Heat supply pipe network pipe burst analysis method oriented to geographic information pipe network data topology conversion

Abstract

The invention relates to a heat supply pipe network pipe burst analysis method oriented to geographic information pipe network data topological conversion. The method comprises the following steps: instantiating geographic information pipe network data: determining a global unique identifier; the geographic information pipe network data comprises four types of objects of a valve, a pipeline, a heat exchange station and a heat source, the two types of objects of the valve and the pipeline are stored in the form of point space elements and line space elements, and the two types of objects of the heat exchange station and the heat source are stored in the form of polygon space elements; instantiating the four types of objects according to the three space elements; node conversion of polygon space elements is carried out; segmenting a pipeline, and switching a valve and the pipeline; automatic topology conversion and graph theory modeling are carried out, and a non-directional diagram is constructed; 5, constructing a relational mapping table from the geographic information pipe network data to the non-directional diagram; and 6, carrying out online pipe explosion analysis.

Description

Technical Field

[0001] This invention belongs to the field of emergency repair technology for urban centralized heating systems, and in particular relates to a method for analyzing pipe bursts in heating networks based on geographic information network data topology conversion. Background Technology

[0002] Urban centralized heating systems are critical infrastructure for maintaining the lifeline function of northern cities during winter. With the increasing service life of pipelines, the accumulation of corrosion fatigue, and disturbances from third-party construction, pipe bursts are becoming more frequent and cascading. A single main pipe burst not only causes instantaneous heat and medium loss but can also trigger secondary disasters, leading to large-scale heating outages or even road collapses. Therefore, completing pipe burst analysis within minutes or even seconds to identify potentially affected areas has become a rigid requirement for emergency management by heating companies.

[0003] In early pipe burst emergency analysis, operators often manually drew a circle or entered a radius as a "buffer zone" centered on the accident point in desktop geographic information system (GIS) pipeline data software. Heat exchange stations, buildings, valves, etc., falling within this circle were all classified as "affected areas," which was insufficient to meet practical needs. In recent years, researchers have attempted to convert pipeline networks into graph models, but all have used points as the basic control elements. When a pipe burst event occurs, the corresponding nodes are found in reverse based on the pipeline network information, and then analysis is performed from those nodes. For example, [CN111814289A] uses the graph search algorithm built into graph databases such as PostGIS for calculation; [CN114840627A] also uses a graph database and its built-in algorithm for calculation; [CN112270635A] starts with the closed valve, performs a full graph traversal, and obtains the affected area by extracting the network difference before and after the calculation; [CN120105636A] uses the valves at both ends of the burst pipe as base points, and calculates the valves that need to be closed based on the adjacency list of each node using the BFS algorithm. Meanwhile, [CN113065217A] uses sensor data to generate the pipeline topology, which requires brute-force simulation of all pipe segment combinations in advance to form a "valve closure scheme library", and requires a long calculation time.

[0004] Furthermore, existing research largely relies on direct network construction using point-to-line vector layers, neglecting the fact that heating geographic information network data commonly uses polygonal spatial elements to describe facilities such as valve wells, heat exchange stations, and buildings. Polygonal spatial elements possess complex characteristics at the spatial level, including area, holes, and multiple components, making them unsuitable for direct graph theory calculations. Additionally, the heating network employs a dual-pipe closed system, with water circulating between the supply and return pipes, necessitating targeted algorithm design for this topology.

[0005] This invention treats lines and points equally, using graph theory to mark edges as states to be removed, thus deeming burst pipes as unconnectable. Then, a connectivity component algorithm is run to find component networks without a heating source, i.e., areas that are unheatable and disconnected due to burst pipe events, meeting the needs of emergency analysis. Summary of the Invention

[0006] This invention addresses the field of emergency repairs for urban centralized heating by providing a method for converting geographic information pipeline network data containing polygonal spatial elements into an undirected graph structure and analyzing the impact range of pipe bursts. This method has the advantages of not requiring hydraulic parameters, not being limited by the actual length of the pipeline network, not requiring a database, and supporting the direct participation of polygonal spatial elements in topology calculations.

