A heat network comprehensive monitoring method, device, equipment and medium

Abstract

The application provides a heat supply network comprehensive monitoring method, device, equipment and medium, comprising: obtaining heat exchange station information and heat supply pipe network information of a target area; determining space-time information of the heat exchange station information and the heat supply pipe network information according to multiple monitoring points; performing correlation processing on the heat exchange station information and the heat supply pipe network information according to the space-time information to obtain a topology graph and data flow of each monitoring point on the topology graph; performing feature extraction on the topology graph to obtain first features of each monitoring point; performing feature extraction on the data flow to obtain second features; splicing the first features and the second features to obtain third features according to each monitoring point; performing fault prediction according to the third features to obtain early warning risk values of each monitoring point; and starting emergency processes of different levels according to the data flow and the early warning risk values. The heat exchange station information and the heat supply pipe network information are integrated together, faults are comprehensively judged, and information resource utilization rate and fault early warning accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of smart heating technology, and in particular to a method, device, equipment and medium for comprehensive monitoring of heating networks. Background Technology

[0002] Existing heating network monitoring systems mostly focus on equipment operation monitoring. Even when other systems are included (such as security systems and automatic fire alarm systems), they operate independently, resulting in low resource utilization, high maintenance costs, and low collaboration efficiency. Therefore, there is an urgent need for a method that can comprehensively analyze and judge multi-dimensional and multi-system information from heating network monitoring systems. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method, device, equipment and medium for comprehensive monitoring of heating networks, which integrates multi-source data of a heating network monitoring system in a certain area, comprehensively judges faults, and improves the utilization rate of information resources and the accuracy of fault early warning.

[0004] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: A method for comprehensive monitoring of heating networks, comprising: Based on multiple monitoring points set up in the heat exchange stations and heating pipelines in the target area, information on the heat exchange stations and heating pipelines for a preset time period is obtained, as well as spatiotemporal information of the heat exchange station information and heating pipelines. Based on the spatiotemporal information, the heat exchange station information and the heating pipeline network information are correlated to obtain a topology map and the data stream of each monitoring point on the topology map; the nodes in the topology map are the monitoring points set in the heat exchange station and the heating pipeline network; the time points of the data stream of each monitoring point on the topology map correspond one-to-one. Feature extraction is performed on the topology map to obtain the first feature of each monitoring point; Feature extraction is performed on the data stream to obtain the second feature of each monitoring point; For each monitoring point, the first feature and the second feature are concatenated to obtain the third feature for each monitoring point; Based on the third feature, fault prediction is performed to obtain the early warning risk value for each monitoring point; Based on the information about the heat exchange station, the heating network, and the early warning risk value, different levels of emergency procedures will be initiated.

[0005] Optionally, obtain information on heat exchange stations and heating network in the target area, including: Acquire environmental monitoring information, security information, fire alarm system and equipment information of heat exchange stations in the target area, as well as flow information of heating pipeline network, fiber optic detection information of buried heating pipelines and video monitoring information of overhead pipelines.

[0006] Optionally, the spatiotemporal information of the heat exchange station information and heating network information is obtained, including: Based on the location information of the monitoring points at the heat exchange station, the first spatial information is determined; The second spatial information is determined based on the location information of the monitoring points of the heating pipeline network. Based on the set of time points from which information about the heat exchange station was obtained, determine the first-time information; The second time information is determined based on the set of time points from which heating network information is obtained; The spatiotemporal information is obtained based on the first spatial information, the second spatial information, the first time information, and the second time information.

[0007] Optionally, based on the spatiotemporal information, the heat exchange station information and the heating network information are correlated to obtain a topology map and a data stream for each monitoring point on the topology map, including: Based on the first spatial information and the second spatial information, a topology map is determined; Based on the first time information and the second time information, the heat exchange station information and the heating pipeline network information are time-aligned to obtain the data stream of each monitoring point on the topology map.

[0008] Optionally, feature extraction is performed on the topology map to obtain the first feature of each monitoring point, including: Establish an adjacency list based on the aforementioned topology graph; Based on the adjacency list and the preset target optimization function, the topology graph is divided into partitions centered on the monitoring points; Feature extraction is performed on each partition to obtain the first feature of each monitoring point.

[0009] Optionally, feature extraction is performed on the data stream to obtain a second feature for each monitoring point, including: If the data in the data stream is numerical, then determine the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, and steady-state deviation based on the data stream; If the data in the data stream is an image, then determine the image features based on the data stream; The second feature of each monitoring point is determined based on the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, steady-state deviation, and image features.

[0010] Optionally, based on the heat exchange station information, heating network information, and the early warning risk value, different levels of emergency procedures can be initiated, including: Determine whether the data in the heat exchange station information and heating network information are within a first preset range. If they are within the first preset range, then initiate an alarm process. Based on the preset first threshold, second threshold, and third threshold, the warning level of the warning risk value is determined, and the emergency procedure is initiated according to the warning level.

[0011] A comprehensive monitoring device for heating networks includes: The data acquisition module is used to acquire heat exchange station information and heating pipeline information for a preset time period based on multiple monitoring points set up in the target area, as well as the spatiotemporal information of the heat exchange station information and heating pipeline information. The processing module is used to correlate the heat exchange station information and the heating pipeline network information according to the spatiotemporal information to obtain a topology map and a data stream for each monitoring point on the topology map; the nodes in the topology map are monitoring points set on the heat exchange station and the heating pipeline network; the time points of the data streams of each monitoring point on the topology map correspond one-to-one; feature extraction is performed on the topology map to obtain a first feature for each monitoring point; feature extraction is performed on the data streams to obtain a second feature for each monitoring point; the first and second features are concatenated according to each monitoring point to obtain a third feature for each monitoring point; fault prediction is performed based on the third feature to obtain a warning risk value for each monitoring point; and different levels of emergency procedures are initiated based on the heat exchange station information, the heating pipeline network information, and the warning risk value.

