A method and system for checking project payment based on BIM model quantity calculation results

Abstract

The present invention provides a method and system for checking the payment of engineering funds based on the result of BIM model calculation, which relates to the technical field of engineering management, including: obtaining first information, second information and third information, wherein the first information is the BIM model data of the construction project, the second information is the manual calculation, progress and quality data input by the user on the cloud platform, and the third information includes the engineering payment application of the construction unit; performing topological analysis and component segmentation processing according to the first information to obtain the engineering quantity list; performing data fusion processing according to the second information to obtain the engineering data set; performing association analysis processing according to the engineering quantity list, the engineering data set and the third information to obtain feature data; judging the feature data based on a preset CART decision tree model, and obtaining the payment review result by evaluating the rationality of the payment application. The present invention realizes the automated review of the engineering payment process by automatically judging the rationality of the payment application.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering management, and in particular to a method and system for checking engineering payment based on BIM model quantity calculation results. Background Art

[0002] At present, with the continuous expansion of the scale of the engineering construction industry, the complexity of engineering projects and the need for refined management are also gradually increasing. As an important part of engineering project management, the management of engineering payment directly affects the overall progress and quality control of the project. However, in practice, the management of engineering payment faces many challenges, especially in the calculation of engineering quantities, payment approval, and the correlation analysis between progress and quality. Traditional engineering management methods mostly rely on manual calculations and manual audits, which not only lead to a large number of human errors, but also seriously reduce the efficiency of data processing. Especially in large-scale projects, the accurate calculation of engineering quantities and the approval process of payment applications often require repeated audits and repeated proofreading of data. The whole process is cumbersome and error-prone, which greatly increases the complexity and cost of management.

[0003] Existing technologies attempt to partially alleviate these problems by using spreadsheet tools or semi-automated report generation systems, but these methods are incapable of processing complex BIM model data, correlation analysis between progress and quality, and judgment of payment rationality. They lack the ability to comprehensively analyze quality, progress, and cost data during the payment process, resulting in insufficient transparency in payment approval and low response speed, which ultimately affects the overall progress and financial planning of the project.

[0004] In response to the above problems, the present invention proposes a method and system for checking engineering payment based on BIM model quantity calculation results based on sparse cross-sectional data and multi-source data fusion, so as to solve the efficiency and accuracy problems in the current engineering payment process. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for checking the payment of construction funds based on the calculation results of the BIM model to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0006] In the first aspect, the present application provides a method for checking the payment of engineering funds based on the quantity calculation results of a BIM model, comprising:

[0007] Obtaining first information, second information and third information, wherein the first information is the BIM model data of the construction project, the second information is the manual quantity calculation, progress and quality data input by the user on the cloud platform, and the third information includes the construction unit's application for payment of the project funds;

[0008] Performing topological analysis and component segmentation processing according to the first information, obtaining a bill of quantities by identifying the spatial relationship between components and extracting engineering quantity information related to the characteristics of each construction stage, wherein the bill of quantities includes the name, quantity, unit and construction stage of the component;

[0009] Performing data fusion processing according to the second information, matching the quality data with the construction progress, and embedding the quality data into the progress and cost analysis of each component to obtain an engineering data set;

[0010] Performing correlation analysis processing according to the bill of quantities, the engineering data set and the third information, and obtaining characteristic data by analyzing the correlation relationship between the actual state of the component and the payment conditions;

[0011] The characteristic data is judged based on the preset CART decision tree model, and the payment review result is obtained by evaluating the rationality of the payment application.

[0012] Secondly, this application also provides a project payment inspection system based on BIM model quantity calculation results, including:

[0013] An acquisition module is used to acquire first information, second information and third information, wherein the first information is BIM model data of a construction project, the second information is manual quantity calculation, progress and quality data input by a user on a cloud platform, and the third information includes a payment application for a construction unit's engineering payment;

[0014] a segmentation module, configured to perform topological analysis and component segmentation processing according to the first information, obtain a bill of quantities by identifying the spatial relationship between components and extracting engineering quantity information related to the characteristics of each construction stage, wherein the bill of quantities includes the name, quantity, unit and construction stage of the component;

[0015] A fusion module, used for performing data fusion processing according to the second information, matching the quality data with the construction progress, and embedding the quality data into the progress and cost analysis of each component to obtain an engineering data set;

[0016] An analysis module, used for performing association analysis processing according to the bill of quantities, the engineering data set and the third information, and obtaining characteristic data by analyzing the association relationship between the actual state of the component and the payment conditions;

[0017] The judgment module judges the characteristic data based on a preset CART decision tree model, and obtains the payment review result by evaluating the rationality of the payment application.

[0018] The beneficial effects of the present invention are:

[0019] The present invention automatically judges the rationality of payment applications by adopting algorithms such as topological analysis, data fusion, feature extraction and CART decision tree, realizes automated review of the project payment process, reduces the complexity of manual operations, reduces the incidence of human errors, and thus improves the accuracy and reliability of project payment review.

[0020] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 A flow chart of a method for checking the payment of a project based on the BIM model quantity calculation results described in an embodiment of the present invention;

[0023] Figure 2 It is a schematic diagram of the overall architecture of the construction payment inspection system based on the BIM model quantity calculation results described in an embodiment of the present invention.

[0024] Markings in the figure: 1. BIM model establishment module; 2. BIM model extraction module; 3. Manual quantity calculation entry module; 4. Progress data entry module; 5. Quality data entry module; 6. Construction unit submission of project payment application module; 7. CART decision tree classification module; 8. Automatic approval module. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0026] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance. Embodiment 1:

[0027] This embodiment provides a method for checking project payment based on BIM model quantity calculation results.

[0028] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400 and step S500.

[0029] Step S100, obtaining first information, second information and third information, wherein the first information is the BIM model data of the construction project, the second information is the manual quantity calculation, progress and quality data input by the user on the cloud platform, and the third information includes the construction unit's application for payment of the project funds;

[0030] It is understandable that the first information, namely the BIM model data of the construction project, is an important way to express the digital construction project. Through the BIM model, detailed information about the geometric shape, physical properties, material type and spatial location of the components can be obtained. These data are not only used to describe the physical properties of the building, but also reflect the spatial relationship between the components and their properties at different construction stages. The use of the BIM model has greatly improved the accuracy of the engineering quantity data and reduced the errors that may occur in traditional manual measurement and recording. The second information is the manual quantity calculation, construction progress and quality data entered by the user on the cloud platform. This part of the data records and feedbacks the dynamic changes in the construction process. Through the cloud platform, users can update the manual quantity calculation results, construction progress and quality inspection in the project in real time. These data, combined with the BIM model data, can more accurately describe the current construction status and actual execution. The manually entered progress and quality data can supplement the lack of static information in the BIM model, making the entire information system dynamic and adaptable to the site. The use of the cloud platform ensures the real-time and shareability of the data, so that all stakeholders in the project can make decisions based on the latest data. The third information is the payment application for the project submitted by the construction unit, which includes the payment request for each construction node or stage. The construction unit submits a payment application based on the completed construction work and relevant agreement provisions. The application data provides a basis for subsequent payment inspection and review.

[0031] Step S200: Perform topological analysis and component segmentation processing according to the first information, identify the spatial relationship between components and extract engineering quantity information related to the characteristics of each construction stage, and obtain a bill of quantities, which includes the name, quantity, unit and construction stage of the component;

[0032] It should be noted that this step can not only effectively structure the model data through automated topological analysis and component segmentation, but also deeply integrate with the construction phase to form a dynamic and traceable bill of quantities. Compared with traditional manual measurement and engineering quantity statistics methods, this automated processing based on the BIM model significantly improves the accuracy and consistency of the data, while reducing errors in manual operations.

[0033] Step S300: performing data fusion processing according to the second information, matching the quality data with the construction progress, and embedding the quality data into the progress and cost analysis of each component to obtain an engineering data set;

[0034] It can be understood that the engineering data set formed through data fusion in this step provides a high-quality comprehensive data foundation for subsequent payment inspections, and can provide complete information on each component at a specific stage, including whether the progress is proceeding as planned, whether the quality is qualified, and the impact of these factors on costs, thereby providing more accurate and transparent data support for payment audits, effectively improving payment rationality and management efficiency.