[0007] The present invention provides a method for analyzing pipe bursts in heating pipe networks based on geographic information network data topology conversion, comprising the following steps: Step 1, Instantiation of Geographic Information Network Data: Determine a globally unique identifier; Geographic Information Network Data includes four types of objects: valves, pipelines, heat exchange stations, and heat sources. Valves and pipelines are stored as point spatial features and line spatial features respectively, while heat exchange stations and heat sources are stored as polygon spatial features; Instantiate the four types of objects based on the above three spatial features. Step 2, Node-based transformation of polygonal spatial features: Convert polygonal spatial features into nodes in an undirected graph to form polygonal nodes; Step 3, pipeline segmentation and valve / pipeline conversion, including: (1) Convert the point spatial elements into nodes in the undirected graph to form valve nodes; (2) Convert the two endpoints of the line space element into nodes in the undirected graph to form virtual nodes; (3) Convert the line space elements into edges in the undirected graph to form edges; Step 4, Automatic Topology Transformation and Graph Theory Modeling: Set a distance threshold, and based on the distance threshold and the geographical location information of all geographic information network data, combined with the characteristics of spatial elements, traverse all nodes and edges in the undirected graph to construct the undirected graph; Step 5: Construct a mapping table from geographic information network data to undirected graphs; Step Six, Online Pipe Burst Analysis: Mark the edges in the undirected graph where the pipe burst event occurred as pending removal. Run the Connected Component Analysis algorithm to find independent network components that do not contain heat source objects. Return the globally unique identifiers of all edges and nodes in this independent network component. The method is as follows: (1) Read the pre-configured globally unique identifier of the heat source object; (2) The network attributes of the pipeline network include water supply network or return network. The system receives user requests, extracts the globally unique identifier of all pipes that have burst in the user request, and iterates through all input pipeline numbers to verify whether the network attributes of all pipes that have burst are consistent. If inconsistency is found, the system refuses to process and prompts an error. (3) In the undirected graph, the edges corresponding to the pipes that burst are marked as to be removed. When running the connected component algorithm, the edges of the to be removed state are skipped, and all network components formed due to the inability to connect due to the burst are identified. (4) Check whether the globally unique identifier of all edges and points in each connected component is duplicated with the globally unique identifier of the heat source object. Determine the independent network component that does not contain any globally unique identifier of the heat source object as the affected area disconnected from the main heat source and return it, that is, return the affected area.

[0008] Furthermore, the method for step two is as follows: (1) Calculate the arithmetic mean of the coordinates of all vertices of the outer ring of the polygonal spatial feature, and use the obtained centroid latitude and longitude coordinates as the node coordinates of the polygonal spatial feature in the undirected graph; (2) Store the outer ring coordinates of the polygon spatial features in the node attributes; (3) For a polygonal spatial feature containing an inner ring hole, calculate the polygon difference set between the polygonal region and the hole region, and use the centroid latitude and longitude coordinates of the polygon difference set as the node coordinates of the polygonal spatial feature in the undirected graph. (4) For polygonal spatial features with nested structures, calculate the centroid coordinates of each outer ring independently and assign a sub-identity code.

[0009] Furthermore, in step three, a pipe object with multiple bends between two endpoints is defined as a multi-line spatial element. The line spatial element includes multi-line spatial elements, and the processing of multi-line spatial elements is as follows: (1) The multi-line spatial elements are split into several independent line spatial elements, and each of the several independent line spatial elements is specially identified in the global unique identifier; (2) Traverse all independent line space features, calculate the arc length in meters, and write the length value into the "Actual Length" attribute field of the independent line space feature; (3) Convert the two endpoints of all independent line space elements into nodes without direction graphs to form virtual nodes.

[0010] Furthermore, in step four, the method for constructing the undirected graph is as follows: (1) Calculate the distance between all virtual nodes. When the distance is less than or equal to the distance threshold, merge the two virtual nodes. (2) Calculate the distance between the valve node and the virtual node. When the distance is less than or equal to the distance threshold, attach the virtual node to the corresponding valve node. (3) Calculate the vertical projection distance from the virtual node and the valve node to the outer ring of the polygonal spatial feature. When the vertical projection distance is less than or equal to the distance threshold, attach the virtual node or the valve node to the corresponding polygonal spatial feature node. (4) Detect the distance between nodes of the same type. When the distance is less than or equal to the distance threshold, add an edge to the undirected graph to connect the two nodes.