[0012] A computing device includes: one or more processors; and a storage device for storing one or more programs that, when executed by the one or more processors, cause the one or more processors to implement the method.

[0013] A computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0014] The above-described solution of the present invention has at least the following beneficial effects: The above-described solution of the present invention acquires heat exchange station information and heating pipeline network information of the target area based on multiple monitoring points in the target area, as well as spatiotemporal information of the heat exchange station information and heating pipeline network information; based on the spatiotemporal information, the heat exchange station information and heating pipeline network information are correlated to obtain a topology map and a data stream for each monitoring point on the topology map; features are extracted from the topology map to obtain a first feature for each monitoring point; features are extracted from the data stream to obtain a second feature for each monitoring point; the first and second features are concatenated for each monitoring point to obtain a third feature for each monitoring point; fault prediction is performed based on the third feature to obtain a warning risk value for each monitoring point; and different levels of emergency procedures are initiated based on the data stream and the warning risk value. This application integrates heat exchange station information and heating pipeline network information together to comprehensively judge faults, improve information resource utilization and fault warning accuracy. Attached Figure Description

[0015] Figure 1 This is a schematic flowchart of an embodiment of the integrated monitoring method for heating networks of the present invention; Figure 2 This is a schematic diagram of the structure of the integrated monitoring device for the heating network of the present invention; Figure 3 This is a schematic diagram of an application scenario of the present invention. Detailed Implementation

[0016] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0017] like Figure 1 As shown, an embodiment of the present invention proposes a method for comprehensive monitoring of a heating network, comprising: Step 11: Based on multiple monitoring points set up in the heat exchange stations and heating pipe networks in the target area, obtain heat exchange station information and heating pipe network information for a preset time period, as well as obtain the spatiotemporal information of the heat exchange station information and heating pipe network information. Step 12: Based on the spatiotemporal information, the heat exchange station information and the heating network information are correlated to obtain a topology map and the data stream of each monitoring point on the topology map; the nodes in the topology map are the monitoring points set in the heat exchange station and the heating network; the time points of the data stream of each monitoring point on the topology map correspond one-to-one. Step 13: Extract features from the topology map to obtain the first feature of each monitoring point; Step 14: Extract features from the data stream to obtain the second feature for each monitoring point; Step 15: For each monitoring point, the first feature and the second feature are concatenated to obtain the third feature for each monitoring point; Step 16: Based on the third feature, perform fault prediction to obtain the early warning risk value for each monitoring point; Step 17: Initiate emergency procedures of different levels based on the data stream and the early warning risk value.

[0018] In this embodiment, the target area is equipped with multiple heat exchange stations and heating pipes connecting these stations, forming a heating network. Each heat exchange station has multiple monitoring points, such as access control and temperature controllers. Similarly, the heating network has multiple monitoring points, such as pressure and temperature sensors on each section of the pipe. By combining data from all monitoring points in the target area, a fault in the heating system can be determined, improving the accuracy of the assessment.

[0019] In some alternative implementations, step 11 involves obtaining information on heat exchange stations and heating networks in the target area, including: Step 111: Obtain environmental monitoring information, security information, fire alarm system and heat exchange station equipment information of the target area, as well as flow information of the heating network, fiber optic detection information of buried heating pipelines and video monitoring information of overhead pipelines.

[0020] In this embodiment, reference Figure 3 For step 111, the heat exchange station in the target area includes an environmental monitoring module, a security module, an equipment management module, and a fire alarm module.

[0021] The monitoring points in the environmental monitoring module include: water immersion sensors, lighting controllers, temperature and humidity controllers, and fan controllers; the monitoring points in the security module include: video surveillance systems, access control systems, and intrusion alarm systems; the monitoring points in the equipment management module include: data acquisition devices for circulating pump controller boxes, water supply pump control boxes, water supply tanks, and pressure sensors; and the monitoring points in the fire alarm module include: fire detectors, fire emergency broadcast / alarm devices, and linkage units.

[0022] Within the target time period, the collection of data streams from each monitoring point in the environmental monitoring module constitutes environmental monitoring information; the collection of data streams from each monitoring point in the security module constitutes security information; the collection of data streams from each monitoring point in the equipment management module constitutes equipment management information; and the collection of data streams from each monitoring point in the fire alarm module constitutes fire alarm information.

[0023] In this embodiment, each pipeline in the heating network is equipped with data acquisition devices related to flow rate, fiber optic monitoring, and video surveillance. During the target time period, flow rate information, fiber optic detection information for directly buried heating pipelines, and video surveillance information for overhead pipelines are acquired based on the data acquisition devices installed on each pipeline.

[0024] In some alternative implementations, step 11, obtaining the spatiotemporal information of the heat exchange station information and the heating network information, includes: Step 112: Determine the first spatial information based on the location information of the monitoring points of the heat exchange station; Step 113: Determine the second spatial information based on the location information of the monitoring points of the heating pipeline network; Step 114: Determine the first time information based on the set of time points from which the heat exchange station information was obtained; Step 115: Determine the second time information based on the set of time points from which heating network information was obtained; Step 116: Obtain the spatiotemporal information based on the first spatial information, the second spatial information, the first time information, and the second time information.

[0025] In this embodiment, for step 112, the first spatial information is determined based on the coordinates of multiple monitoring points in the heat exchange station; for step 113, the second spatial information is determined based on the coordinates of multiple monitoring points in the heating network.

[0026] For step 114, considering that the frequency of data collection at each monitoring point may be inconsistent within the preset time period, or that there may be short-term shutdowns due to maintenance, it cannot be guaranteed that heat exchange station information can be obtained at every time point within the preset time period. Therefore, it is necessary to determine the first time information based on the set of time points of the obtained heat exchange station information.

