Intelligent water supply network grading and partitioning method based on multi-source data fusion

Abstract

The invention relates to the field of water supply system optimization, in particular to a water supply network intelligent grading and partitioning method based on multi-source data fusion. The invention provides a multi-source data fusion-based intelligent grading and partitioning method for a water supply pipe network, which comprises the following steps of: inputting pipe network topological data, flow pressure data and user water consumption historical data, and constructing a pipe network diagram G; according to the data input into the pipe network diagram G, calculating topological characteristics, hydraulic characteristics and user characteristics of each node in the pipe network diagram G, and calculating comprehensive similarity measurement between every two nodes; performing multi-constraint spectral clustering processing on all nodes, and outputting a first-level partition; and performing hierarchical recursive partitioning on the first-level partition, and outputting a second-level partition. Through multi-dimensional data fusion and constraint optimization, the partition reasonability is improved, the first-level partition and the second-level partition are finally output by preprocessing each group of data, the manual partition process is reduced, and the partition division accuracy is improved.

Description

Technical Field

[0001] This invention relates to the field of water supply system optimization, and more specifically, to a method for intelligent hierarchical zoning of water supply networks based on multi-source data fusion. Background Technology

[0002] Pipeline zoning metering is a key tool for leakage management and control in the water supply industry. The rationality of pipeline zoning directly affects the water supply company's management and control of leakage. Unreasonable zoning makes it difficult to detect leakage points, resulting in a large amount of water waste and high operating costs. Traditional zoning methods are mainly implemented based on experience and the current status of the pipeline network, which has the following defects: 1. It relies on manual experience for zoning and lacks quantitative basis; 2. Single parameter decision-making leads to insufficient rationality of zoning; 3. Static zoning is difficult to adapt to dynamic changes in the pipeline network. Summary of the Invention

[0003] To address the technical problems, one objective of this invention is to provide a method for intelligent hierarchical zoning of water supply networks based on multi-source data fusion.

[0004] To achieve the above objectives, embodiments of the present invention provide a method for intelligent hierarchical zoning of water supply networks based on multi-source data fusion, comprising: S1. Input the pipeline topology data, flow and pressure data, and user water usage history data to construct the pipeline diagram G; S2. Based on the data input into the pipeline network diagram G, calculate the values ​​for each node in the pipeline network diagram G. Topological features Hydraulic characteristics User characteristics Calculate the comprehensive similarity measure between every two nodes; S3. Perform multi-constraint spectral clustering on all nodes and output the first-level partition. ; S4. For the first-level partition Perform hierarchical recursive partitioning and output the second-level partitions. .

[0005] In the above technical solution, in step S2, for each node Define multidimensional feature vectors Multidimensional feature vectors Including topological features Hydraulic characteristics and user characteristics .

[0006] In the above technical solution, the topological feature include: Node degree ; Betweenness centrality ,in: It is the total number of shortest paths from node s to node t. It is the number of shortest paths through node i; Proximity centrality ,in: It is a node Average distance to all other nodes It is the total number of nodes; Based on node degree Betweenness centrality and proximity centrality Obtain the data matrix ].

[0007] In the above technical solution, the hydraulic characteristics include nodal pressure. Traffic demand and head loss gradient ; ,in: It is the Hayzen-Williams coefficient; Based on node pressure Traffic demand and head loss gradient Obtain the data matrix .

[0008] In the above technical solution, the user characteristics include: User type ,in: For industry, For business, For residents; Water standard deviation , of which: For the first Water consumption per user This refers to the user's average water consumption over a certain period. Count the number of times a user uses water; Water usage time distribution ,in: Indicates user Water consumption per hour.

[0009] Based on user type Water usage standard deviation and water usage time distribution Obtain the data matrix ].

[0010] In the above technical solution, the steps for calculating the comprehensive similarity metric are as follows: Define the similarity matrix between nodes. Weighting coefficient ,satisfy ; All three types of data were calculated using the cosine similarity formula, namely: ; ; .

[0011] In the above technical solution, step S3 specifically includes: S31. When the above Construct the Laplacian similarity matrix The normalized Laplace matrix is ​​obtained ; S32. Perform eigenvalue decomposition and calculate... The eigenvectors corresponding to the first k smallest eigenvalues Construct the feature matrix ]; S33. To Clustering the vectors yields the first-level partitions. .

