Civil engineering cost evaluation optimization system

Abstract

The invention discloses a civil engineering cost evaluation optimization system, and relates to the technical field of civil engineering cost management. The system comprises a PLM full-life-cycle cost data closed-loop management module, a PDM cross-organization collaborative management and control module, a PLM supply chain real-time data coupling module and a PDM cost process dynamic optimization module. The PLM full life cycle module realizes standardized acquisition, storage, analysis and continuous optimization of multi-source heterogeneous data, the PDM cross-organization module ensures efficient and safe cooperation of multiple participants through a unified data standard and a three-dimensional dynamic permission model, and the PLM supply chain module completes deep fusion and dynamic correction of supply chain real-time data and cost data. And the PDM cost optimization module realizes automatic identification and accurate quantification of changes and intelligent generation and rapid examination and approval of multiple sets of optimization schemes. According to the method, cost evaluation is promoted to be transformed from static lag to dynamic real-time and from dispersed independence to collaborative closed loop, the management and control precision and efficiency are improved, and technical support is provided for efficient landing of large EPC projects.

Description

Technical Field

[0001] This invention relates to the field of civil engineering cost management technology, specifically a civil engineering cost evaluation and optimization system. Background Technology

[0002] In large-scale civil engineering EPC (Engineering, Procurement, Construction) general contracting projects such as urban rail transit, cost assessment and control span the entire project lifecycle, from planning and design to procurement, construction, and operation and maintenance. Its accuracy and efficiency directly determine the project's investment returns and implementation quality. Currently, most mainstream cost assessment systems in the industry adopt a single-module, independent operation mode. For example, some systems rely solely on PLM (Product Lifecycle Management) for full lifecycle data storage, but lack standardized data interaction mechanisms and dynamic access control in cross-organizational collaborative scenarios. Other systems handle cost changes during the construction phase through PDM (Product Data Management), but are not deeply coupled with real-time supply chain data, resulting in a lack of real-time data support for quantifying the impact of changes.

[0003] The aforementioned technological status quo directly leads to core problems in the cost assessment of large-scale EPC projects, including insufficient data linkage throughout the entire lifecycle, low efficiency of cross-organizational collaboration, delayed response to supply chain fluctuations, and inaccurate quantification of change impacts. On the one hand, due to inconsistent data formats and rigid permission allocation among multiple participants such as owners, design units, construction parties, and suppliers, "information silos" often occur, resulting in data sharing delays of up to several days and lengthy change approval cycles. On the other hand, cost assessments rely heavily on historical static data and cannot keep up with dynamic changes in the supply chain in real time. When raw material prices fluctuate by more than 10% or logistics costs rise abnormally, cost adjustments lag behind the actual scenario, easily leading to the risk of cost overruns. At the same time, the impact of changes is estimated solely based on manual experience, with an error rate as high as 15%-20%, seriously affecting the scientific and accurate nature of project cost control and becoming a key bottleneck restricting the efficient implementation of large-scale civil engineering EPC projects.

[0004] In view of the above, this application is hereby submitted. Summary of the Invention

[0005] The purpose of this invention is to provide a civil engineering cost assessment and optimization system to solve the problems mentioned in the background art.

[0006] To address the aforementioned technical issues, the civil engineering cost assessment and optimization system provided by this invention comprises a PLM (Product Lifecycle Management) full-lifecycle cost data closed-loop management module, a PDM (Product Data Management) cross-organizational collaborative control module, a PDM cost process dynamic optimization module, and a PLM supply chain real-time data coupling module. These four modules are deeply coupled to form a full-process cost control architecture. The PLM full-lifecycle cost data closed-loop management module establishes a unified data resource pool covering all stages of project planning, design, procurement, construction, and operation and maintenance, achieving a complete data collection-storage-analysis-feedback-optimization process through a data closed-loop algorithm. The PDM cross-organizational collaborative control module establishes a unified cross-organizational data standard and dynamic access control model, enabling seamless integration between multiple stakeholders and the data resource pool. The PLM supply chain real-time data coupling module captures real-time supply chain data through an API interface and updates the cost value in the data resource pool in real time based on a cost correction model. The PDM cost process dynamic optimization module identifies project changes and quantifies their impact, generates cost adjustment plans, and, after approval through a cross-organizational collaborative process, feeds them back to the data resource pool, completing the closed-loop update of cost data. The expression of the cost correction model is: ,in for Value is created through constant adjustments. To create value from the beginning The coefficient representing the influence of material prices. for Real-time material prices The initial material price, The impact coefficient of logistics costs. for Real-time logistics costs The initial logistics cost is calculated; through the deep coupling of four modules, a full lifecycle, cross-organization, real-time, and dynamic cost control system is constructed, which solves the problems of data fragmentation, inefficient collaboration, disconnect between cost and actual scenario, and delayed change response in traditional systems, and greatly improves the accuracy of cost assessment and the efficiency of full-process control.