[0027] For each module within the heat exchange station, iterate through all time points of data collected by monitoring points within that module in the time point set of heat exchange station information. If all monitoring points within that module detect valid heat exchange station information at a certain time point, then that time point is determined as a candidate time point. Then, compare the candidate time points of all modules within the heat exchange station, filter out time points with missing values, and determine the remaining time points as valid time points. The set of valid time points of the heat exchange station is the first time information. This ensures that valid heat exchange station information is collected at all valid time points of the heat exchange station.

[0028] Similarly, by iterating through all time points in the set of time points for the heating network information, if all monitoring points within the heating network detect valid values ​​at a certain time point, then that time point is determined as a valid time point. This yields all valid time points for the heating network within the preset time period, and the set of valid time points for the heating network constitutes the second time information. This ensures that valid data is collected for all valid time points within the heating network.

[0029] In this embodiment, effective time points are selected in modules and steps, and a large amount of data is distributed and analyzed in parallel to improve data processing efficiency.

[0030] For step 115, the spatiotemporal information is obtained based on the first spatial information, the second spatial information, the first time information, and the second time information.

[0031] This embodiment determines the first spatial information and the second spatial information, ensuring that all time points within the first time information correspond to the monitoring values ​​of all monitoring points within the heat exchange station, and ensuring that all time points within the second time information correspond to the monitoring values ​​of all monitoring points within the heating pipeline network, thus providing data support for subsequent calculations.

[0032] In some optional implementations, in step 12, the heat exchange station information and the heating network information are correlated based on the spatiotemporal information to obtain a topology map and a data stream for each monitoring point on the topology map, including: Step 121: Determine the topology map based on the first spatial information and the second spatial information; Step 122: Based on the first time information and the second time information, align the heat exchange station information and the heating network information in time to obtain the data stream of each monitoring point on the topology map.

[0033] For step 121, the monitoring points of the heat exchange station and pipeline are used as nodes, and the pipelines in the heating network are used as edges. The monitoring points are then connected to obtain a topology graph.

[0034] For step 122, iterate through all valid time points in the first time information and the second time information, compare the valid time points in the first time information and the second time information, and if a valid time point is missing a valid value, then remove the valid time point. Align the heat exchange station information and heating network information corresponding to the remaining valid time points to obtain the data stream of each monitoring point on the topology map.

[0035] In this embodiment, the effective time points in the first and second time information are filtered and aligned again to ensure that each time point in the data stream has corresponding heat exchange station information and heating network information time points.

[0036] In some alternative implementations, in step 13, feature extraction is performed on the topology map to obtain the first feature of each monitoring point, including: Step 131: Establish an adjacency list based on the topology graph; Step 132: Based on the adjacency list and the preset target optimization function, the topology graph is divided into partitions centered on the monitoring points; Step 133: Extract features from each partition to obtain the first feature of each monitoring point.

[0037] Adjacency lists are the standard storage structure for graph structures. They are used to represent other monitoring points that are directly connected to each monitoring point, providing a basic data structure for subsequent graph partitioning and graph feature extraction, and are adapted to the connectivity topology of heating pipe networks.

[0038] In this embodiment, for step 131, the adjacency list is a standard storage structure for graph structures, used to represent other monitoring points directly connected to each monitoring point, providing a basic data structure for subsequent graph partitioning and graph feature extraction, and adapting to the heating pipe network connectivity topology. The steps for establishing the adjacency list include: 1. Traverse all nodes in the topology graph and initialize node indices. Let the topology graph be... All monitoring points in China , Represent the unique number of each monitoring point; establish a node-number mapping table: The mapping relationship associates the monitoring point type, location, and module (heat exchange station environment / security / equipment / fire module, pipeline flow / fiber optic / video monitoring point), while retaining business attributes.

[0039] 2. Traverse the topological edges and determine the direct connectivity between nodes. According to rule 121: a pipe is an edge connecting two monitoring points. For any two nodes... , (In the topology diagram, a monitoring point is considered a node): If there is a direct physical connection between two monitoring points via a heating pipe, it is denoted as... If there is no direct pipe connection, it is recorded as... The topology is an undirected graph (the heating network is bidirectionally connected), which satisfies... Nodes themselves do not form self-loops. . , The set is denoted as .

[0040] 3. Construct the adjacency list data structure. The adjacency list uses an array + linked list / list format. Define the adjacency list. : ,in, Represents a node The set of adjacent nodes, storing all nodes that are adjacent to each other. The node numbers of directly connected nodes, i.e.: .

[0041] Additionally, project attributes can be attached to each record in the adjacency list, specifically including: adjacent edge length. (Actual pipe length); Node type identifier (internal monitoring point of heat exchange station / pipeline monitoring point); Region / Heat exchange station number.

[0042] In this embodiment, for step 132, each monitoring point The purpose of dividing the local topology into sub-regions (neighborhood subgraphs) around the center is to extract the local connectivity structure features (first feature) of a single monitoring point. The partitioning rule is: partition = center node + its topological neighboring nodes. The objective optimization function ensures that the partitions are compact and conform to the physical connectivity logic of the heating network.

[0043] Specifically, for each monitoring point Execute the following partitioning process independently: 1. With Let the indicator variable be the central node. Initial partition ; in, , Represents a node Transferred to the center partitions ; Not included.

[0044] 2. Neighborhood filtering: based on adjacency list traverse all candidate nodes Calculate the value of each candidate node ; in, Represents a node With the center Topological hop cost (hop count: the number of edges in the shortest path between two nodes). Represents a node With the center The cost of laying pipelines over distances.

[0045] according to Determine the topology hop count cost. This indicates that the longer the topological path from the central node, the greater the cost; therefore, directly or nearest-neighbor connected nodes are preferred. Represents a node With the center The shortest topological hop count.

[0046] according to Determine the cost of pipeline laying distance. This indicates that the greater the physical distance, the greater the spatial cost; therefore, geographically proximate measurement points are prioritized. Represents a node With the center The actual length of the heating pipeline between them This represents the maximum pipe length among all monitoring points within the target area, used for normalization. Interval.

