CN111353202A - Partitioning method for underground pipe network general investigation in municipal administration - Google Patents

Abstract

The invention discloses a partitioning method facing to underground pipe network general survey in municipal administration, which comprises the following steps of (1) collecting pipe network data of an area to be general surveyed, obtaining pipe point data and pipe section data according to the pipe network information of the area to be general surveyed, and constructing a topology structure chart of a pipe network; (2) extracting a main network of the underground pipe network based on a cutting algorithm; (3) based on a community discovery algorithm, dividing an underground pipe network into a plurality of minimum partition units by taking the length of the pipe network as a convergence condition; (4) inputting the number of appointed general investigation partitions, and aggregating the minimum partition units of each pipe network by taking the length and the area of the minimum partition unit as constraint conditions; (5) and outputting pipe network partition results, including pipe network vector data and statistical information such as pipe network length, area and the like of each partition. The invention overcomes the defects that the traditional regular grid partitioning method damages the continuity of the pipe network, the workload of general survey is uneven and the like in the general survey work of urban underground pipe network management.

Description

A zoning method for underground pipe network census in municipal management

technical field

The invention relates to a method for partitioning an urban water supply pipe network, in particular to a method for partitioning an urban underground water supply pipe network census in municipal management, and belongs to the technical field of municipal engineering management.

Background technique

Underground pipelines are an important urban infrastructure, known as the "lifeline" of the city. With the continuous expansion of the city, the old and new underground pipelines are mixed, and the network structure is becoming more and more complicated. Due to the lack of comprehensive pipeline information, especially accurate spatial location information, various safety accidents often occur and even lead to major economic losses. While strengthening the scientific planning and information management of pipelines, more and more attention has been paid to the census and detection of existing urban underground pipelines. In the traditional pipe network census, the census area is usually demarcated manually or a regular grid based on map sheets is used for division. It not only destroys the overall topology connectivity of the pipe network, but also causes the discontinuity of the census pipe sections. At the same time, the structure of the pipeline network within the region is ignored, which leads to an unbalanced distribution of the workload of the census based on partitions. Therefore, in the face of the census of the underground pipeline network, how to scientifically and effectively divide the pipeline network is the key issue in the specific operation of the pipeline network census.

At present, the technology of pipe network partitioning is mainly aimed at the research of various problems such as pipe network design and analysis. In order to reduce the scale and complexity of the pipe network, it is usually based on the simplification of the pipe network, and the backbone of the pipe network is mainly discussed. It is applied to the planning of new urban pipeline networks and the optimization of pressure zoning for existing pipeline networks. However, starting from the census of the current pipeline network, the analysis and modeling of the urban underground complex pipeline network are carried out, and the research on the partition of the pipeline network is rarely involved. Considering the integrity and topology characteristics of the pipeline network to meet the needs of the number of specific work areas in the census division, there is a lack of targeted research.

SUMMARY OF THE INVENTION

Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides a zoning method for the census of the underground pipe network in municipal management, which realizes that the total length and area of the census of the zoning of the pipe network are approximately equal, and satisfies the reasonable distribution of the census workload. demand. Applicable to the vast majority of pipe network census zoning problems. The physical properties of the pipe network are fully considered, the good connectivity of the pipe network in the partition is maintained, and the split of the pipe section after the partition is avoided.

Technical scheme: In order to realize the above-mentioned purpose, the technical scheme adopted in the present invention is:

A zoning method for underground pipe network census in municipal management, comprising the following steps:

Step 1: Collect pipe network information in the area to be surveyed, and obtain pipe point data and pipe section data according to the pipe network information in the area to be surveyed. The pipeline network equipment has important characteristics in the census of the underground pipeline network, focusing on the pipeline network equipment (such as tee, valve well, venting and other facilities and equipment), and the relevant information of the pipe section (such as material, pipe diameter, etc.). This method requires two abstractions. The first time abstracts the pipe network equipment as vector points and the pipe segments as vector lines, and integrates the pipe point data and the pipe segment data to form the pipe network vector data. The second abstraction takes the pipe point and the pipe segment as a unified spatial entity object, and projects from the Euclidean space to the topological space as a whole. According to the pipe network vector data, the pipe points are abstracted into nodes, the pipe segments are abstracted into the edges connecting the nodes, and the length of the pipe segments is used as the weight of the edges to construct the undirected topological graph of the pipe network.