[0007] Furthermore, the PLM full lifecycle cost data closed-loop management module includes a data acquisition submodule, a data storage submodule, a data analysis submodule, and a feedback optimization submodule. The data acquisition submodule collects design drawing data, construction plan data, historical cost data, policy and regulatory standards, and supplier basic information through standardized interfaces, and transmits them to the data storage submodule after format conversion. The data storage submodule adopts a distributed storage architecture and classifies and stores data according to project stage and data type. The data analysis submodule uses multivariate statistical analysis algorithms to extract features and mine patterns in the stored data. The feedback optimization submodule converts the analysis results into cost optimization instructions, synchronizes them to the data resource pool, and drives other modules to perform adjustments. This achieves standardized integration and full lifecycle flow of multi-source heterogeneous data, ensuring data integrity and availability, providing a solid data foundation for subsequent cross-organizational collaboration, real-time correction, and dynamic optimization, and improving data utilization efficiency and the scientific nature of cost assessment.

[0008] Furthermore, the PDM cross-organizational collaborative management module includes a data standardization submodule and a dynamic access control submodule. The data standardization submodule establishes a common data source format and data transmission protocol for multiple participants, enabling seamless integration between the PLM data resource pool and the systems of owners, design units, construction parties, and suppliers. The dynamic access control submodule constructs a data access, modification, and approval permission allocation mechanism based on project stage and user identity, ensuring the information security of the data resource pool. It solves the problem of cross-organizational data format incompatibility, achieves efficient data sharing among multiple participants, and balances collaborative efficiency and data security through precise access control, reducing the risk of sensitive information leakage.

[0009] Furthermore, the PLM supply chain real-time data coupling module includes a real-time data acquisition submodule and a data coupling processing submodule. The real-time data acquisition submodule connects to building materials e-commerce platforms, supplier ERP systems, and logistics tracking systems via API interfaces to capture raw material prices, supplier inventory status, logistics transportation costs, and production capacity fluctuation data in real time. The data coupling processing submodule substitutes the real-time data into the cost correction model to calculate the real-time cost value, synchronizes it to the PLM data resource pool, and updates the cost assessment results. This achieves dynamic linkage between cost data and the real-time status of the supply chain, solving the prediction bias problem caused by the reliance on historical data in traditional cost assessment, significantly improving the real-time performance and accuracy of cost assessment, and reducing the risk of cost overruns during the material procurement stage.

[0010] Furthermore, the PDM cost process dynamic optimization module includes a change identification submodule, an impact quantification submodule, and a scheme generation submodule. The change identification submodule receives design change requests through a cross-organizational collaboration platform and identifies the change type and trigger source by combining the anomaly monitoring results from the PLM supply chain real-time data coupling module. The impact quantification submodule calculates the scope and magnitude of the change's impact on the cost based on a change correlation analysis algorithm, and its quantification formula is as follows: ,in This represents the total cost change. For the first The value created by each related sub-project For the first The impact weight of changes to each sub-item of the project. For the first The change degree coefficient of each sub-project; the scheme generation sub-module uses an improved genetic algorithm to generate multiple cost adjustment schemes, which are then submitted to the cross-organizational collaboration platform to initiate the approval process; this enables rapid identification and accurate quantification of changes, generating scientific and reasonable cost adjustment schemes, solving the problems of traditional change handling relying on manual calculation, delayed response, and inaccurate impact assessment, and improving the efficiency of change handling and the flexibility of cost control.