[0012] In the above technical solution, step S4 specifically includes: processing each first-level partition. Recalculate the local similarity matrix Recursively apply the MC-SC algorithm to generate secondary partitions. The termination condition is the pressure variance within the partition. , User feature similarity is set to a preset variance threshold. , Minimum partition size is set to the preset similarity threshold. , For the number of partitions, This is the preset minimum number of partitions.

[0013] Additional aspects and advantages of the invention will become apparent in the following description or may be learned by practice of the invention. Attached Figure Description

[0014] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of the invention; Figure 2 This is a schematic diagram illustrating the effect of the present invention without partitioning; Figure 3This is a schematic diagram illustrating the effect of the first-level partitioning of the present invention; Figure 4 This is a schematic diagram illustrating the effect of the two-level partitioning method of this invention; Detailed Implementation To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0015] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0016] The following reference Figures 1 to 4 This invention describes a method for intelligent hierarchical zoning of water supply networks based on multi-source data fusion, according to some embodiments of the present invention.

[0017] like Figures 1 to 4 As shown, an embodiment of the present invention provides a method for intelligent hierarchical zoning of water supply networks based on multi-source data fusion, including: The specific steps for S1 are as follows: Input pipeline topology data, flow and pressure data, and historical user water usage data into the pipeline network modeling software to model the pipeline network as a graph. ,in , represents the set of all nodes; , represents the set of all edges, i.e., the pipe segment. ,in Let be the starting and ending points of the edge. Indicates the length of the pipe section, Indicates the diameter of the pipe section, This is the roughness coefficient of the pipe section.

[0018] The specific steps for S2 are as follows: For each node Define multidimensional feature vectors , including topological features Hydraulic characteristics User characteristics .

[0019] The topological features Including node degree Betweenness centrality and proximity centrality ; Node degree This refers to the node possessing Adjacent nodes number; Betweenness centrality ,in: It is the total number of shortest paths from node s to node t. Betweenness centrality refers to the number of shortest paths passing through node i; it measures the importance of a node as a "mediator" or "bridge" in a network. A higher betweenness centrality indicates that more information, resources, or traffic flows through that node. We use mathematical transformations to scale betweenness centrality to a standard range, i.e., standardization without loss of accuracy. ,in Yes Standardized output value, It is a node betweenness centrality, It is the minimum value of the centrality of the number of intermediate nodes. It is the maximum value of the intermediate centrality of all nodes. After the transformation, the values ​​of all nodes are linearly mapped to the interval [0, 1], which preserves the relative order and distance ratio between the values ​​in the original data. The larger the value, the higher the betweenness centrality of a node, indicating that more information, resources, or traffic is flowing through this node.

[0020] Proximity centrality ,in: It is a node Average distance to all other nodes The total number of nodes is the centrality of a node. Closeness centrality is defined as the reciprocal of the average distance between a node and all other nodes. The shorter the distance, the higher the closeness centrality. It measures how close a node is to all other nodes. Closeness centrality is the average shortest distance from a node to all other nodes in the network. The higher the closeness centrality of a node, the more it is topologically located at the "center" of the network and can reach all other nodes very quickly.

[0021] Based on node degree Betweenness centrality and proximity centrality Obtain the data matrix ].

[0022] The hydraulic characteristics include nodal pressure. Traffic demand and head loss gradient ; Node pressure and traffic demand Data is taken from hydraulic simulations and sensor acquisitions; The head loss gradient uses the Hayzen-Williams formula, i.e. ,in: The Hayzen-Williams coefficient (dimensionless) reflects the roughness of the pipe's inner wall. New steel pipes and plastic pipes: Used steel pipes and cast iron pipes: Concrete pipe: ; Based on node pressure Traffic demand and head loss gradient Obtain the data matrix .

[0023] The user characteristics include user type Water usage standard deviation and water usage time distribution ; User type ,in: For industry, For business, For residents, the calculated user type Perform normalization processing; Water standard deviation , of which: For the first Water consumption per user This refers to the user's average water consumption over a certain period. The number of times a user uses water is counted, and the calculated standard deviation of water usage is used. Perform normalization processing; Water usage time distribution ,in: Indicates user Water consumption per hour, and the calculated water consumption distribution during different time periods. Perform normalization processing; Based on the normalized user type Water usage standard deviation and water usage time distribution Obtain the data matrix ].