[0047] 3. Substitute into the optimization function: Enumerate all legal node combinations and solve for the objective function. The minimum value is used to obtain the optimal node combination; Specifically, the objective function is: ; Its constraints are: ; Among them, if For internal partitioning of the heat exchange station, then , ,like For long-distance pipeline network zoning, then , . Indicates an indicator variable. This indicates the number of nodes in the topological graph. Represents the topology hop count cost. This indicates the cost of pipeline laying distance. Indicates the maximum number of nodes in the partition, preferred. , Represents the topology hop count threshold, preferred .

[0048] according to Determine the minimum overall cost of the partition, and set it as the central monitoring point under constraints of the number of nodes and topology hops. Select neighboring nodes that are topologically connected and physically close to each other to form a local partition. .

[0049] 4. Constraint verification: If the number of nodes within the partition... Topological hop count If so, the node that violated the rules will be removed. 5. Create partitions: Finally determine the partitions. Local partitions centered on ; 6. Traverse all nodes: Repeat steps 1 to 5 to obtain the independent partitions corresponding to all monitoring points.

[0050] In this embodiment, for step 133, the second feature of each monitoring point is determined based on the time point after alignment of the first time information and the second time information: according to The node density characteristics of the partition are determined, which represent the density of monitoring points in a local area; among them, This indicates the actual number of nodes in the partition. Indicates the maximum number of nodes in the partition. The value ranges from 0 to 1. The larger the value, the denser the monitoring points around the measuring point. according to The degree value characteristic of the central node of the partition is determined. This characteristic represents the number of directly connected adjacent nodes and characterizes the topological connectivity importance of the measurement point; among which, Represents the adjacency list The number of adjacent nodes, The larger the degree value, the more the measuring point is located at a key point in the pipeline network, and the wider the range of impact from the fault. according to The average topological hop count characteristic of a partition is determined, which represents the overall topological proximity of nodes within the partition relative to the central measurement point; among which, Indicates partition Actual number of nodes Relative to Other nodes, Indicated by monitoring point The central partition, express To other nodes within the partition The shortest topological hop count; according to The average pipeline distance characteristic of a zone is determined, which represents the overall physical distribution distance of measuring points within the zone; where represents the zone. Actual number of nodes Relative to Other nodes, Indicated by monitoring point The central partition, express and The actual length of the heating pipeline between them This indicates the maximum pipe length between all monitoring points within the target area; according to The average network distance characteristic of each partition is determined, which represents the internal connectivity density of the local subgraph; where, Indicates partition The actual number of topological edges that exist internally. This represents the theoretical maximum number of undirected edges with m nodes; according to The first feature of each monitoring point. The first feature is a one-dimensional feature.

[0051] In an optional embodiment of the present invention, step 14, which involves feature extraction of the data stream to obtain a second feature for each monitoring point, includes: Step 141: If the data in the data stream is numerical, then determine the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, and steady-state deviation based on the data stream. Step 142: If the data in the data stream is an image, then determine the image features based on the data stream; Step 143: Determine the second feature of each monitoring point based on the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, steady-state deviation, and image features.

[0052] In this embodiment, the data corresponding to the effective time points of each monitoring point constitute a data stream. Since the data types monitored by each monitoring point are different, corresponding methods need to be used for feature extraction for different types of data.

[0053] Step 141: If the data in the data stream is numerical, then determine the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, and steady-state deviation based on the data stream, specifically including: 1. If the data monitored by the monitoring point is numerical, then: according to Determine the time-series average values ​​(used to determine whether equipment / pipelines have deviated from normal operating conditions for an extended period, such as average network pressure, average temperature of heat exchange stations, and average flow rate; these are fundamental indicators for judging steady-state operating conditions); among which, express The sampled value at time t, This indicates the number of time points after aligning the first time information and the second time information; according to Determine the time series standard deviation (characterizing the overall fluctuation amplitude of the data stream; excessive fluctuations typically correspond to faults such as pipeline leaks, pump vibrations, valve malfunctions, and sudden temperature changes; the larger the value, the more severe the data oscillation and the higher the risk of fault); among which... This is expressed as the time series mean at that measurement point. express The sampled value at time t; according to Determine the extreme value fluctuation coefficient (characterizing the relative deviation between the maximum and minimum values ​​of the sequence, capturing short-term extreme conditions such as over-temperature, over-pressure, and flow interruption, reflecting the maximum fluctuation of the measured value over a period of time; pipeline blockage, instantaneous pressure relief, and fire alarm triggering will all cause this value to increase sharply); where i=1,…, T, This represents the maximum value in the data stream. This is expressed as the time series mean at that measurement point. This indicates the minimum value in the data stream. This represents the number of time points after aligning the first and second time information. according to Determine the absolute value of the average first-order difference (characterizing the average rate of change per unit time, identifying sudden changes in operating conditions; the larger the value, the more drastic the change in operating conditions, often used to identify sudden faults); among which... This represents the difference between sampled values ​​at adjacent time points. express The sampled value at time t, Indicate The sampled value at time t-. This represents the number of time points after aligning the first and second time information. according to Determine the steady-state deviation (introducing the rated standard value of equipment / pipeline network, characterizing the deviation between the current overall operating condition and the design rated value, which is the core indicator for fault diagnosis; the higher the deviation, the longer the measuring point has been operating in an abnormal range, and the greater the risk of aging, imbalance, and pipeline network instability); among which... Indicates monitoring point Rated standard value / Normal design value It is represented as the time series mean of the measurement point i; Step 142: If the data in the data stream is an image, then determine the image features based on the data stream, specifically including: Based on principal component analysis, feature extraction is performed on images in the data stream. Specifically, each image within a preset time period is converted into a matrix, then stretched into a one-dimensional vector, resulting in multiple one-dimensional vectors. A covariance matrix is ​​constructed based on these vectors. Eigenvalues ​​are obtained by performing eigenvalue decomposition on the covariance matrix. The first s largest eigenvalues ​​are retained, and the vectors corresponding to these s largest eigenvalues ​​are used to construct a projection basis. The matrix obtained from each image is mapped to the projection basis to obtain the extracted feature vectors. These extracted feature vectors are then integrated into a one-dimensional vector to obtain the image features. .