[0011] Furthermore, the connected component algorithm is calculated based on a disjoint-set data structure, as follows: (1) Initialize the disjoint set, establish a parent node pointer and rank attribute for each vertex in the undirected graph. In the initial state, the parent node of each vertex points to itself and the rank is 0, which means that each vertex is an independent set. (2) Traverse all edges in the graph that have not been removed by the burst pipe, and perform a merge operation on the two endpoints of each edge; find the root node of the set where the two endpoints are currently located. If the root nodes are different, merge the tree with the smaller rank into the tree with the larger rank as a subtree; if the two trees have the same rank, select any root node and increment the rank by 1. (3) Traverse upwards from the query vertex to the root node, and directly modify the parent pointer of all the nodes passed through to point to the root node; (4) Traverse all vertices, find their final root node, group vertices with the same root node into the same connected component, and output several independent vertex sets, each set representing a connected subnetwork.

[0012] Compared with the prior art, the present invention has the following advantages: 1. No hydraulic parameters such as pipe diameter and friction are required for pipe burst analysis applicable to any heating network; 2. Polygonal spatial features are automatically converted into nodes in an undirected graph, fundamentally solving the problem that polygons cannot participate in graph theory operations; 3. Using the concept of "virtual nodes", the endpoints of edges are regarded as virtual nodes, which solves the problem of no line-to-line connection in graph theory while fully preserving the original geographic information network data. 4. By using the "mark-skip" strategy and connected component algorithm, the problem of edges not being able to be set as disconnected in geographic information pipeline network data is solved, avoiding the modification of graph structure, ensuring the read-only characteristics of the original network topology, supporting high-concurrency multiple pipeline burst simulation analysis, reducing the time complexity of edge removal operations, and improving the computational efficiency under large-scale pipeline networks.

[0013] 5. During runtime, modifications to graph features are performed entirely in memory, eliminating the need for a complex database and ensuring concurrency safety. Furthermore, an asynchronous reloading mechanism allows for updates to pipeline data without restarting the service.

[0014] 6. Through a closed loop of "spatial resolution—automatic connection—graph structure persistence—online connectivity analysis," the system can quickly provide a quantitative answer to "if a pipe section ruptures, which heat exchange stations will shut down due to loss of heat source." Dispatchers can then use this information to proactively activate emergency heat sources and send notifications to users, shortening accident response time.

[0015] 7. Since all intermediate results are persisted to the server cache, once the original geographic information is updated, the graph structure only needs to be regenerated once to enjoy the rapid analysis capability again, providing efficient and scalable technical support for the safe operation of urban heating. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method and system of the present invention.

[0017] Figure 2 This is a schematic diagram of the geographic information network data of the present invention.

[0018] Figure 3 This is a schematic diagram of the backflow graph network of the present invention.

[0019] Figure 4 This is a schematic diagram of the water supply graph network of the present invention. Detailed Implementation

[0020] The present invention will now be described in conjunction with the accompanying drawings and embodiments.

[0021] The present invention provides a method for analyzing pipe bursts in heating pipe networks based on topology transformation of geographic information pipe network data. This method automatically performs topology transformation and graph theory modeling on geographic information pipe network data in GeoJSON format containing polygon spatial elements, and quickly identifies the pipe network affected by a pipe burst event by marking states to be removed and using connected component algorithms.

[0022] 1. Data input, character decoding, and network partitioning The system asynchronously traverses the local disk storage directory, batch reading geographic information files conforming to the GeoJSON format. It attempts to parse the files sequentially using GB2312, GBK, and UTF-8 encoding formats. If a Unicode exception is triggered, it automatically switches to the next encoding scheme until all feature attribute fields are correctly recognized. After successful decoding, the DH field is extracted from the pipeline network data as a globally unique identifier. The spatial fields are then instantiated into three types of spatial features—points, lines, and polygons—by the Shapely spatial engine. Based on the WL field, all spatial features are divided into water supply networks or return water networks, stored separately in the FEATS and R_FEATS dual-hash index structures for physical isolation. The water supply and return water networks are serialized into Python pickle binary files.