[0011] Furthermore, the data acquisition submodule is equipped with a data cleaning unit, which uses an outlier detection algorithm to remove invalid data and a data completion algorithm to fill in missing data. The data analysis submodule is equipped with a data fusion unit, which uses a weighted average fusion algorithm to integrate multi-source basic data and a feature extraction algorithm to extract key cost influencing factors. This improves the quality and purity of basic data throughout the entire lifecycle, avoids interference from invalid and missing data on cost analysis results, and enhances the decision-making value of basic data through accurate data fusion and feature extraction, providing reliable data support for subsequent cost control.

[0012] Furthermore, the dynamic access control submodule constructs a three-dimensional access control model based on roles, stages, and data types. Roles include owner representatives, design engineers, construction project managers, supplier specialists, and auditors; stages include planning, design, procurement, construction, and operation and maintenance; and data types include basic cost data, sensitive cost parameters, supplier quotation data, and change approval documents. The three-dimensional access control model assigns operation permissions for corresponding data types based on the matching relationship between roles and stages. This achieves refined access control, ensuring that different participants can only access and operate necessary data at different project stages, thus guaranteeing collaborative needs while minimizing the risk of data leakage and improving system security and compliance.

[0013] Furthermore, the PLM full lifecycle cost data closed-loop management module also includes a data traceability submodule. The data traceability submodule assigns a unique identifier to each piece of cost data, records the data collection source, modification records, flow path, and associated approval information, forming a full lifecycle data traceability chain. The cost data details of any node can be queried through the identifier; this achieves full-process traceability of cost data, clarifies the data responsibility entities at each stage, reduces audit risks and the difficulty of dispute resolution, and improves the credibility and compliance of cost assessment results.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. By deeply coupling the PLM full lifecycle cost data closed-loop management module, the PDM cross-organizational collaborative control module, the PLM supply chain real-time data coupling module, and the PDM cost process dynamic optimization module, this system overcomes the limitations of existing technologies where a single module operates independently or is simply superimposed. It constructs a complete cost control system covering all stages and multiple stakeholders, fundamentally solving the core technical challenges of cost assessment in large-scale EPC projects. The PLM full lifecycle data closed-loop management module achieves standardized collection of multi-source heterogeneous data through multi-interface adaptation and data cleaning technologies. Combined with distributed hybrid storage and deep analysis algorithms, it constructs a full-cycle data resource pool. Furthermore, through closed-loop control algorithms, it achieves continuous data optimization, completely breaking down the barriers of insufficient data linkage throughout the entire lifecycle. This allows cost data to flow seamlessly and iterate dynamically at each stage, providing high-quality data support for subsequent control processes.

[0015] 2. The PDM cross-organizational collaborative management module innovatively establishes unified data standards and transmission protocols, eliminating data format differences among multiple participants. Coupled with a three-dimensional dynamic permission model based on role, stage, and data type, it achieves precise allocation and real-time adjustment of permissions. This ensures both the efficiency of cross-organizational data sharing and strengthens data security, effectively overcoming the dilemmas of information silos and inefficient collaboration. The PLM supply chain real-time data coupling module connects to multi-dimensional supply chain data sources through multiple API interfaces. It employs a weighted fusion algorithm to deeply integrate real-time data such as raw material prices, inventory status, and logistics costs with cost data. A dedicated cost correction model then dynamically updates cost value, enabling cost assessment to respond to supply chain fluctuations in real time. This completely changes the traditional reliance on static historical data in cost assessment, significantly improving the real-time nature and accuracy of cost control.

[0016] 3. The PDM cost process dynamic optimization module integrates multi-source change trigger data, and realizes automatic identification and accurate classification of changes through natural language processing and anomaly detection algorithms. It completes the scientific calculation of the impact of changes based on the correlation network model and quantitative formula, and then generates multiple optimization schemes through an improved genetic algorithm and combines them with cross-organizational collaborative approval processes for rapid implementation. This replaces the inefficient mode of traditional manual identification of changes, experience-based estimation of impact, and single-scheme approval, significantly improving the efficiency and rationality of change processing.

