A BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis

Abstract

The application discloses a BIM five-dimensional auxiliary bid evaluation method based on light-weight analysis, and belongs to the field of bid evaluation in the bid process. The method comprises the following steps: obtaining a BIM bid document of a project and a project bid evaluation standard, and analyzing the BIM bid document into structured data; constructing a five-dimensional BIM model of the current project according to the structured data; wherein the five-dimensional BIM model comprises a time sub-model, a cost sub-model and a three-dimensional space sub-model; determining a plurality of target bid evaluation items to be called by a user according to the project bid evaluation standard; when a target bid evaluation item is called by the user, responding to bid evaluation data of the target bid evaluation item through the five-dimensional BIM model, and performing bid evaluation weighting; obtaining the bid evaluation data after bid evaluation weighting, and generating a bid evaluation report. The application converts the bid document into structured data through light-weight analysis technology, and constructs a five-dimensional BIM model combining time, cost and three-dimensional space, automatically adapts the bid evaluation scene, realizes multi-dimensional collaborative bid evaluation, and improves the bid evaluation efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of engineering construction, and in particular relates to a BIM five-dimensional auxiliary bid evaluation method based on light-weight analysis. BACKGROUND

[0002] With the continuous expansion of the construction industry, the difficulties in the construction process are increasing, and the market bidding has entered a new stage of development. BIM technology, as a new digital means, can transmit and store the building information in the whole life cycle process of the project from planning, design, bidding, construction to operation, which can greatly improve the efficiency and quality of engineering construction. The application of BIM technology brings major changes to bid evaluation experts and project evaluation, mainly in the aspects of visual scheme evaluation, evaluation standards more in line with project needs, complete and accurate evaluation factors, improved professional degree of evaluation experts, and high degree of digitalization of intelligent auxiliary bid evaluation software, and the quality of project evaluation can be greatly improved.

[0003] In traditional engineering project construction, the pricing scheme in the economic bid should be calculated according to the actual construction plan, labor, material and machine input, and combined with the pricing strategy. However, in the bidding, the economic bid and the technical bid are completely separated and independent, which leads to an unobjective and unscientific evaluation result.

[0004] In the prior art, the bid evaluation experts and the bidders of the BIM bid evaluation system realize online bidding operation through the Internet, and the network dependency is strong. In the bid evaluation process on the client side, complete bid evaluation data needs to be downloaded, so the data volume is large. It is easy to cause lag and data loss in the bid evaluation process. SUMMARY

[0005] The present application provides a BIM five-dimensional auxiliary bid evaluation method based on light-weight analysis. In the bid evaluation process, time and cost are introduced, the cost, project progress and three-dimensional model are linked, and five-dimensional bid evaluation is realized. In this process, the BIM five-dimensional bid evaluation model is cooperated through the light-weight calling of structured data, to prevent data redundancy and key data loss. Based on the data depth adaptation of the five-dimensional model in the bid evaluation scene, the bid evaluation user can output more accurate bid evaluation results.

[0006] In the first aspect, the present application provides a BIM five-dimensional auxiliary bid evaluation method based on light-weight analysis, comprising:

[0007] Step A1: obtaining the BIM bid document file and the project bid evaluation standard of the current project;

[0008] Step A2: parsing the BIM bid document file into structured data;

[0009] Step A3: constructing a five-dimensional BIM model of the current project according to the structured data; wherein the five-dimensional BIM model comprises a time sub-model, a cost sub-model and a three-dimensional space sub-model;

[0010] Step A4: determining a plurality of target evaluation items to be called by the user according to the project evaluation standard;

[0011] Step A6: when the target evaluation item is called by the user, responding to the evaluation data of the target evaluation item through the five-dimensional BIM model, and performing evaluation weighting;

[0012] Step A7: obtaining the evaluation data after evaluation weighting, and generating an evaluation report.

[0013] The above scheme integrates multi-dimensional data (space, time, cost) and dynamic response mechanism, constructs a five-dimensional BIM model based on lightweight structured data, improves the precision and intelligence of the evaluation process, and overcomes the defects of the traditional BIM evaluation, such as disconnection between progress and cost, subjectivity of weight allocation, and single evaluation dimension. The traditional evaluation system mainly evaluates through web login, and when the bid document is large, the evaluation may be slow. The lightweight structure of the present application can also prevent the evaluation process from being slow.

[0014] In combination with the first aspect, step A2 comprises:

[0015] The non-structured description text in the BIM bid document is subjected to semantic analysis to determine the logical association relationship between components, and associated metadata is generated;

[0016] According to the associated metadata and the non-structured description text, the component data is compressed into structured data with a data integrity not lower than a target threshold, a data volume not higher than a target volume threshold, and an error not exceeding a target error value, and the structured data is stored in a target database;

[0017] The structured data of the components in the target database is provided with a unique component ID, and an index association operation is responded to the unique component ID.

[0018] In the above embodiment, during the bid document uploading process, the model parameter deficiency is made up through unstructured text semantic analysis, the data quality and storage efficiency are balanced through double-target data compression, and the five-dimensional BIM model cross-dimension association is realized based on the unique ID index.

[0019] In combination with the first aspect, step A2 further comprises:

[0020] According to the BIM bid document and the project evaluation standard, determining the evaluation key points;

[0021] According to the evaluation key points and the structured data, determining a semantic tag system associated with the BIM data;

[0022] According to the semantic tag system, the structured data is cut into a plurality of independently loaded target module data; wherein the target module data is indexed and associated in response to the unique component ID.

[0023] The above embodiment, in the process of BIM data management, realizes accurate matching of data and bid evaluation criteria through the bid evaluation scene driven semantic tag system; improves data loading efficiency through modular cutting; realizes multi-dimensional collaborative review through cross-module indexing, and solves the problem of disconnection between module division and bid evaluation requirements.

[0024] In combination with the first aspect, before step A3 is performed, the method further includes:

[0025] Obtaining construction plan data of the current project, and dividing the structured data into data modules corresponding to construction nodes; wherein the data modules include time data, space data and cost data;

[0026] Constructing a time-space evolution data flow that maps the data modules and the BIM model to each other, and performing five-dimensional division on the module data according to the time-space evolution data flow.

[0027] The above embodiment, through the multi-dimensional data module division driven by the construction plan, the dynamic time-space data flow and the five-dimensional division of the bid evaluation adaptation, prevents the problem that the construction plan and the multi-dimensional data are disconnected in the bid evaluation process, and cannot reflect the dynamic construction process.

[0028] In combination with the first aspect, step A3 includes:

[0029] Pre-configuring a quantization port for time vectorization, cost vectorization and IFC three-dimensional vectorization of the five-dimensional BIM model;

[0030] Vectorizing the structured data through the quantization port to generate a first information guide vector;

[0031] Matching the first information guide vector according to the model dimensions of the five-dimensional BIM model to generate a second information guide vector based on data matching degree; wherein the second information guide vector is used to represent the model parameter vector of the five-dimensional BIM model;

[0032] Generating the five-dimensional BIM model of the current project according to the model parameter vector.

[0033] The above embodiment, through the special quantization port, ensures the uniformity of multi-dimensional data format; through the two-stage vector mapping, improves the adaptability of data and model; through the matching degree optimization, guarantees the accuracy of model parameters, and further prevents the problems of error in multi-dimensional data fusion and disconnection between model parameters and actual data in the bid evaluation process.