[0023] 2. Node-based transformation of polygonal spatial features Convert polygonal spatial features into nodes in an undirected graph: (1) Calculate the arithmetic mean of the coordinates of all vertices of the outer ring of the polygonal spatial feature, and use the obtained centroid latitude and longitude coordinates as the node coordinates of the polygonal spatial feature in the undirected graph; (2) The outer ring coordinate string of the polygonal spatial feature is completely preserved and stored in the node attributes; (3) For polygonal spatial features containing inner ring holes, first calculate the polygon difference set between the polygonal region and the hole region, and then recalculate the centroid latitude and longitude coordinates on the remaining effective region as the node coordinates of the polygonal spatial feature in the undirected map. (4) For multi-polygon spatial features, calculate the centroid coordinates of each outer ring independently and assign a sub-identity code, in the format of the original globally unique identifier plus the _Part_ serial number suffix; (5) The centroid encoding uses a combination of six decimal places and MD5 hash to generate a unique string key value. If a hash collision occurs, an auto-incrementing sequence number is appended to the end.

[0024] 3. Pipeline segmentation, valve and pipeline conversion Convert point features to nodes in an undirected graph, and line features to edges in an undirected graph: (1) Automatically decompose multi-line elements into several independent single-line string elements, and make special identification in the global unique identifier, and add the _SubDH_ sequence number suffix; (2) Traverse all single-line string features, use the WGS-84 ellipsoidal earth model, calculate the real distance of the great circle between adjacent vertices using the Vincenty formula, accumulate the length of the entire pipe in meters by segment, and write it into the length_m attribute field of the line feature; (3) Convert the start and end point coordinates of all single-line string elements into virtual nodes without direction graphs, and make special identification in the global unique identifier, in the format of virtual_xxx; (4) Convert the valve element directly into a node without a direction graph and mark it as a solid node.

[0025] 4. Automatic topology transformation and graph theory modeling A graph theory model is constructed based on the dynamic grading threshold and the geographic location information of all pipeline network data: (1) Set the search radius max_distance_m to 1 meter; (2) Calculate the ellipsoidal distance between all virtual nodes. When the distance is less than or equal to max_distance_m, merge the two virtual nodes into the same node number. (3) Calculate the distance between the valve node and the virtual node. When the distance is less than or equal to max_distance_m, attach the virtual node to the corresponding valve node and share the valve node number. (4) Calculate the vertical projection distance from the virtual node and valve node to the outer ring of the polygonal spatial feature. When the distance is less than or equal to max_distance_m, snap the virtual node or valve node to the corresponding polygonal spatial feature node.

[0026] 5. Construction and persistence of the mapping table Save the transformed undirected graph and its mapping relationships to disk: (1) Construct the nodes_dict node table, with fields including node_id node number, node type, longitude lon, latitude lat, original feature attribute original_attrs, global unique identifier, and adsorb distance adsorb_dist; (2) Construct the lines_dict line segment table, with fields including line_id (line segment number), start and end nodes from_node / to_node, actual length length_m, globally unique identifier, coordinate string coords, and sub_lines (sub-segment list); (3) Construct the connections table, with fields including line_id (segment number), start and end nodes from_node / to_node, match_method, start distance from_dist, and end distance to_dist; (4) Package the above three tables into a unified JSON file, use UTF-8 encoding and perform GZIP compression storage, and record the generation timestamp, total number of nodes, total number of edges, number of unmatched lines, and the threshold parameter max_distance_m used; (5) Serialize the undirected graph along with the node attribute dictionary and edge attribute dictionary into a Pickle binary file to form a binary image of the topology graph.

[0027] 6. Online pipe burst analysis process (1) Creation and Deployment of Pipe Burst Analysis Service ① Create a RESTful-style / burst interface service based on an asynchronous event-driven architecture, i.e., a burst analysis service, bind it to a network port and start listening; ② A multi-instance deployment mode is adopted to support concurrent processing. Each instance corresponds to an independent heating area dataset, which isolates computing resources and allows for horizontal scaling. Furthermore, the system's burst monitoring service employs a thread pool to manage concurrent requests, setting a maximum of 32 worker threads to prevent unlimited thread creation from exhausting system resources. Each request is allocated independent memory space, and thread-local variables ensure data isolation, preventing data pollution between requests. (2) Analysis of request reception and format parsing ① Accepts POST requests. The request body is in JSON format with a standard structure of {"bh_list": ["B011"],"net":"water supply"}, where bh_list is a required field. It supports the analysis of multiple pipe identifiers at the same time, but all pipes must belong to the same water supply or return network. ② The net field is optional and is used to explicitly specify the network type. If omitted, the server will automatically identify it, simplifying user calls. Furthermore, the interface supports batch request processing, allowing a maximum of 100 pipelines to be analyzed in a single request. If this limit is exceeded, the requests are processed in batches and the results are merged. A checksum field is added to the request body to prevent data corruption during transmission and improve request reliability. (3) Triple verification mechanism for request data ① Network consistency check: If both the supply and return water identifiers exist in bh_list, return an HTTP 400 status code and indicate "illegal cross-network request"; ② Spatial element existence verification: Verify the existence of each identifier in bh_list in the corresponding network. If no pipeline is hit, return an HTTP 404 status code and list the missing identifiers.