[0024] The steps for calculating the comprehensive similarity metric are as follows: Define the similarity matrix between nodes. Weighting coefficient ,satisfy ; All three types of data were calculated using the cosine similarity formula, namely: ; ; .

[0025] The specific steps for S3 are as follows: S31. Computation Node similarity If not connected, then Construct the Laplacian similarity matrix The normalized Laplace matrix is ​​obtained ; S32. Perform eigenvalue decomposition and calculate... The eigenvectors corresponding to the first k smallest eigenvalues Construct the feature matrix ]; S33. To Clustering the vectors yields the first-level partitions. .

[0026] The specific steps for S4 are as follows: For each first-level partition Recalculate the local similarity matrix Recursively apply the MC-SC algorithm to generate secondary partitions. The termination condition is the pressure variance within the partition. , User feature similarity is set to a preset variance threshold. , Minimum partition size is set to the preset similarity threshold. , For the number of partitions, This is the preset minimum number of partitions.

[0027] After completing steps S1, S2, S3, and S4, the post-processing stage begins. This stage includes connectivity correction and hydraulic balance verification. Connectivity correction primarily determines if there are disconnected subgraphs in the clustering results and merges the nearest neighbor nodes. Hydraulic balance verification involves adjusting the partition boundary valves to ensure pressure balance.

[0028] The following describes the implementation process of this invention in detail using a portion of a city's water supply network (taking "City A" as an example, including 200 nodes and 200 pipe sections), covering data preparation, algorithm implementation, parameter optimization, and result verification.

[0029] 1. Data preparation and preprocessing 1.1 Data Sources and Formats (1) Pipeline topology data: Export nodes.csv (fields: node ID, coordinate X, coordinate Y, elevation, pipe segment connection list) and edges.csv (fields: start ID, end ID, length, pipe diameter, material) from the A City GIS system.

[0030] (2) Hydraulic parameter data: real-time data hydraulic.csv (fields: node ID, timestamp, pressure (MPa), flow (m³ / h)) is obtained from the SCADA system of City A and stored in hourly granularity.

[0031] (3) User characteristic data: Export users.csv (fields: node ID, user type code, daily water consumption standard deviation, 24-hour water consumption curve) from the metering system of City A.

[0032] 1.2 Data Preprocessing Steps (1) Topology modeling Graph structure construction: Use the networkx library to convert nodes and pipe segments into a directed graph G, calculate topological features, and save the results to nodes_features.csv, with fields: node ID, degree, betweenness centrality, and proximity centrality.

[0033] (2) Hydraulic simulation Steady-state simulations were performed using the EPANET tool to generate node pressure and flow requirements. The results were saved to hydraulic_features.csv, with fields including node ID, pressure (MPa), and flow requirement (m³ / h).

[0034] (3) User feature normalization User types (industrial=0, commercial=1, residential=2) are coded using unique thermal encoding, and the 24-hour water usage curves are normalized using Min-Max. The results are saved to users_normalized.csv.

[0035] 2. Algorithm Implementation Steps 2.1 First-level partition Input: Merge nodes_features.csv, hydraulic_features.csv, and users_normalized.csv into merged_data.csv.

[0036] Overall similarity calculation: based on the formula above. Calculate the overall similarity and set the weights. .

[0037] First-level partitioning is performed using multi-constraint spectral clustering, such as... Figure 2 As shown in (b).

[0038] 2.2 Second-level partitioning: Hierarchical recursive spectral clustering Recursive logic: For each first-level sub-map, recalculate the local similarity matrix (focusing on flow balance and terrain elevation differences): set weights. Spectral clustering yields sub-partitions, with termination conditions being pressure variance < 0.1 MPa², user similarity > 0.85, and minimum number of nodes ≥ 50. Figure 2 As shown in (c).

[0039] 3. Post-processing optimization 3.1 Connectivity Correction Detecting isolated subgraphs: Use nx.connected_components to identify disconnected clustering results.

[0040] Merging strategy: Calculate the shortest distance between subgraphs and merge the subgraphs with the closest distance (based on Dijkstra's algorithm).