[0034] In combination with the first aspect, step A4 includes:

[0035] According to the second information guide vector, determine the textual description data representing different dimensions of the five-dimensional BIM model;

[0036] According to the textual description data, build an evaluation interface of the five-dimensional evaluation scene, and the evaluation interface includes at least one built structured component;

[0037] Obtain a first position specified by a user in the evaluation interface, and the first position includes a target coordinate point in a model space, a target time node on a progress time axis, or a target hierarchical position in a cost classification tree;

[0038] In response to the first position and information of the built structured component, display N recommended structured components in the building page through an association analysis algorithm;

[0039] In response to a selection operation of the user on at least one of the N recommended structured components, including clicking and dragging, deploy structured data of the selected component to the first position in the five-dimensional scene, and establish an association relationship between the selected component and the built structured component, and generate a corresponding target evaluation item.

[0040] The above embodiment realizes accurate mapping of data and interfaces through five-dimensional parameter-driven textual description; improves component selection efficiency through multi-dimensional position positioning and algorithm recommendation; realizes cross-dimensional data linkage review through dynamic association deployment; prevents the problem that components deployed in the BIM evaluation interface are disconnected from the evaluation demand and cannot reflect the association relationship between different components or data.

[0041] In combination with the first aspect, step A5 includes:

[0042] Receive a first evaluation instruction of a user, and determine a target evaluation item; wherein the first evaluation instruction includes at least one target evaluation item;

[0043] According to the target evaluation item, build an evaluation rule pool; wherein the evaluation rule pool includes a plurality of association rule subsets, and each association rule subset includes a plurality of evaluation rules of different dimensions;

[0044] Match the evaluation rules and the five-dimensional BIM model through a rule engine, and determine a responsive evaluation item.

[0045] The above embodiment solves the problems of scattered rules, the need for manual experience, and the need for manual switching for multi-dimensional verification in BIM evaluation through the association rule subsets of the evaluation scene, the multi-dimensional matching of the five-dimensional model, and the dynamic calling driven by the target.

[0046] In combination with the first aspect, step A5 further includes:

[0047] The target evaluation item is determined through the five-dimensional BIM model, and the target evaluation item is called through the five-dimensional BIM model.

[0048] According to each evaluation data stream, an evaluation index system and a target key dimension are constructed; wherein the evaluation index system is used to evaluate at least one measurement index of each target evaluation item and key data of the measurement index in the five-dimensional BIM model;

[0049] According to the evaluation index system and the target key dimension, the target evaluation item is determined, and the scheme adaptation degree is determined with respect to the target evaluation item;

[0050] According to the scheme adaptation degree of at least one key data, the evaluation result of the corresponding target evaluation item is determined.

[0051] The above embodiment realizes multi-dimensional data linkage through five-dimensional data stream cooperative calling; improves evaluation accuracy through key dimension constraint index system; and realizes evaluation result comparability through scheme adaptation quantification, thereby solving the problems of index redundancy focus and result subjectivity in BIM evaluation.

[0052] In combination with the first aspect, the evaluation weighting includes:

[0053] The spatial conflict level, the time deviation value and the cost overrun proportion are identified from the evaluation data;

[0054] Based on the spatial conflict level and the component type correlation, high, medium and low risk intervals are divided;

[0055] In combination with the time deviation value and the cost overrun proportion, the comprehensive influence level of progress delay and budget overconsumption is determined;

[0056] According to the project type, a predefined rule library is called to assign a basic weight coefficient;

[0057] The spatial weight factor, the time cost influence value and the basic weight coefficient are fused to output a weight list of each target evaluation item.

[0058] The above embodiment can solve the problems of poor scene adaptability and disconnection between weight and actual risk through component type differentiated risk division, time-cost coupling calculation, project type dynamic adaptation and nonlinear fusion algorithm.

[0059] In combination with the first aspect, the step A6 includes:

[0060] The evaluation weighting range is extracted through the permission recognition technology, and the core attention dimension is determined to construct a semantic mapping system associated with the evaluation target;

[0061] Integrate the scoring results, detailed basis, scheme comparison and risk labeling data in the evaluation weighting range to determine the structured evaluation data set;

[0062] Based on the semantic mapping system, the key information of each evaluation project is automatically extracted to generate an evaluation report containing conclusion judgment, advantage analysis and risk prompt.

[0063] The beneficial effects of the present application are:

[0064] The present application improves the efficiency of processing BIM bid documents through lightweight analysis technology, and realizes three-dimensional geometric structure by constructing a structured five-dimensional BIM model, fuses time and cost, and extracts accurate data for specific evaluation items and dynamically weights, which not only improves the efficiency of evaluation, but also can flexibly set different evaluation standards for different types of bid documents for each evaluation item, so that the multidimensional correlation review of project progress and cost is reflected in the evaluation process.

[0065] The five-dimensional BIM model of the present application deeply associates time and cost with three-dimensional model, so that the progress node and cost list are automatically mapped to realize the early warning effect of automatically triggering cost deviation due to progress delay. In step A5, the user can automatically call the sub-model in the five-dimensional BIM model according to the demand, so as to realize the review of cost dimension without loading irrelevant data such as time data, and reduce the interference of invalid data. In the process of evaluation weighting, the five-dimensional BIM model automatically calculates the weight based on the evaluation data to realize objective weighting driven by data and reduce the deviation of artificial subjective weighting.

[0066] In the whole evaluation process, the present application builds a five-dimensional BIM evaluation model through structured analysis of bid documents to realize static display of evaluation data. Through the called evaluation items, comprehensive evaluation can be carried out, and single evaluation can also be carried out to realize dynamic selection and intelligent decision of evaluation, when there is evaluation exception, the five-dimensional model automatically associates and displays the cost deviation data of the corresponding node, and adjusts the weight proportion of the two through weighting algorithm to realize multidimensional linkage review, improve the efficiency of evaluation, and ensure the objectivity of evaluation.

[0067] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application can be achieved and obtained by the structures specifically pointed out in the written description and drawings.

[0068] The technical solutions of the present application will be further described in detail below through the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0069] The accompanying drawings are included to provide a further understanding of the application, and are incorporated in and constitute a part of this specification, illustrate embodiments of the application, and together with the description serve to explain the application, and do not limit the application.

[0070] In the drawings:

[0071] Figure 1 A diagram of a BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis in an embodiment of the application;

[0072] Figure 2 A diagram of the bidding system of the present application in an embodiment of the application;

[0073] Figure 3 A diagram of the traditional bid evaluation system in an embodiment of the application;

[0074] Figure 4 A functional diagram of the structured analysis processor in an embodiment of the application;

[0075] Figure 5 A process diagram of the lightweight loading of bid evaluation data in an embodiment of the application;

[0076] Figure 6 A diagram of the full-cycle data call of the five-dimensional BIM model in an embodiment of the application;

[0077] Figure 7 A process diagram of the determination of the bid evaluation project in an embodiment of the application;

[0078] Figure 8 A process diagram of the output of the bid evaluation result in an embodiment of the application.

[0079] 10 is a first device, 20 is a structured analysis processor, 30 is a third device, 40 is a fourth device, and 50 is a fifth device. DETAILED DESCRIPTION

[0080] The preferred embodiments of the application are described below with reference to the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the application, and do not limit the application.

