Progressive topological geometry matching method for three-dimensional pipe isomodel

Abstract

The progressive topological geometry matching method of three-dimensional pipeline heterogeneous model relates to the technical fields of industrial data processing and three-dimensional model mapping. The method solves the problem that the existing pipeline design model and stress analysis model cannot be automatically and efficiently mapped due to data structure heterogeneity, topological expression difference and geometric shape deviation. The method is aimed at two types of heterogeneous models of three-dimensional design pipeline model and stress analysis pipeline model without preset mapping relationship for the same pipeline system. A directed topological graph is first constructed. Then, threshold filtering is performed to remove secondary elements and simplify the topology. Subsequently, path compression is carried out to encapsulate the continuous two-degree node chain into a super-edge carrying the original node sequence. The cross-model super-edge group topological matching is completed by relying on the node color iterative refinement algorithm, and the two-level precise matching of super-edge group and super-edge is realized layer by layer. The method can realize the automatic consistency checking of two sets of heterogeneous models of chemical and power pipeline networks.

Description

Technical Field

[0001] This invention relates to the field of industrial data processing and 3D model mapping technology. Background Technology

[0002] In the design and operation phases of large industrial facilities (such as thermal power plants, chemical plants, and nuclear power plants), there are typically two types of core data models:

[0003] 3D design model: Built with software such as PDMS (Plant Design Management System), it contains the complete geometric topology of the piping system, equipment layout and design parameters, and is used for visualization and construction guidance.

[0004] Stress analysis model: Constructed using stress analysis software such as CAESAR II, it includes the mechanical parameters, load conditions and constraint information of the pipeline elements, and is used for pipeline strength verification and safety assessment.

[0005] Due to significant differences in the construction goals, data formats, and modeling granularity of the two types of models, the following technical problems are encountered in practical engineering:

[0006] (1) Data heterogeneity problem: The design model and the analysis model use different data structures and naming rules. The same pipeline system is expressed in different ways in the two models, and it is impossible to directly establish a correspondence.

[0007] (2) Topology difference problem: The design model contains complete pipe components (elbows, tees, valves, reducers, etc.), while the analysis model usually simplifies or segments the pipe according to the needs of mechanical calculation, resulting in the topology of the two models not being completely matched.

[0008] (3) Geometric shape deviation problem: Even pipes with similar topological structures may have different vector representations of their geometric paths due to differences in modeling accuracy, coordinate reference or simplification strategy.

[0009] (4) Mapping efficiency problem: Traditional manual comparison methods are time-consuming and laborious, and it is difficult to guarantee comprehensiveness and accuracy; while precise algorithms such as global graph isomorphism (such as VF2 algorithm) have NP-hard complexity and exponential explosion risk in industrial-grade large-scale pipeline networks (the number of nodes can reach tens of thousands), which cannot meet the requirements of engineering practicality.

[0010] There is a lack of existing technologies that can automatically, efficiently, and robustly establish the mapping relationship between design models and analysis models. Summary of the Invention

[0011] The purpose of this invention is to address the problem of the inability to automatically and efficiently map existing pipeline design models and stress analysis models due to heterogeneous data structures, differences in topological representation, and geometric deviations. This invention provides a progressive topological geometry matching method for heterogeneous three-dimensional pipeline models.

[0012] A progressive topological geometry matching method for 3D pipeline heterogeneous models includes:

[0013] Acquire first pipeline model data and second pipeline model data, and construct corresponding first and second directed topology graphs respectively. Nodes in the topology graphs represent pipeline elements, and edges represent the connection relationship and relative displacement vector between pipeline elements. The first pipeline model and the second pipeline model are respectively the three-dimensional design pipeline model and stress analysis pipeline model of the same target pipeline system, and the two do not have a preset mapping relationship.

[0014] Based on the first pipeline model data and the second pipeline model data, corresponding first directed topology graphs and second directed topology graphs are constructed respectively; the nodes in the directed topology graphs represent pipeline components, and the edges represent the connection relationships between components and the relative displacement vectors in the connection directions.