[0028] ③ Instance lock verification: After the verification is successful, subsequent operations are locked through a thread-local variable to apply to a single graph instance, preventing data confusion in a multi-instance environment.

[0029] Furthermore, the verification phase records detailed logs, including the request source IP, timestamp, identifier list, and verification result, facilitating API call auditing and troubleshooting. For client IPs that frequently request invalid identifiers, a circuit breaker mechanism is automatically triggered to limit their access frequency.

[0030] (4) Graph structure loading and cache reconstruction ① Read the graph data file in Pickle format and the corresponding JSON format file from persistent storage, and construct the mapping table of undirected graph topology and pipeline attributes in memory through deserialization mechanism.

[0031] ② Mount the deserialized graph object to the global variable gis.G (water supply network) or gis.R_G (return water network) to achieve independent memory management for the two networks.

[0032] Furthermore, the loading process uses an asynchronous I / O mechanism to read the file stream in a non-blocking manner, shortening the file loading time; after loading is completed, the basic integrity of the graph object is verified (node ​​set and edge set non-empty verification), and a reconstruction process is triggered when an exception occurs.

[0033] (5) Simulation of pipe burst removal and topology disconnection ① Receive the list of globally unique identifiers of pipelines to be exploded from the request, query the pre-built pipeline-edge mapping index (pipe_to_edges), extract all ordered pairs of graph edge endpoints corresponding to each identifier, and construct the set of edges to be removed (removed_pairs).

[0034] ② During the connected component calculation stage, when traversing the complete edge set of the graph model, edges whose endpoints exist in the set of edges to be removed are automatically skipped. Through the differentiated processing of the union of sets, the topological disconnection state of the pipeline is simulated at the logical level.

[0035] Furthermore, the set of edges to be removed uses a hash set structure to achieve an existence query with an O(1) time complexity. The entire process does not modify the underlying graph structure, ensuring the read-only characteristics of the original network topology and the reusability of multiple analyses. The mapping index supports one-to-many relationships, and a single pipeline identifier can be mapped to multiple geometric edges.

[0036] (6) Calculation of connected components ① Initialize the disjoint-set data structure, and establish a parent node pointer and rank attribute for each vertex in the graph. In the initial state, each vertex forms its own independent set.

[0037] ② Traverse all edges in the graph and perform a merge operation on edges that are not marked as to be removed: find the root node of the set containing the two endpoints. If the root nodes are different, merge them by rank (merge the tree with the smaller rank under the tree with the larger rank). If the two trees have the same rank, select any root node and increment its rank by 1. Path compression is performed during the merging process, and the parent pointers of all nodes on the query path are directly pointed to the root node.

[0038] ③ Traverse all vertices, find their final root node, group vertices with the same root node into the same connected component, and output several independent vertex sets.

[0039] Furthermore, the disjoint-set data structure operation employs a dual-set parallel processing mechanism to maintain the connectivity status of the water supply network and the return water network respectively.

[0040] (7) Determination of the correlation between heat source objects and identification of the scope of influence ① Read the pre-configured set of globally unique identifiers for heat source objects, which contains globally unique identifiers for objects such as heat sources, heat exchange stations, and pipe networks.

[0041] ② For each connected component, collect all vertex identifiers and their original pipeline identifiers mapped through correspondences to form an identifier set for that component; determine whether this set overlaps with the globally unique identifier set of the heat source object. If not, determine that the connected component is an isolated region disconnected from the heat source, and extract the globally unique pipeline identifiers corresponding to the non-virtual nodes within the isolated region as the affected objects.

[0042] Furthermore, the internal node identifiers are converted into globally unique identifiers of the business layer's original pipeline through the original number mapping table (orig_map) to ensure the consistency between the returned results and the input identifier system; the duplicate determination is based on set disjointness detection, rather than element-by-element traversal comparison.

