption that the measured value of a certain data indicator is... Calculate the mean and standard deviation Calculate the Grubbs statistic for each measurement. ,like The measured value is then determined to be an outlier and removed. This is the Grubbs critical value. The weighted mean of data from the same node in adjacent acquisition periods to fill in missing values ​​refers to the process of using this method to fill in missing values ​​in the dataset after removing outliers. Complete the task, including Data from the previous period. This is for the data in the next period; the standardization process involves converting all cleaned data into a unified dimension, through... Obtain standardized data, among which This is the historical average of the indicator. The standard deviation of this indicator is given, and the standardized data follows a standard normal distribution. However, the collected data may contain outliers or missing values ​​due to equipment errors or environmental interference. If left unprocessed, these outliers will directly affect the accuracy of subsequent analysis results. Standardization eliminates dimensional differences between different indicators, making multi-parameter correlation analysis feasible. The Grubbs criterion used is a mature and reliable method for outlier removal. Weighted mean imputation of missing values ​​takes into account the correlation between data from different periods. The standardization formula conforms to general statistical analysis standards and effectively improves data quality.

[0021] S3. Construct a multi-dimensional correlation model for heating network data and obtain the operational coordination index; In this embodiment, it should be specifically noted that: the multi-dimensional correlation model of the heating network data is a model constructed by integrating heating network operation data, heating area environmental data, and user-side heating data to quantify the degree of coordination and matching between various parameters; the operation coordination index This indicator is used to characterize the level of coordination and adaptation between the environmental conditions and user heating demand at each node of the heating network. Heating network operation is a systematic process involving the interaction of multiple parameters; a single parameter cannot reflect the overall adaptation status of heating network operation, environmental conditions, and user demand. Constructing a correlation model and quantifying the degree of coordination can objectively present the comprehensive operational level of each node. The model integrates three key data categories: heating network, environment, and users. Its calculation logic follows the actual impact of each parameter on heating network operation, accurately characterizing the coordination and adaptation of each node and providing a scientific basis for subsequent risk assessment.

[0022] S4. Use the K-means clustering algorithm to extract three typical heating patterns, and evaluate the matching degree of user heating patterns. In this embodiment, it should be specifically explained that: the K-means clustering algorithm extracts three typical heating patterns for clustering historical user heating data, setting the number of clusters to three corresponding high, medium, and low heating demand patterns, thus obtaining feature vectors for the three typical heating patterns; the feature vectors of the typical heating patterns include a set temperature. Heating duration Number of users Specifically, it is: Mode 1 The user's heating mode matching degree This indicator is used to characterize the degree of fit between actual heating behavior and typical patterns. Among these, common patterns exist in user heating behavior; extracting typical heating patterns simplifies complex user data; matching degree calculation can quickly distinguish the heating characteristics of users in different regions, providing direction for targeted operation and management; the K-means clustering algorithm is suitable for pattern extraction; the division of the three types of patterns conforms to the hierarchical characteristics of user heating needs in most scenarios; and the matching degree calculation comprehensively considers the differences in key feature parameters, accurately reflecting the degree of fit between actual heating behavior and typical patterns.

[0023] S5. Integrate multi-dimensional data to construct a fault risk assessment model and obtain a fault risk index; In this embodiment, it should be specifically noted that the fault risk index ,in The average of the coordination indexes across all nodes. This represents the historical average pressure of the water supply pipe. This represents the historical average pressure of the return water pipe. It's important to understand that heating network failures are not caused by a single factor; a comprehensive consideration of factors such as coordination and adaptation levels, pressure stability, and environmental impact is necessary. Constructing a multi-factor assessment model can improve the accuracy of risk prediction and avoid the limitations of single-indicator warnings. The model integrates the previously calculated coordination index with basic operational data, and its calculation logic aligns with the inherent patterns of heating network failures. It can quantify the potential failure probability of each node, providing a scientific quantitative basis for subsequent risk warnings.