[0035] Step S400: performing correlation analysis processing according to the bill of quantities, the engineering data set and the third information, and obtaining characteristic data by analyzing the correlation relationship between the actual state of the component and the payment conditions;

[0036] It can be understood that this step quantitatively analyzes the relationship between the actual status of the components and the payment conditions to form feature data with in-depth information, providing a basis for judging the rationality of the payment application. The correlation analysis not only significantly reduces the complexity of manual review, but also improves the accuracy and fairness of payment review, effectively prevents overpayment or payment that does not meet the conditions, and improves the overall transparency and rationality of project payment management.

[0037] Step S500: judge the characteristic data based on the preset CART decision tree model, and obtain the payment review result by evaluating the rationality of the payment application.

[0038] It should be noted that the judgment model based on the CART decision tree can efficiently evaluate the rationality of payment applications, thereby automatically generating payment review results. This method greatly improves the efficiency and accuracy of the review process, ensures that payment review decisions have good transparency and explainability, and reduces subjective bias and human errors in manual judgment, thereby achieving more fair payment management.

[0039] Further, step S200 includes steps S210 to S230.

[0040] Step S210: Perform topological analysis processing according to the first information, establish a connection network between components by using a spatial relationship graph algorithm, and analyze the interaction between different components during construction, treating components with similar functions or interdependence as a whole, and obtaining a topological analysis result;

[0041] It should be explained that this step uses a spatial relationship graph algorithm to establish a connection network containing the relationships between components, and analyzes the interactions and interdependencies of the components in this network to identify which components can be regarded as a whole. First of all, topological analysis is a process of in-depth study of the geometric and spatial relationships of components in the BIM model. In the building model, each component (such as columns, beams, plates, etc.) has a unique spatial position and geometric characteristics, but these components do not exist independently. They realize the overall function of the building through complex spatial connections and functional dependencies. Through topological analysis processing, these implicit associations can be made explicit, which is particularly important for evaluating construction status and progress in payment audits.

[0042] Specifically, the spatial relationship graph algorithm can effectively establish a connection network between components. The algorithm regards each component as a node in the graph and constructs connection lines through geometric contact, overlap or adjacency between components, thereby forming a global graph structure. Through this graph structure, the dependencies between different components can be effectively captured, such as the connection between beams and columns, support relationships, etc. These spatial relationships can reflect the real synergy in the construction process, thus providing a scientific basis for subsequent engineering quantity assessment and payment review.

[0043] After the connection network is established, further analysis is to study the interaction of these components. By analyzing the functions and interaction characteristics of components in construction, it is possible to identify which components have similar functions or obvious dependencies during the construction process. For example, some components must be constructed synchronously to ensure the integrity of the structure, or some components share materials and construction equipment in the same stage. This information can be discovered through in-depth analysis of the graph structure. Here, a method similar to spectral clustering is used to group components, and components with similar functions or high dependence in construction are regarded as a logical whole. This method can significantly reduce the complexity in the bill of quantities, thereby simplifying the subsequent analysis and payment review process. The uniqueness of topological analysis and component grouping is that it defines the relationship between components in terms of spatial location and construction logic, rather than relying solely on geometric characteristics. Especially in construction projects, the connection between construction stages and the coordination between components are very important. Traditional methods are difficult to systematically capture these implicit interactions. Through automated topological analysis, this information can be integrated into the data structure at an early stage, so that the subsequent payment review is not only supported by quantitative engineering data, but also takes into account the logical dependencies in construction, ensuring that each payment application is in line with the actual progress of the project.

[0044] Step S220: Classify components according to the topological analysis results, classify components into different categories based on material type, function and spatial structure, and merge interrelated components in the same construction stage into logical components to obtain component group information;

[0045] Specifically, first, the results of the topological analysis are classified, which requires in-depth property analysis of each component in the building project. This process involves classifying components based on multiple dimensions, including but not limited to material type, function, and spatial structure. Through material type, it is possible to identify which components use the same building materials, such as reinforced concrete, prefabricated parts, or steel structures; based on functional analysis, it is possible to determine the function of the component in the entire building, such as load-bearing components, partition components, or decorative components; and based on the analysis of spatial structure, components in the same area or with specific geometric relationships can be classified. These multi-dimensional classifications ensure that the system's identification of components not only stays at the geometric or material level, but also understands the function of each component and its position in the overall structure.

[0046] After the components are classified, the next step is to merge the related components in the same construction stage to form a logical component group. The construction process is often carried out in stages, and each stage involves a group of components with common goals and dependencies. By merging the interrelated components in the same stage, these components are regarded as a logical whole, thereby simplifying the bill of quantities and management complexity. For example, in the foundation construction stage, there are components such as cushions, foundation beams, and foundation slabs. These components must be completed in a specific order to form a complete infrastructure, so these components can be merged into a logical component group. The formation of a logical component group is not only convenient for management, but also can be analyzed at a higher level in payment review to ensure the consistency between the quantity of work in each stage and the payment application.

[0047] Preferably, the classification and merging process technically adopts a method based on hierarchical clustering combined with logical rules. Hierarchical clustering is used to automatically identify and merge corresponding components based on the similarity between components, while logical rules are used to ensure that the specific phase requirements and logical dependencies in the construction are considered during the merging process. For example, during the clustering process, if two components are similar in material type, spatial location and function, and they belong to the same construction stage, they will be merged into a logical component group. This logical rule-based grouping method effectively avoids the merging of inappropriate components, thereby ensuring the accuracy of the bill of quantities.

[0048] Step S230, performing feature extraction processing according to the component group information, and calculating the actual engineering quantity of each component group using a preset engineering quantity calculation formula to obtain engineering quantity information;

[0049] It should be noted that in the feature extraction process, the component group information includes the geometric features, functional type, construction stage, etc. of the components. First, the geometric and physical features of these component groups are extracted to ensure that the basic data required for the calculation of engineering quantities is accurate and comprehensive. For example, for a logical component group, including a combination of several columns, beams and foundations, it is necessary to extract the geometric information such as the size, volume, area, etc. of each component, as well as the physical properties such as material density and structural function. The feature extraction of each logical component group must take into account the interactions and common parts between components to ensure that the engineering quantities are not repeated or omitted.

[0050] Next, the engineering quantities of these component groups are calculated using the preset engineering quantity calculation formula. The engineering quantity calculation formula is based on industry standards, such as the "Construction Engineering Quantity List Pricing Specification" (GB50500) and the "Urban Rail Transit Engineering Quantity Calculation Specification" (GB50861). These specifications provide clear guidance for the project code, project name, project characteristics, measurement unit and engineering quantity calculation rules of different sub-projects. For example, when the system calculates the engineering quantity of the concrete foundation component group, it needs to distinguish the project characteristics according to the location, cross-sectional form, and concrete strength grade, and calculate it by volume according to the design drawing size. In the calculation process, the material uniformity and functional synergy of all components in the logical component group are considered, so that the calculation results reflect not only the characteristic quantity of each component, but also the characteristic quantity of the entire component group. Furthermore, the engineering quantity calculation here is not only a simple addition calculation, but also needs to take into account the spatial overlap and geometric interaction relationship within the component group. Through the spatial relationship diagram and topological information, the possible geometric overlap in the component group can be automatically detected, such as crossed beams or shared support structures. This information must be deducted during the calculation, otherwise it will lead to repeated calculation of the engineering quantity, which will affect the accuracy of the payment review. Through geometric Boolean operations, it is possible to accurately identify which parts need to be repeatedly deducted and which parts belong to shared engineering quantities.

[0051] The resulting engineering quantity information includes the overall quantitative data of each logical component group, which is used for comparison and confirmation in the subsequent payment application review. This data includes quantity (such as the number of components), volume, area, weight, and other specific engineering quantity information defined by material type and function. Each engineering quantity information is automatically calculated by the system and generated through multiple verifications to ensure its reliability and accuracy.

[0052] Step S240: According to the engineering quantity information, the name, quantity, unit and construction stage of the component group are structured to obtain a bill of quantities.

[0053] First, the quantity information includes the name, quantity, unit, volume, area, weight, and construction stage of each logical component group. This information is relatively complex and contains multiple data types. In this step, the system needs to organize this quantity information in an orderly manner to facilitate the subsequent payment review and use by stakeholders. Through structured processing, the system converts the quantity information of each logical component group into a standard bill of quantities, in which each entry contains the name of the component group (such as "foundation beam group"), quantity (such as "5"), unit (such as "root"), construction stage (such as "foundation construction stage") and other information.