[0015] Threshold filtering is performed on the first directed topology graph and the second directed topology graph respectively to remove nodes or edges in their respective models that do not meet the preset importance conditions, thereby obtaining the first smooth directed topology graph and the second smooth directed topology graph.

[0016] Path compression is performed on the first smoothed directed topology graph and the second smoothed directed topology graph respectively, compressing the continuous node chains with a degree of 2 in each graph into hyperedges, thereby generating the first simplified graph and the second simplified graph respectively; each hyperedge stores the original node sequence it passes through;

[0017] In each simplified graph, all the hyperedges between two adjacent nodes that have hyperedges constitute a hyperedge group;

[0018] Hyperedge pair matching: Based on the node color iterative refinement algorithm, the nodes in the first and second simplified graphs are classified by topological color. The mapping relationship between the nodes in the first and second simplified graphs is determined according to the node color. Based on the positional relationship between each node and its adjacent hyperedge pair in the mapping relationship, the mapping relationship between the hyperedge pairs in the first and second simplified graphs is determined. Hyperedge pairs in the first and second simplified graphs that have a mapping relationship form hyperedge pairs, and the number of hyperedges in each hyperedge pair is the same.

[0019] Hyperedge pair matching: For hyperedge pairs with a known mapping relationship, extract the displacement vectors between adjacent nodes in the original node sequence stored in each hyperedge to form a path vector group; for hyperedge pairs with a known mapping relationship, calculate the geometric similarity between the path vector group corresponding to each hyperedge in the first simplified graph and the path vector group corresponding to each hyperedge in the second simplified graph, and establish a one-to-one mapping relationship between the hyperedge pairs with the highest geometric similarity scores, thereby obtaining the correspondence between nodes and hyperedges in the first and second simplified graphs.

[0020] Preferably, in the threshold filtering, the preset importance condition includes: comprehensively judging the pipe diameter, temperature and pressure of the pipe element to obtain a comprehensive judgment result, and removing the pipe element whose comprehensive judgment result is lower than the preset threshold from the corresponding directed topology graph.

[0021] Preferably, the method for obtaining the comprehensive judgment result is as follows:

[0022] ;

[0023] in: It is a comprehensive judgment result;

[0024] It refers to the temperature of the pipe components;

[0025] It refers to the pressure of the pipeline components;

[0026] It refers to the pipe diameter of the piping component;

[0027] and These are the highest and lowest temperatures in the piping system, respectively.

[0028] and These are the highest and lowest pressures in the pipeline system, respectively.

[0029] and These are the maximum and minimum pipe diameters in the piping system, respectively.

[0030] These are the normalization coefficients for temperature, pressure, and pipe diameter, respectively.

[0031] Preferably, the path compression specifically involves:

[0032] Nodes with a degree of 1 and those with a degree greater than or equal to 3 are taken as boundary nodes. All continuous node chains with a degree of 2 between any two boundary nodes are extracted. Each continuous node chain is compressed into a superedge, and the names of all the original nodes traversed by each continuous node chain are stored in the attributes of the superedge in order.

[0033] Preferably, the pipe component types in the first pipe model data and the second pipe model data include, but are not limited to, pipe segment ends, elbows, tees, valves, and reducers.

[0034] Preferably, for hyperedge pairs with a known mapping relationship, the method for calculating the geometric similarity between the path vector groups corresponding to each hyperedge in the first simplified graph and the path vector groups corresponding to each hyperedge in the second simplified graph is as follows:

[0035] Perform a tolerance-based longest common subsequence alignment on the path vector groups of two hyperedges to be matched, dividing the path vector groups into common segments and difference segments;

[0036] Calculate the similarity of the common segment and the similarity of the difference segment separately, and then weight and combine the two to obtain the geometric similarity score of the hyperedge pair.