[0024] S6. Generate a 3D topology map of the heating network to visualize spatial relationships; generate a heat map of heating network risks to visualize risk distribution; draw dynamic time series curves of key parameters to visualize trend evolution; and use radar charts to visualize pattern matching. In this embodiment, it should be specifically noted that: the spatial correlation visualization of the three-dimensional topology map of the heat network is based on generating the three-dimensional topology map of the heat network using a spatial coordinate system and labeling the coordinates of each node. The operation coordination index is represented by a color gradient; the heat network risk heat map visualizes the risk distribution by generating a heat network risk heat map, radiating outwards from the core of each node, with the radiation range positively correlated with the fault risk index, and the color intensity indicating the level of risk, overlaying to display the specific values ​​of the fault risk index; the key parameter time series dynamic curve visualizes the trend evolution by plotting the water supply pipe temperature. , return water pipe temperature Pipeline flow Operational Coordination Index Failure Risk Index The system features dynamic time-series curves that support viewing by day, month, and week. Curve inflection points are automatically marked and associated with corresponding environmental data changes for each time period. The radar chart, used for pattern matching visualization, displays the matching degree between user heating patterns at each node and three typical patterns, with each dimension of the radar chart corresponding to characteristic parameters. Management personnel require intuitive and efficient information acquisition methods; a single presentation format cannot meet the information needs of different scenarios. Multi-level visualization can convey data information from multiple dimensions, including spatial distribution, risk level, trend changes, and pattern matching. The 3D topology map matches the spatial distribution characteristics of the heating network, the risk heat map highlights high-risk areas, the dynamic curves clearly show parameter change trends, and the radar chart intuitively presents pattern matching differences. These various visualization methods are highly targeted, ensuring efficient and accurate information transmission.

[0025] S7. Based on the fault risk index, three levels of early warning are divided, and corresponding level early warning prompts are triggered and pushed to mobile devices. Based on the operation coordination index, user heat usage pattern matching degree, fault risk index, and historical operation data, three types of decision suggestions are generated. In this embodiment, it should be specifically noted that the three-level early warning level classification is based on a fault risk index, dividing risks into three levels: low risk, medium risk, and high risk. Low risk is... Medium risk High risk is The corresponding risk level warning prompts are triggered and pushed to mobile devices by a visual display terminal that pops up a warning prompt corresponding to the risk level: low risk is indicated by a blue prompt box, medium risk by a yellow prompt box, and high risk by a red prompt box. Simultaneously, the prompts are pushed to the mobile devices of operations management personnel. The generation of the three types of decision suggestions is based on the operational coordination index. User hot mode matching degree and failure risk index Based on historical operational data, resource allocation suggestions, inspection priority suggestions, and heat usage guidance suggestions are generated; the resource allocation suggestions are based on the operational synergy index. And user hot mode matching degree The node simultaneously obtains the resource allocation adjustment coefficient. Used to guide the optimization of heating plans, among which The average of the coordination index for all nodes; the inspection priority suggestion: obtain the inspection priority. The inspection plan recommendation list is generated by sorting the data from highest to lowest priority; the heating guidance suggestions are based on the matching degree of the user's heating mode. The system generates user heating habit optimization guidance plans by combining typical pattern characteristics at key nodes. Different risk levels correspond to different response priorities, and tiered early warnings allow managers to quickly focus on key issues. Decision recommendations must be based on multi-dimensional analysis results to provide specific and feasible guidance for operational optimization, meeting the core needs of digital decision-making. The risk grading standard is based on a reasonable range of risk indices, and the early warning method ensures that information reaches managers in a timely manner. The three types of decision recommendations are designed for resource allocation, inspection planning, and heating usage guidance scenarios, respectively, and are logically closely linked to the previously analyzed indicators, making the recommendations scientific and feasible.

[0026] S8. Stores full-process data, provides multi-condition combined query and data export functions, generates standardized data reports, and uses an LSTM neural network model to construct a heating network operation trend prediction model and calculate the predicted value of the coordination index for the next three days. In this embodiment, it should be specifically noted that: the full-process data storage refers to the data storage center storing collected data, preprocessing results, calculated indicators, early warning records, and decision suggestions; the multi-condition combined query and data export function generates standardized data reports by supporting multi-condition combined queries based on time intervals, node numbers, parameter types, and risk levels, exporting corresponding raw data and analysis results, and generating standardized data reports; the LSTM neural network model used to construct the heating network operation trend prediction model is a model built based on historical data using an LSTM neural network model to predict the heating network operation trend; the calculation of the predicted value of the coordination index for the next three days involves inputting the operation coordination index for the past seven days. Environmental data and user heat consumption data, through The predicted values ​​of the synergy index for the next three days are obtained, among which... , The weight matrix, For the bias vector, for Activation function The hyperbolic tangent activation function is used. This is the hidden layer state. For the input vector, The system is structured in a cellular manner. Historical data serves as a crucial foundation for problem tracing and operational pattern discovery. Trend prediction enables advance assessment of the heating network's operational status, supporting short-term adjustments and long-term planning, aligning with the continuous and forward-looking requirements of heating network operation. Data storage covers key information across the entire process, ensuring the integrity and traceability of the tracing process. The employed LSTM neural network model excels at processing time-series data, adapting to the forecasting needs of heating network operational trends. The input and output parameters are designed to fit actual forecasting scenarios, effectively improving forecast accuracy.