[0081] The automatic bidding system can input the bidding file, automatically classify the bidding file, and realize automatic bid evaluation. The traditional bid evaluation system is as follows: Figure 3As shown, it includes an application layer, a service platform, a data layer, and a cloud platform. The application layer sets up multiple bid evaluation robots and bid evaluation management terminals. Multiple bid evaluation robots can evaluate tenders through various bid evaluation schemes, thereby realizing multi-dimensional evaluation to score and filter from the perspectives of average analysis or bias (protecting the environment and minimizing costs), etc., to determine the best scoring result. The bid evaluation robot can simulate the evaluation logic of a real person or an expert. The bid evaluation management terminal is used to input the bid evaluation scheme. The service platform is the main body of the bid evaluation system, including a BIM design evaluation module for performing project evaluation and BIM module evaluation. A supply chain evaluation module is used to evaluate the stability of material supply and whether the material is reliable. A project collaborative management evaluation module is used to evaluate the project quality, project progress, and project intelligent control process during project implementation. A quality evaluation module is used to evaluate project materials and construction technology. An equipment evaluation module is used to evaluate the types of construction equipment and the maintenance status of the construction equipment of the bid project. The quality of the equipment is related to the quality of the project. A logistics evaluation module is used to evaluate the timeliness of vehicle transportation and material transportation during project implementation. A big data analysis module performs comprehensive evaluation based on project content, cost, and time, and outputs the tender evaluation result. The data layer is used to scan the tender, and realize tender input. The private cloud platform is used to control the uploading of the tender and the calling of the cloud computing resources.

[0082] However, the data layer can only input two-dimensional CAD drawings + electronic documents. Most projects use two-dimensional CAD drawings + electronic documents (PDF / OFD format) to input tenders. The bid evaluation system mainly uses text and table data, and lacks three-dimensional model interaction function. Such review method is low in efficiency, strongly dependent on personal experience, and poor in review effect.

[0083] Cloud computing has a delay. BIM models usually have a large amount of data. The loading and rendering of BIM models require high hardware performance support. Pure cloud resources cannot be used. There is no correlation and coordination mechanism for the overall bid evaluation.

[0084] As shown in Figure 2 The application provides a bid evaluation system for performing a BIM five-dimensional auxiliary bid evaluation method with light parsing. The bid evaluation system is composed of a first device 10, a structured parsing processor 20, a third device 30, a fourth device 40, and a fifth device 50.

[0085] Embodiment 1:

[0086] Referring to Figure 1 The application provides a BIM five-dimensional auxiliary bid evaluation method based on light parsing, which comprises the following steps:

[0087] Step A1: obtaining a BIM tender file of a current project and a bid evaluation standard of the project;

[0088] Step A1 is performed by the first device 10, which receives the BIM tender file and determines the project evaluation criteria by uploading the file through the web page of the fixed client or the client-side web page. The BIM tender file is an electronic document containing building information model (BIM) data, usually stored in IFC (Industry Foundation Class) format or proprietary format; the project evaluation criteria are the evaluation rules specified in the tender document, such as technical evaluation and commercial weight evaluation. The BIM tender file is obtained through file reading techniques such as dedicated data interface or XML / JSON parsing, IFC parsing library, etc.

[0089] Step A2: parsing the BIM tender file into structured data;

[0090] Step A2 is performed by the structured parsing processor 20. The BIM tender file (such as IFC format) is essentially semi-structured data, which contains entity, attribute, relationship, etc. data of different project construction in the entire project, and the data stream is large, which is easy to cause lag and data loss in the review process. In order to reduce the amount of data transmission and reduce the situation of lag and data loss, the application extracts key information such as component parameters, progress plan, and cost list through lightweight parsing tools, and converts them into database tables, JSON, etc. Structured format, stored in the review database.

[0091] Step A3: constructing a five-dimensional BIM model of the current project according to the structured data; wherein the five-dimensional BIM model includes a time sub-model, a cost sub-model, and a three-dimensional space sub-model;

[0092] Step A3 is performed by the third device 30, which uses data correlation to perform five-dimensional modeling based on the structured data of step A2, and uses algorithms such as time series analysis, cost-space mapping, and three-dimensional geometric reconstruction to integrate scattered evaluation-related data into three independent but related sub-models: a three-dimensional space sub-model for modeling the size and spatial layout of building components in the evaluation project, focusing on geometric information; a time sub-model for constructing a model of related project progress data based on key node duration and construction phase division in the project; and a cost sub-model for integrating budget and actual expenditure data through material cost, labor cost, and change order cost. In the operation process, each sub-model can independently store and call the structured data set, so that there is no redundant operation in the full model loading process of the entire system.

[0093] The five-dimensional BIM model built by the application is used for fast data acquisition in the evaluation process. The five-dimensional model and the evaluation requirements are bound, and the internal sub-models respond to each evaluation project, achieving dynamic and fast data calling.

[0094] Step A4: According to the project evaluation standard, determine the target evaluation item to be called by the user;

[0095] Step A4 is executed by the fourth device 40, which parses the key evaluation points in the evaluation standard through a rule engine or natural language processing technology, and maps them to the target evaluation items identified by the entire evaluation system; this process converts abstract evaluation standards into operational specific evaluation items, preventing evaluation rules from being ambiguous and inconsistent, which can lead to subjective judgment bias in the evaluation process. In actual implementation, different evaluation standards can be set according to different projects, different environmental requirements, economic requirements, or terrain restrictions.

[0096] Step A5: When there is a target evaluation item called by the user, respond to the evaluation data of the target evaluation item through the five-dimensional BIM model and perform evaluation weighting.

[0097] Step A5 is executed by the fifth device 50, which extracts relevant data from the key node duration of the time sub-model or the construction sequence of the three-dimensional space sub-model according to the target evaluation item called by the user; permission control is achieved through role permission tables, dynamic tokens, etc., and data access permissions are allocated to different users. In specific implementation, only the sub-model data related to the target evaluation item is loaded, and the material details of the three-dimensional model that do not need to be displayed during evaluation of the duration do not need to be displayed, and there is no interference from irrelevant information. Based on the weighting mechanism, the access range of sensitive data is limited to reduce the risk of leakage. Through the coupling of evaluation standards and five-dimensional models, the traditional artificial-dominated evaluation process can be converted into a data-driven process (BIM bid documents and evaluation items corresponding to project evaluation standards). In the evaluation process, after receiving the user's evaluation instruction, the evaluation data is directly responded to, the required evaluation data is called, and the evaluation project is weighted, achieving response and weighting at the same time, and improving the overall evaluation efficiency.

[0098] Step A6: Obtain the evaluation data after weighting, and generate an evaluation report.

[0099] The evaluation report is a textual and tabular evaluation result data, which is based on template engine and data visualization technology, and automatically fills in the evaluation data such as time deviation rate and cost overrun proportion into a structured report according to the preset format.

[0100] It can be understood that the present application improves the efficiency of processing BIM bid documents through lightweight parsing technology, and realizes three-dimensional geometric structure by constructing a structured five-dimensional BIM model, integrating time and cost, and extracting precise data and dynamically weighting for specific evaluation items, which improves the efficiency of evaluation while enabling flexible evaluation of different types of bid documents, with different evaluation standards for each evaluation item, so that the project progress and project cost are correlated in the same dimension during the evaluation process.