[0054] Step 143: Based on the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, steady-state deviation, and image features, determine the second feature of each monitoring point, which may specifically include: according to By combining the above features, the second feature of each monitoring point is obtained.

[0055] For step 15, the first feature and the second feature are spliced ​​together for each monitoring point. Let i = 1, ..., T, where T is the number of time points after aligning the first and second time information, to obtain the third feature of each monitoring point. .

[0056] In some alternative implementations, in step 16, fault prediction is performed based on the third feature to obtain a warning risk value for each monitoring point, including: In this embodiment, the third feature is input into a preset model to obtain the early warning risk value for each monitoring point, specifically including: 1. According to Construct the covariance matrix, where, This represents an 11×11 dimensional feature covariance matrix, characterizing the correlation between features in each dimension. This indicates the total number of monitoring points in the topology graph. This represents the standardized third feature vector of the i-th monitoring point. This represents the mean vector of the standardized third feature of all monitoring points.

[0057] 2. According to Perform eigenvalue decomposition. Among them, This represents the eigenvector matrix (projection basis matrix), corresponding to the projection basis logic of image feature extraction in step 142 of the original text. , Represents an eigenvalue diagonal matrix. The feature values ​​are arranged in descending order, representing the amount of feature information.

[0058] 3. Based on the cumulative variance contribution rate, core principal components are selected, and effective features for identifying heating network faults are retained. Determine the cumulative variance contribution rate of the first m principal components, where, This represents the cumulative variance contribution rate of the first m principal components. This represents the number of core principal components obtained through screening. The top m optimal principal components, selected using this formula, constitute the core feature vector after dimensionality reduction. This replaces the original redundant third feature.

[0059] 4. According to Determine the risk weight of the k-th principal component, where, This represents the risk weight of the k-th principal component. This represents the eigenvalue corresponding to the k-th principal component.

[0060] 5. According to Determine the initial fault risk score for the i-th monitoring point, where, This represents the initial fault risk score for the i-th monitoring point. This represents the eigenvalue of the i-th monitoring point and the k-th principal component.

[0061] 6. According to Normalizing the initial score yields a standard risk value in the range of 0 to 1. in, This represents the final warning risk value for the i-th monitoring point. This represents the maximum and minimum initial risk scores for all monitoring points. 0 represents no risk, and 1 represents the highest risk of failure.

[0062] In some alternative implementations, step 17 involves initiating different levels of emergency procedures based on the data stream and the warning risk value, including: Step 171: Determine whether the data in the data stream is within a first preset range. If it is within the first preset range, then initiate the alarm process. Step 172: Determine the warning level of the warning risk value based on the preset first threshold, second threshold and third threshold, and initiate the emergency procedure according to the warning level.

[0063] For step 171, the effective range of each monitoring point is determined based on the historical data collected for each monitoring point. For each monitoring point, it is determined whether each data monitored by the monitoring point is within the effective range within the preset time period. If not, it indicates that the monitoring point has malfunctioned, and an audible and visual alarm is immediately activated.

[0064] For step 172, the early warning risk value is a comprehensive prediction for a single monitoring point when no obvious anomalies are found in the data (i.e., within the valid range). The aim is to predict potential faults at the monitoring point using comprehensive data (including location information and data stream). For example, even if all data in the data stream of a certain monitoring point is within the valid range, significant fluctuations between data points may indicate a potential fault at that monitoring point. Thresholds are set for the parameters that classify different levels of early warning: First threshold X1, Second threshold X2, Third threshold X3: If the warning risk value is 0-X1, a Level IV (weak) warning is triggered: status alert; If the warning risk value is X1~X2, a Level III (minor) warning will be triggered: the maintenance personnel will be notified to conduct on-site inspection and pay close attention. If the risk value is X2~X3, a Level II warning (general) will be triggered: a warning alarm will be issued, the person in charge will be notified to go to the site, video verification will be initiated, and the emergency plan will be activated. If the warning risk value is greater than X3, a Level I warning (severe) is triggered, the highest level alarm. Based on the alarm data, the system will link up with the on-site equipment and remotely organize on-site accident handling. After completing the corresponding emergency procedures, the system was restored to normal monitoring status.

[0065] In one application scenario, refer to Figure 3 This application scenario includes a monitoring front-end layer, a network transmission layer (IoT or communication protocol), a central management layer, and a terminal application layer (central control room). Specifically: The monitoring front-end layer is used for information collection of multiple heat exchange stations and multiple sections of heating pipeline within the monitoring range. The information in each heat exchange station includes: environmental monitoring, security, fire alarm system and heat exchange station equipment information, etc.; the information in the heating pipeline includes: flow meter, direct buried heating pipeline fiber optic detection system, overhead pipeline video monitoring system, etc.; among them, (1) Environmental monitoring module: mainly to detect the environmental parameters of the heat exchange station in real time and continuously. Ensure the safe and stable operation of the heat exchange station. The content monitored by this module includes temperature and humidity parameters in the heat exchange station, water immersion sensors installed in cable trenches and cable wells, intelligent lighting control, fan and air conditioning operation monitoring. (2) Security module: this module mainly provides technical security for the heat exchange station. It includes video monitoring system, access control system, intrusion alarm system, video monitoring system, real-time video detection, real-time video monitoring, real-time image display, video intrusion alarm function, video fire detection function, etc. Access control system: mainly solves the management of the entrance and exit of the heat exchange station. Intrusion alarm system: used to detect unauthorized entry into the heat exchange station. Front-end detectors: Glass break detectors, electronic fences, smart cameras, etc. are responsible for detecting intrusion behavior. The heat exchange station is monitored 24 hours a day to detect intrusion behavior in a timely manner. The signal is sent to the central management level. The integrated management host judges whether it is a valid alarm according to the preset rules. (3) Equipment management module: This module mainly monitors, controls and manages the operation and safety status of various equipment in the heat exchange station automatically. Circulating water pump / makeup water pump: The circulating water pump / makeup water pump control box (including frequency converter) collects the circulating pump operation signal. Valve: The valve opening and closing status is collected by the integrated valve control box. The makeup water tank is equipped with a liquid level sensor to collect the makeup water tank liquid level signal. The primary and secondary water inlet and return pipes in the heat exchange station are equipped with temperature sensors, pressure sensors, heat sensors, etc. to collect relevant parameters on the pipes. The water pipes in the heat exchange station are equipped with leakage detection alarms, and the data is transmitted to the central management level remotely. (4) Fire alarm module: This module mainly arranges fire detectors, manual fire alarm buttons, fire sound and light alarms and fire alarm controllers in the heat exchange station. Used for early monitoring and warning of fires in heat exchange stations.