[0027] S9. Set an optimization period, extract optimization samples, use gradient descent to minimize the sum of squared errors between the model predictions and actual values, and update the relevant model parameters.

[0028] In this embodiment, the following details need to be explained: the optimization cycle is set to 30 days; the optimization sample extraction involves extracting all data within each optimization cycle as optimization samples; the gradient descent method minimizes the sum of squared errors by using the gradient descent method to minimize the sum of squared errors between the model's predicted and actual values; and the relevant model parameter update updates the feature weights in the operational coordination index calculation, the parameter coefficients of the fault assessment model, and the weight matrix and bias vector of the LSTM model, thereby improving the accuracy of model analysis and prediction. The heating network operating environment and user heating demand change dynamically over time. Fixing model parameters will lead to a decrease in analysis and prediction accuracy. Regular optimization allows the model to adapt to these changes and maintain the long-term effectiveness of the method. The set optimization cycle conforms to the accumulation pattern of heating network operating data. Gradient descent is a mature and effective method for parameter optimization. By updating parameters by minimizing the sum of squared errors, the model can be continuously calibrated, improving the model's analysis and prediction accuracy and ensuring the long-term applicability of the entire monitoring method.

[0029] The above scheme is as follows (attached). Figure 2 As shown, the present invention also provides a data visualization and monitoring system for the heating network in the heating area of ​​a heating station, comprising: Data acquisition and positioning module: synchronously acquires multi-dimensional data, establishes a spatial coordinate system for the heating network to complete the positioning of acquisition nodes, and stores the coordinate data in the data storage center; Data cleaning and preprocessing module: Uses Grubbs' criterion to remove outliers, uses the weighted average of data from the same node in adjacent collection periods to fill in missing values, and performs standardization processing on the cleaned data; Association Model Calculation Module: Constructs a multi-dimensional association model for heating network data and obtains the operational synergy index; Heating pattern matching module: Uses K-means clustering algorithm to extract three typical heating patterns, and evaluates the matching degree of user heating patterns; Fault Risk Assessment Module: Integrates multi-dimensional data to construct a fault risk assessment model and obtain a fault risk index; The data intelligent visualization module generates a 3D topology map of the heating network to visualize spatial relationships, generates a heat map of heating network risks to visualize risk distribution, plots dynamic time series curves of key parameters to visualize trend evolution, and uses radar charts to visualize pattern matching. Early warning decision generation module: Based on the fault risk index, it divides the early warning level into three levels, triggers the corresponding level of early warning prompts and pushes them to the mobile terminal, and generates three types of decision suggestions based on the operation coordination index, user heating mode matching degree and fault risk index combined with historical operation data; Data storage and prediction module: Stores full-process data, provides multi-condition combined query and data export functions, generates standardized data reports, and uses an LSTM neural network model to build a heating network operation trend prediction model and calculate the predicted value of the coordination index for the next three days; Model parameter optimization module: Set the optimization period, extract optimization samples, use gradient descent to minimize the sum of squared errors between the model predictions and actual values, and update the relevant model parameters.

[0030] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for visually monitoring heating network data in a heating station's heating area, characterized in that... ,include: S1. Synchronously collect multi-dimensional data, establish a spatial coordinate system for the heating network to complete the positioning of the collection nodes, and store the coordinate data in the data storage center; S2. Outliers are removed using the Grubbs criterion, and missing values ​​are filled in using the weighted average of data from the same node in adjacent collection periods. The cleaned data is then standardized. S3. Construct a multi-dimensional correlation model for heating network data and obtain the operational coordination index; S4. Use the K-means clustering algorithm to extract three typical heating patterns, and evaluate the matching degree of user heating patterns. S5. Integrate multi-dimensional data to construct a fault risk assessment model and obtain a fault risk index; S6. Generate a 3D topology map of the heating network to visualize spatial relationships; generate a risk heat map of the heating network to visualize risk distribution; draw dynamic time series curves of key parameters to visualize trend evolution; and use radar charts to visualize pattern matching. S7. Based on the fault risk index, three levels of early warning are divided, and corresponding level early warning prompts are triggered and pushed to mobile devices. Based on the operation coordination index, user heating mode matching degree, fault risk index and historical operation data, three types of decision suggestions are generated. S8. Stores full-process data, provides multi-condition combined query and data export functions, generates standardized data reports, and uses an LSTM neural network model to construct a heating network operation trend prediction model and calculate the predicted value of the coordination index for the next three days. S9. Set an optimization period, extract optimization samples, use gradient descent to minimize the sum of squared errors between the model predictions and actual values, and update the relevant model parameters.