[0101] It can also be understood that the five-dimensional BIM model of the present application deeply associates time and cost with the three-dimensional model, so that the progress node and the cost list are automatically mapped, and the progress delay automatically triggers the early warning effect of the cost deviation. In step A5, the user can automatically call the sub-model in the five-dimensional BIM model according to the demand, and when reviewing the cost dimension, no irrelevant data such as time data is loaded, reducing the interference of invalid data. In the process of bid evaluation weighting, the five-dimensional BIM model automatically calculates the weight based on the bid evaluation data, realizes data-driven objective weighting, and reduces the deviation of artificial subjective weighting.

[0102] It can also be understood that in the entire bid evaluation process, the present application builds a five-dimensional BIM bid evaluation model through structured analysis of the bid document, and realizes static display of bid evaluation data. Through the called bid evaluation item, comprehensive bid evaluation can be performed, or single bid evaluation can be performed, realizing dynamic selection and intelligent decision-making of bid evaluation. When there is bid evaluation anomaly, the five-dimensional model automatically associates and displays the cost deviation data of the corresponding node, and adjusts the weight proportion of the two through the weighting algorithm, realizes multi-dimensional linkage review, improves the efficiency of bid evaluation, and at the same time guarantees the objectivity of bid evaluation.

[0103] Embodiment 2:

[0104] The step A2 of the present application is to solve the problems of low conversion efficiency of unstructured text, unbalanced data precision and data volume, and cross-dimension data linkage in the BIM bid evaluation scene. The present application proposes a scheme of converting BIM bid document into structured data.

[0105] Referring to Figure 4 The structured analysis processor 20 includes three parts of an analysis module, a lightweight compression module and a dynamic indexing module. The analysis module is used for semantic analysis of unstructured description text in the BIM bid document, determines the logical association relationship between components, and generates associated metadata;

[0106] The BIM bid document usually contains structured three-dimensional model data, and the three-dimensional model data includes IFC format geometric information. In specific implementation, the project modeling in the bid document is mainly realized through modeling software. The modeling software includes but is not limited to SZ-IFC, IFC, IGMS, GFC, SKP, 3DS and other formats made by market mainstream modeling software Revit, ArchiCAD, Tekla, MagiCAD, BIMMAKE, GTJ, Rhino, Bentley, etc. The existing analysis method mainly performs structured conversion through text mining, but cannot analyze and determine the dependent relationship such as process connection and spatial position association between components.

[0107] The application can identify the logical association relationship between components based on semantic analysis, and then determine the specific construction sequence rationality or the improvement of cost collection accuracy of components by combining BIM bid document and component association analysis for this specific scene of bid evaluation. It can be determined that the application can combine associated metadata and component logical relationship to automatically fill in the fault data when there is a fault in the model parameters and text description. For example, when analyzing the strength grade of the concrete base, the cost data corresponding to the strength of the concrete base is automatically associated.

[0108] The non-structural description text is the text content of the design description, construction requirements, etc. in the BIM bid document. The traditional analysis method only processes structured model data, and the analysis module of the application performs semantic analysis on unstructured text through natural language processing technology, and then determines the association relationship between different components. This logical association relationship or association relationship data in the construction process will be converted into machine-recognized associated metadata, which includes but is not limited to RDF triple data, column, support components and beams of building components. Semantic analysis can extract the implicit logical relationship in the text, while traditional keyword recognition cannot recognize this implicit logical relationship. Therefore, the semantic analysis of the application can make the five-dimensional BIM model (three-dimensional model + time, cost sub-model) and the components have dynamic association factors, for example, the construction delay of a certain beam may be caused by the incomplete support of the column, which prevents the top of the building from being constructed.

[0109] After the associated metadata is determined, the lightweight compression module will compress the component data into structured data with a data integrity not lower than a target threshold, a data volume not higher than a target volume threshold, and an error not exceeding a target error value according to the associated metadata and the non-structural description text, and store the structured data into a target database; wherein the structured data of the components in the target database is provided with a unique component ID, and responds to an index association operation of the unique component ID.

[0110] The application sets two constraint conditions that the target database has a data integrity not lower than a target error value and a data volume not higher than a target data volume threshold. The target error value and the target data volume threshold are both preset by the user.

[0111] Based on the associated metadata, the application filters the material strength, construction period constraints and other component data of the key stress components related to bid evaluation from the BIM bid document, and eliminates irrelevant redundant information, including the color of decorative components, the texture details of non-load-bearing components, etc. An adaptive compression algorithm is used to compress the data volume to a target threshold while ensuring data integrity, balancing the dual requirements of data accuracy and processing efficiency in the bid evaluation scene.

[0112] In the process of compressing component data to a target database completeness that is not lower than a target error value, and structured data whose data volume is not higher than a target data volume threshold: in order to prevent the loss of key component data, through the constraints of completeness and data volume, lightweight is realized under the premise of ensuring the integrity of core data such as cost and time, and distortion of subsequent bid evaluation data caused by compression is avoided.

[0113] The compressed data is stored in a structured format of a relational database or a graph database, and then a unique ID is assigned to each component, and an index is established through the ID to support fast association query across sub-models, for example: through the ID, the construction progress in the time sub-model is associated with the budget data in the cost sub-model. The unique ID also has cross-sub-model association operations, for example: through the component ID, the three-dimensional spatial coordinates, the corresponding time dimension node, and the cost accounting are indexed at the same time, and then fast data retrieval and cross-dimension association of a specific component are realized.

[0114] In the present application, the unique ID is a combination of encoding and component type, and the spatial data, time data and cost data of the same component can be retrieved at one key in the five-dimensional model, thereby realizing multi-dimensional analysis of each project.

[0115] Traditional BIM parsing technology. For example: IFC standard parsing only processes structured geometric data, the present application introduces unstructured text semantic parsing (NLP technology) into the BIM parsing process to extract logical association metadata between components. Through the completion of semantic information, the subsequent five-dimensional model can associate the dynamic influence relationship of the components.

[0116] It can be understood that the present application completes unstructured data through semantic parsing to ensure information integrity, controls the volume of data and the completeness and lightweight of storage through double-target compression, and realizes cross-dimension indexing through unique ID. Further, the obtained data has the effect of lightweight.

[0117] Embodiment 3:

[0118] In order to realize lightweight fast loading of bid evaluation data and precise loading of bid evaluation data, refer to Figure 5 The following methods are also used:

[0119] Determine the bid key points according to the BIM bid document and the project bid evaluation standard;

[0120] Firstly, the key indicators of bid evaluation need to be determined, therefore, the key information extraction of demand driving (project bid evaluation standard, bid evaluation demand can be determined, and the corresponding bid data in BIM bid document) is used to identify the core evaluation indicators that need to be focused on in the bid evaluation process. Through the core evaluation indicators, the bid evaluation standard can be semantically analyzed, the key evaluation items can be extracted, and the key points of bid evaluation that are strongly related to the evaluation items can be selected in combination with the actual data in the BIM bid document.

[0121] After determining the bid evaluation key points, the semantic label system associated with the BIM data can be determined according to the bid evaluation key points and the structured data;

[0122] Through data-demand mapping, the bid evaluation key points are associated with the structured data, and based on the component association data generated in step A2 above, a machine-recognized semantic label system is generated. For each component or parameter in the structured data, based on its association degree with the bid evaluation key points, a semantic label such as Figure 4 the parsing module's process-coordinate-property label is assigned, and a semantic label component based on BIM bid evaluation scene is realized. The semantic label is not customized, so when there are different projects, the semantic label can also judge the differences between the key points; the label system adopts hierarchical design, for example: the first-level label is cost control, and the second-level label is material cost-key item.