[0043] (8) Result encapsulation and structured return ① The set of globally unique identifiers of the pipelines identified as affected is deduplicated and sorted in ascending lexicographical order, and then encapsulated into a JSON payload and returned, containing the following fields: list of affected identifiers (affected), count, network type (net), and algorithm performance metrics (performance).

[0044] ② The performance metrics include runtime statistics such as connected component calculation time, total number of components, number of affected components, and number of edges removed.

[0045] Furthermore, if there is an input but no matching pipeline identifier is found on any edge of the network, a warning field is added to the returned load, listing the invalid identifiers; the returned format supports data interface interface with the downstream emergency repair scheduling system.

[0046] (9) Resource cleanup and status maintenance ① After the analysis is complete, release the memory occupied by the set of edges to be removed and clean up temporary variables; ② Keep the graph structure (gis.G / gis.R_G) and edge cache in memory unchanged so that it can respond immediately to subsequent analysis requests without repeating file loading and deserialization operations.

[0047] Furthermore, to enhance system flexibility, multiple / burst interfaces, such as / burst1 and / burst2, are created according to usage requirements, each corresponding to a different heating area.

[0048] Furthermore, an update file interface / resset is set up and triggered using the GET method. When the host machine's data file is updated, this interface is called to reload the cache without restarting the service, thus achieving hot update capability.

[0049] Furthermore, create the / datasets interface and use the GET method to view the read status of all datasets. The returned JSON format contains complete information such as the water supply and return network status of each instance, the number of elements, the data directory, the file list, and the KEY_GUID configuration, which is convenient for operation and maintenance monitoring.

[0050] Furthermore, here's an example of the information returned by the / datasets interface: { "1": { "Water supply network": "Non-empty" "Return Water Network": "Empty" Number of water supply elements: 16 "Number of return water elements": 0, "Data Directory":"burst1 / Data", "List of JSON files": ["TEST_gong.json"], "KEY_GUID file":"key_bh.txt", "KEY_GUID value": ["ABC005"] }, "2": { "Water supply network": "empty" "Return Water Network": "Non-empty", Number of water supply elements: 0, "Number of return water elements": 16, "Data Directory":"burst2 / Data", "List of JSON files": ["TEST_hui.json"], "KEY_GUID file":"key_bh.txt", "KEY_GUID value": ["ABC005"] }, "3": { "Water supply network": "Non-empty" "Return Water Network": "Non-empty", Number of water supply elements: 16 "Number of return water elements": 16, "Data Directory":"burst3 / Data", "List of JSON files": ["TEST.json"], "KEY_GUID file":"key_bh.txt", "KEY_GUID value": ["ABC005"] }, "4": { "Water supply network": "Non-empty" "Return Water Network": "Non-empty", Number of water supply elements: 16 "Number of return water elements": 16, "Data Directory":"burst4 / Data", "List of JSON files": ["TEST.json"], "KEY_GUID file":"key_bh.txt", "KEY_GUID value": ["ABC005"] } }

Claims

1. A method for analyzing pipe bursts in heating pipe networks based on geographic information network data topology conversion, comprising the following steps: Step 1, Instantiation of Geographic Information Network Data: Determine a globally unique identifier; Geographic Information Network Data includes four types of objects: valves, pipelines, heat exchange stations, and heat sources. Valves and pipelines are stored as point spatial features and line spatial features respectively, while heat exchange stations and heat sources are stored as polygon spatial features; Instantiate the four types of objects based on the above three spatial features. Step 2, Node-based transformation of polygonal spatial features: Convert polygonal spatial features into nodes in an undirected graph to form polygonal nodes; Step 3, pipeline segmentation and valve / pipeline conversion, including: (1) Convert the point spatial elements into nodes in the undirected graph to form valve nodes; (2) Convert the two endpoints of the line space element into nodes in the undirected graph to form virtual nodes; (3) Convert the line space elements into edges in the undirected graph to form edges; Step 4, Automatic Topology Transformation and Graph Theory Modeling: Set a distance threshold, and based on the distance threshold and the geographical location information of all geographic information network data, combined with the characteristics of spatial elements, traverse all nodes and edges in the undirected graph to construct the undirected graph; Step 5: Construct a mapping table from geographic information network data to undirected graphs; Step Six, Online Pipe Burst Analysis: Mark the edges in the undirected graph where the pipe burst event occurred as pending removal. Run the Connected Component Analysis algorithm to find independent network components that do not contain heat source objects. Return the globally unique identifiers of all edges and nodes in this independent network component. The method is as follows: (1) Read the pre-configured globally unique identifier of the heat source object; (2) The network attributes of the pipeline network include water supply network or return network. The system receives user requests, extracts the globally unique identifier of all pipes that have burst in the user request, and iterates through all input pipeline numbers to verify whether the network attributes of all pipes that have burst are consistent. If inconsistency is found, the system refuses to process and prompts an error. (3) In the undirected graph, the edges corresponding to the pipes that burst are marked as to be removed. When running the connected component algorithm, the edges of the to be removed state are skipped, and all network components formed due to the inability to connect due to the burst are identified. (4) Check whether the globally unique identifier of all edges and points in each connected component is duplicated with the globally unique identifier of the heat source object. Determine the independent network component that does not contain any globally unique identifier of the heat source object as the affected area disconnected from the main heat source and return it, that is, return the affected area.