[0037] Preferably, in the longest common subsequence alignment with tolerance, the condition for determining that two vectors are equal is: the difference in magnitude between the two vectors is less than a preset magnitude tolerance threshold, and the angle between the directions of the two vectors is less than a preset angle tolerance threshold.

[0038] Preferably, the method for calculating the similarity of common segments is as follows: calculate the similarity between displacement vectors at the same position in two common segments for each displacement vector, and combine the similarities of each vector to obtain the similarity of the common segments;

[0039] The similarity of the difference segments is calculated as follows: each displacement vector in one difference segment is compared with the displacement vector in the same position in another difference segment, and the magnitude difference and direction angle are calculated respectively; the difference penalty value of each displacement vector is calculated by weighted summation based on the magnitude difference and direction angle, and then accumulated after being mapped by the exponential decay function, and then normalized by dividing by the total number of vectors in the difference segment to obtain the difference segment similarity.

[0040] The beneficial effects of this invention are:

[0041] (1) The algorithm complexity is controllable: by path compression, the global large-scale graph matching is reduced to finite simple chain matching, avoiding the NP-hard exponential explosion problem of global subgraph isomorphism, and has linear engineering practicality.

[0042] (2) High robustness of matching: First coarse matching, then fine matching. Specifically, first perform super edge pair matching, then perform super edge pair matching. In the process of super edge pair matching, the longest common subsequence alignment with tolerance is performed. The alignment has good adaptability to small offsets of geometric paths and segment granularity differences.

[0043] (3) Balance between accuracy and efficiency is controllable: the decoupling of topological screening and geometric confirmation is realized, so that the computing resources are concentrated on high-confidence hyperedge pair matching and hyperedge pair matching, maximizing computing efficiency while ensuring matching accuracy.

[0044] This invention reduces NP-hard global graph matching to linear complexity hyperedge matching, balancing matching robustness and computational efficiency, and can realize automated consistency verification of two heterogeneous models, namely chemical engineering and power pipeline network. Attached Figure Description

[0045] Figure 1 This is a flowchart of the progressive topological geometry matching method for the three-dimensional pipeline heterogeneous model described in this invention; Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0048] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0049] Specific Implementation Method 1: Combination Figure 1 This embodiment describes a progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model, which includes:

[0050] S1. Obtain the first pipeline model data and the second pipeline model data, and construct the corresponding first directed topology graph and the second directed topology graph respectively. The nodes in the topology graph represent pipeline elements, and the edges represent the connection relationship and relative displacement vector between pipeline elements. The first pipeline model and the second pipeline model are respectively the three-dimensional design pipeline model and the stress analysis pipeline model of the same target pipeline system, and the two do not have a preset mapping relationship.

[0051] When applied, the 3D design pipeline model data includes: PDMS macro format output file, line-by-line parsing instructions (NEW, END, HPOS, TPOS, POS, ORI, SPRE, LSTU, etc.), constructing a hierarchical tree structure (PIPE → BRANCH → specific component hierarchy), nodes encapsulating component type, coordinates, nominal diameter, and direction matrix, and edges encapsulating the relative displacement vector between adjacent components.

[0052] The extraction of stress analysis pipe model data supports two data formats:

[0053] For CAESAR II neutral text format, stress analysis pipe model data includes: reading node coordinates line by line, element connection relationships, and load condition parameters.

[0054] For Access database format (.mdb), the stress analysis piping model data includes: connecting to the database and reading node coordinates, element connection relationships, rigid element information, stress enhancement coefficients, and other data from tables such as INPUT_BASIC_ELEMENT_DATA, INPUT_REDUCERS, INPUT_RIGIDS, and INPUT_SIFTEES.

[0055] S2. Construct corresponding first and second directed topological graphs based on the first and second pipeline model data, respectively. Nodes in the directed topological graphs represent pipeline components, and edges represent the connection relationships between components and the relative displacement vectors in the connection direction. S3. Perform threshold filtering on the first and second directed topological graphs respectively to remove nodes or edges that do not meet preset importance conditions, resulting in a first smoothed directed topological graph and a second smoothed directed topological graph. S4. Perform path compression on the first and second smoothed directed topological graphs respectively, compressing continuous node chains with a degree of 2 into hyperedges, generating a first simplified graph and a second simplified graph. Each hyperedge stores the original node sequence it traverses.