[0017] 4. Overall, by integrating the aforementioned unique technical means, this invention has achieved a transformation in cost assessment from static and lagging to dynamic and real-time, from decentralized and independent to collaborative and closed-loop, and from experience-driven to data-driven. This not only improves the accuracy and scientific nature of cost assessment but also optimizes cross-organizational collaboration efficiency, reduces the risk of cost overruns, and enhances the systematicness and controllability of cost management throughout the project lifecycle, providing strong technical support for the efficient implementation of large-scale civil engineering EPC projects. Attached Figure Description

[0018] Figure 1 A schematic diagram of the principle of a civil engineering cost assessment and optimization system. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 This invention provides a technical solution: a civil engineering cost assessment and optimization system. To address the problems of data fragmentation, inefficient collaboration, insufficient real-time performance, and delayed change response in traditional cost assessment systems, this system constructs a full-process control architecture with four deeply coupled core modules: a PLM (Product Lifecycle Management) closed-loop cost data management module, a PDM (Product Data Management) cross-organizational collaborative control module, a PLM supply chain real-time data coupling module, and a PDM cost process dynamic optimization module. These modules do not operate independently but are linked in real-time with a unified data resource pool through data interfaces, forming a complete control chain of "data acquisition - collaborative sharing - real-time correction - dynamic optimization - closed-loop update".

[0021] I. Application Scenarios and Existing Technological Shortcomings: This implementation method addresses the cost control needs of large-scale civil engineering EPC general contracting projects, using urban rail transit line construction projects as a specific application scenario. This project involves seven types of participants, including owners, design units, general contractors, building material suppliers, logistics companies, supervision units, and auditing units. The project cycle is 5 years, covering five stages: planning, design, procurement, construction, and operation and maintenance. It is characterized by a large volume of data, frequent cross-organizational collaboration, significant impact from supply chain fluctuations, and diverse change scenarios.

[0022] Existing publicly available technologies include a PLM-based engineering cost management system, but it only achieves full-cycle data management within a single enterprise and does not involve cross-organizational collaboration or supply chain data linkage. Publicly available documents propose a PDM-driven cost change processing method, but lack full-lifecycle data support and real-time data empowerment. Publicly available research on cross-organizational engineering project collaborative management only focuses on collaborative process optimization and does not incorporate dynamic correction and process optimization of cost data. The unique technical approach of this system lies in the deep integration of PLM's full-lifecycle data management capabilities, PDM's process control and collaboration capabilities, and the dynamic empowerment of real-time supply chain data. This breaks through the limitations of existing technologies' single-module applications or simple layering, constructing a cost control system covering multiple participants, all stages, and all data dimensions.

[0023] II. PLM Full Lifecycle Cost Data Closed-Loop Management Module: (I) Data Acquisition Submodule: Cost data for large-scale EPC projects comes from diverse sources, involving various heterogeneous data such as design drawings, construction plans, historical project data, policy standards, and supplier basic information. Existing technologies mostly rely on manual data entry or single-system integration for data collection, resulting in issues such as inconsistent formats, missing data, and low collection efficiency, leading to a lack of reliable data support for subsequent cost assessments. This step utilizes standardized interfaces and data cleaning techniques to achieve efficient acquisition and standardized processing of multi-source heterogeneous data. Specific technical methods are as follows: The data acquisition submodule employs multi-interface adaptation technology, configuring dedicated standardized interfaces for different data source types: It uses the IFC data interface to connect with the CAD / BIM system of design units, collecting data such as component dimensions, material specifications, and process requirements from the 3D model; it uses the RESTful interface to connect with the project management system of construction parties, collecting data such as construction organization design, schedule plans, and resource allocation; it uses the JDBC interface to connect with industry databases, collecting data such as historical cost data, policies, regulations, standards, and basic building material prices; and it uses the WebService interface to connect with supplier systems, collecting data such as supplier qualifications, product specifications, and basic quotations.