[0054] In the process of structured sorting, first classify and sort all logical component groups to ensure that the list is logically clear and well-organized. For example, sort by construction stage, first list all component groups in the main structure construction stage, including foundation construction, frame construction and other parts; then list the component groups in the mechanical and electrical installation construction stage; and finally list the component groups in the decoration and renovation construction stage. Classifying and sorting in a scientific construction sequence can help accurately track the engineering quantities of each construction stage in payment audits, and ensure the consistency and operability of data logic and management processes. In this way, the bill of quantities can reflect the sequence of construction, so that auditors and managers can quickly locate the content of the current construction stage and easily compare payment applications and construction plans. Secondly, for each component group, it is also necessary to ensure the accuracy and standardization of its name, quantity and unit. This step is mainly completed through automated format specifications and data consistency checks. The system will standardize the names, for example, unify different names such as "foundation beam group" or "beam foundation group" into standard names to avoid information confusion. In addition, the system verifies the correctness of quantities and units, for example ensuring that all components within the same component group have the same units and that quantities match previously calculated quantities.

[0055] Another important aspect is to ensure traceability when generating the bill of quantities. Each item is linked to the previous BIM model data, topology analysis, feature extraction and other steps, so that each quantity can be reverse verified during subsequent review and audit. In this way, even if there is a discrepancy between the quantity and the payment application at a later stage, the bill of quantities can be traced back to the corresponding logical component group and construction stage.

[0056] The final bill of quantities is a standardized document that not only contains the quantitative data of each component group, but also organizes this data into construction phase items through a structured presentation, forming a logically smooth and clear table.

[0057] Further, step S210 includes steps S211 to S214.

[0058] Step S211, performing component extraction processing according to the first information, extracting each component in the model and its geometric information by using a geometric feature recognition algorithm, and obtaining a component list;

[0059] It is understandable that the geometric feature recognition algorithm is completed using methods in computer vision and image processing technology. Preferably, the geometric boundaries of each component are accurately found through three-dimensional edge detection and vertex matching; through region growth and cluster analysis, the various faces of the component can be segmented and combined to ensure the integrity of the component. In the BIM model, since the components have different complex geometric shapes, the recognition algorithm needs to be adjusted and optimized accordingly according to different geometric features. For example, for linear components such as columns or beams, the main axis recognition method is used to determine their axial and extension lengths, while for planar components such as plates or walls, contour recognition and area filling methods are used to obtain their geometric information.

[0060] Furthermore, component extraction is not limited to simple geometric shape extraction, but also involves the spatial attributes and associations of each component. For example, the position (coordinates) of a component in the model and its spatial relationship with other components are also very important information in the extraction process. These spatial attributes provide important data support for subsequent topological analysis, because the spatial position of a component determines its role in the overall building and its relationship with other components.

[0061] After the extraction is completed, all component information will be organized into a component list. This component list contains the geometric description of each component (such as volume, area, shape parameters), material properties, spatial coordinates, and other properties related to the component itself. The component list is the basis of the entire project payment inspection process because it provides a complete description of the construction project and effectively decomposes a large BIM model into independent components that can be analyzed one by one.

[0062] Step S212, performing spatial relationship analysis and processing according to the component list, generating a graph structure by establishing virtual connection lines to represent the contact, overlap or adjacent relationship between each other, and obtaining a connection network graph;

[0063] It should be noted that in the process of constructing the connection network diagram, some techniques in graph theory are preferably used, such as the Delaunay triangulation algorithm and the Voronoi diagram to help divide the space and ensure that the construction of the connection lines is both accurate and can fully cover all possible relationships. Specifically, the Delaunay triangulation algorithm is used to calculate the nearest neighbor relationship between components and build a minimum connection network based on these relationships, while the Voronoi diagram is used to define the influence range of each component to ensure that the construction of the connection line conforms to the geometric boundary of the component. In addition, the spatial adjacency between components is judged by calculation based on the Euclidean distance to ensure that even components that are not in direct contact but need to be considered for coordination in construction can be correctly connected.

[0064] The unique value of the connection network diagram is that it can fully reflect the physical dependencies and spatial constraints between components, which is particularly important for coordination during the construction process. For example, by analyzing the connection network diagram, it is possible to identify possible collision problems of certain components during the construction process, or to clarify the dependency order of certain components. This information plays a very important role in the subsequent construction plan preparation and payment review, because they ensure that the components are constructed and checked in the correct order and logic.

[0065] Step S213, performing interaction analysis processing according to the connection network diagram, using the spectrum characteristics of the spectrum clustering algorithm to analyze the diagram to cluster components with associated relationships in the same construction stage to obtain component groups;

[0066] It is understandable that the nodes in the connection network diagram represent individual components, and the edges represent the spatial relationships between them (such as contact, adjacency, or overlap). However, it is difficult to clearly distinguish which components have closer interactions in the actual construction process by directly observing the connection network diagram. Therefore, it is necessary to further analyze the structural characteristics of the connection network diagram to find those components with higher correlations. These correlations are usually reflected in the construction sequence, functional logic, or physical dependencies. For example, some components must work together during construction to complete a specific structural part, or because of mutual support, their completion order during the construction process must be strictly followed.

[0067] To perform these analyses, a spectral clustering algorithm was used. Spectral clustering is a graph-based clustering method that discovers natural clustering structures in a graph by analyzing the eigenvalues ​​and eigenvectors of the graph's Laplacian matrix. Specifically, spectral clustering represents a connected network graph as a Laplacian matrix, and the spectral features of the matrix (i.e., eigenvalues ​​and eigenvectors) reflect the closeness between components. Using these spectral features, nodes that have strong connections to each other can be effectively clustered together. This method is suitable for data structures such as connected network graphs because it can perform clustering based on the structural features of the entire graph, rather than relying solely on single geometric or distance information.

[0068] Specifically, in the cluster analysis of the construction stage, the spectral clustering algorithm automatically groups those components with strong associations in the same construction stage into one group. The spectral clustering algorithm can cluster these components into a group by analyzing the connection weights and topological characteristics of these components in the connection network diagram, ensuring that these components are considered as a whole in the subsequent engineering quantity calculation and payment review. Such a component group not only helps the synchronization of construction plans, but also helps to judge the rationality of payment applications, because these components are often completed in the same time period and share construction equipment and resources. The advantage of spectral clustering is that it is not limited to the spatial distance of components, but can also complete clustering based on the diversified interactions between components. For example, although some components are far apart in physical space, they need to be completed at the same time due to the functional relationship in construction. Spectral clustering can identify them as related components and cluster them into the same group by analyzing the indirect paths in the connection network. This clustering result helps to discover hidden dependencies in construction, optimize construction plans, and reduce coordination costs during construction. It also ensures the rationality and comprehensiveness of payment review and prevents deviations in review caused by omissions or improper splitting of individual components.

[0069] Step S214: Perform topological analysis processing according to the component group, perform quantitative analysis on the components based on the interaction characteristics, calculate the connection strength and interaction frequency between the components, and obtain the topological analysis results.

[0070] It should be noted that the calculation of connection strength is mainly based on spatial relationships and physical dependencies. Specifically, the connection strength reflects the importance of the interaction between two components in the building structure. For example, the connection between a column and a beam is more critical than the connection between a wall decorative panel and a supporting structure, so their connection strength will be assigned a higher value. Preferably, when calculating the connection strength, a weighted formula based on structural importance and spatial overlap is used. This formula takes into account the direct contact area between components, the physical support function, and the functional logical dependency. For example, if two components have a large contact surface and support each other mechanically, their connection strength will be correspondingly higher. Such a quantification process ensures that the system can clearly reflect the role of each component in the group and its relative importance.

[0071] The calculation of interaction frequency is to describe the number of times components collaborate during the construction process. This includes direct interactions in construction operations (such as the common use of support tools or materials) as well as repeated interactions in multiple construction stages. The interaction frequency is calculated by analyzing the task scheduling logs and resource sharing between components during the construction process. For example, if two components need to be constructed synchronously in multiple construction stages, or must be operated simultaneously when using the same equipment or materials, their interaction frequency will increase accordingly. Technically, the interaction frequency is calculated by combining the construction time points and cross-operation data of the components through time series analysis to ensure that every interaction is accurately recorded. The final topological analysis results not only include whether there is a connection between components, but also describe the tightness of this connection and the number of interactions in actual construction. These quantitative indicators can help better identify the critical paths and key components in construction, ensuring that these important structural elements can be considered in the review of project payment.

[0072] Further, step S220 includes steps S221 to S224.