The undirected graph of the pipe network topology obtained in step 2 and step 1 is the actual pipe network structure and has not been simplified. The pipe network is structurally divided using a specific tailoring algorithm. The division results include the backbone skeleton network structure and the branch network structure. The backbone skeleton network structure is the core of the entire pipeline network, which is mainly composed of the main pipeline network and some branch pipeline networks involving key topological positions. The branch network structure has obvious predecessor and successor relationship, forming a hierarchical tree structure, in which there is only one predecessor node, and the leaf node without successor node. Extract the backbone network structure and branch network structure of the water supply pipe network in the undirected graph of the pipe network topology. In the undirected graph of pipe network topology, there are many nodes with degree 1. After deleting these nodes, new nodes with degree 1 will appear. After iteration, the remaining topology is the backbone network structure. The order in which nodes are deleted is recorded, and the reverse order output is constructed as a tree structure.

Step 3: Divide the water supply pipe network into nc units according to the obtained backbone network structure and branch network structure.

In step 31, each node in the backbone skeleton network structure is regarded as an independent unit, and the number of initial units is the same as the number of nodes. And the weight of each unit includes all length weights of the tree structure mounted on the node. Because the workload of the pipe network census is mainly the length of the pipe network, the length attribute of the pipe segment is set as the weight here.

Step 32: For each node i , try to allocate node i to the unit where its adjacent nodes are located, calculate the modularity difference ΔQ before and after allocation, and record the adjacent node with the largest modularity difference ΔQ. If the largest ΔQ>0, then assign node i to the unit where the adjacent node with the largest ΔQ is located, otherwise give up the division. The formula for calculating the modularity difference is as follows:

是单元C内的权重之和，

表示单元C的内 部权重，

是连接单元C的外部连接的权重之和，

表示连接单元C的外部权重，C表 示邻接节点n所在的单元，

是连接节点i的外部连接的权重之和，

是节点i与内部 节点的连接权重相加的和，

为整个网络的权重之和。 Among them, ΔQ represents the modularity difference, TopologySim represents the topology similarity of two adjacent nodes, i,n is the node number, i represents the ith node, n is the adjacent node of i ,

is the sum of the weights within cell C,

represents the internal weight of unit C,

is the sum of the weights of the outer connections of the connection unit C,

represents the external weight of the connection unit C, C represents the unit where the adjacent node n is located,

is the sum of the weights of the outer connections connecting node i ,

is the sum of the connection weights of node i and internal nodes,

is the sum of the weights of the entire network.

Step 33: Repeat step 32 until the units to which all nodes belong are no longer changed.

In step 34, the node with the same attribution obtained in step 33 is regarded as a new node, and the subgraph is reconstructed, and the weight between the two new nodes is the sum of the weights of all the edges between the corresponding two units.

Step 35: Given the minimum community subgroup resolution parameter, use the subgraph obtained in step 34 as input, and re-execute steps 31 to 34 until the given minimum subgroup resolution parameter is satisfied, and obtain the minimum community subgroup and community. subgroup.

Step 36 , output the minimum community sub-group and the connection relationship of the community sub-group, and use the minimum community sub-group as the minimum partition unit of the pipe network. This step realizes the most refined segmentation of the pipe network within the specified resolution.

Step 4: On the basis of the subcluster connection relationship, according to the specified number of partitions, the length is the main convergence condition, and the area is the secondary convergence condition, and the minimum partition unit aggregation of the pipe network is realized to reach the specified number of partitions. Taking the new node as the possibility of a community subgroup, the specific formula for calculating the possibility of node i being selected as a community subgroup is as follows:

表示节点i被选为社区子团的概率，D（i）定义为节点i到邻居子团欧氏空 间的质心的距离，

表示节点i的所有邻接节点n到节点i的距离平方和。此外，不 同于一般的社区发现算法的收敛条件，本方法还从面积分配的角度，定义空间面积上的收 敛条件： in,

represents the probability that node i is selected as a community subgroup, D ( i ) is defined as the distance from node i to the centroid of the Euclidean space of neighbor subgroups,

Represents the sum of squared distances from all adjacent nodes n of node i to node i . In addition, different from the convergence conditions of general community discovery algorithms, this method also defines the convergence conditions on the space area from the perspective of area allocation:

表示每个分区面积的最小值，

表示分区数，

为候选节点，

是 第

个分区，m _i 为邻接节点的质心。根据不同的收敛数，选择

取极值点处的数值， 作为最优的子团分配方式。 in,

represents the minimum value of the area of each partition,

represents the number of partitions,

is a candidate node,

is the first

partitions, where m _i is the centroid of adjacent nodes. According to different convergence numbers, choose

The value at the extreme point is taken as the optimal subcluster allocation method.

Step 41: Assign the weight of the branch network structure to the node of the backbone network structure to which it is mounted, and assign the length of the pipe section of the branch network structure to the unit according to the unit to which the node belongs.

Step 42 , select the node with the smallest weight in the community subgroup, merge the node with the smallest weight among its adjacent nodes as a new node, and the new node weight is obtained by adding the sum of the edge weights between the original node and the original node. When the two nodes with the smallest weight do not have direct topological associations, the area of the outer rectangle when the two nodes are added to the neighbors are calculated respectively, and the merging scheme with a larger area change is selected.