[0024] After data collection, standardization processing is required. Data format conversion algorithms are used to unify different data formats into JSON format. The conversion rules are as follows: For structured data (such as historical cost tables in Excel format), it is directly mapped to key-value pairs; for semi-structured data (such as BIM model data in XML format), core fields are extracted through XPath parsing; for unstructured data (such as policy documents in PDF format), OCR recognition technology is used to extract text information and then perform structure conversion.

[0025] To ensure data quality, a data cleaning unit is configured, employing a method based on... The outlier detection algorithm of the criterion removes invalid data, and the formula is as follows: in For a single data value, This is the mean of this type of data. The standard deviation of this type of data is used. Data that satisfies this formula are considered outliers and removed. For missing data, missing value imputation is performed using the K-nearest neighbor algorithm, with the following formula: ; in The missing data values ​​are the completed values. The number of nearest neighbor data. For the first The weights of the nearest neighbor data, For the first The actual value of each nearest neighbor, weight It is calculated using the reciprocal of the distance.

[0026] Example: During the design phase of an urban rail transit project, the design unit uploads BIM model data of the station's main structure to this system via the IFC interface. The data format is STEP format. The system converts it to JSON format using a format conversion algorithm, extracting core cost-related data such as frame column cross-sectional dimensions, concrete strength grade, and steel reinforcement ratio. Simultaneously, the system collects station main structure cost data from the industry's historical cost database for the past 5 years from similar urban rail transit projects via the JDBC interface. A total of 20 valid projects were collected, with one project showing a concrete unit cost of 3200 yuan / cubic meter, significantly higher than the average of 2100 yuan / cubic meter for other projects. This was calculated using the 3σ criterion. , , The data was identified as an outlier and removed. Additionally, data on the unit cost of steel reinforcement for three projects was missing; this was identified using the K-nearest neighbor algorithm. The data for the most similar items were used to calculate the completed values ​​as 850 yuan / cubic meter, 880 yuan / cubic meter, and 920 yuan / cubic meter.

[0027] Compared to existing technologies, this technical solution improves manual data collection efficiency, enhances data format uniformity, and strengthens the ability to handle abnormal and missing data, effectively solving the problems of difficult and low-quality collection of multi-source heterogeneous data. Existing publicly available documents only support single-format data collection and do not address multi-interface adaptation and data cleaning techniques. The uniqueness of this system lies in its multi-interface adaptation covering all types of data sources, combined with outlier detection and missing value completion algorithms to ensure data quality, providing high-quality basic data support for subsequent cost assessments.

[0028] (II) Data Storage Submodule: Large-scale EPC projects generate massive amounts of cost data throughout their entire lifecycle, with diverse data types including structured, semi-structured, and unstructured data. Existing technologies primarily employ a single relational database storage method, which suffers from insufficient storage capacity, low query efficiency, and difficulty in data association, failing to meet the long-term storage and rapid retrieval requirements of data throughout its entire lifecycle. This step adopts a distributed storage architecture to achieve categorized storage and efficient association of multiple data types. Specific technical methods are as follows: The data storage submodule adopts a hybrid storage architecture that combines the Hadoop Distributed File System (HDFS) and the MySQL database. HDFS is used to store unstructured data (such as policy documents in PDF format and BIM model files) and semi-structured data (such as real-time data in JSON format), while the MySQL database is used to store structured data (such as historical cost data tables and sub-item project value tables).

[0029] To enable rapid data retrieval and association, a data partitioning storage strategy is adopted, dividing the data into storage partitions based on project phases: planning phase, design phase, procurement phase, construction phase, and operation and maintenance phase. Each partition is further divided into sub-partitions based on data type: basic data, cost data, collaborative data, and change data. Simultaneously, a data indexing mechanism is established, using a B+ tree index algorithm to construct the data index table. Index fields include project number, data type, timestamp, and associated sub-project number. The index construction formula is as follows: ;in For index items, The index key consists of the project number, data type, and timestamp. This is a pointer to the data storage path.

[0030] To ensure data security, a data redundancy storage strategy is adopted, storing three copies of core cost data on different nodes. The copies are synchronized using a consistent hashing algorithm, with the following formula: ;in The hash value of the index key is used to distribute data replicas evenly across different storage nodes, ensuring that no data is lost when a single node fails.