[0073] Step S221, identifying component relationships based on the topological analysis results, evaluating the interaction between components by using a connection strength algorithm and connecting components that are interdependent during construction to obtain an association relationship network between components;

[0074] It can be understood that this step first begins to identify the specific relationship between components based on the component group data obtained from the topological analysis. The topological analysis has provided the spatial relationship and basic dependencies between components, but these dependencies need to be further quantified in order to clarify which components are closely connected during the construction process and have actual construction synergy. In this step, the connection strength algorithm is used to calculate the degree of dependency between components. The connection strength can be regarded as a measure of the importance and interaction of two components in construction.

[0075] The connection strength algorithm evaluates the relationship between components through multiple dimensions, first considering the physical contact area and functional coupling. The larger the physical contact area, the higher the connection strength between the two components, because their construction operations and mechanical support are closer. Secondly, functional coupling refers to the interaction between components in the building system. For example, the connection between load-bearing components (such as columns and beams) is more important than the connection between decorative components. When evaluating the connection strength, the system also combines information on the construction stage to determine which components must be constructed in the same stage to achieve specific construction goals. For example, if two components have multiple collaborative operations in different construction stages, their connection strength will be increased accordingly.

[0076] After calculating the connection strength between components, these relationships are made explicit to generate an association network. In this network, nodes represent individual components and edges represent the connection strength between components. Unlike the previous connection network diagram, the association network here further refines and quantifies the relationships between components, clearly identifying those components with high connection strength and strong construction dependence for subsequent construction management and payment review. When constructing the association network, component pairs with connection strength exceeding a certain threshold are automatically selected for connection to ensure that the network only contains those relationships that have a significant impact on the construction progress and the overall structure, thereby effectively filtering out those minor connections that have no critical impact on payment review.

[0077] The unique value of this associative network is that it can be dynamically updated and adjusted at different stages of construction. For example, the connection strength of some components may be very important at the foundation stage, but their dependence will decrease at the structural capping stage. Through this dynamic associative network, we can always maintain a focus on key components at different construction stages, ensuring that the allocation and approval of resources in payment review and construction management are based on the most critical components and their relationships.

[0078] Step S222: performing interaction analysis according to the association relationship network, identifying the components that are interrelated in each construction stage by introducing the construction stage logic rule, and determining the cooperation relationship between the related components to obtain the interrelated components;

[0079] It should be noted that the association network describes the dependencies and interactions between the components in the building, but these relationships need to be refined according to the construction stage, because in the actual construction process, the coordination requirements for components in different stages are different. For example, in the foundation construction stage, all components involved in the foundation (such as the cushion layer, foundation beam, and foundation slab) need to work together, while in the decoration construction stage, only the relationship between the wall and the doors and windows is involved. In order to clarify the association relationship between components in each stage, the components in the association network are matched and screened with the construction stage by introducing the logical rules of the construction stage. The logical rules of the construction stage are formulated based on the construction plan and management experience of the building project. These rules include the goals of each construction stage, the main components involved, and the relative construction sequence between the components. For example, in the foundation stage, the construction logic requires that the cushion layer must be completed before the foundation beam. Through these logical rules, it is possible to identify which components in each construction stage are interrelated and need to be constructed at the same time or in a specific order. The logical rules can also analyze the interdependence of adjacent components, such as some components supporting other components during the construction process to ensure that their construction sequence can be correctly connected.

[0080] When performing interaction analysis, components in each phase are analyzed one by one based on the logical rules and association network of the construction phase. In the same construction phase, components with strong associations are automatically identified as component groups that need to work together. By analyzing the cooperative relationship between them, the specific roles and task allocations of these components in construction are clarified. For example, if some components have a mutual support effect in the structure, these components will be identified as cooperative components and the corresponding collaborative relationship will be established. This cooperative relationship is not only reflected in the construction sequence, but also includes the sharing of resources (such as equipment and materials) and the common impact on the construction progress.

[0081] In addition, the interaction analysis also takes into account the dynamic changes of components during the construction process. For example, if there is some rework or adjustment during construction, the association relationship of the components can be dynamically adjusted according to the construction log. This dynamic adjustment ensures that the component cooperation relationship during the construction phase can reflect the actual progress of the construction in real time, thereby providing accurate and real-time association data for payment review. For payment review, this means that the rationality of the payment application can be judged based on the actual cooperation situation during the construction phase, such as whether the payment application is consistent with the phased construction progress and whether it involves all necessary cooperation components.

[0082] Step S223: performing merging processing according to the interrelated components, merging the interdependent components in the same construction stage into a single logical component by applying a weighted merging method, and obtaining preliminary logical component information;

[0083] First, based on the interrelated components determined in step S222, the component groups that need to be constructed collaboratively in the same construction phase are identified. These components are closely related in function and construction operation, such as structural support, material sharing, construction synchronization, etc. In order to further improve the efficiency of construction management and payment review, these interdependent components are merged into a logical component to form a higher-level abstract representation. The definition of a logical component is to treat multiple physical components with common construction goals, functional dependencies, and structural interactions as a whole, thereby simplifying subsequent engineering quantity calculations and status monitoring.

[0084] Preferably, the merging process adopts a weighted merging method. The uniqueness of this method is that it does not simply treat all interrelated components equally, but performs weighted merging according to their specific importance and the closeness of their interaction in the construction stage. The weight of weighted merging is determined by several key factors:

[0085] Association strength: Set a weight for the association strength between each pair of components. The higher the association strength, the higher the priority of the component merging. For example, in foundation construction, the association strength of columns and foundation beams is high because they must provide structural support together, so in the weighted merging process, these components will be preferentially merged into one logical component.

[0086] Functional dependency: Analyze the functional dependency of components, including mechanical effects, and the relationship between support and support. For example, there is a mechanical support relationship between beams and columns, and beams need to rely on columns to achieve overall structural stability. Therefore, these dependency characteristics are taken into account during weighted merging, and components that are highly coupled in function are merged to ensure that the merged logical components can represent the actual structural behavior during construction.

[0087] Construction synchronization: Consider the synchronization between components during the construction process. For example, in a construction phase, some components need to be completed synchronously to achieve the phased goals (such as the synchronous construction of the foundation and foundation beams). The synchronization between these components will also be reflected in the weighting coefficient. If two components need to be completed in the same time period, their weights will increase, making them more likely to be merged into the same logical component.

[0088] The weighted merging method allows the merging process to respect the actual physical and functional relationships between the components, while also being able to systematically simplify these components according to the needs of project management. Ultimately, the preliminary logical component information obtained through merging is a higher-level abstract representation that represents the overall performance of the original component group in terms of structure, function, and construction. The logical component not only contains the engineering quantity information of the original component (such as volume, area, weight, etc.), but also carries the overall construction phase attributes, so that subsequent construction management and payment audits can be based on simpler and more intuitive data.

[0089] Step S224: Perform overall performance analysis based on the preliminary logical component information, and obtain component group information by constructing a logical relationship model between components during the construction process and identifying connectivity, construction sequence, and functional interaction between components.

[0090] Specifically, connectivity describes how logical components support and connect to each other in the building as a whole, such as the node connection between columns and beams or the contact between walls and floors. This analysis uses the adjacency matrix in graph theory to represent the physical connection between components. The value of the adjacency matrix determines the connection strength between each logical component, ensuring that the system has a clear understanding of the overall connection status of the components. The construction sequence refers to the installation and construction order of each logical component during the construction phase. The analysis of the construction sequence is based on the construction logic rules, and the construction sequence of each logical component is clarified by constructing a topological sort. For example, it analyzes which logical components must be constructed first to ensure safety (such as the foundation must be constructed before the superstructure), and which logical components can be constructed simultaneously. Topological sorting not only provides the construction sequence of logical components, but also provides a basis for judging the matching degree between the progress of a certain logical component and the overall construction plan in payment review.

[0091] Further, step S300 includes steps S310 to S330.

[0092] Step S310: Perform quality and progress matching processing according to the manual quantity calculation data in the second information, calculate the weighted Euclidean distance and establish a corresponding relationship between the quality data and the construction progress data to obtain a matching result;

[0093] It is understandable that the manual quantity calculation data in the second information includes multiple characteristics of construction quality and construction progress. These characteristics usually reflect the specific construction status of each component, such as quality inspection results (such as strength, flatness, verticality, etc.) and construction progress (such as completion percentage, node completion time, etc.). In actual project management, quality and progress are two closely related elements that must be reasonably coordinated. If the construction quality does not meet the requirements, the progress will be ineffective and will lead to rework. Therefore, evaluating the degree of matching between quality and progress is crucial to ensure the smooth implementation of the construction plan and improve the efficiency of capital use. The purpose of the matching process is to calculate the matching distance of each component and use it as a basis to judge the coordination of quality and progress. If the matching distance of a component is large, it means that the construction progress fails to reflect the expected quality level, and there may be a risk that the construction is not carried out according to the standard, the progress is rushed, or the construction quality does not meet the standard, which will become the focus of subsequent payment review.