Step 43: Repeat step 42 until all nodes are merged to obtain each partition under different partition schemes.

Step 44: Calculate the variance of each partition length under different partition schemes, and select the partition scheme with the smallest variance. When the length variances are equal, the area of the enclosing rectangle is used as the judgment condition.

Step 5, output the pipe network partition result, including the vector data and the pipe network length information of the partition.

Preferably: the method for constructing the undirected graph of the pipe network topology in step 1 is as follows:

Step 11: Collect pipe network equipment (such as tee, valve well, venting and other facilities and equipment), and related information of pipe sections (such as material, pipe diameter, etc.) point data and pipe segment data) for conceptual abstraction. First, check the pipe point data and pipe segment data to ensure that each pipe segment is a connection between two pipe points and that there is no redundant pipe point equipment inside the pipe segment. Correct pipe segments and pipe points that do not meet the requirements, delete pipe segments with missing end point pipe points, and break pipe segments with other pipe points inside. On the basis of completing the inspection, the pipe points are abstracted into nodes, the pipe segments are abstracted into edges connecting nodes, and the length of the pipe segments is used as the weight of the edges to construct the topological graph of the pipe network.

与节点

相连，则邻接表中增加一条记录

，

是节点编 号，

、

分别表示第

个节点。 Step 12: Analyze each node in turn, find the adjacent nodes that share the same edge with the node, and construct the node's adjacency list.

with node

If connected, add a record to the adjacency list

,

is the node number,

,

respectively represent the

node.

Step 13: Since an undirected graph is constructed for the pipe network, the pipe points do not distinguish between in-degree and out-degree. Therefore, according to the node number size rule, the duplicate node adjacency information is deleted, and the pipe network topology undirected graph G is formed according to the adjacency table.

Preferably: the method for extracting the backbone network structure and branch network structure of the water supply pipe network in the undirected graph of the pipe network topology in step 2 is as follows:

Step 21, traverse each node in the undirected graph of the pipe network topology, if the node has no adjacent node, record the node as a discrete value, and delete the node.

，

，

是边的编号，

表示第

条边，

是节点编号，

表示第

个节点，若节点

的邻接节点数大于 节点

的邻接节点数，且节点

仅和节点

相邻，则

为

的父节点，更新父子关系表， 并删除边

和节点

。 Step 22, traverse each edge in the undirected graph of the pipe network topology

,

,

is the edge number,

means the first

stripe,

is the node number,

means the first

nodes, if the node

The number of adjacent nodes is greater than the node

the number of adjacent nodes, and the node

and node only

adjacent, then

for

the parent node, update the parent-child relationship table, and delete the edge

and node

.

In step 23, steps 21 and 22 are repeated until no nodes can be simplified, and the remaining structure in the undirected graph of the pipe network topology is the backbone skeleton network structure R of the pipe network.

Step 24: According to the parent-child relationship table, for any node V _g1 in the parent-child relationship table, find its predecessor node V _j1 , if V _j1 is still in the parent-child relationship table, continue to search for the predecessor node V _k1 of V _j1 , and so on. , until the precursor node does not exist or is a point in the backbone skeleton network structure, build a tree structure T according to this rule, the tree structure T is the branch network structure, g1 , j1 , k1 are the node numbers, representing the first g1 , j1 and k1 nodes.

Preferably: the method for outputting the pipe network partition result in step 5 is as follows:

Step 51: Assign the partition number to the smallest partition unit and all nodes inside the unit to obtain the partition number table of the pipe point.

Step 52, associate the partition number table with the vector data according to the identification field of the pipe point.

Step 53: Count the lengths of pipe sections in different partitions, and if two pipe points in one pipeline section are in the same partition, record the length of this pipeline section as the length of this partition. Otherwise this segment length is ignored.

Step 54, outputting the vector data associated with the partition table and the length statistics of each partition.

Preferably: in step 3, the pipe network structure diagram G=G(V, E), V is the node abstracted by the pipe point, E is the edge abstracted by the pipe segment, and the community discovery is to determine nc units in the pipe network structure diagram G , nc≥1, so that the node set of each unit constitutes a cover of node V.

Compared with the prior art, the present invention has the following beneficial effects:

(1) The process of dividing the area in the present invention is a "discrete-aggregation" process. The community discovery algorithm is used to find the smallest division unit that maintains connectivity, and the division is generated in consideration of constraints such as length and area. Therefore, the total length of the divided pipeline network, that is, the census workload is basically the same, and the optimization of the census work area will facilitate the rational distribution of the workload during the census.

(2) Taking into account the connection relationship of pipeline entities, the integrity of the internal pipeline network in the region is ensured. The traditional grid division and partitioning methods separate the pipeline sections and increase the redundancy of the census work.