2. The method for visual monitoring of heating network data in a heating station's heating area according to claim 1, characterized in that... The multidimensional data includes heating network operation data, heating area environmental data, and user-side heat consumption data; the heating network operation data includes the water supply pipe temperature at each data collection node. Water supply pipe pressure , return water pipe temperature , return water pipe pressure Pipeline flow Thermal conductivity of pipe material Pipe wall thickness The environmental data for the heating area includes the ambient temperature corresponding to each data collection node. Ambient humidity Wind speed Sunshine duration The user-side heating data includes the indoor temperature of the users covered by each data collection node. Heating duration Set temperature Number of users The spatial coordinate system of the heating network has its origin at the geographical center of the heating station, with the extension direction of the main pipeline of the heating network as the X-axis, the horizontal direction perpendicular to the main pipeline as the Y-axis, and the vertical direction perpendicular to the XY plane as the Z-axis. The location of the data acquisition nodes involves assigning the spatial coordinates of each data acquisition node to the corresponding data acquisition node number. .

3. The method for visual monitoring of heating network data in a heating station's heating area according to claim 1, characterized in that... The Grubbs criterion for removing outliers is based on the assumption that the measured value of a certain data indicator is... Calculate the mean and standard deviation Calculate the Grubbs statistic for each measurement. ,like The measured value is then determined to be an outlier and removed. This is the Grubbs critical value. The weighted mean of data from the same node in adjacent acquisition periods to fill in missing values ​​is a formula used to fill in missing values ​​in the dataset after removing outliers. Complete the task, including Data from the previous period. The data is for the next period; the standardization process is to convert all the cleaned data into a uniform dimension, and the standardized data follows a standard normal distribution.

4. The method for visual monitoring of heating network data in a heating station's heating area according to claim 2, characterized in that... The multi-dimensional correlation model of the heating network data is a model constructed by integrating heating network operation data, heating area environmental data, and user-side heat consumption data to quantify the degree of coordination and matching among various parameters; the operation coordination index An indicator used to characterize the level of coordination and adaptation between the environmental conditions of the heating network operation at each node and the user's heating demand.

5. A method for visual monitoring of heating network data in a heating station's heating area according to claim 2, characterized in that... The K-means clustering algorithm extracts three typical heating patterns for clustering historical user heating data. The number of clusters is set to three, corresponding to high, medium, and low heating demand patterns, resulting in feature vectors for the three typical heating patterns. The feature vectors of the typical heating patterns include a set temperature. Heating duration Number of users Specifically, it is: Mode 1 The user heating pattern matching degree is an indicator used to characterize the degree of fit between actual heating behavior and typical patterns; the fault risk index ,in The average of the coordination indexes across all nodes. The historical average pressure of the water supply pipe. This represents the historical average pressure of the return water pipe.

6. The method for visual monitoring of heating network data in a heating station's heating area according to claim 1, characterized in that... The spatial visualization of the three-dimensional topology map of the heating network is achieved by generating the three-dimensional topology map of the heating network based on a spatial coordinate system and labeling the coordinates of each node. The operation coordination index is represented by a color gradient; the heat network risk heat map visualizes the risk distribution by generating a heat network risk heat map, radiating outwards from the core of each node, with the radiation range positively correlated with the fault risk index, and the color intensity indicating the level of risk, overlaying to display the specific values ​​of the fault risk index; the key parameter time series dynamic curve visualizes the trend evolution by plotting the water supply pipe temperature. , return water pipe temperature Pipeline flow Operational Coordination Index Failure Risk Index The time series dynamic curve supports switching between day, month, and week for viewing. The curve inflection points are automatically marked and associated with the corresponding environmental data changes for the time period. The radar chart realizes pattern matching visualization by using a radar chart to show the matching degree between the user's heat usage pattern at each node and three typical patterns. Each dimension of the radar chart corresponds to the feature parameters.