[0041] 3.2 Hydraulic Balance Verification EPANET was used for dynamic simulation, and the opening of the boundary valves was adjusted to ensure that the pressure fluctuation within the zone was less than 5%.

[0042] The present invention has the following advantages: 1. This application improves the rationality of partitioning by multi-dimensional data fusion and constraint optimization. It preprocesses each group of data and finally outputs first-level and second-level partitions, reducing the manual partitioning process and improving the accuracy of partitioning. 2. This application solves the problem that static partitioning is difficult to adapt to dynamic changes in the pipeline network by inputting multi-dimensional data in real time and then outputting first-level and second-level partitions in real time.

[0043] In this invention, the terms "installation," "connection," "linking," and "fixing" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; "linking" can be a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of these terms in this invention according to the specific circumstances.

[0044] In the description of this invention, it should be understood that the terms "inner" and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or unit referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0045] In the description of this specification, the terms "one embodiment," "some embodiments," "specific embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0046] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent hierarchical zoning of water supply networks based on multi-source data fusion, characterized in that, include: S1. Input the pipeline topology data, flow and pressure data, and user water usage history data to construct the pipeline diagram G; S2. Based on the data input into the pipeline network diagram G, calculate the values ​​for each node in the pipeline network diagram G. Topological features Hydraulic characteristics User characteristics Calculate the comprehensive similarity measure between every two nodes; S3. Perform multi-constraint spectral clustering on all nodes and output the first-level partition. ; S4. For the first-level partition Perform hierarchical recursive partitioning and output the second-level partitions. .

2. The intelligent hierarchical zoning method for water supply networks based on multi-source data fusion according to claim 1, characterized in that, In step S2, for each node Define multidimensional feature vectors Multidimensional feature vectors Including topological features Hydraulic characteristics and user characteristics .

3. The intelligent hierarchical zoning method for water supply networks based on multi-source data fusion according to claim 2, characterized in that, The topological features include: Node degree ; Betweenness centrality ,in: It is the total number of shortest paths from node s to node t. It is the number of shortest paths through node i; Proximity centrality ,in: It is a node Average distance to all other nodes It is the total number of nodes; Based on node degree Betweenness centrality and proximity centrality Obtain the data matrix ].

4. The intelligent hierarchical zoning method for water supply networks based on multi-source data fusion according to claim 2, characterized in that, The hydraulic characteristics include nodal pressure. Traffic demand and head loss gradient ; ,in: It is the Hayzen-Williams coefficient; Based on node pressure Traffic demand and head loss gradient Obtain the data matrix .

5. The intelligent hierarchical zoning method for water supply networks based on multi-source data fusion according to claim 2, characterized in that, The user characteristics include: User type ,in: For industry, For business, For residents; Water standard deviation , of which: For the first Water consumption per user This refers to the user's average water consumption over a certain period. Count the number of times a user uses water; Water usage time distribution ,in: Indicates user Water consumption per hour. Based on user type Water usage standard deviation and water usage time distribution Obtain the data matrix ].

6. The intelligent hierarchical zoning method for water supply networks based on multi-source data fusion according to claim 2, characterized in that, The steps for calculating the comprehensive similarity metric are as follows: Define the similarity matrix between nodes. Weighting coefficient ,satisfy ; All three types of data were calculated using the cosine similarity formula, namely: ； ； 。 7. The intelligent hierarchical zoning method for water supply networks based on multi-source data fusion according to claim 2, characterized in that, Step S3 specifically includes: S31. When the above Construct the Laplacian similarity matrix The normalized Laplace matrix is ​​obtained ; S32. Perform eigenvalue decomposition and calculate... The eigenvectors corresponding to the first k smallest eigenvalues Construct the feature matrix ]; S33. To Clustering the vectors yields the first-level partitions. .

8. The intelligent hierarchical zoning method for water supply networks based on multi-source data fusion according to claim 2, characterized in that, Step S4 specifically includes: for each first-level partition Recalculate the local similarity matrix Recursively apply the MC-SC algorithm to generate secondary partitions. The termination condition is the pressure variance within the partition. , User feature similarity is set to a preset variance threshold. , Minimum partition size is set to the preset similarity threshold. , For the number of partitions, This is the preset minimum number of partitions.

Images