(3) The traditional pipe network partition method is characterized by the end-to-end connection of the pipe network, and the annular pipe segment is extracted as the backbone network. In the census of the pipe network in local areas of the city, this feature is not applicable due to the lack of data. The clipping algorithm of the invention can overcome this defect and obtain the primary and secondary pipe sections of the pipe network topology structure.

(4) The number of divided areas can be flexibly set according to the user, which has strong flexibility, while the traditional grid division method can only take an even number.

Description of drawings

Figure 1 is a schematic diagram of the backbone network structure of the pipe network.

FIG. 2 is a schematic diagram of the tree structure of the pipe network.

Figure 3 is a schematic diagram of the polymerization process of the minimum unit.

Figure 4 shows the backbone skeleton network of a city's water supply network.

Fig. 5 is a schematic diagram when a city's pipeline network is divided into two partitions.

Fig. 6 is a schematic diagram when a city's pipeline network is divided into 4 partitions.

Fig. 7 is a schematic diagram when a city's pipeline network is divided into 5 partitions.

Detailed ways

Below in conjunction with the accompanying drawings and specific embodiments, the present invention will be further clarified. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. Modifications in the form of valence all fall within the scope defined by the appended claims of the present application.

A zoning method for underground pipe network census in municipal management, comprising the following steps:

Step 1: Collect pipe network equipment (such as tee, valve well, venting and other facilities and equipment), and related information of pipe sections (such as material, pipe diameter, etc.). This method requires secondary abstraction. For the first time, the pipe network equipment is abstracted as a vector point, and the pipe segment is abstracted as a vector line, and the pipe network vector data is formed according to the comprehensive pipe point data and pipe segment data. The second abstraction takes the pipe point and the pipe segment as a unified spatial entity object, and projects from the Euclidean space to the topological space as a whole. According to the pipe network vector data, the pipe points are abstracted into nodes, the pipe segments are abstracted into the edges connecting the nodes, and the length of the pipe segments is used as the weight of the edges to construct the undirected topological graph of the pipe network.

The method of constructing the undirected graph of the pipe network topology is as follows:

Step 11, the pipe network vector data (shapefile format) includes pipe point data and pipe segment data. Check the pipe point data and pipe segment data to ensure that each pipe segment is a line connecting two pipe points and that there are no pipe points inside the pipe segment. Correct pipe segments and pipe points that do not meet the requirements, delete pipe segments with missing end point pipe points, and break pipe segments with other pipe points inside. On the basis of completing the inspection, the pipe points are abstracted into nodes, the pipe segments are abstracted into edges connecting nodes, and the length of the pipe segments is used as the weight of the edges to construct the topological graph of the pipe network.

与节点

相连，则邻接表中增加一条记录

，

是节点编 号，

、

分别表示第

个节点。 Step 12: Analyze each node in turn, find the adjacent nodes that share the same edge with the node, and construct the node's adjacency list.

with node

If connected, add a record to the adjacency list

,

is the node number,

,

respectively represent the

node.

，若

节点编号大，则删除邻接表

，依据邻接表形成管网拓扑无向 图G。 Step 13: Delete duplicate node adjacency information, such as two records in the adjacency table, according to the node number size rule

,like

If the node number is large, delete the adjacency list

, and form the undirected graph G of the pipe network topology according to the adjacency list.

Step 2, the undirected graph of the pipe network topology obtained in step 1 includes the backbone network structure and the branch network structure. The backbone skeleton network structure is the core of the entire pipe network, which is mainly composed of the main line pipe network and some branch pipe networks involving key topological positions. . The branch network structure has obvious predecessor and successor relationship, forming a hierarchical tree structure, in which there is only one predecessor node, and the leaf node without successor node. As shown in Figures 1 and 2, the spatial structure of the pipe network can be divided into a backbone network structure and a tree structure. The main pipe section carries the main load functions of the entire pipe network, as shown in Figure 1. When the pipeline is laid into a residential area or a new area, it often presents a tree-like structure, as shown in Figure 2. Using this feature, the cutting algorithm is executed to delete leaf nodes iteratively, and the backbone network structure and branch network structure of the water supply pipe network in the undirected graph of the pipe network topology can be extracted.

The method of extracting the backbone network structure and branch network structure of the water supply pipe network in the undirected graph of the pipe network topology is as follows:

Step 21, traverse each node in the undirected graph of the pipe network topology, if the node has no adjacent node, record the node as a discrete value, and delete the node.