[0066] The network transmission layer is used for reliable data transmission, responsible for securely, reliably, and stably transmitting data from the monitoring front-end layer to the remote monitoring center's integrated management system. The integrated management system consists of two parts. The network transmission layer includes IoT gateways and a communication network. IoT gateways are deployed at the heat exchange stations and heating pipeline sites to aggregate and convert data collected by various detectors. The communication network is a ubiquitous IoT network composed of a dedicated fiber optic network, 4G / 5G, LoRa, and NB-IoT.

[0067] The central management layer is responsible for processing data, making judgments, analyzing decisions, and triggering alarms. The main function of the terminal application layer is to provide an interactive interface for operation and maintenance personnel, offering real-time preview, playback, and control operations.

[0068] Based on the above application scenarios: (1) First, based on the monitoring points of each module in the environmental monitoring module, security module, equipment management module, and fire alarm module, obtain the heat exchange station information, heating network information, and the spatiotemporal information of the heat exchange station information and heating network information; (2) Determine the first spatial information and the second spatial information based on the location information of all monitoring points in the target area; For each module, the monitoring points are filtered according to their time points to ensure that all monitoring points in the module have collected valid data at the filtered time points. Then, the filtered time points of all modules are compared to ensure that all monitoring points in all modules have collected valid data at the filtered time points, thus obtaining first-time information. Similarly, by iterating through all time points in the set of time points for heating network information, we can ensure that all monitoring points in the heating network have collected valid data at the selected time points, thus obtaining the second time information.

[0069] (3) Determine the topology map based on the first spatial information and the second spatial information. Based on the first time information and the second time information, align the heat exchange station information and the heating network information in time to obtain the data stream of each monitoring point on the topology map.

[0070] (4) Based on the objective function Feature extraction is performed on the topology map to obtain the first feature of each monitoring point. ; (5) If the data in the data stream is numerical, then determine the time-series mean, time-series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, and steady-state deviation based on the data stream; if the data in the data stream is an image, then determine the image features based on the data stream; and determine the second feature of each monitoring point based on the time-series mean, time-series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, steady-state deviation, and image features. ; (6) For each monitoring point, splice the first feature and the second feature together. Let i = 1, ..., T, where T is the number of time points after aligning the first and second time information, to obtain the third feature of each monitoring point. .

[0071] (7) The preset model can be principal component analysis or other pre-trained classification models, and the third feature will be used. Input the model to obtain the early warning risk value for each monitoring point; (8) Determine the effective range of each monitoring point based on the historical data collected for each monitoring point. For each monitoring point, determine whether each data monitored by the monitoring point is within the effective range within the preset time period. If not, it means that the monitoring point has malfunctioned and immediately activate the audible and visual alarm. Thresholds are set for the parameters that classify different levels of early warning: first threshold X1, second threshold X2, and third threshold X3. If the warning risk value is 0-X1, a Level IV (weak) warning is triggered: status alert; If the warning risk value is X1~X2, a Level III (minor) warning will be triggered: the maintenance personnel will be notified to conduct on-site inspection and pay close attention. If the risk value is X2~X3, a Level II warning (general) will be triggered: a warning alarm will be issued, the person in charge will be notified to go to the site, video verification will be initiated, and the emergency plan will be activated. If the warning risk value is greater than X3, a Level I warning (severe) is triggered, the highest level alarm. Based on the alarm data, the system will link up with the on-site equipment and remotely organize on-site accident handling. After completing the corresponding emergency procedures, the system was restored to normal monitoring status.

[0072] like Figure 2 As shown, an embodiment of the present invention provides a comprehensive monitoring device 200 for a heating network, comprising: The acquisition module 201 is used to acquire heat exchange station information and heating pipeline information for a preset time period based on multiple monitoring points set in the heat exchange station and heating pipeline network in the target area, as well as to acquire the spatiotemporal information of the heat exchange station information and heating pipeline network information. The processing module 202 is used to correlate the heat exchange station information and the heating pipeline network information according to the spatiotemporal information to obtain a topology map and a data stream for each monitoring point on the topology map; the nodes in the topology map are monitoring points set on the heat exchange station and the heating pipeline network; the time points of the data streams of each monitoring point on the topology map correspond one-to-one; feature extraction is performed on the topology map to obtain a first feature for each monitoring point; feature extraction is performed on the data streams to obtain a second feature for each monitoring point; the first feature and the second feature are concatenated according to each monitoring point to obtain a third feature for each monitoring point; fault prediction is performed based on the third feature to obtain a warning risk value for each monitoring point; and different levels of emergency procedures are initiated according to the heat exchange station information, the heating pipeline network information, and the warning risk value.

[0073] Optionally, obtain information on heat exchange stations and heating network in the target area, including: Acquire environmental monitoring information, security information, fire alarm system and equipment information of heat exchange stations in the target area, as well as flow information of heating pipeline network, fiber optic detection information of buried heating pipelines and video monitoring information of overhead pipelines.