[0056] In each simplified graph, all the hyperedges between two adjacent nodes that have hyperedges constitute a hyperedge group;

[0057] S5. Hyperedge Pair Matching: Based on the node color iterative refinement algorithm, the nodes in the first and second simplified graphs are classified by topological color. The mapping relationship between the nodes in the first and second simplified graphs is determined according to the node colors. Based on the positional relationship between each node and its adjacent hyperedge pairs in the mapping relationship, the mapping relationship between the hyperedge pairs in the first and second simplified graphs is determined. Hyperedge pairs in the first and second simplified graphs that have a mapping relationship are formed, and the number of hyperedges in each hyperedge pair is the same.

[0058] S6. Hyperedge Pair Matching: For hyperedge pairs with a determined mapping relationship, extract the displacement vectors between adjacent nodes in the original node sequence stored in each hyperedge to form a path vector group; for hyperedge pairs with a determined mapping relationship, calculate the geometric similarity between the path vector group corresponding to each hyperedge in the first simplified graph and the path vector group corresponding to each hyperedge in the second simplified graph, and establish a one-to-one mapping relationship between the hyperedge pairs with the highest geometric similarity scores, thereby obtaining the correspondence between nodes and hyperedges in the first and second simplified graphs.

[0059] This invention uses "hyperedge groups" as topological matching units and "hyperedges" as geometric fine matching units. Based on global dimensionality reduction, it further groups the data, raising the matching granularity from "single hyperedge" to the "hyperedge group" level. This avoids mismatches between multiple parallel paths within the same topological branch. The introduction of the "hyperedge group" concept is one of the core technical contributions of this invention, solving the technical problem in the background art where multiple parallel paths at T-junctions / forks are difficult to accurately correspond one-to-one.

[0060] This invention reflects the complete mapping relationship between two heterogeneous pipeline models through the mapping relationship between the first and second simplified diagrams, providing a directly usable mapping benchmark for downstream engineering applications (stress back-substitution, change assessment, consistency verification). Furthermore, the node color iterative refinement algorithm used is existing technology.

[0061] In practical applications, the pipe component types in the first and second pipe model data include, but are not limited to, pipe segment endpoints, elbows, tees, valves, and reducers.

[0062] Furthermore, in the threshold filtering, the preset importance condition includes: comprehensively judging the pipe diameter, temperature and pressure of the pipe element to obtain a comprehensive judgment result, and removing the pipe element whose comprehensive judgment result is lower than the preset threshold from the corresponding directed topology graph.

[0063] Specifically, the method for obtaining the comprehensive judgment result is as follows:

[0064] ;

[0065] in: It is a comprehensive judgment result;

[0066] It refers to the temperature of the pipe components;

[0067] It refers to the pressure of the pipeline components;

[0068] It refers to the pipe diameter of the piping component;

[0069] and These are the highest and lowest temperatures in the piping system, respectively.

[0070] and These are the highest and lowest pressures in the pipeline system, respectively.

[0071] and These are the maximum and minimum pipe diameters in the piping system, respectively.

[0072] These are the normalization coefficients for temperature, pressure, and pipe diameter, respectively.

[0073] In this preferred approach, the "preset importance condition" is specified as a comprehensive evaluation of pipe diameter / temperature and pressure, which makes threshold filtering clearly feasible in engineering, ensures the consistency of the two models in terms of precision, and provides clean input data for subsequent path compression.

[0074] Furthermore, the path compression specifically involves:

[0075] Nodes with a degree of 1 and those with a degree greater than or equal to 3 are taken as boundary nodes. All continuous node chains with a degree of 2 between any two boundary nodes are extracted. Each continuous node chain is compressed into a superedge, and the names of all the original nodes traversed by each continuous node chain are stored in the attributes of the superedge in order.