，

，

是边的编号，

表示第

条边，

是节点编号，

表示第

个节点，若节点

的邻接节点数大于 节点

的邻接节点数，且节点

仅和节点

相邻，则

为

的父节点，更新父子关系表， 并删除边

和节点

。 Step 22, traverse each edge in the undirected graph of the pipe network topology

,

,

is the edge number,

means the first

stripe,

is the node number,

means the first

nodes, if the node

The number of adjacent nodes is greater than the node

the number of adjacent nodes, and the node

and node only

adjacent, then

for

the parent node, update the parent-child relationship table, and delete the edge

and node

.

Step 23: Repeat steps 21 and 22 until no nodes can be deleted. The remaining structure in the undirected graph of the pipe network topology is the backbone skeleton structure R of the pipe network, which is the core of the entire pipe network and is mainly managed by the main line. It consists of a network and some branch pipelines involving key topological positions. It is a structure after subtracting branch pipelines that can be deleted. There are some edge trunk lines that are not topologically important. On the contrary, some branch lines have important topology. The status of the structure, through the above steps, the influence of the pipeline level on the network structure can be reduced, and the pipeline network structure most suitable for the census division can be formed.

Step 24: According to the parent-child relationship table, for any node V _g1 in the parent-child relationship table, its predecessor node V _j1 can be found. If V _j1 is still in the parent-child relationship table, continue to search for the predecessor node V _k1 of V _j1 , and sequentially. By analogy, until the predecessor node does not exist or is a point in the backbone skeleton structure, a tree structure T is constructed according to this rule. The tree structure T is the branch network structure, and g1 , j1 , k1 are the node numbers, representing g1 , j1 and k1 nodes.

Step 3: Divide the water supply pipe network into nc units according to the obtained backbone network structure and branch network structure.

Pipe network structure graph G=G(V, E), V is the node abstracted by the pipe point, E is the edge abstracted by the pipe segment, community discovery is to determine nc units in the pipe network structure graph G, nc≥1, so that The set of nodes of each element constitutes a cover of node V.

The method of dividing the water supply network into nc units is as follows:

In step 31, each node in the backbone network structure is regarded as an independent unit, and the number of initial units is the same as the number of nodes.

Step 32: For each node i , try to allocate node i to the unit where its adjacent nodes are located, calculate the modularity difference ΔQ before and after allocation, and record the adjacent node with the largest modularity difference ΔQ. If the largest ΔQ>0, then assign node i to the unit where the adjacent node with the largest ΔQ is located, otherwise give up the division. The formula for calculating the modularity difference is as follows:

是单元C内的权重之和，

表示单元C的内 部权重，

是连接单元C的外部连接的权重之和，

表示连接单元C的外部权重，C 表示邻接节点n所在的单元，

是节点i与内部节点的连接权重相加的和，

是连接节 点i的外部连接的权重之和，

为整个网络的权重之和。 where ΔQ represents the modularity difference, TopologySim represents the topology similarity of two adjacent nodes, i,n is the node number, i represents the ith node, n is the adjacent node of i ,

is the sum of the weights within cell C,

represents the internal weight of unit C,

is the sum of the weights of the outer connections of the connection unit C,

represents the external weight of the connection unit C, C represents the unit where the adjacent node n is located,

is the sum of the connection weights of node i and internal nodes,

is the sum of the weights of the outer connections connecting node i ,

is the sum of the weights of the entire network.

Step 33: Repeat step 32 until the units to which all nodes belong are no longer changed.

In step 34, the node with the same attribution obtained in step 33 is regarded as a new node, and the subgraph is reconstructed, and the weight between the two new nodes is the sum of the weights of all the edges between the corresponding two units.

Step 35: Given the minimum community subgroup resolution parameter, take the subgraph obtained in step 34 as input, and re-execute steps 31 to 34 until the given minimum community subgroup resolution parameter is reached, and obtain the minimum community subgroup and community subgroup.

Step 36, take the smallest community subgroup as the smallest partition unit of the pipe network, and output its connection relationship with the community subgroup.

Step 4: According to the specified number of partitions, and with the length equivalent as the convergence condition, realize the aggregation (dynamic aggregation) of the smallest partition unit of the pipe network. When there are multiple adjacent nodes with smaller weights, calculate the area size of the outer rectangle when this node is added to the neighbor respectively, and assign the node to the adjacent node with smaller area. Through the binary judgment of length and area, the attribution of the subcluster is finally determined. The specific formula is as follows:

表示节点i被选为社区子团的概率，D（i）定义为节点i到邻居子团欧氏空 间的质心的距离，

表示节点i的所有邻接节点n到节点i的距离平方和。此外，不 同于一般的社区发现算法的收敛条件，本方法还从面积分配的角度，定义空间面积上的收 敛条件： in,

represents the probability that node i is selected as a community subgroup, D ( i ) is defined as the distance from node i to the centroid of the Euclidean space of neighbor subgroups,

Represents the sum of squared distances from all adjacent nodes n of node i to node i . In addition, different from the convergence conditions of general community discovery algorithms, this method also defines the convergence conditions on the space area from the perspective of area allocation:

表示每个分区面积的最小值，

表示分区数，

为候选节点，

是第

个分区，m _i 为邻接节点的质心。根据不同的收敛数，选择

取极值点处的 数值，作为最优的子团分配方式。 of which,

represents the minimum value of the area of each partition,

represents the number of partitions,

is a candidate node,

is the first

partitions, where m _i is the centroid of adjacent nodes. According to different convergence numbers, choose

The value at the extreme point is taken as the optimal subcluster allocation method.