[0123] The semantic label system of the present application maps the bid evaluation data to the data layer automatically through the bid evaluation key points, combined with semantic labels and data classification, without interference data.

[0124] According to the semantic label system, the structured data is cut into multiple independent target module data; wherein the target module data responds to the index association operation of the unique component ID.

[0125] Finally, through the modular cutting and association technology, the structured data is divided into multiple independent modules based on the semantic label, and the independent modules include: structural safety module, construction period control module, cost optimization module, etc. The logical association between modules is maintained through the unique component ID (defined in step A2). Independent modules can avoid full data loading, and only the modules related to the current evaluation item need to be loaded when evaluating on the Web side. The data loading time is shortened, and the unique ID index mechanism ensures the logical continuity of the data between modules, realizes fast loading, and realizes on-demand loading of the associated bid evaluation scene.

[0126] When loading data, a single evaluation dimension data can be loaded to solve the problem of traditional full-quantity model loading causing lag. When responding to data, a unique component ID is used to achieve one-time indexing and multi-module response. Further, component-level multi-dimensional penetrating evaluation is achieved. Through evaluation key point guidance, a label system is constructed, a semantic label system determines specific rules for module segmentation, target module data is associated through a unique ID, and deep linkage of cost and progress is achieved.

[0127] Embodiment 4:

[0128] Before the step A3 is performed, i.e., before the five-dimensional BIM model of the current project is constructed according to the structured data, in order to enable the five-dimensional BIM model to have full-cycle data calling capability, the following steps are adopted according to the structured data: Figure 6

[0129] Obtain the construction plan data of the current project, and divide the structured data into data modules corresponding to the construction nodes; wherein the data modules include time data, space data, cost data, design data and evaluation data;

[0130] The construction plan data includes construction nodes and their time intervals. After each construction node extracts the construction plan data through a data interface with XML / CSV parsing protocol and the like, the structured data (second target data from step A2) is divided into data of a three-dimensional space model, data of a time sub-model and data of a cost sub-model according to the construction nodes.

[0131] Time data: progress information strongly associated with the time interval of the construction node, for example: the actual completion time of the pile foundation construction phase;

[0132] Space data: three-dimensional space information corresponding to the construction node, for example: the component coordinates of the pile foundation construction phase;

[0133] Cost data: cost expenditure corresponding to the construction node, for example: the concrete procurement cost of the pile foundation construction phase;

[0134] Design data: component design scheme corresponding to the construction node, for example: specific geometric structure data such as length, width and height of the pile foundation;

[0135] ​Evaluation data: component evaluation item data and evaluation standard data corresponding to the construction node, such as pile foundation material, pile foundation bearing capacity, etc. Through analysis of construction plan data, discrete structured data can be organized according to actual construction process, avoiding data disconnection with construction stage, and data module and construction node can also realize dynamic association, which can respond to construction plan in tender at any time. The structured data of the application is dynamically divided according to construction node, for example, all five-dimensional data of foundation construction stage are cut into independent modules to realize binding of scattered multi-dimensional data and construction node.

[0136] It can be determined that the application realizes multi-dimensional data aggregation of single node by binding time, space, cost, design and evaluation five-dimensional data to construction node, preventing cross-module data from being unable to query. The specific performance is that in actual implementation, by clicking the foundation construction node, the progress deviation, cost overrun, design change and quality evaluation data of the node are displayed synchronously.

[0137] The time-space evolution data flow of the data module and the BIM model is constructed, and the module data is divided into five dimensions according to the time-space evolution data flow.

[0138] The time-space evolution data flow is a set of dynamic mapping rules for defining the association mode of data module and BIM model. Time dimension mapping is the binding of time interval of construction node and time axis of BIM model. For example, the construction project is pile foundation construction, corresponding to the time axis of BIM model 0-30 days; the time-space evolution data flow is not formed by data superposition, but by evolving the data of construction node, determining the data module, and finally determining the BIM model, forming a dynamic mapping logic, realizing automatic evolution, and the data flow responds to the change of construction node in real time. At the same time, the construction node of the application can obtain the spatial coordinates, cost consumption, design change record and quality evaluation result of all components under the node at the same time, expanding the cross-module index dimension.

[0139] The space dimension mapping is the coordinate association between the spatial data of the data module and the three-dimensional components (such as pile foundation components) of the BIM model;

[0140] The cost dimension mapping is the binding of the cost data (such as pile foundation concrete cost) of the data module and the cost attribute (such as cost budget of pile foundation component) of the BIM model. The specific implementation can be realized by event-driven data flow engine, which can monitor the change of construction node in real time and trigger the synchronous update of data and model. The dynamic data in the evolution process is associated with the construction process, and the dynamic response ability is enhanced.

[0141] It can be determined that the application simulates the dynamic data change in the construction process through the specific data of the data module, so that the result of bid evaluation is more in line with the actual construction scene. Through five-dimensional division, the user can quickly locate the bid evaluation risk point by directly corresponding to the bid evaluation focus. For example: the node exists progress delay and cost overrun at the same time, improves the bid evaluation efficiency, and realizes the bid evaluation data processing of one leading to another.

[0142] In actual implementation, bid evaluation traces the root data of cost overrun through construction evolution data, determines whether the delay of the main structure construction node or other delay causes the price change or space change of material supply, and is associated to the evaluation data to realize accurate positioning of bid evaluation risk.

[0143] Embodiment 5:

[0144] The five-dimensional BIM model is a quick data calling and data quantification processing model. Step A3 of the application proposes to construct the five-dimensional BIM model of the current project according to the structured data, and the steps include:

[0145] Pre-configure the quantization ports of time vectorization, cost vectorization and IFC three-dimensional vectorization of the five-dimensional BIM model;

[0146] The quantization port is a set of pre-defined interface rules. Unlike the existing general format data port, the quantization port can customize the quantization parameters of the bid evaluation scene, and then be used to convert the structured data into mathematical vectors. The structured data includes time nodes, cost amounts and IFC component coordinates in JSON format. The time vectorization port is used to define the vector dimension of time data, and the vector dimension includes planned events, actual events and deviation time. The cost vectorization port is used to define the vector dimension of cost data, and the vector dimension includes budget cost, actual cost and overrun / surplus. The IFC three-dimensional vectorization port is used to define the vector dimension of IFC space data, and the vector dimension includes X / Y / Z coordinates, component type code and component type code. The specific configuration can be realized by pre-set metadata template to ensure that the structured data of different projects can be vectorized according to unified rules. In actual implementation, the dimension-specific data verification is realized through port pre-configuration to reduce data format conversion error. For example: the cost port automatically checks the compliance of the list valuation specification.

[0147] Vectorize the structured data through the quantization port to generate a first information guide vector;

[0148] The structured data is converted into a numerical vector through the rules of the quantification port. The vectorization process can convert discrete data into a continuous vector through machine learning feature extraction algorithms. For example, similarity calculation between vectors can be used to determine the matching degree of data and model dimensions. The first information guiding vector is the vectorization of structured data, which is used to realize dynamic matching of data and model dimensions with the second information guiding vector (vector matched with model dimensions), so that there is no deviation between data and model dimensions.