[0031] Example: For an urban rail transit project, the steel reinforcement procurement contract (PDF format, unstructured data), real-time concrete price data (JSON format, semi-structured data), and sub-item cost calculation sheets (Excel format, structured data) generated during the construction phase are stored separately using a hybrid storage architecture: the steel reinforcement procurement contract is stored in the collaborative data sub-partition of the construction phase storage partition in HDFS, with the storage path / hdfs / construction / collaboration / contract_steel_20240510.pdf; the real-time concrete price data is stored in the cost data sub-partition of the construction phase storage partition in HDFS, with the storage path / hdfs / construction / cost / concrete_price_20240510.json; and the sub-item cost calculation sheets, after being converted to structured data, are stored in the cost data sub-partition of the construction phase storage partition in a MySQL database, with the table name t_construction_cost_item. Meanwhile, a B+ tree index is built for the sub-item cost calculation table. The index key is project number CRT202401 + data type cost_item + timestamp 20240510. The index table records the storage path pointer of this data in MySQL. When it is necessary to query the cost data of this sub-item, the storage location can be quickly located by the index key, thus shortening the query time.

[0032] Compared to existing technologies, this solution offers increased storage capacity, improved data query efficiency, and enhanced data security, achieving orderly storage and rapid association of multiple data types throughout their entire lifecycle. Current publicly available research on engineering cost data storage and management technologies only uses a single relational database to store structured data, without addressing distributed or hybrid storage architectures. The uniqueness of this system lies in its use of a hybrid storage architecture combining HDFS and MySQL to adapt to diverse data storage needs, combined with partitioned storage and B+ tree indexing algorithms to improve query efficiency, and redundant storage to ensure data security, thus solving the technical challenges of large-scale cost data storage and retrieval.

[0033] (III) Implementation of the Data Analysis Submodule: Existing technologies for analyzing cost data mostly employ simple statistical methods, only capable of basic analyses such as summation and mean calculation. They fail to uncover potential correlations and patterns of change among data points, resulting in a lack of in-depth data support for cost assessments and hindering accurate prediction and optimization. This step utilizes multivariate statistical analysis and machine learning algorithms to achieve in-depth mining and pattern extraction of cost data. Specific technical methods are as follows: The data analysis submodule uses a multiple linear regression algorithm to construct a cost prediction model, which is used to predict the initial cost of each sub-item of the current project based on historical data. The model formula is as follows: ;in To create value for sub-projects For constant terms, For regression coefficients, Key factors affecting construction costs (such as component dimensions, material specifications, construction techniques, market prices, etc.) This represents the random error term. The regression coefficients are solved using the least squares method, with the following formula: ;in These are the estimated values ​​of the regression coefficients. For the influence factor matrix, Create value vectors for history.

[0034] The Apriori algorithm was used to mine the association relationships between different sub-projects and between influencing factors and creation value. The minimum support was set to 20%, and the minimum confidence was set to 60%. The association rules were represented as follows: ,in For example, using C40 concrete. For subsequent components (such as increased unit weight of steel reinforcement), support Indicates simultaneous inclusion and Data percentage, confidence level Indicates inclusion The data also contains The percentage of data.

[0035] Example: In the design phase of an urban rail transit project, it is necessary to predict the initial cost of the station's main structure. Five influencing factors are selected as independent variables: component cross-sectional dimensions, concrete strength grade, steel reinforcement ratio, construction area, and construction season. We collected cost data for individual components of 20 similar historical projects as the dependent variable. The regression coefficients are solved using a multiple linear regression algorithm to obtain the prediction model: ;in The dimensions of the frame column cross-section (unit: square meters) are shown. For concrete strength grades (C30 is 3, C40 is 4, C50 is 5). Reinforcement ratio (unit: %) For construction areas (1 for first-tier cities, 0 for second-tier cities). The construction season is represented by 1 for winter and 0 for other seasons. The current project's impact factor values ​​are then substituted into the model. , , , , The initial value of the building was calculated. Yuan per square meter.