The aggregation process of the smallest unit is as follows:

Step 41: Assign the weight of the branch network structure to the node of the backbone network structure to which it is mounted, and assign the length of the pipe section of the branch network structure to the unit according to the unit to which the node belongs.

Step 42 , select the node with the smallest weight in the community subgroup, merge the node with the smallest weight among its adjacent nodes as a new node, and the new node weight is obtained by adding the sum of the edge weights between the original node and the original node.

Step 43: Repeat step 42 until all nodes are merged to obtain each partition under different partition schemes.

Step 44: Calculate the variance of each partition length under different partition schemes, and select the partition scheme with the smallest variance. When the length variances are equal, the area of the enclosing rectangle is used as the judgment condition.

As shown in Figure 3, a topology network composed of a unit has a total of 8 small units a~h, and each node level is determined successively according to the defined rules.

In the first iteration, select the node with leaf nodes mounted on the backbone skeleton node, and then merge the units according to the cutting method, so c belongs to b, g and h belong to e, at this time, the entire backbone skeleton network is divided into four parts .

In the second iteration, select the smallest node a, which belongs to the adjacent node b or d with the lowest weight. By analogy, d and e are combined to form a dichotomous scheme of the pipe network.

Eventually all the small units converge to a node representing the whole pipe network.

Step 5: Output the pipe network partition result, including vector data, the pipe network length information of the partition, and the area information of the outer rectangle.

The method of outputting the pipe network partition results is as follows:

Step 51: Assign the partition number to the smallest partition unit and all nodes inside the unit to obtain the partition number table of the pipe point.

Step 52, associate the partition number table with the vector data according to the identification field of the pipe point.

Step 53: Count the lengths of pipe sections in different partitions, and if two pipe points in one pipeline section are in the same partition, record the length of this pipeline section as the length of this partition. Otherwise this segment length is ignored.

Step 54, outputting the vector data associated with the partition table and the length statistics of each partition.

Example

Taking the sample water supply pipe network data as an example, its coverage area is about 10.2 square kilometers, the pipe network data includes 12,807 pipe points, 15,114 pipe sections, and the total length is 102.78km. The pipeline network is divided into 2 points, 4 points and 5 points respectively. The pipe network structure is shown in Figure 4.

As shown in Figure 5, when the pipeline network is divided into two partitions: partition 1 contains a total of 6676 pipeline points and 7917 pipeline segments, and the total length of the pipeline segments is 58.172km. Division 2 contains a total of 6131 pipe points and 7197 pipe sections with a total length of 44.608km and an area of 4.04km ² and 3.98km ² respectively.

As shown in Figure 6, when the pipe network is divided into 4 partitions: partitions 1 to 4 contain 3488, 2833, 3145 and 3341 pipe points respectively. 3644, 3698, 3759, 4034 pipe sections. The total lengths of the pipe sections are 24.39km, 25.61km, 23.34km and 30.66km respectively. The areas are 2.01km ² , 1.99km ² , 2.15km ² and 1.99km ² respectively.

As shown in Figure 7, when the pipe network is divided into 5 partitions: partitions 1 to 5 contain 2760, 2253, 3328, 2106 and 2360 pipe points respectively. 2704, 2890, 3892, 2874, 2934 pipe sections. The total lengths of the pipe sections are 25.04km, 25.26km, 25.21km and 25.81km respectively. The areas are 1.24km ² , 1.16km ² , 1.33km ² , 1.40km ² and 1.52km ² respectively.

The invention overcomes the defects of the traditional regular grid partition method in the census work of the urban underground pipeline network, which destroys the continuity of the pipeline network and the census workload is uneven, and realizes the reasonable pipeline network partition with the number of census kilometers as the index.

The above is only the preferred embodiment of the present invention, it should be pointed out that: for those skilled in the art, without departing from the principle of the present invention, several improvements and modifications can also be made, and these improvements and modifications are also It should be regarded as the protection scope of the present invention.