[0149] The first information guiding vector is matched with the model dimensions of the five-dimensional BIM model to generate a second information guiding vector based on data matching degree; wherein the second information guiding vector is used to represent the model parameter vector of the five-dimensional BIM model;

[0150] In this application, through the original data of the first information guiding vector and the model parameters of the second information guiding vector, data dimension reduction and accurate mapping are realized. In specific implementation, threshold screening is also performed through data matching degree to ensure high adaptability of the model parameter vector to the five-dimensional model.

[0151] According to the model parameter vector, a five-dimensional BIM model of the current project is generated.

[0152] Each dimension (time, cost, IFC three dimensions) of the five-dimensional BIM model has a preset model parameter range, and the model parameter range includes a time deviation allowed range and a cost deviation allowed range. Through a vector matching algorithm (such as cosine similarity calculation), the first vector is compared with the parameter range of the model dimension to generate a corrected second vector. For example: after the original time vector is matched with the allowed time deviation range of the model, a second vector (effective) is generated; if the original cost vector exceeds the allowed cost deviation range of the model, it is corrected as over budget and needs to be marked. The matching result is weighted by a weight coefficient to finally form a second vector representing the model parameters. The model parameter vector (second vector) is input into the five-dimensional model construction engine to map the vector data to the time axis, cost axis and three-dimensional space axis of the model, so that the vectorization interface and the dynamic matching algorithm are integrated to generate a five-dimensional model, preventing the operation of manually associating data and model dimensions in the bid evaluation process from being too cumbersome. Through vectorization processing and multi-level matching, the construction precision of the five-dimensional model is improved.

[0153] In one embodiment, after the cost data is verified through a special port, it is accurately matched to the corresponding component of the three-dimensional model through a two-stage vector mapping, and the matching degree optimization ensures that the cost deviation analysis error is ≤1%. This improves the model precision and data matching degree.

[0154] Embodiment 6:

[0155] Reference Figure 7In the specific process of bid evaluation, for the bid evaluation project, step A4 is to determine a plurality of target bid evaluation items to be called by the user according to the project bid evaluation standard, and specifically:

[0156] According to the second information guide vector, text description data representing different dimensions of the five-dimensional BIM model is determined.

[0157] The second information guide vector (model parameter vector) is a mathematical numerical vector, which needs to be converted into text description through natural language generation technology. Specific implementation can be achieved by mapping the vector data into structured text through a pre-trained text generation model combined with a pre-defined template. The text description data is to convert abstract vector data into natural language that can be directly understood by users (bid evaluation experts), reducing the interaction threshold, so that experts can quickly grasp the core information of the model dimension without the need to interpret vector values. The text description generated by the second information guide vector is directly related to the bid evaluation dimension, and users can quickly locate the key data of bid evaluation through the text, reducing the complexity of interface operation.

[0158] According to the text description data, a bid evaluation interface of the five-dimensional bid evaluation scene is built, and the bid evaluation interface includes at least one structured element;

[0159] The bid evaluation interface is rendered into a visual interface based on the text description data through a front-end framework, including three core areas: a three-dimensional space view for displaying IFC three-dimensional models and target coordinate points; a progress timeline for displaying time nodes and deviations; and a cost classification tree for displaying cost levels and overruns / surpluses. The structured element refers to a pre-defined bid evaluation related component integrated into the interface through a component library, so that experts can switch between components and independent tools in multiple independent tools, and the overall interface is dynamically generated by the components, so that abstract data can be converted into concrete interactive elements.

[0160] A first position specified by a user in the bid evaluation interface is obtained, and the first position includes a target coordinate point in a model space, a target time node on a progress timeline, or a target hierarchical position in a cost classification tree;

[0161] The user of the present application specifies the first position through mouse clicking, dragging or voice input, and the system captures the position information through an event listening interface; the target coordinate point in the model space obtains the X / Y / Z coordinates of the clicked position through a three-dimensional rendering engine (such as Cesium); the target time node on the progress timeline obtains the time stamp corresponding to the clicked position through a timeline component (such as D3.js); and the target hierarchical position in the cost classification tree obtains the hierarchical path of the clicked node through a tree component.

[0162] The first position is input through multiple dimensions such as space, time and cost, and is associated with components, so that invalid components are not filtered out.

[0163] In response to the first position and information of the built structured element, N recommended structured components are displayed in the building page through an association analysis algorithm; the association analysis algorithm combines real-time operation position and five BIM model parameters to form a chained search condition, generates a recommended result, and exceeds the conventional recommendation in cross-association processing of five-dimensional data.

[0164] In response to the user's operation, specifically, in response to the selection operation of at least one of the N recommended structured components, including clicking, dragging, deploying the structured data of the selected element to the first position in the five-dimensional scene, and establishing the association relationship with the built structured component, a corresponding evaluation item is generated. After the user selects the recommended element by clicking or dragging, the structured data of the selected element is embedded into the first position, the association relationship between the element and the built element is established through the database foreign key, and then the associated data is packaged as an evaluation item. The component can automatically associate semantic tags, unique IDs and five-dimensional vectors, and realize precise dragging and other operations.

[0165] After dynamic deployment, the association is automatically established through the unique component ID and five-dimensional parameter vector, avoiding manual configuration errors. The evaluation interface is converted from static data display to dynamic scene construction. For example, the user selects the "material cost overruns" node in the cost classification tree, and the system recommends the supplier qualification component through association analysis. After deployment, the procurement time of the material is automatically associated with the spatial use position, realizing three-dimensional evaluation of cost-time-space.

[0166] Embodiment 7:

[0167] In order to solve the problems of scattered evaluation rules, weak multi-dimensional rule coordination, and low correlation between evaluation rules and models in the evaluation process, in step A5 of the present application, if the target evaluation item is called by the user, the five-dimensional BIM model responds to the evaluation data of the target evaluation item and performs evaluation weighting, specifically including:

[0168] Receiving a first evaluation instruction of a user to determine a target evaluation item; wherein the first evaluation instruction includes at least one target evaluation item;

[0169] The present application supports multiple input methods, such as interface clicking, voice input, API calling, obtaining the first evaluation instruction input by the user, analyzing the instruction through natural language processing technology, extracting the target evaluation item, and finally verifying the legality of the target item according to the project evaluation standard (step A1). In addition, different user habits are adapted, the use scenarios are expanded, and the user can dynamically specify the target project to be evaluated, so that the evaluation of a certain project does not exist one-sidedness.

[0170] According to the target evaluation item, an evaluation rule pool is built; wherein the evaluation rule pool includes multiple association rule subsets, and the association rule subsets include multiple evaluation rules of different dimensions;

[0171] The rule library retrieval of the application is to retrieve rules associated with the target evaluation item from a pre-stored evaluation rule library; the evaluation rule library includes a duration rule library, a cost rule library, and a space rule library; the rule subset construction is to group the retrieved rules according to dimensions to form association rule subsets. For example, a rule subset of duration compliance may include a time dimension-node delay rate, a resource dimension-labor input sufficiency, and a risk dimension-weather impact plan; the rule conflict detection is to ensure the consistency of rules in the rule pool through logical verification.

[0172] In the possible implementation process, the association rule subset can trigger a rule chain after determining the project to be evaluated, and generate multiple evaluation rules. For example, a node progress delay automatically triggers a cost deviation rate rule check to realize automatic missed detection.

[0173] The evaluation rules are matched with the five-dimensional BIM model through the rule engine to determine the responsive evaluation item.