[0094] The use of weighted Euclidean distance enables the system to calculate the matching degree based on the relative importance of each feature, which not only takes into account the direct comparison between quality and progress, but also incorporates the influence of external factors and the nonlinear relationship between each feature. This matching method makes the entire process more flexible and adaptable, especially in the face of complex and dynamically changing construction environments, by reasonably adjusting weights and adjustment coefficients to obtain more accurate matching results. The specific formula is:

[0095] ;

[0096] in, Indicates the serial number of the component; Indicates the characteristic serial number of external influencing factors; Represents the total number of external factor characteristics; Characteristic serial numbers representing quality and progress data; Indicates the total number of quality and progress characteristics; Representation component The comprehensive weighted matching distance between the quality data and the construction progress data; Representation component In the Quality data on each feature; Representation component In the Construction progress data on each feature; Indicates The weighting coefficient of each feature; Indicates that for A nonlinear transformation function for the quality and progress data characteristics; Representation component In the External influence characteristics on each feature; represents the external influence adjustment coefficient; Indicates that for The weight of the external factor characteristics.

[0097] Step S320: Perform data association analysis based on the matching results, and obtain a correlation matrix by calculating the Pearson correlation coefficient between each pair of quality indicators and construction progress;

[0098] It should be noted that the Pearson correlation coefficient can quantify the linear relationship between characteristics, so that complex construction quality and progress data can be presented in a form that is easy to understand and apply. This quantification can help quickly identify problems or optimization opportunities in construction. For example, if certain quality characteristics have a high positive correlation with progress, it can be inferred that paying attention to these characteristics in construction management will help improve the overall progress control efficiency and thus arrange the payment plan reasonably. The calculation formula is:

[0099] ;

[0100] in, Indicates the quality characteristic number; Indicates the sequence number of the progress feature; Representing quality characteristics With progress characteristics Pearson correlation coefficient between ; Representation component In the Quality data on each feature; Representing quality characteristics The mean of Representation component In the Construction progress data on each feature; Indicates progress characteristics The mean of Indicates the total number of components.

[0101] Step S330: Perform data fusion processing according to the correlation matrix, identify the main sources of variation in the quality and progress data by using principal component analysis, and merge multiple related variables into at least two comprehensive indicators to form a new engineering data set.

[0102] Specifically, principal component analysis can combine multiple quality and progress features into a few principal components. For example, the original data may contain multiple quality indicators such as concrete strength, verticality, surface flatness, and multiple progress indicators such as completion time of different construction nodes and progress percentage of each process. Through principal component analysis, the system may find that there is a high degree of covariation among these features, and thus combine them into a comprehensive indicator - such as "construction quality and progress synergy", which can describe the overall performance between construction progress and quality. The selected principal components are generally at least two or more, which can not only cover most of the information in the original data, but also significantly reduce the dimension of the data, reduce the redundancy and mutual interference between features. For example, there are often redundant features between construction quality and progress. If each feature is used as an independent variable in payment review, it will lead to inefficient data processing and increase the risk of misjudgment. The comprehensive indicators after principal component analysis can use fewer indicators to fully describe the overall status of the original features, making the payment review process simpler and more intuitive. An important feature of principal component analysis is that it transforms the complex relationship of the original features into a few independent principal components, which are orthogonal, that is, they are not correlated with each other. Therefore, the comprehensive indicators after principal component analysis can avoid the model instability problem caused by multicollinearity, making payment audit and construction progress control more reliable and scientific. Principal component analysis not only extracts the most important information in the data, but also eliminates redundancy and noise through data fusion, enhancing the ability to grasp the key construction status in payment audit.

[0103] Further, step S400 includes steps S410 to S440.

[0104] Step S410, performing status assessment processing according to the bill of quantities and the engineering data set, and classifying the actual status of the components by using a K-nearest neighbor algorithm to obtain a component status classification result;

[0105] It should be noted that due to the high dimensionality and diversity of component features in engineering construction, the K-nearest neighbor algorithm can learn directly based on existing feature data without making too many assumptions and modeling on the data, which makes it suitable for such scenarios with diverse data and complex structures. It can efficiently use existing data features and historical labels to perform new classifications and ensure the accuracy of the evaluation results.

[0106] Step S420: performing payment condition analysis based on the component status classification result and the third information, evaluating the relationship between the component status and the payment condition by implementing polynomial logistic regression, and obtaining the payment compliance prediction probability corresponding to each status;

[0107] Specifically, multinomial logistic regression is a statistical model used to handle multi-classification problems. In this embodiment, the component status is associated with the payment decision (payment or non-payment). It predicts the payment rationality under each component status by calculating the weighted sum of each input feature (such as quality, progress, construction stage), and then converting it into a probability value through a logistic function.

[0108] Multinomial logistic regression works as follows:

[0109] Step S421, extract relevant feature data from the component status classification results, and establish a payment evaluation model in combination with the payment application data in the third information. The feature data includes the classification status of the component (such as progress delay, compliance with standards, etc.), and other factors related to the payment application (such as payment time point, component importance, etc.). These features will be used as inputs to the model to evaluate their impact on payment conditions. The application of multinomial logistic regression can effectively handle the complex relationships between multiple status categories. For example, the "compliance with quality standards" of a component may be affected by multiple factors such as progress indicators, quality indicators, and construction environment, and there may be nonlinear and interactive relationships between these factors. By establishing a weight relationship between multiple categories, the polynomial logistic regression model can more accurately evaluate the impact of these complex features on payment conditions, thereby providing a scientific basis for payment decisions. The expression of the polynomial logistic regression model is described as:

[0110] ;

[0111] in, Indicates the target category of payment, that is, whether the relationship between the component status to be predicted and the payment condition is satisfied; Indicates a specific payment category; Represents a feature vector, which is composed of multiple features, such as component quality data, progress information, external environmental conditions, engineering quantities, etc.; A serial number indicating a category; Indicates the total number of categories; For and category The associated coefficient vector; For and category The associated coefficient vector; Represents a transpose operation; represents a natural constant; Represents the component status in multinomial logistic regression Belongs to category probability.

[0112] Step S422: The payment evaluation model is trained based on historical payment data to determine the weight of each feature in the payment conditions. Once the training is completed, the model is used to predict the payment conditions for the new component status to obtain the payment compliance prediction probability corresponding to each component status. For example, for a component classified as "meets quality standards", the model may predict that its payment compliance is 90%, indicating that the status of the component is very consistent with the payment requirements; while for a component classified as "needs rework", the payment compliance may be only 20%, which means that direct payment is unreasonable.

[0113] Step S423: For each component, a payment compliance prediction probability is output. This probability value reflects the rationality of payment under the current state of the component. The decision on whether to approve the payment request is based on this probability value. For example, the system can set a payment threshold. If the payment compliance probability of a component is higher than 80%, the payment can be approved directly. Components below a certain threshold require further review, or payment is temporarily refused until the construction status meets the requirements.

[0114] Step S430: extracting feature data according to the predicted probability of payment compliance, evaluating the importance of each feature in the classification, and selecting the most influential component features based on a preset threshold to obtain a preliminary feature data set;

[0115] This process can simplify the feature space and highlight the factors that have a key impact on payment rationality and construction quality, thereby providing more refined and effective data support for subsequent payment audits.

[0116] Step S440: Perform comprehensive analysis and processing based on the preliminary feature data set, and build a multivariate linear regression model to model the relationship between the feature data and the actual state of the component to obtain a final feature data set.

[0117] This step uses the preliminary feature data set for comprehensive analysis to further establish an accurate relationship model between the feature data and the actual status of the component. For this purpose, a multivariate linear regression model is constructed to establish a clear linear relationship between the various features of the component (such as quality indicators, progress data, structural characteristics, etc.) and its actual construction status, so as to further verify and obtain a clearer feature data set, which will provide direct support for the final payment decision.

[0118] Further, step S500 includes steps S510 to S540.

