CN-122022734-A - Automatic engineering quantity list compiling and auditing method and system based on industrial big data

CN122022734ACN 122022734 ACN122022734 ACN 122022734ACN-122022734-A

Abstract

The invention relates to the technical field of data processing and engineering cost, and discloses an automatic compiling and auditing method for an engineering quantity list based on industrial big data, which comprises the following steps: S1, constructing an industrial big data pool, and converging heterogeneous data sources from engineering material Internet of things sensors, project BIM design model databases, historical engineering databases and market price information platforms in a mode that the industrial big data pool is real-time or delayed by not more than a preset time threshold. According to the invention, through constructing an industrial big data pool and integrating a natural language processing model and a BIM technology, the project characteristics can be automatically analyzed and extracted, and forced verification is carried out by utilizing three-dimensional space calculation quantity, so that conventional manual intervention is obviously reduced, understanding deviation and calculation errors caused by human negligence are avoided, and meanwhile, the history data is called by an intelligent matching engine to automatically correct and complement a list, so that the normalization and the integrity of the description of the project characteristics of the list are ensured.

Inventors

LEI YASHE
ZHENG LINA
JIANG SONGLEI
FENG GUOHAI
LI XIANG
Lei deyi
DING MING
LI PEI
Zhu Gaihuan

Assignees

筑建方城(北京)建筑设计有限公司
商丘工学院

Dates

Publication Date: 20260512
Application Date: 20260202

Claims (10)