[0074] Optionally, the spatiotemporal information of the heat exchange station information and heating network information is obtained, including: Based on the location information of the monitoring points at the heat exchange station, the first spatial information is determined; The second spatial information is determined based on the location information of the monitoring points of the heating pipeline network. Based on the set of time points from which information about the heat exchange station was obtained, determine the first-time information; The second time information is determined based on the set of time points from which heating network information is obtained; The spatiotemporal information is obtained based on the first spatial information, the second spatial information, the first time information, and the second time information.

[0075] Optionally, based on the spatiotemporal information, the heat exchange station information and the heating network information are correlated to obtain a topology map and a data stream for each monitoring point on the topology map, including: Based on the first spatial information and the second spatial information, a topology map is determined; Based on the first time information and the second time information, the heat exchange station information and the heating pipeline network information are time-aligned to obtain the data stream of each monitoring point on the topology map.

[0076] Optionally, feature extraction is performed on the topology map to obtain the first feature of each monitoring point, including: Establish an adjacency list based on the aforementioned topology graph; Based on the adjacency list and the preset target optimization function, the topology graph is divided into partitions centered on the monitoring points; Feature extraction is performed on each partition to obtain the first feature of each monitoring point.

[0077] Optionally, feature extraction is performed on the data stream to obtain a second feature for each monitoring point, including: If the data in the data stream is numerical, then determine the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, and steady-state deviation based on the data stream; If the data in the data stream is an image, then determine the image features based on the data stream; The second feature of each monitoring point is determined based on the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, steady-state deviation, and image features.

[0078] Optionally, based on the heat exchange station information, heating network information, and the early warning risk value, different levels of emergency procedures can be initiated, including: Determine whether the data in the heat exchange station information and heating network information are within a first preset range. If they are within the first preset range, then initiate an alarm process. Based on the preset first threshold, second threshold, and third threshold, the warning level of the warning risk value is determined, and the emergency procedure is initiated according to the warning level.

[0079] All implementations in the above method embodiments are applicable to the embodiments of the device, and similarly, all implementations in the above device embodiments are applicable to the embodiments of the above method, achieving the same technical effect.

[0080] The present invention also provides a computing device, comprising: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method.

[0081] In another aspect, the present invention also provides a computer-readable storage medium storing a program that, when executed by a processor, implements the above-described method. All implementations in the above method embodiments are applicable to the embodiments of this computer-readable storage medium and can achieve the same technical effects.

[0082] All implementations in the above method embodiments are applicable to the embodiments of the computing device and medium. Similarly, all implementations in the above computing device and medium embodiments are applicable to the embodiments of the above method, and can achieve the same technical effect.

[0083] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for comprehensive monitoring of a heating network, characterized in that, include: Based on multiple monitoring points set up in the heat exchange stations and heating pipelines in the target area, information on the heat exchange stations and heating pipelines for a preset time period is obtained, as well as spatiotemporal information of the heat exchange station information and heating pipelines. Based on the spatiotemporal information, the heat exchange station information and the heating pipeline network information are correlated to obtain a topology map and a data stream for each monitoring point on the topology map; The nodes in the topology diagram are monitoring points set up in heat exchange stations and heating pipelines; The time points of the data stream at each monitoring point on the topology diagram correspond one-to-one. Feature extraction is performed on the topology map to obtain the first feature of each monitoring point; Feature extraction is performed on the data stream to obtain the second feature of each monitoring point; For each monitoring point, the first feature and the second feature are concatenated to obtain the third feature for each monitoring point; Based on the third feature, fault prediction is performed to obtain the early warning risk value for each monitoring point; Based on the heat exchange station information, heating network information, and the early warning risk value, different levels of emergency procedures are initiated; This includes obtaining information on heat exchange stations and heating network in the target area, including: Acquire environmental monitoring information, security information, fire alarm system and heat exchange station equipment information of the target area, as well as flow information of the heating network, fiber optic detection information of buried heating pipelines and video monitoring information of overhead pipelines; Specifically, feature extraction is performed on the topology map to obtain the first feature of each monitoring point, including: Establish an adjacency list based on the topology graph; Based on the adjacency list and the preset objective optimization function, the topology graph is divided into partitions centered on monitoring points; wherein, according to Determine the minimum overall cost of partitioning, and set it as the central monitoring point under constraints of the number of nodes and topology hops. Select neighboring nodes that are topologically connected and physically close to each other to form a local partition. The objective function is: ； Its constraints are: ； Among them, if For internal partitioning of the heat exchange station, then , ,like For long-distance pipeline network zoning, then , ; Indicates an indicator variable. This indicates the number of nodes in the topological graph. Represents the topology hop count cost. This indicates the cost of pipeline laying distance. Indicates the maximum number of nodes in the partition. Indicates the topology hop count threshold; Feature extraction is performed on each partition to obtain the first feature of each monitoring point; where, according to The first characteristic of each monitoring point, where, Indicates the density of monitoring points in a local area; This represents the number of directly connected adjacent nodes, characterizing the importance of the topological connections of the measurement point; This indicates the overall topological distance of nodes within a partition relative to the central measurement point; Indicates the overall physical distribution distance of the measuring points within the partition; This indicates the degree of internal connectivity of a local subgraph; In this process, feature extraction is performed on the data stream to obtain the second feature for each monitoring point, including: If the data in the data stream is numerical, then determine the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, and steady-state deviation based on the data stream; If the data in the data stream is an image, then determine the image features based on the data stream; Based on the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, steady-state deviation, and image features, determine the second feature of each monitoring point; Among them, according to Normalizing the initial score yields a standard risk value in the range of 0 to 1, where... This represents the final warning risk value for the i-th monitoring point. This represents the maximum and minimum initial risk scores for all monitoring points. 0 represents no risk, and 1 represents the highest risk of failure; Determine the initial fault risk score for the i-th monitoring point, where, This represents the initial fault risk score for the i-th monitoring point. This represents the eigenvalue of the i-th monitoring point and the k-th principal component; Determine the risk weight of the k-th principal component, where, This represents the risk weight of the k-th principal component. This represents the eigenvalue corresponding to the k-th principal component; The emergency response procedures, based on the heat exchange station information, heating network information, and the early warning risk value, are initiated at different levels, including: Determine whether the data in the heat exchange station information and heating network information are within a first preset range. If they are within the first preset range, immediately initiate the alarm process. Based on the preset first threshold, second threshold, and third threshold, the warning level of the warning risk value is determined, a comprehensive prediction is made for a single monitoring point, and an emergency procedure is initiated according to the warning level.