[0119] Step S510: Based on the feature data, a correlation matrix is ​​constructed, the state features of the components at each project stage are extracted, and the features are combined with the basic features of the payment application to obtain a comprehensive feature set;

[0120] First, the construction of the association matrix is ​​to clarify the relationship between each feature in the feature data set. These feature relationships include the relationship between construction status features, the relationship between the amount allocation in the payment application and the component features, etc. By constructing the association matrix, the coupling between the various features can be intuitively understood. Preferably, the association matrix takes the form of a two-dimensional matrix, in which each element represents the relationship between two features (such as correlation or interdependence). The process of constructing the association matrix can lay the foundation for the construction of a comprehensive feature set, so that the subsequent feature fusion can reflect the deep connection between the features.

[0121] Secondly, each component will go through different project stages during the construction process, and each stage has its own unique construction requirements and status characteristics. In order to quantify the performance of each component in different stages, it is necessary to extract the characteristics of each project stage. For example, the foundation stage may focus on the quality of concrete pouring, while the structure capping stage focuses more on characteristics such as verticality and bearing capacity. By extracting these staged status characteristics, the different performances of each component throughout the construction process can be fully captured to ensure that no key information is missed during payment review. These status characteristics are obtained by analyzing the construction records of different project stages. They reflect the progress of the component in each stage, the changing trend of quality indicators and other information, thereby providing support for the comprehensive evaluation of the construction status of the component.

[0122] Finally, the basic features in the payment application need to be spliced ​​with the feature data extracted previously. The basic features of the payment application include payment amount, payment time, payment node, etc. These features directly affect the results of the payment review, so they must be combined with the construction status features for unified analysis. The splicing process is to merge the component status features with the payment application features into a comprehensive feature vector. Each feature vector not only contains the construction quality and progress information of the component, but also includes all information related to the payment request, such as the matching degree between the payment plan of the component at different stages and its actual status. This splicing closely combines the payment conditions and the component status, so that the payment review is not only based on the rationality of the construction quality and progress, but also combines the feasibility of financial conditions.

[0123] Step S520: Perform data enhancement processing according to the comprehensive feature set, analyze and extract abnormal behavior features in the payment application, and combine the abnormal behavior features with the comprehensive feature set to construct an enhanced data set;

[0124] It should be noted that the comprehensive feature set includes the state characteristics of the components at each stage of the construction process, the performance at different stages, and various detailed information in the payment application. However, these conventional features alone cannot effectively identify all abnormal behaviors in the payment application, such as abnormal payment amount, repeated payment applications, or payment applications without sufficient construction progress support. In order to ensure the comprehensiveness and rationality of payment review, it is necessary to further enhance the feature data, especially to expand and deeply mine the features in identifying abnormal payment behaviors. Specifically, during the payment review process, the abnormal behavior characteristics of each payment application are analyzed to identify unreasonable phenomena, such as excessive payment amount, abnormal payment frequency, or mismatch between progress and payment. By quantifying these abnormal behavior characteristics (such as payment rationality index, abnormal frequency score), they are merged with the original feature set to generate an enhanced data set. The enhanced data set contains construction status and abnormal risk information, making payment review more accurate, improving the system's sensitivity to abnormal behavior, and ensuring the scientific nature of payment review and effective prevention of financial risks.

[0125] Step S530: Perform scenario simulation and feature interaction analysis processing according to the enhanced data set, create scenarios of at least two construction stages, simulate the impact of different component states on the payment application, generate corresponding interaction features, and obtain interaction feature information;

[0126] It is understandable that actual construction is often a dynamic process, and the environment, resource allocation, and component status at different construction stages may have different impacts on payment conditions. Scenario simulation and feature interaction analysis provide a method to simulate these complex dynamics and capture the interactions between features.

[0127] Further, step S530 includes steps S531 to S534:

[0128] Step S531, creating a construction phase scenario;

[0129] First, create scenarios for at least two construction phases to simulate the changes in the characteristics of the project at different construction phases. For example, scenarios can include "foundation construction phase" and "main structure construction phase". These scenarios simulate possible changes in actual construction by defining different construction conditions, resource distribution, and component status. Each scenario contains quality and schedule requirements that match that phase, as well as payment conditions. For example, in the foundation construction phase, the focus is on simulating the foundation bearing capacity and the size and elevation of earthwork excavation, while in the main structure construction phase, more attention is paid to the formwork installation size, steel bar installation connection, and concrete pouring quality.

[0130] Step S532: Simulate the impact of different component states on payment applications:

[0131] Specifically, in each scenario, the impact of the construction status of different components on the payment application is simulated by enhancing the feature data in the dataset. The simulation process includes multiple component states, such as "construction quality qualified", "progress delayed" or "quality rework". By adjusting the state variables of the component, its impact on the payment conditions in a specific scenario is simulated. For example, if the concrete strength of a component fails to meet expectations during the foundation construction phase, the compliance of the payment application in this case will be simulated. This simulation can reveal which features have the most important impact on the rationality of payment in a specific stage, thereby better guiding payment review.

[0132] Step S533: feature interaction analysis;

[0133] Feature interaction analysis is performed on the basis of scenario simulation, with the focus on evaluating the interaction between features. For example, concrete strength and progress completion may interact with each other in some cases. It is particularly important to ensure that the progress is as planned while the strength meets the standard. Through this interactive analysis, it is possible to identify which features have important synergistic effects or conflicting situations in specific scenarios. Preferably, the relationship between such features is quantified by constructing a feature interaction model (for example, using cross-product terms), thereby forming interactive feature information. For example, an interactive feature such as "concrete strength × progress completion" can reflect the collaborative state between construction quality and progress, which is very valuable for payment audits.

[0134] Step S534: Generate interaction feature information;

[0135] Finally, through scenario simulation and feature interaction analysis, interactive feature information was generated. The interactive feature not only describes the performance of a single feature at a certain stage, but also captures the dynamic relationship and interaction between different features under different construction conditions. For example, the interactive feature in a scenario may reveal that in the foundation construction stage, the rationality of the payment application depends not only on the strength of the concrete, but also on whether the construction progress is carried out as planned. Such interactive feature information can provide a more in-depth analysis basis to ensure that the payment review not only considers a single construction status, but also considers the dynamic impact between various factors.

[0136] Step S540: Train and predict the interactive feature information based on the preset CART decision tree model, and obtain the payment review result by analyzing each payment application.

[0137] This step combines multi-dimensional features and interactive information to form a review decision for the payment application, including specific decisions such as whether the payment is approved, suspended or rejected. To achieve this goal, the system first selects the appropriate algorithm and determines the data model requirements through a series of steps, and then uses the CART decision tree for analysis and prediction.

[0138] First, the selection of the CART decision tree model is based on the actual needs and data characteristics of payment review. The CART decision tree can segment multi-dimensional data layer by layer and ultimately classify payment applications into multiple categories. In this system, based on the available data model dimensions, the classification targets of the CART decision tree are defined as six types, namely:

[0139] Progress not associated: indicates that the construction progress data cannot be associated with the project part and cannot be reviewed;

[0140] Progress is not completed: the construction progress has not reached the expected requirements and cannot enter the payment stage;

[0141] Quality acceptance not completed: Quality acceptance not completed, not meeting payment conditions;

[0142] Contract overpayment warning: If the cumulative payment amount exceeds the contract agreement, the contract will enter the overpayment warning state;

[0143] Doubts about the comparison with the model: There is a discrepancy between the actual construction status and the expected engineering quantity in the BIM model, which requires further review;

[0144] Meets requirements: Construction progress, quality and contract amount all meet requirements and the payment application can be reviewed.

[0145] The setting of these classification goals enables the CART decision tree to accurately classify each payment request into the category that best matches its current status based on the construction status and various information of the payment application.

[0146] Determining the data model requirements is an important step in implementing the decision tree algorithm. In this embodiment, the data model covers a variety of business data, including progress, quality, cost, and BIM model engineering quantity. These data dimensions are organized into a data model for data mining and Gini coefficient calculation. Specific data dimensions include:

[0147] Unique identification (primary key): used to identify each piece of data and ensure the association and uniqueness between data.

[0148] Progress column item name: an entry that identifies progress data and is used to determine whether the progress data is associated with a specific construction site.

[0149] Associated parts: Indicates the components related to each part, broken down in the form of a part tree to ensure that the list of parts can be found in the bill of quantities.

[0150] Model engineering quantity: The engineering quantity in the model is extracted through part-related data, which needs to be pre-calculated in the decision tree.

[0151] Fill in the engineering quantity: The engineering quantity data filled in by the construction unit is the basic data for payment application.

[0152] Cumulative Project Volume: Indicates the cumulative amount of work that has been paid, used to determine whether there is overpayment in the contract.