1. The automatic engineering quantity list compiling and auditing method based on industrial big data is characterized by comprising the following steps: S1, constructing an industrial big data pool, and converging heterogeneous data sources from engineering material Internet of things sensors, project BIM design model databases, historical engineering databases and market price information platforms in a mode that the industrial big data pool is real-time or delayed by not more than a preset time threshold; S2, analyzing design documents and bid-inviting documents of the project to be compiled based on a natural language processing model, extracting project feature structure trees and key process parameters, and carrying out three-dimensional space calculation quantity reference verification by combining BIM model data in an industrial large data pool to generate an initial engineering quantity definition item; S3, calling an intelligent matching engine, and performing multi-dimensional similarity matching on the initial engineering quantity clearing items and the historical engineering data in the industrial big data pool, wherein the multi-dimensional degree comprises structural feature dimensions, process engineering method dimensions and material specification dimensions; S3a, dynamically adjusting the comprehensive unit price of the history item related and recommended in the step S3 based on real-time market price information in the industrial big data pool, generating a standard unit price which accords with the current market quotation, and taking the standard unit price as a comparison standard of unbalanced quotation risk early warning in the step S4; s4, constructing a dynamic auditing rule base based on the reinforced learning model, wherein training data of the reinforced learning model comprises historical auditing records, change visa data and final settlement data, and the dynamic auditing rule base carries out real-time compliance auditing, logic contradiction detection and unbalanced quotation risk early warning on the inventory items generated in the step S3; And S5, outputting the audited final engineering quantity list, and synchronously generating a list compiling and auditing process report, wherein the report visually displays the matching source, auditing basis and risk prompt of the key item.
2. The method according to claim 1, wherein the processing of the heterogeneous data source in the step S1 includes the steps of marking the working condition context of the sensor data of the internet of things, carrying out standardized cleaning and reconstruction based on a unified project decomposition structure on the historical engineering data, and carrying out trend filtering and outlier rejection on market price information by adopting a time sequence analysis model.
3. The method of claim 1, wherein the three-dimensional space computation standard verification in step S2 is specifically that a list item obtained by natural language analysis is spatially mapped and compared with a component engineering quantity automatically extracted by a BIM model, and when the deviation exceeds a preset threshold, a manual review prompt is triggered and a deviation reason is recorded to an industrial big data pool for optimizing a subsequent analysis model.
4. The method according to claim 1, wherein the step S3 of performing the multi-dimensional similarity matching by the intelligent matching engine includes: s31, screening a candidate item set from a historical engineering database according to the characteristic structure tree of the current item; S32, carrying out semantic vectorization processing on the corresponding description of the current menu item and the candidate item set by utilizing a pre-trained engineering semantic model, and generating a matching index reflecting the semantic association degree of the current menu item and the candidate item set; The fine tuning process comprises the steps of using a text which is extracted from a historical engineering database and contains paired synonymous or near-ambiguous Cheng Miaoshu workers as positive samples and different engineering description texts which are paired randomly as negative samples, and fine tuning the model by adopting a contrast learning loss function so as to optimize vector representation of the model and maximize cosine similarity between the positive samples and minimize cosine similarity between the negative samples; s33, integrating semantic association degree matching indexes, engineering quantity proportional relations and matching degrees of a process engineering method, and outputting an integrated matching confidence degree; And S34, enabling historical data of the associated recommendation to correct and complement the current menu item only when the comprehensive matching confidence is higher than a set threshold.
5. The method according to claim 1, wherein the process of constructing the dynamic auditing rule base in step S4 includes: s41, generating a basic compliance rule through supervised learning based on decisions and modification suggestions in the historical auditing records; s42, mining potential risk association modes through unsupervised learning based on association analysis of change visa data and corresponding menu items in historical engineering to form a logic contradiction detection rule; And S43, dynamically optimizing the triggering condition of unbalanced quotation risk early warning through the reinforcement learning model based on backtracking comparison of the final settlement data and the winning bid list data.
6. The method according to claim 1, wherein the step S5 is followed by a closed loop optimization step S6 of streaming back the verified actual resource consumption data collected by the Internet of things sensor during the actual construction process and the inventory adjustment data finally determined during the engineering settlement stage as a feedback data stream to the industrial big data pool for incrementally updating the historical engineering database and training the reinforcement learning model.
7. A system for implementing the industrial big data based engineering quantity inventory automatic compiling and auditing method of any one of claims 1-6, comprising: The industrial big data aggregation and management module is used for constructing and maintaining the industrial big data pool and realizing the access, cleaning, fusion and storage of multi-source heterogeneous data; The intelligent analysis and initial programming module is integrated with a natural language processing model and a BIM interface and is used for analyzing a design document and a bid-inviting document of a project to be programmed, extracting a project characteristic structure tree and key process parameters, and carrying out three-dimensional space calculation benchmark verification by combining BIM model data to generate an initial engineering quantity definition item; The intelligent data matching and list optimizing module is integrated with the intelligent matching engine and is used for carrying out multidimensional similarity matching on the initial engineering quantity clearing items and the historical engineering data in the industrial big data pool, and adaptively correcting the engineering quantity and item feature description of the initial engineering quantity clearing items and the comprehensive unit price and resource consumption index of the associated recommended historical items according to the matching result; The dynamic intelligent auditing and risk early warning module is integrated with the dynamic auditing rule base based on the reinforcement learning model and is used for carrying out real-time compliance auditing, logic contradiction detection and unbalanced quotation risk early warning on the optimized inventory item; And the list output and visualization report module is used for outputting the audited final engineering quantity list and synchronously generating a list compiling and auditing process report.
8. The system according to claim 7, wherein the industrial big data aggregation and management module comprises a working condition context labeling unit, a data standardization cleaning unit and a market information processing unit, and the data intelligent matching and list optimizing module comprises a candidate item screening unit, an engineering semantic model unit and a confidence calculation and decision unit.
9. The system of claim 7, wherein the dynamic intelligent audit and risk early warning module includes a rule generation unit, a pattern mining unit, and a rule dynamic update unit for generating basic compliance rules based on historical audit records, mining risk association patterns based on changing visa data to form logic contradiction detection rules, and dynamically optimizing triggering conditions for unbalanced bid risk early warning based on final settlement data, respectively.
10. The system of claim 7, further comprising a system collaboration and data bus module; The industrial big data aggregation and management module, the intelligent analysis and initial compiling module, the data intelligent matching and list optimizing module, the dynamic intelligent auditing and risk early warning module and the list output and visual reporting module are all connected with the data bus module through system cooperation and data exchange; The system collaboration and data bus module is configured to: Receiving an initial engineering quantity list item from an intelligent analysis and initial compilation module, and distributing the initial engineering quantity list item and BIM verification data related to the initial engineering quantity list item to a data intelligent matching and list optimization module in an oriented manner based on a preset flow rule; Receiving optimized list items and associated historical data recommendation credibility from a data intelligent matching and list optimizing module, and triggering a dynamic intelligent auditing and risk early warning module to load a corresponding auditing rule subset for auditing; And receiving an auditing result and a risk mark from the dynamic intelligent auditing and risk early warning module, and controlling the inventory output and visual report module to generate visual reports with different detailed degrees according to the grade of the risk mark, or feeding inventory items and auditing basis of the high risk mark back to the intelligent analyzing and initial compiling module to start a manual review process.