[0174] The rule engine of the application converts the evaluation rules into vector conditions, and then matches the vector space with the parameter limits of the five-dimensional BIM model to realize accurate matching of the rules and the model and two-level rule checking. In the output result, not only whether it meets the rules will be output, but also the quantified response of the associated dimension will be generated based on the five-dimensional model parameters, for example, the cost overrun and the duration delay are related, that is, the result and the reason correspond.

[0175] In the specific implementation, the rule engine of the application can realize:

[0176] Rule analysis: converting the rules in the rule pool into logical expressions executable by the rule engine;

[0177] Data extraction: extracting data associated with the rules from the five-dimensional BIM model (step A3);

[0178] Matching execution: the rule engine applies the rules to the extracted data through reasoning to output the matching result;

[0179] Response evaluation item generation: the matching result is summarized as the response data of the target evaluation item.

[0180] The application realizes conflict identification across latitude through five-dimensional parameter parallel matching by a rule engine, for example, cost overrun caused by spatial design change. The application mainly automatically calls corresponding rule subsets through target evaluation items, and can also reduce the operation steps of users. In the application, when the user selects the progress-cost linkage review target, the progress and cost association rule subsets are automatically called, the time delay (time dimension) of the construction node and the corresponding cost overrun (cost dimension) are verified synchronously through the rule engine, and the association (material price rise caused by delay) of the two is identified.

[0181] Embodiment 8:

[0182] In the process of bid evaluation, there are many bid evaluation indexes, and there is no association between different bid evaluation indexes. The bid evaluation scheme may not be suitable for the corresponding bid item or the problem of the bid document, see Figure 8 The application also proposes:

[0183] The target evaluation item in the application is determined by the five-dimensional BIM model, and the multiple evaluation data streams are called by the five-dimensional BIM model through the to-be-evaluated component. Multiple evaluation data streams can realize single construction multi-dimensional data aggregation, and prevent tedious traceability query; in actual implementation, the construction progress, cost consumption, spatial position and quality score of a certain column component can be displayed synchronously by clicking the column component.

[0184] The to-be-evaluated component is determined based on the target evaluation item, for example, main structure construction compliance, and the components directly related to the bid item are selected from the five-dimensional BIM model, such as frame column and load-bearing wall. For each to-be-evaluated component, the multiple sub-model data of the five-dimensional model are associated through the unique component ID (defined in step A2), and multiple data streams related to the evaluation item are called. The traditional index system is a preset template, and the application determines multiple associated factors in time and space by calling the real-time data stream of the five-dimensional BIM model, dynamically generates an index system, and can determine multiple-dimensional evaluation data, for example, time series data, spatial geometric data and text evaluation data.

[0185] According to each evaluation data stream, an evaluation index system and a target key dimension are constructed; wherein the evaluation index system is used to evaluate at least one measurement index of each target evaluation item and the key data of the measurement index in the five-dimensional BIM model;

[0186] The statistical analysis algorithm is adopted in the application to reduce the dimension of the measurement indexes in the evaluation index system, the dimension with the highest variance contribution is extracted as the target key dimension, and then data and index mapping is realized, the main purpose is to realize the automatic extraction of the key dimension characteristics. The data-index mapping is to extract the quantifiable measurement indexes from the bid evaluation data stream, the size error = | actual size - design size | / design size x 100%; based on the bid evaluation target, the dimensions that have the greatest impact on the target are screened, and the measurement indexes and the key dimensions are combined into a hierarchical system.

[0187] According to the evaluation index system and the target key dimension, the target bid item reaching result is determined, and the scheme adaptation degree relative to the reaching result is determined; wherein, according to the scheme adaptation degree of at least one key data, the bid evaluation result of the corresponding target bid item is determined. The application quantifies the adaptation degree by the key data reaching degree, realizes the quantitative sorting of multiple bid schemes, and the adaptation degree score can be traced back to specific data.

[0188] Finally, the reaching result is determined: the measurement indexes are compared with the bid evaluation standard, and the reaching or not reaching conclusion is output; the scheme adaptation degree calculation is calculated by weighted summation or machine learning model, combined with the weight of the key dimension, the matching degree of the scheme and the bid evaluation target is calculated, the bid evaluation result is generated based on the reaching result and the adaptation degree, and the final bid evaluation result is output, the calculation of the scheme adaptation degree can prevent the existence of index conflict, and improve the effect of comprehensive decision.

[0189] It can be determined that the application upgrades the bid evaluation from experience judgment to data-driven decision, when there are multiple bid schemes, key data is extracted from multiple data streams, the adaptation of each scheme is calculated respectively, and data-driven comparative bid evaluation is realized.

[0190] Embodiment 9:

[0191] For different bid evaluation projects or different bid evaluation instructions in the bid evaluation process, there may be subjective problems, the application proposes: according to the bid evaluation data of each bid item, the bid evaluation weighting of each bid item is executed, which specifically includes:

[0192] The spatial conflict level, time deviation value and cost overrun ratio are identified from the bid evaluation data, the key indexes are extracted from the bid evaluation data through data mining algorithm, the purpose is to dynamically adjust the risk interval threshold through the material strength, connection mode and other conditions of the component type, prevent the problem that the risk sensitivity of different components is different, and in actual implementation, the three-dimensional spatial data of the BIM model and the component attribute library need to be combined.

[0193] Spatial conflict level: based on the collision detection result of the IFC three-dimensional model, divided by a preset threshold;

[0194] Time deviation value: Calculate the actual time-plan time difference of each construction node;

[0195] Cost overrun ratio: Calculate (actual cost-budget cost) / budget cost x 100%.

[0196] The identification process is implemented through data cleaning tools such as the Pandas library in Python to ensure uniform data format.

[0197] Based on the spatial conflict level and the association of component type, high, medium and low risk intervals are divided;

[0198] The influence degree of spatial conflict in the present application is strongly related to the component type. By constructing the "component type-conflict level" association matrix, combined with expert experience or historical evaluation data to train the classification model (such as decision tree), the spatial conflict is divided into high, medium and low risk intervals.

[0199] Combined with the time deviation value and the cost overrun ratio, the comprehensive influence level of schedule delay and budget overconsumption is determined;

[0200] There is a linkage between time delay and cost overrun. By constructing a linkage influence model, combined with the preset level threshold, the comprehensive influence level is determined.

[0201] According to the type of the project, the pre-defined rule base is called to allocate the basic weight coefficient;

[0202] The pre-defined rule base stores the basic weight template of different project types, such as residential projects or bridge projects, which have different costs.

[0203] The output of the evaluation item weight list is fused with the spatial weight factor, time cost influence value and basic weight coefficient. Unlike the traditional technology which uses a general template weight rule, the rule base based on project type is dynamically called and multiple factors are fused to determine the weight of the evaluation item.

[0204] The maximum weight is calculated by a multi-factor fusion formula, and the evaluation item weight = spatial weight factor (0.3) x risk interval coefficient (high = 1.5 / medium = 1 / low = 0.5) + time cost influence value (0.4) x comprehensive influence level coefficient (high = 1.2 / medium = 1 / low = 0.8) + basic weight coefficient (0.3).

[0205] Example 10:

[0206] In view of the problems of low efficiency, non-uniform semantics, and poor adaptability of permissions and content of the evaluation report, the step A6 of the present application includes:

[0207] The evaluation weighting range is extracted through permission recognition technology, and the core attention dimension is determined to construct a semantic mapping system associated with the evaluation target.