[0153] Contract Quantity: Calculated based on the associated data of the contract and the bill, used to clarify the scope of the construction contract and the corresponding quantity of work.

[0154] Percentage of completion: The calculation formula is (reported project quantity + cumulative project quantity) / contract project quantity, which is used to determine whether there is overpayment.

[0155] Contract warning threshold: a single value used to determine whether the completion percentage exceeds the threshold, thereby entering the overpayment warning state.

[0156] Whether the progress implementation has been completed: used to evaluate the completion of the progress.

[0157] Whether the quality acceptance has been completed: Determine whether the construction quality has been accepted and meets the standards.

[0158] The training process of the CART decision tree is based on these data dimensions. By continuously splitting the input features to minimize the error, the decision tree divides the data set into subsets layer by layer. For each split, the CART model selects the most discriminating feature for splitting based on the Gini coefficient (used to evaluate the impurity of the data set) until the payment application is finally classified into a specific category. For example, for a payment application, the CART model may first check whether the "progress column name" exists. If not, it will be directly classified as "progress not associated"; if progress data exists, it will further check the relationship between the "cumulative project volume" and the "contract warning threshold" to determine whether there is a risk of overpayment.

[0159] The prediction and classification of payment applications is done through the layer-by-layer segmentation path of the model, and ultimately the payment applications are classified into specific payment review categories. The system generates review results for each payment application, including automatic approval of payment applications in the "compliant" category, and marking other categories such as "progress not completed" or "quality not accepted" as requiring further review or directly rejected. These classification results provide a clear basis for the payment review process, allowing reviewers to quickly understand the current status of each payment application and the system's judgment logic.

[0160] The integration of the automatic approval module is the last step of this step. The automatic approval module receives the classification results of the CART decision tree and generates corresponding decisions according to the preset approval rules. For example, for applications that "meet the requirements", the system will automatically approve the payment; for applications marked as "contract overage warning" or "progress not completed", the system will take corresponding measures according to the preset rules, which may be marked as requiring manual review or directly refusing payment. The implementation of automatic approval greatly improves the efficiency of the payment review process, and at the same time ensures the accuracy and rationality of the payment through a detailed analysis of the classification results.

[0161] Finally, by using the CART decision tree model to train and predict interactive feature information, the automated review of payment applications is realized, and the decision logic and results of each payment application are explained in detail. By comprehensively analyzing multi-dimensional features such as payment progress, quality acceptance, contract amount, and abnormal behavior, the CART decision tree can reasonably classify payment applications based on complex input features and provide clear payment review decisions. It realizes an efficient and transparent payment review process, ensures the rationality of each payment, reduces financial risks, and provides managers with a clear basis for decision-making. Embodiment 2:

[0162] This embodiment provides a system for checking the payment of engineering funds based on the BIM model quantity calculation results, and the system includes:

[0163] An acquisition module is used to acquire first information, second information and third information, wherein the first information is BIM model data of a construction project, the second information is manual quantity calculation, progress and quality data input by a user on a cloud platform, and the third information includes a payment application for a construction unit's engineering payment;

[0164] a segmentation module, configured to perform topological analysis and component segmentation processing according to the first information, obtain a bill of quantities by identifying spatial relationships between components and extracting engineering quantity information related to the characteristics of each construction stage, wherein the bill of quantities includes the name, quantity, unit and construction stage of the component;

[0165] A fusion module is used to perform data fusion processing according to the second information, and obtain an engineering data set by matching the quality data with the construction progress and embedding the quality data into the progress and cost analysis of each component;

[0166] An analysis module is used to perform correlation analysis based on the bill of quantities, the engineering data set and the third information, and obtain characteristic data by analyzing the correlation between the actual status of the component and the payment conditions;

[0167] The judgment module judges the feature data set based on the preset CART decision tree model, and obtains the payment review result by evaluating the rationality of the payment application.

[0168] In a specific embodiment disclosed in the present application, the segmentation module includes:

[0169] A first analysis unit is used to perform topological analysis processing according to the first information, establish a connection network between components by using a spatial relationship graph algorithm, and analyze the interaction of different components during construction, treating components with similar functions or interdependence as a whole, and obtaining a topological analysis result;

[0170] The first classification unit is used to perform classification processing according to the topological analysis results, classify the components into different categories based on material type, function and spatial structure, and merge the interrelated components in the same construction stage into logical components to obtain component group information;

[0171] The first extraction unit is used to perform feature extraction processing according to the component group information, and calculate the actual engineering quantity of each component group through a preset engineering quantity calculation formula to obtain engineering quantity information;

[0172] The first processing unit is used to obtain a bill of quantities by structuring the name, quantity, unit and construction stage of the component group according to the engineering quantity information.

[0173] In a specific embodiment disclosed in the present application, the first analysis unit includes:

[0174] A second extraction unit performs component extraction processing according to the first information, and extracts each component in the model and its geometric information by using a geometric feature recognition algorithm to obtain a component list;

[0175] The second analysis unit performs spatial relationship analysis processing according to the component list, generates a graph structure by establishing virtual connection lines to represent the contact, overlap or adjacent relationship between them, and obtains a connection network graph;

[0176] The third analysis unit performs interaction analysis based on the connection network diagram, and uses the spectrum characteristics of the spectrum clustering algorithm to cluster the components with correlation relationships in the same construction stage to obtain component groups;

[0177] The fourth analysis unit performs topological analysis processing according to the component group, performs quantitative analysis on the components based on the interaction characteristics, calculates the connection strength and interaction frequency between the components, and obtains the topological analysis results. Embodiment 3:

[0178] like Figure 2 As shown, this embodiment provides an overall architecture of a project payment inspection system based on BIM model quantity calculation results.

[0179] As shown in the figure, the overall architecture is divided into two parts: client and cloud, each part contains different modules, each of which undertakes specific tasks.

[0180] The client consists of two main modules: BIM Model Building Module 1 and BIM Model Extraction Module 2. BIM Model Building Module 1 is the starting point for BIM model construction, which involves importing the design drawings of the building project into professional BIM software to build a three-dimensional model that can reflect all the physical and functional characteristics of the project. The task of BIM Model Extraction Module 2 is to extract detailed engineering quantity data from the BIM model, including the name, quantity, unit and construction stage of the component.

[0181] The cloud part is divided into two main functional areas: information entry and automatic processing. The cloud is mainly used for data storage, processing, review and payment decision execution, including: manual quantity entry module 3, progress data entry module 4, quality data entry module 5, construction unit submission of engineering payment application module 6, CART decision tree classification module 7 and automatic approval module 8. The construction party or the supervisor can manually input the engineering quantity data through the manual quantity entry module 3 in the cloud system, and these data will be compared and analyzed with the automatically extracted BIM engineering quantity. Users enter the progress data of the project through the progress data entry module 4 in the cloud, and enter the quality data through the quality data entry module 5. The construction unit submission of engineering payment application module 6 is used to submit the engineering payment application. The CART decision tree classification module 7 uses the CART decision tree model to classify the payment application. The automatic approval module 8 performs automatic approval based on the classification results of the CART decision tree.

[0182] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for checking the payment of engineering funds based on the quantity calculation results of a BIM model, characterized in that: include: Obtaining first information, second information and third information, wherein the first information is the BIM model data of the construction project, the second information is the manual quantity calculation, progress and quality data input by the user on the cloud platform, and the third information includes the construction unit's application for payment of the project funds; Performing topological analysis and component segmentation processing according to the first information, obtaining a bill of quantities by identifying the spatial relationship between components and extracting engineering quantity information related to the characteristics of each construction stage, wherein the bill of quantities includes the name, quantity, unit and construction stage of the component; Performing data fusion processing according to the second information, matching the quality data with the construction progress, and embedding the quality data into the progress and cost analysis of each component to obtain an engineering data set; Performing correlation analysis processing according to the bill of quantities, the engineering data set and the third information, and obtaining characteristic data by analyzing the correlation relationship between the actual state of the component and the payment conditions; The characteristic data is judged based on a preset CART decision tree model, and the payment review result is obtained by evaluating the rationality of the payment application; The topological analysis and component segmentation processing are performed according to the first information, and the bill of quantities is obtained by identifying the spatial relationship between the components and extracting the engineering quantity information related to the characteristics of each construction stage, including: Performing topological analysis processing according to the first information, establishing a connection network between components by using a spatial relationship graph algorithm, and analyzing the interaction between different components during construction, treating components with similar functions or interdependence as a whole, and obtaining a topological analysis result; Classification processing is performed according to the topological analysis results, components are divided into different categories based on material type, function and spatial structure, and interrelated components in the same construction stage are merged into logical components to obtain component group information; Performing feature extraction processing according to the component group information, calculating the actual engineering quantity of each component group using a preset engineering quantity calculation formula to obtain engineering quantity information; According to the engineering quantity information, a bill of quantities is obtained by structuring the names, quantities, units and construction stages of the component groups; The characteristic data is judged based on the preset CART decision tree model, and the payment review result is obtained by evaluating the rationality of the payment application, including: According to the feature data, a correlation matrix is ​​constructed, state features of components at each project stage are extracted and spliced ​​with basic features of the payment application to obtain a comprehensive feature set; Performing data enhancement processing according to the comprehensive feature set, analyzing and extracting abnormal behavior features in the payment application, and combining the abnormal behavior features with the comprehensive feature set to construct an enhanced data set; Performing scenario simulation and feature interaction analysis processing according to the enhanced data set, by creating scenarios of at least two construction stages, simulating the impact of different component states on payment applications and generating corresponding interaction features, thereby obtaining interaction feature information; The interactive feature information is trained and predicted based on a preset CART decision tree model, and the payment review result is obtained by analyzing each payment application.