2. The integrated monitoring method for heating networks according to claim 1, characterized in that, The spatiotemporal information for obtaining the heat exchange station information and heating network information includes: Based on the location information of the monitoring points at the heat exchange station, the first spatial information is determined; The second spatial information is determined based on the location information of the monitoring points of the heating pipeline network. Based on the set of time points from which information about the heat exchange station was obtained, determine the first-time information; The second time information is determined based on the set of time points from which heating network information is obtained; The spatiotemporal information is obtained based on the first spatial information, the second spatial information, the first time information, and the second time information.

3. The integrated monitoring method for heating networks according to claim 2, characterized in that, Based on the spatiotemporal information, the heat exchange station information and the heating network information are correlated to obtain a topology map and a data stream for each monitoring point on the topology map, including: Based on the first spatial information and the second spatial information, a topology map is determined; Based on the first time information and the second time information, the heat exchange station information and the heating pipeline network information are time-aligned to obtain the data stream of each monitoring point on the topology map.

4. A comprehensive monitoring device for a heating network, characterized in that, include: The data acquisition module is used to acquire heat exchange station information and heating pipeline information for a preset time period based on multiple monitoring points set up in the target area, as well as the spatiotemporal information of the heat exchange station information and heating pipeline information. The processing module is used to associate the heat exchange station information and the heating network information according to the spatiotemporal information to obtain a topology map and a data stream of each monitoring point on the topology map; The nodes in the topology diagram are monitoring points set up in heat exchange stations and heating pipelines; The time points of the data stream at each monitoring point on the topology diagram correspond one-to-one. Feature extraction is performed on the topology map to obtain the first feature of each monitoring point; feature extraction is performed on the data stream to obtain the second feature of each monitoring point; the first and second features are concatenated for each monitoring point to obtain the third feature of each monitoring point; fault prediction is performed based on the third feature to obtain the early warning risk value of each monitoring point. Based on the heat exchange station information, heating network information, and the early warning risk value, different levels of emergency procedures are initiated; This includes obtaining information on heat exchange stations and heating network in the target area, including: The system acquires environmental monitoring information, security information, fire alarm system and equipment information of heat exchange stations in the target area, as well as flow information of the heating network, fiber optic detection information of buried heating pipelines, and video monitoring information of overhead pipelines. Specifically, it extracts features from the topology map to obtain the first feature of each monitoring point, including: Establish an adjacency list based on the topology graph; Based on the adjacency list and the preset objective optimization function, the topology graph is divided into partitions centered on monitoring points; wherein, according to Determine the minimum overall cost of partitioning, and set it as the central monitoring point under constraints of the number of nodes and topology hops. Select neighboring nodes that are topologically connected and physically close to each other to form a local partition. The objective function is: ； Its constraints are: ； Among them, if For internal partitioning of the heat exchange station, then , ,like For long-distance pipeline network zoning, then , ; Indicates an indicator variable. This indicates the number of nodes in the topological graph. Represents the topology hop count cost. This indicates the cost of pipeline laying distance. Indicates the maximum number of nodes in the partition. Indicates the topology hop count threshold; Feature extraction is performed on each partition to obtain the first feature of each monitoring point; where, according to The first characteristic of each monitoring point, where, Indicates the density of monitoring points in a local area; This represents the number of directly connected adjacent nodes, characterizing the importance of the topological connections of the measurement point; This indicates the overall topological distance of nodes within a partition relative to the central measurement point; Indicates the overall physical distribution distance of the measuring points within the partition; This indicates the degree of internal connectivity of a local subgraph; In this process, feature extraction is performed on the data stream to obtain the second feature for each monitoring point, including: If the data in the data stream is numerical, then determine the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, and steady-state deviation based on the data stream; If the data in the data stream is an image, then determine the image features based on the data stream; Based on the time series mean, time series standard deviation, extreme value fluctuation coefficient, average first-order difference absolute value, steady-state deviation, and image features, determine the second feature of each monitoring point; Among them, according to Normalizing the initial score yields a standard risk value in the range of 0 to 1, where... This represents the final warning risk value for the i-th monitoring point. This represents the maximum and minimum initial risk scores for all monitoring points. 0 represents no risk, and 1 represents the highest risk of failure; Determine the initial fault risk score for the i-th monitoring point, where, This represents the initial fault risk score for the i-th monitoring point. This represents the eigenvalue of the i-th monitoring point and the k-th principal component; Determine the risk weight of the k-th principal component, where, This represents the risk weight of the k-th principal component. This represents the eigenvalue corresponding to the k-th principal component; The emergency response procedures, based on the heat exchange station information, heating network information, and the early warning risk value, are initiated at different levels, including: Determine whether the data in the heat exchange station information and heating network information are within a first preset range. If they are within the first preset range, immediately initiate the alarm process. Based on the preset first threshold, second threshold, and third threshold, the warning level of the warning risk value is determined, a comprehensive prediction is made for a single monitoring point, and an emergency procedure is initiated according to the warning level.

5. A computing device, comprising: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to perform the method as described in any one of claims 1 to 3.

6. A computer-readable storage medium storing a program that, when executed by a processor, implements the method as claimed in any one of claims 1 to 3.

Images