Claims (5)

1. A partitioning method facing underground pipe network general survey in municipal administration is characterized by comprising the following steps:

step 1, collecting pipe network information of an area to be generally surveyed, obtaining pipe network data and pipe section data according to the pipe network information of the area to be generally surveyed, abstracting twice for pipe network equipment in underground pipe network general survey, wherein the pipe network equipment is abstracted into vector points for the first time, the pipe sections are abstracted into vector lines, and the pipe network vector data are formed by integrating the pipe network data and the pipe section data; the pipe section with pipe points is used as a uniform space entity object in the second abstraction, the pipe section with the pipe points is projected to a topological space from an Euclidean space as a whole, the pipe points are abstracted into nodes according to pipe network vector data, the pipe section is abstracted into edges connecting the nodes, the length of the pipe section is used as the weight of the edges, and a pipe network topological undirected graph is constructed;

step 2, the pipe network topological undirected graph obtained in the step 1 is an actual pipe network structure, is not simplified, and is divided by using a cutting algorithm; the division result comprises a main skeleton network structure and a branch network structure, wherein the main skeleton network structure is the core of the whole pipe network and mainly comprises a main pipe network and a part of branch pipe networks related to key topological positions; the branch network structure has obvious predecessor and successor relations to form a hierarchical tree structure, wherein only one predecessor node is provided, and leaf nodes are provided without successor nodes; extracting a main network structure and a branch network structure of a water supply network in a pipe network topology undirected graph; in the pipe network topology undirected graph, a plurality of nodes with the degree of 1 exist, new nodes with the degree of 1 appear after the nodes are deleted, and after iteration, the remaining topological structure is a backbone skeleton network structure; recording the sequence of deleting nodes, and outputting in a reverse order to construct a tree structure;

step 3, dividing the water supply network into nc units according to the obtained main network structure and branch network structure;

step 31, regarding each node in the backbone network structure as an independent unit, wherein the number of the initial units is the same as the number of the nodes;

step 32, for each nodeiTry to get the nodeiDistributing the data to a unit where the adjacent node is located, calculating a modularity difference value delta Q between the data before distribution and the data after distribution, and recording the adjacent node with the largest modularity difference value delta Q; if maximum Δ Q>0, then nodeiDistributing the unit where the adjacent node with the maximum delta Q is located, and otherwise, abandoning the division; the calculation formula of the modularity difference value is as follows:

wherein, the Delta Q represents the modularity difference,TopologySimindicating the topological similarity of two adjacent nodes,i,nis the number of the node to which the node is connected,iis shown asiThe number of the nodes is one,nis composed ofiThe adjacent node of (a) is,

is the sum of the weights within the cell C,

the internal weight of the cell C is represented,

is the sum of the weights of the external connections of the connection unit C,

represents the external weight of the connection unit C, C represents the adjacent nodenThe unit in which the device is located,

is a nodeiSum of connection weights with internal nodes，

Is a connecting nodeiThe sum of the weights of the external connections of (2),

is the sum of the weights of the entire network;

step 33, repeating step 32 until all the belonged units of the nodes are not changed;

step 34, regarding the node with the same attribution obtained in the step 33 as a new node, and reconstructing a subgraph, wherein the weight between two new nodes is the sum of the weights of all edges between two corresponding units;

step 35, giving a minimum community sub-group resolution parameter, taking the sub-image obtained in the step 34 as an input, and re-executing the steps 31 to 34 until the given minimum community sub-group resolution parameter is reached to obtain a minimum community sub-group and a community sub-group;

step 36, outputting the minimum community sub-group and the connection relation of the community sub-groups, and taking the minimum community sub-group as a minimum partition unit of the pipe network;

step 4, according to the number of the specified partitions, the lengths are equivalent and used as main convergence conditions, the areas are used as secondary convergence conditions, and the minimum partition unit aggregation of the pipe network is realized to reach the number of the specified partitions; possibility to treat new node as community sub-group, nodeiThe probability of being selected as a community subgroup is calculated by the following specific formula:

wherein,

representing nodesiThe probability of being selected as a community sub-group,D（i) Is defined as a nodeiThe distance to the centroid of the neighbor's subgroup euclidean space,

representing nodesiAll adjacent nodes ofnTo the nodeiAlso from the area allocation point of view, defines the convergence condition on the spatial area:

wherein,

the minimum value of the area of each partition is represented,

the number of the partitions is represented by,

are the nodes of the candidate nodes, and are the nodes,

is the first

The number of the sub-areas is equal to that of the sub-areas,m _i selecting the centroids of adjacent nodes according to different convergence numbers

Taking the numerical value at the extreme value point as an optimal sub-cluster distribution mode;

step 41, giving the weight of the branched network structure to the node of the main network structure mounted by the branched network structure, and giving the length of the pipe section of the branched network structure to the unit to which the node belongs according to the unit to which the node belongs;

step 42, selecting the node with the minimum weight in the community sub-group, combining the node with the minimum weight in the adjacent nodes as a new node, wherein the new node weight is obtained by adding the sum of the edge weights between the original node and the original node; when the two nodes with the minimum weight do not have direct topological correlation, the area of the outsourcing rectangle when the two nodes are added into the neighbor is respectively calculated, and a combination scheme with large area change is selected;

43, repeating the step 42 until all the nodes are combined to obtain each partition under different partition schemes;

step 44, respectively calculating the variance of the length of each partition under different partition schemes, selecting the partition scheme with the minimum variance, and using the area of the outsourcing rectangle as a judgment condition when the length variances are equal;

and 5, outputting pipe network partition results, wherein the pipe network partition results comprise vector data, pipe network length to be generally searched in the general search partition and partition area information.