[0208] In the present application, the core attention dimension can determine the role division in the bid evaluation scene, realize the dimension differentiation identification under the participation of multiple parties. Through the permission identification technology, the bid evaluation weighting range is determined, and then the necessary bid evaluation data can be determined, and the review efficiency of the bid document is improved. The semantic mapping is converted into a structured index, which improves the utilization rate of unstructured data.

[0209] The permission identification of the present application adopts role-based access control or attribute-based permission control technology to identify user permissions, such as bid evaluation experts, supervision administrators, system administrators, and then extract the accessible bid evaluation weighting range; through core dimension screening combined with bid evaluation targets, the core attention dimensions that have the greatest impact on the target are screened through importance analysis; the semantic mapping is constructed based on ontology or knowledge graph technology, the mapping relationship of bid evaluation target-core dimension-data field is established, and the semantic mapping system that can be understood by machine is formed.

[0210] Integrate the bid evaluation weighting range of the score result, the detailed basis, the scheme comparison and the risk annotation data to determine the structured bid evaluation data set;

[0211] In the process of determining the structured bid evaluation data set, first, the score result, the detailed basis, the scheme comparison, and the risk annotation are extracted from different data sources in the weighting range; then, based on data cleaning, through deduplication, error correction, and standardization processing, data noise is eliminated; through the bidirectional mapping rule, the disconnection between data and bid evaluation targets can be prevented.

[0212] Based on the semantic mapping system, the key information of each attention dimension is automatically extracted, and the bid evaluation report containing conclusion judgment, advantage analysis, and risk prompt is generated.

[0213] The key information extraction of the present application is to extract key information from the structured data set according to the semantic mapping system; the report template matches and calls the pre-defined report template, the template contains the conclusion judgment, the advantage analysis, the risk prompt and other modules; the natural language generation converts the extracted key information into natural language through the NLG technology.

[0214] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis, characterized in that, The method comprises the following steps: Step A1: obtaining the BIM tender file of the current project and the project evaluation standard; Step A2: parsing the BIM tender file into structured data; Step A3: constructing a five-dimensional BIM model of the current project according to the structured data; wherein the five-dimensional BIM model comprises a time sub-model, a cost sub-model and a three-dimensional space sub-model; Step A4: determining a plurality of target evaluation items to be called by the user according to the project evaluation standard; Step A5: when there is a target evaluation item called by the user, responding to the evaluation data of the target evaluation item through the five-dimensional BIM model and performing evaluation weighting; Step A6: obtaining the evaluation data after evaluation weighting, and generating an evaluation report; The step A3 comprises: Pre-configuring the quantization ports of time vectorization, cost vectorization and IFC three-dimensional vectorization of the five-dimensional BIM model; Vectorizing the structured data through the quantization port to generate a first information guide vector; Data matching the first information guide vector according to the model dimensions of the five-dimensional BIM model to generate a second information guide vector based on the data matching degree; wherein the second information guide vector is used to represent the model parameter vector of the five-dimensional BIM model; Generating the five-dimensional BIM model of the current project according to the model parameter vector; The step A4 comprises: According to the second information guide vector, determining the textual description data representing different dimensions of the five-dimensional BIM model; According to the textual description data, building an evaluation interface of the five-dimensional evaluation scene, and the evaluation interface comprises at least one built structured component; Obtaining a first position specified by the user in the evaluation interface, wherein the first position comprises a target coordinate point in the model space, a target time node on the progress time axis or a target hierarchical position in the cost classification tree; In response to the first position and the information of the built structured component, N recommended structured components are displayed in the building page through an association analysis algorithm; In response to the selection operation of at least one of the N recommended structured components, the structured data of the selected component is deployed to the first position in the five-dimensional scene, and an association relationship with the built structured component is established, and a corresponding target evaluation item is generated.

2. The BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis according to claim 1, wherein, The step A2 comprises: Performing semantic analysis on the non-structured description text in the BIM tender file to determine the logical association relationship between components and generate association metadata; According to the association metadata and the non-structured description text, the component data is compressed into structured data with a data integrity not lower than a target threshold, a data volume not higher than a target volume threshold and an error not exceeding a target error value, and the structured data is stored in a target database; Wherein, the structured data of the component in the target database is provided with a unique component ID, and responds to the index association operation of the unique component ID.

3. The BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis according to claim 2, characterized in that, The step A2 further comprises: Determining the evaluation key points according to the BIM tender file and the project evaluation standard; According to the evaluation key points and the structured data, determining the semantic tag system associated with the BIM data; According to the semantic label system, the structured data is cut into a plurality of independently loaded target module data; wherein, the target module data is indexed and associated in response to the unique component ID.

4. The BIM five-dimensional auxiliary bid evaluation method based on lightweight resolution according to claim 3, characterized in that, Before the step A3 is executed, it further includes: Obtain the construction plan data of the current project, and divide the structured data into data modules corresponding to the construction nodes; wherein, the data modules include time data, space data, cost data, design data and evaluation data; Construct a time-space evolution data flow that maps the data modules and the BIM model to each other, and divide the module data according to the time-space evolution data flow.

5. The BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis according to claim 1, characterized in that, The step A5 includes: Receive the first evaluation instruction of the user, and determine the target evaluation item; wherein, the first evaluation instruction includes at least one target evaluation item; According to the target evaluation item, build an evaluation rule pool; wherein, the evaluation rule pool includes a plurality of associated rule subsets, and each associated rule subset includes a plurality of evaluation rules of different dimensions; Match the evaluation rules and the five-dimensional BIM model through the rule engine to determine the responsive evaluation item.

6. The BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis according to claim 1, wherein, The step A5 further includes: Determine the evaluation component in the target evaluation item through the five-dimensional BIM model, and call a plurality of evaluation data streams through the evaluation component and the five-dimensional BIM model; According to each evaluation data stream, construct an evaluation index system and a target key dimension; wherein, the evaluation index system is used to evaluate at least one measurement index of each target evaluation item and the key data of the measurement index in the five-dimensional BIM model; According to the evaluation index system and the target key dimension, determine the compliance result of the target evaluation item and the scheme adaptation degree relative to the compliance result; wherein, according to the scheme adaptation degree of at least one key data, determine the evaluation result of the corresponding target evaluation item.

7. The BIM five-dimensional auxiliary bid evaluation method based on lightweight resolution according to claim 1, characterized in that, The evaluation weighting includes: Identify the spatial conflict level, the time deviation value and the cost overrun proportion from the evaluation data; Based on the spatial conflict level and the component type correlation, divide the high, medium and low risk intervals; Combine the time deviation value and the cost overrun proportion to determine the comprehensive influence level of progress delay and budget overconsumption; Call the pre-defined rule library according to the project type to allocate the basic weight coefficient; Fuse the spatial weight factor, the time cost influence value and the basic weight coefficient to output the weight list of each target evaluation item.

8. The BIM five-dimensional auxiliary bid evaluation method based on lightweight analysis according to claim 1, wherein, The step A6 includes: Extract the evaluation weighting range and determine the core attention dimension through the permission recognition technology, and construct a semantic mapping system associated with the evaluation target; Integrate the evaluation result, the detailed basis, the scheme comparison and the risk annotation data in the evaluation weighting range to determine the structured evaluation data set; Based on the semantic mapping system, automatically extract the key information of each evaluation item to generate an evaluation report containing conclusion judgment, advantage analysis and risk prompt.

Images