Description

Automatic engineering quantity list compiling and auditing method and system based on industrial big data Technical Field The invention relates to the technical field of data processing and engineering cost, in particular to an automatic compiling and auditing method and system for an engineering quantity list based on industrial big data. Background The engineering quantity list is a core file for construction engineering cost management, the accuracy and the high efficiency of programming and auditing are directly related to project investment control, bidding fairness and subsequent construction and settlement, and currently, the programming and auditing of the engineering quantity list mainly depend on manual or primary informatization tools, so that two technical problems which cannot be effectively solved for a long time exist: firstly, the programming process excessively depends on manual experience, the efficiency is low and the consistency is poor, the existing method is usually implemented by manually reviewing massive design drawings and bidding documents by a manufacturing cost engineer, project characteristics are manually extracted, project quantity is calculated and rated, the process is time-consuming and labor-consuming and is limited by personal experience and business level of the engineer, deviation of drawing understanding, omission or repetition of project characteristic description and error of project quantity calculation are easily caused, BIM-based calculation software (such as Revit and widely-reached BIM calculation) and a list programming auxiliary system exist in the current market, but the tools still have significant limitations that on one hand, the tools mainly focus on automatic calculation of geometric quantity, semantic information cannot be intelligently extracted from unstructured texts (such as design specifications and technical specifications), the integrity and compliance of project characteristics still need to be manually judged, on the other hand, even if a historical database is integrated by the system, the matching mode is mainly keyword retrieval or simple classification, deep understanding of project semantics cannot be realized, meanwhile, intelligent correction and recommendation based on multi-dimensional similarity cannot be realized, and meanwhile, how the intelligent correction and correction of the intelligent reference data and the intelligent correction system are provided for the massive project data are effectively accumulated inside the projects, and the intelligent production system is difficult to be realized. Secondly, the auditing link lacks intelligent and prospective risk early warning capability, the current inventory auditing mainly relies on auditing personnel to check against norms and experiences for post-hoc manual inspection, and belongs to static and passive quality control, the current auditing system is mostly based on a fixed rule base (such as a price regulation item), or simple numerical range verification, logic contradiction hidden in the inventory (such as repeated or missing of working contents), unbalanced quotation risks which are disjointed with market price and high risk items which are predictable based on historical change data cannot be systematically found, the auditing quality is seriously dependent on personal capability and energy of auditing specialists, an accumulated and iteratable auditing knowledge system which can be automatically executed cannot be formed, and particularly, closed-loop correlation analysis cannot be carried out on subsequent change evidence data, actual consumption data and final settlement data of the project and an original inventory, so that auditing rules are dynamically optimized, and the transition from 'post-hoc correction' to 'pre-precaution' is realized. Disclosure of Invention The invention aims to provide an automatic engineering quantity list compiling and auditing method and system based on industrial big data, so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the automatic programming and auditing method for the engineering quantity list based on industrial big data comprises the following steps: S1, constructing an industrial big data pool, and converging heterogeneous data sources from engineering material Internet of things sensors, project BIM design model databases, historical engineering databases and market price information platforms in a mode that the industrial big data pool is real-time or delayed by not more than a preset time threshold; S2, analyzing design documents and bid-inviting documents of the project to be compiled based on a natural language processing model, extracting project feature structure trees and key process parameters, and carrying out three-dimensional space calculation quantity reference verification by combining BIM model data in an industria