2. The method for analyzing pipe bursts in heating networks based on geographic information network data topology conversion according to claim 1, characterized in that, The method for step two is as follows: (1) Calculate the arithmetic mean of the coordinates of all vertices of the outer ring of the polygonal spatial feature, and use the obtained centroid latitude and longitude coordinates as the node coordinates of the polygonal spatial feature in the undirected graph; (2) Store the outer ring coordinates of the polygon spatial features in the node attributes; (3) For a polygonal spatial feature containing an inner ring hole, calculate the polygon difference set between the polygonal region and the hole region, and use the centroid latitude and longitude coordinates of the polygon difference set as the node coordinates of the polygonal spatial feature in the undirected graph. (4) For polygonal spatial features with nested structures, calculate the centroid coordinates of each outer ring independently and assign a sub-identity code.

3. The method for analyzing pipe bursts in heating networks based on geographic information network data topology conversion according to claim 1, characterized in that, In step three, a pipe object with multiple bends between its two endpoints is defined as a multiline spatial element. The line spatial elements include multiline spatial elements, and the processing of multiline spatial elements is as follows: (1) The multi-line spatial elements are split into several independent line spatial elements, and each of the several independent line spatial elements is specially identified in the global unique identifier; (2) Traverse all independent line space features, calculate the arc length in meters, and write the length value into the "Actual Length" attribute field of the independent line space feature; (3) Convert the two endpoints of all independent line space elements into nodes without direction graphs to form virtual nodes.

4. The method for analyzing pipe bursts in heating networks based on geographic information network data topology conversion according to claim 1, characterized in that, In step four, the method for constructing the undirected graph is as follows: (1) Calculate the distance between all virtual nodes. When the distance is less than or equal to the distance threshold, merge the two virtual nodes. (2) Calculate the distance between the valve node and the virtual node. When the distance is less than or equal to the distance threshold, attach the virtual node to the corresponding valve node. (3) Calculate the vertical projection distance from the virtual node and the valve node to the outer ring of the polygonal spatial feature. When the vertical projection distance is less than or equal to the distance threshold, attach the virtual node or the valve node to the corresponding polygonal spatial feature node. (4) Detect the distance between nodes of the same type. When the distance is less than or equal to the distance threshold, add an edge to the undirected graph to connect the two nodes.

5. The method for analyzing pipe bursts in heating networks based on geographic information network data topology conversion according to claim 1, characterized in that, The connected component algorithm is based on a disjoint-set data structure and is performed as follows: (1) Initialize the disjoint set, establish a parent node pointer and rank attribute for each vertex in the undirected graph. In the initial state, the parent node of each vertex points to itself and the rank is 0, which means that each vertex is an independent set. (2) Traverse all edges in the graph that have not been removed by the burst pipe, and perform a merge operation on the two endpoints of each edge; find the root node of the set where the two endpoints are currently located. If the root nodes are different, merge the tree with the smaller rank into the tree with the larger rank as a subtree; if the two trees have the same rank, select any root node and increment the rank by 1. (3) Traverse upwards from the query vertex to the root node, and directly modify the parent pointer of all the nodes passed through to point to the root node; (4) Traverse all vertices, find their final root node, group vertices with the same root node into the same connected component, and output several independent vertex sets, each set representing a connected subnetwork.

Images