2. The zoning method facing underground pipe network census in municipal management according to claim 1, wherein: the method for constructing the topological undirected graph of the pipe network in the step 1 comprises the following steps:

step 11, the pipe network vector data comprises pipe section data and pipe section data; checking the pipe point data and the pipe section data to ensure that each pipe section is a connecting line of two pipe points and no pipe point exists in the pipe section; correcting the pipe sections and pipe points which do not meet the requirements, deleting the pipe sections lacking the end point pipe points, and breaking the pipe sections with other pipe points inside; on the basis of finishing the inspection, abstracting a pipe point into a node, abstracting a pipe section into an edge connecting the node, and constructing a topological graph of the pipe network by taking the length of the pipe section as the weight of the edge;

step 12, analyzing each node in turn, finding the adjacent node sharing the same edge with the node, constructing the adjacent table of the node, if the node is

And node

If connected, add a record in the adjacency list

，

Is the number of the node to which the node is connected,

、

respectively represent

A node;

and step 13, deleting repeated node adjacency information according to the node number size rule, and forming a pipe network topology undirected graph G according to the adjacency list.

3. The zoning method facing to underground pipe network general investigation in municipal management according to claim 2, characterized in that: the method for extracting the main network structure and the branch network structure of the water supply network in the pipe network topology undirected graph in the step 2 comprises the following steps:

step 21, traversing each node in the topological undirected graph of the pipe network, if the node has no adjacent node, recording the node as a discrete value, and deleting the node;

step 22, traversing each edge in the topological undirected graph of the pipe network

，

，

Is the number of the edge or edges,

is shown as

The number of the edges is one,

is the number of the node to which the node is connected,

is shown as

A node, if node

Has more adjacent nodes than nodes

Number of adjacent nodes of, and node

Only sum node

Adjacent to each other, then

Is composed of

The parent node updates the parent-child relationship table and deletes the edge

And node

；

Step 23, repeating the step 21 and the step 22 until no node can be deleted, wherein the rest structure in the topological undirected graph of the pipe network is a main skeleton network structure R of the pipe network, and the structure is the core of the whole pipe network and mainly comprises a main pipe network and a part of branch pipe networks related to key topological positions;

step 24, according to the parent-child relationship table, for any node in the parent-child relationship tableV _g1 Find its predecessor nodeV _j1 If, ifV _j1 If the data is still in the parent-child relationship table, the search is continuedV _j1 Is a precursor nodeV _k1 And the analogy is repeated until the precursor node does not exist or is a point in the backbone framework network structure, a tree structure T is constructed according to the rule, the tree structure T is a branch network structure,g1、j1、k1is a node number, each representingg1、j1、k1And (4) each node.

4. The zoning method facing underground pipe network census in municipal management according to claim 3, wherein: the method for performing the aggregation of the minimum partition units as required in the step 4 comprises the following steps:

step 41, giving the weight of the branched network structure to the node of the main network structure mounted by the branched network structure, and giving the length of the pipe section of the branched network structure to the unit to which the node belongs according to the unit to which the node belongs;

step 42, selecting the node with the minimum weight in the graph, combining the node with the minimum weight in the adjacent nodes as a new node, wherein the weight of the new node is obtained by adding the sum of the edge weights between the original node and the original node;

43, repeating the step 42 until all the nodes are combined to obtain each partition under different partition schemes;

and 44, respectively calculating the variance of the length of each partition under different partition schemes, and selecting the partition scheme with the minimum variance.

5. The zoning method facing underground pipe network census in municipal management according to claim 4, wherein: the method for outputting the pipe network partition result in the step 5 comprises the following steps:

step 51, assigning the partition number to the minimum partition unit and all nodes in the unit to obtain a partition number table of the management point;

step 52, associating the partition numbering table with the vector data according to the identification field of the control point;

step 53, counting the lengths of the pipe sections of different partitions, and recording the length of the pipe section into the length of the partition if two pipe points of one pipe section are in the same partition; otherwise, neglecting the length of the pipe section;

and step 54, outputting the vector data related to the partition table and the statistical result of the length and the occupied area of each partition.

Images