[0076] This preferred method provides a clear definition of boundary nodes (degree ≠ 2), which clearly defines the objects and methods of path compression, ensuring that the compression process is complete and without omissions or repetitions; the original node names are stored in an orderly manner as the data source for the geometric matching stage.

[0077] Furthermore, for hyperedge pairs with established mapping relationships, the method for calculating the geometric similarity between the path vector groups corresponding to each hyperedge in the first simplified graph and the path vector groups corresponding to each hyperedge in the second simplified graph is as follows:

[0078] First, the path vector groups of the two hyperedges to be matched are aligned using the longest common subsequence with tolerance, dividing the path vector groups into common segments and difference segments. In the alignment of the longest common subsequence with tolerance, the conditions for determining that two vectors are equal are: the difference in magnitude between the two vectors is less than a preset magnitude tolerance threshold, and the angle between the directions of the two vectors is less than a preset angle tolerance threshold. This transforms the "tolerance" from a qualitative description into a specific dual judgment condition (magnitude + direction), ensuring the feasibility of LCS alignment in engineering. The "AND" relationship ensures that both conditions must be met simultaneously, maintaining matching accuracy.

[0079] Secondly, the similarity of the common segment and the similarity of the differing segment are calculated separately, and the two are weighted and combined to obtain the geometric similarity score of the hyperedge pair. Specifically:

[0080] The method for calculating the similarity of common segments is as follows: calculate the similarity between displacement vectors at the same position in two common segments for each displacement vector, and combine the similarity of each vector to obtain the similarity of the common segment.

[0081] The similarity of the difference segments is calculated as follows: each displacement vector in one difference segment is compared with the displacement vector in the same position in another difference segment, and the magnitude difference and direction angle are calculated respectively; the difference penalty value of each displacement vector is calculated by weighted summation based on the magnitude difference and direction angle, and then accumulated after being mapped by the exponential decay function, and then normalized by dividing by the total number of vectors in the difference segment to obtain the difference segment similarity.

[0082] The geometric similarity calculation is refined from "overall comparison" to "quantification of common segments and difference segments separately". The common segment represents the positive similarity contribution of the aligned part, and the difference segment represents the similarity contribution of the unaligned part through the penalty term, so that the score can better reflect the actual geometric matching degree between hyperedges.

[0083] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A progressive topological geometry matching method for three-dimensional heterogeneous pipeline models, characterized in that, include: Acquire first pipeline model data and second pipeline model data, and construct corresponding first and second directed topology graphs respectively. Nodes in the topology graphs represent pipeline elements, and edges represent the connection relationship and relative displacement vector between pipeline elements. The first pipeline model and the second pipeline model are respectively the three-dimensional design pipeline model and stress analysis pipeline model of the same target pipeline system, and the two do not have a preset mapping relationship. Based on the first pipeline model data and the second pipeline model data, corresponding first directed topology graphs and second directed topology graphs are constructed respectively; the nodes in the directed topology graphs represent pipeline components, and the edges represent the connection relationships between components and the relative displacement vectors in the connection directions. Threshold filtering is performed on the first directed topology graph and the second directed topology graph respectively to remove nodes or edges in their respective models that do not meet the preset importance conditions, thereby obtaining the first smooth directed topology graph and the second smooth directed topology graph. Path compression is performed on the first smoothed directed topology graph and the second smoothed directed topology graph respectively, compressing the continuous node chains with a degree of 2 in each graph into hyperedges, thereby generating the first simplified graph and the second simplified graph respectively; each hyperedge stores the original node sequence it passes through; In each simplified graph, all the hyperedges between two adjacent nodes that have hyperedges constitute a hyperedge group; Hyperedge pair matching: Based on the node color iterative refinement algorithm, the nodes in the first and second simplified graphs are classified by topological color. The mapping relationship between the nodes in the first and second simplified graphs is determined according to the node color. Based on the positional relationship between each node and its adjacent hyperedge pair in the mapping relationship, the mapping relationship between the hyperedge pairs in the first and second simplified graphs is determined. Hyperedge pairs in the first and second simplified graphs that have a mapping relationship form hyperedge pairs, and the number of hyperedges in each hyperedge pair is the same. Hyperedge pair matching: For hyperedge pairs with a known mapping relationship, extract the displacement vectors between adjacent nodes in the original node sequence stored in each hyperedge to form a path vector group; for hyperedge pairs with a known mapping relationship, calculate the geometric similarity between the path vector group corresponding to each hyperedge in the first simplified graph and the path vector group corresponding to each hyperedge in the second simplified graph, and establish a one-to-one mapping relationship between the hyperedge pairs with the highest geometric similarity scores, thereby obtaining the correspondence between nodes and hyperedges in the first and second simplified graphs.