2. The method for checking the construction payment based on the BIM model quantity calculation results according to claim 1 is characterized in that: A topological analysis is performed based on the first information. A connection network between components is established by using a spatial relationship graph algorithm. The interaction between different components in construction is analyzed. Components with similar functions or interdependence are considered as a whole to obtain topological analysis results, including: Perform component extraction processing according to the first information, and extract each component in the model and its geometric information by using a geometric feature recognition algorithm to obtain a component list; Performing spatial relationship analysis and processing according to the component list, generating a graph structure by establishing virtual connection lines to represent the contact, overlap or adjacent relationship between the components, and obtaining a connection network graph; Performing interaction analysis and processing according to the connection network diagram, clustering components with associated relationships in the same construction stage using the spectrum characteristics of the analysis diagram using a spectrum clustering algorithm to obtain a component group; A topological analysis process is performed on the component group, a quantitative analysis is performed on the components based on the interaction characteristics, the connection strength and interaction frequency between the components are calculated, and a topological analysis result is obtained.

3. The method for checking the construction payment based on the BIM model quantity calculation results according to claim 1 is characterized in that: According to the topological analysis results, classification processing is performed, and components are divided into different categories based on material type, function and spatial structure, and interrelated components in the same construction stage are merged into logical components to obtain component group information, including: Identify component relationships based on the topological analysis results, evaluate the interactions between components by using a connection strength algorithm, and connect components that are mutually dependent during construction to obtain an association relationship network between components; Performing interaction analysis according to the association relationship network, identifying the components that are interrelated in each construction stage by introducing the construction stage logic rule, and determining the cooperation relationship between the related components to obtain the interrelated components; Performing merging processing according to the interrelated components, merging the interdependent components in the same construction stage into a single logical component by applying a weighted merging method, and obtaining preliminary logical component information; An overall performance analysis is performed based on the preliminary logical component information, and component group information is obtained by constructing a logical relationship model between components during the construction process and identifying connectivity, construction sequence and functional interaction between components.

4. The method for checking the construction payment based on the BIM model quantity calculation results according to claim 1 is characterized in that: Data fusion processing is performed according to the second information, and the quality data is matched with the construction progress, and the quality data is embedded into the progress and cost analysis of each component to obtain an engineering data set, including: Performing quality and progress matching processing according to the manual quantity calculation data in the second information, calculating the weighted Euclidean distance and establishing a corresponding relationship between the quality data and the construction progress data, to obtain a matching result; Performing data association analysis according to the matching results, and obtaining a correlation matrix by calculating the Pearson correlation coefficient between each pair of quality indicators and construction progress; Data fusion processing is performed according to the correlation matrix, and a new engineering data set is formed by identifying the main sources of variation in the quality and progress data through the use of principal component analysis, and merging multiple related variables into at least two comprehensive indicators.

5. The method for checking construction payment based on BIM model quantity calculation results according to claim 1 is characterized in that: Performing correlation analysis processing according to the bill of quantities, the engineering data set and the third information, and obtaining characteristic data by analyzing the correlation relationship between the actual state of the component and the payment conditions, including: Performing status assessment processing according to the bill of quantities and the engineering data set, and classifying the actual status of the components by using a K-nearest neighbor algorithm to obtain a component status classification result; Perform payment condition analysis based on the component status classification result and the third information, evaluate the relationship between the component status and the payment condition by implementing multinomial logistic regression, and obtain the payment compliance prediction probability corresponding to each status; Performing feature data extraction processing according to the payment compliance prediction probability, evaluating the importance of each feature in the classification, and screening out the most influential component features based on a preset threshold to obtain a preliminary feature data set; A comprehensive analysis is performed on the preliminary feature data set, and a multiple linear regression model is constructed to model the relationship between the feature data and the actual state of the component, so as to obtain a final feature data set.

6. A construction payment inspection system based on BIM model quantity calculation results, characterized in that: include: An acquisition module is used to acquire first information, second information and third information, wherein the first information is BIM model data of a construction project, the second information is manual quantity calculation, progress and quality data input by a user on a cloud platform, and the third information includes a payment application for a construction unit's engineering payment; a segmentation module, configured to perform topological analysis and component segmentation processing according to the first information, obtain a bill of quantities by identifying the spatial relationship between components and extracting engineering quantity information related to the characteristics of each construction stage, wherein the bill of quantities includes the name, quantity, unit and construction stage of the component; A fusion module, used for performing data fusion processing according to the second information, matching the quality data with the construction progress, and embedding the quality data into the progress and cost analysis of each component to obtain an engineering data set; An analysis module, used for performing association analysis processing according to the bill of quantities, the engineering data set and the third information, and obtaining characteristic data by analyzing the association relationship between the actual state of the component and the payment conditions; A judgment module judges the characteristic data based on a preset CART decision tree model, and obtains a payment review result by evaluating the rationality of the payment application; Wherein, the segmentation module includes: A first analysis unit is used to perform topological analysis processing according to the first information, establish a connection network between components by using a spatial relationship graph algorithm, and analyze the interaction between different components during construction, treating components with similar functions or interdependence as a whole, and obtaining a topological analysis result; A first classification unit is used to perform classification processing according to the topological analysis result, classify the components into different categories based on material type, function and spatial structure, and merge the interrelated components in the same construction stage into logical components to obtain component group information; A first extraction unit is used to perform feature extraction processing according to the component group information, and calculate the actual engineering quantity of each component group by a preset engineering quantity calculation formula to obtain engineering quantity information; The first processing unit is used to obtain a bill of quantities by structuring the names, quantities, units and construction stages of the component groups according to the engineering quantity information; The feature data is judged based on the preset CART decision tree model, and the payment review result is obtained by evaluating the rationality of the payment application, including: According to the feature data, a correlation matrix is ​​constructed, state features of components at each project stage are extracted and spliced ​​with basic features of the payment application to obtain a comprehensive feature set; Performing data enhancement processing according to the comprehensive feature set, analyzing and extracting abnormal behavior features in the payment application, and combining the abnormal behavior features with the comprehensive feature set to construct an enhanced data set; Performing scenario simulation and feature interaction analysis processing according to the enhanced data set, by creating scenarios of at least two construction stages, simulating the impact of different component states on payment applications and generating corresponding interaction features, thereby obtaining interaction feature information; The interactive feature information is trained and predicted based on a preset CART decision tree model, and the payment review result is obtained by analyzing each payment application.

7. The construction payment inspection system based on BIM model quantity calculation results according to claim 6 is characterized in that: The first analysis unit comprises: A second extraction unit is used to perform component extraction processing according to the first information, and extract each component in the model and its geometric information by using a geometric feature recognition algorithm to obtain a component list; A second analysis unit is used to perform spatial relationship analysis processing according to the component list, generate a graph structure by establishing virtual connection lines to represent the contact, overlap or adjacent relationship between each other, and obtain a connection network graph; A third analysis unit is used to perform interaction analysis processing according to the connection network diagram, and cluster components with associated relationships in the same construction stage using the spectrum characteristics of the spectrum clustering algorithm analysis diagram to obtain a component group; The fourth analysis unit is used to perform topological analysis processing according to the component group, perform quantitative analysis on the components based on the interaction characteristics, calculate the connection strength and interaction frequency between the components, and obtain the topological analysis result.

Images