7. A method for visual monitoring of heating network data in a heating station's heating area according to claim 1, characterized in that... The aforementioned three-level early warning classification is based on a fault risk index, dividing risks into three levels: low risk, medium risk, and high risk. Low risk is... Medium risk High risk is The corresponding risk level warning prompts are triggered and pushed to mobile devices by a visual display terminal that pops up a warning prompt corresponding to the risk level: low risk is indicated by a blue prompt box, medium risk by a yellow prompt box, and high risk by a red prompt box. Simultaneously, the prompts are pushed to the mobile devices of operations management personnel. The generation of the three types of decision suggestions is based on the operational coordination index. User hot mode matching degree and failure risk index Based on historical operational data, resource allocation suggestions, inspection priority suggestions, and heat usage guidance suggestions are generated; the resource allocation suggestions are based on the operational synergy index. And user hot mode matching degree The node simultaneously obtains the resource allocation adjustment coefficient. Used to guide heating plan optimization; the inspection priority suggestion: obtain inspection priority. The inspection plan recommendation list is generated by sorting the data from highest to lowest priority; the heating guidance suggestions are based on the matching degree of the user's heating mode. The nodes, combined with typical pattern characteristics, generate user heating habit optimization guidance schemes.

8. A method for visual monitoring of heating network data in a heating station's heating area according to claim 1, characterized in that... The full-process data storage refers to the data storage center storing collected data, preprocessing results, calculated indicators, early warning records, and decision-making suggestions. The multi-condition combined query and data export function generates standardized data reports by supporting multi-condition combined queries based on time intervals, node numbers, parameter types, and risk levels, exporting corresponding raw data and analysis results, and generating standardized data reports. The LSTM neural network model used to construct the heating network operation trend prediction model is based on historical data and uses an LSTM neural network model to predict the heating network operation trend. The calculation of the predicted value of the coordination index for the next three days involves inputting the operation coordination index for the past seven days. Environmental data and user heat consumption data are used to obtain the predicted value of the synergy index for the next three days through a prediction formula.

9. A method for visual monitoring of heating network data in a heating station's heating area according to claim 1, characterized in that... The optimization cycle setting is set to 30 days per cycle; the optimization sample extraction is to extract all data within each optimization cycle as optimization samples; the gradient descent method to minimize the sum of squared errors is to use the gradient descent method to minimize the sum of squared errors between the model's predicted value and the actual value; the relevant model parameter update is to update the feature weights in the calculation of the collaborative index, the parameter coefficients of the fault assessment model, and the weight matrix and bias vector of the LSTM model, thereby improving the accuracy of model analysis and prediction.

10. A data visualization and monitoring system for a heating network in a heating station's heating area, used to implement the data visualization and monitoring method for a heating network in a heating station's heating area as described in any one of claims 1 to 9, characterized in that... ,include: Data acquisition and positioning module: synchronously acquires multi-dimensional data, establishes a spatial coordinate system for the heating network to complete the positioning of acquisition nodes, and stores the coordinate data in the data storage center; Data cleaning and preprocessing module: Uses Grubbs' criterion to remove outliers, uses the weighted average of data from the same node in adjacent collection periods to fill in missing values, and performs standardization processing on the cleaned data; Association Model Calculation Module: Constructs a multi-dimensional association model for heating network data and obtains the operational synergy index; Heating pattern matching module: Uses K-means clustering algorithm to extract three typical heating patterns, and evaluates the matching degree of user heating patterns; Fault Risk Assessment Module: Integrates multi-dimensional data to construct a fault risk assessment model and obtain a fault risk index; The data intelligent visualization module generates a 3D topology map of the heating network to visualize spatial relationships, generates a heat map of heating network risks to visualize risk distribution, plots dynamic time series curves of key parameters to visualize trend evolution, and uses radar charts to visualize pattern matching. Early warning decision generation module: Based on the fault risk index, it divides the early warning level into three levels, triggers the corresponding level of early warning prompts and pushes them to the mobile terminal, and generates three types of decision suggestions based on the operation coordination index, user heating mode matching degree and fault risk index combined with historical operation data; Data storage and prediction module: Stores full-process data, provides multi-condition combined query and data export functions, generates standardized data reports, and uses an LSTM neural network model to build a heating network operation trend prediction model and calculate the predicted value of the coordination index for the next three days; Model parameter optimization module: Set the optimization period, extract optimization samples, use gradient descent to minimize the sum of squared errors between the model predictions and actual values, and update the relevant model parameters.

Images