2. The progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model according to claim 1, characterized in that, In the threshold filtering, the preset importance condition includes: comprehensively judging the pipe diameter, temperature and pressure of the pipe element to obtain a comprehensive judgment result, and removing the pipe element whose comprehensive judgment result is lower than the preset threshold from the corresponding directed topology graph.

3. The progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model according to claim 2, characterized in that, The method for obtaining the comprehensive judgment result is as follows: ； in: It is a comprehensive judgment result; It refers to the temperature of the pipe components; It refers to the pressure of the pipeline components; It refers to the pipe diameter of the piping component; and These are the highest and lowest temperatures in the piping system, respectively. and These are the highest and lowest pressures in the pipeline system, respectively. and These are the maximum and minimum pipe diameters in the piping system, respectively. These are the normalization coefficients for temperature, pressure, and pipe diameter, respectively.

4. The progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model according to claim 1, characterized in that, The path compression specifically refers to: Nodes with a degree of 1 and those with a degree greater than or equal to 3 are taken as boundary nodes. All continuous node chains with a degree of 2 between any two boundary nodes are extracted. Each continuous node chain is compressed into a superedge, and the names of all the original nodes traversed by each continuous node chain are stored in the attributes of the superedge in order.

5. The progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model according to claim 1, characterized in that, The pipe component types in the first and second pipe model data include, but are not limited to, pipe end points, elbows, tees, valves, and reducers.

6. The progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model according to claim 1, characterized in that, For hyperedge pairs with established mapping relationships, the method for calculating the geometric similarity between the path vector groups corresponding to each hyperedge in the first simplified graph and the path vector groups corresponding to each hyperedge in the second simplified graph is as follows: Perform a tolerance-based longest common subsequence alignment on the path vector groups of two hyperedges to be matched, dividing the path vector groups into common segments and difference segments; Calculate the similarity of the common segment and the similarity of the difference segment separately, and then weight and combine the two to obtain the geometric similarity score of the hyperedge pair.

7. The progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model according to claim 6, characterized in that, In the longest common subsequence alignment with tolerance, the condition for determining that two vectors are equal is: the difference in magnitude between the two vectors is less than a preset magnitude tolerance threshold, and the angle between the directions of the two vectors is less than a preset angle tolerance threshold.

8. The progressive topological geometry matching method for a three-dimensional heterogeneous pipeline model according to claim 6, characterized in that, The method for calculating the similarity of common segments is as follows: calculate the similarity between displacement vectors at the same position in two common segments for each displacement vector, and combine the similarity of each vector to obtain the similarity of the common segment. The similarity of the difference segments is calculated as follows: each displacement vector in one difference segment is compared with the displacement vector in the same position in another difference segment, and the magnitude difference and direction angle are calculated respectively; the difference penalty value of each displacement vector is calculated by weighted summation based on the magnitude difference and direction angle, and then accumulated after being mapped by the exponential decay function, and then normalized by dividing by the total number of vectors in the difference segment to obtain the difference segment similarity.

Images