[0036] By mining association rules using the Apriori algorithm, it was found that when the concrete strength grade is C40 ( And the construction area is a first-tier city ( When the reinforcement ratio is ≥2.5%, the steel reinforcement ratio is ≥2.5%. The support rate is 25% and the confidence rate is 70%. This association rule indicates that under this condition, the amount of steel reinforcement is likely to increase, providing data support for subsequent cost optimization.

[0037] Compared to basic statistical analysis using existing technologies, this step achieves in-depth mining and accurate prediction of cost data, revealing key factors and correlation patterns affecting costs, and providing a scientific basis for cost assessment and optimization. Existing publicly available documents only use simple weighted average algorithms for cost analysis, without involving multiple linear regression and association rule mining. The uniqueness of this system lies in constructing a predictive model through machine learning algorithms and combining it with association rule mining to extract data patterns. This solves the problems of insufficient depth and low prediction accuracy of traditional analysis methods, improving the scientific rigor and accuracy of cost assessment.

[0038] (iv) Feedback Optimization Submodule: Existing cost data flows are mostly unidirectional transmissions, and data analysis results cannot be fed back to the front-end data collection and subsequent cost control stages in a timely manner, resulting in a broken data loop and making it difficult to achieve continuous cost optimization. This step constructs a data feedback optimization mechanism, transforming data analysis results into optimization instructions to drive adjustments in each module, forming a data loop. Specific technical means are as follows: The feedback optimization submodule adopts a closed-loop control algorithm. Based on the data analysis results and the preset cost optimization goals, such as cost reduction and accuracy improvement, it generates corresponding optimization instructions, including data acquisition optimization instructions, cost model adjustment instructions, and collaborative process optimization instructions.

[0039] The data collection optimization command is used to adjust the frequency and scope of data collection. When the correlation of a certain type of influencing factor is higher than 50%, the collection frequency of this type of data is increased from once a week to once a day. When the missing rate of a certain type of data is lower than 5%, the collection scope of this type of data is narrowed, and only the core fields are retained.

[0040] The cost model adjustment command is used to update the parameters of the cost prediction model. When the prediction error exceeds a preset threshold (10%), the regression coefficients are recalculated using a rolling update algorithm. The formula is as follows: ;in For the first The regression coefficients after the next update. This is the smoothing coefficient (value 0.7). For the first The regression coefficient of the second order. These are the regression coefficients calculated based on the newly added data.

[0041] The collaborative process optimization command is used to adjust the permissions and processes for cross-organizational collaboration. When the access frequency of a certain type of data is higher than 10 times per day, the access approval process for that type of data is simplified and the approval nodes are shortened.

[0042] Example: During the construction phase of an urban rail transit project, data analysis revealed a correlation of 65% with steel reinforcement prices, exceeding the preset threshold of 50%. The feedback optimization submodule generated a data collection optimization instruction, adjusting the collection frequency of steel reinforcement price data from once a week to once a day. Simultaneously, the cost model predicted the cost of a certain sub-item to be 1.2 million yuan, while the actual calculated value was 1.35 million yuan, resulting in a prediction error of 12.5%, exceeding the preset threshold of 10%. A rolling update algorithm was then used to update the regression coefficients. , This represents the regression coefficient vector during the design phase. The regression coefficient vector was calculated based on 10 new sets of data added during the construction phase. The updated regression coefficients reduced the subsequent prediction error to 8%. In addition, it was found that the average daily call frequency of basic cost data was 15 times. A collaborative process optimization instruction was generated to simplify the approval process for this type of data from 3 nodes to 2 nodes, shortening the approval time.

[0043] Compared to the unidirectional data flow of existing technologies, this step achieves closed-loop data management and continuous optimization, enabling cost data to be continuously iterated and improved throughout its entire lifecycle, thus enhancing the dynamic adaptability of cost control. No publicly available technology has yet proposed a closed-loop feedback mechanism for cost data. The uniqueness of this system lies in its use of a closed-loop control algorithm to transform analysis results into optimization instructions, driving continuous adjustments at each stage. This solves the problems of data flow fragmentation and inability to continuously optimize in traditional systems, forming a closed-loop optimization system for cost data throughout its entire lifecycle.

[0044] III. PDM Cross-Organizational Collaborative Management and Control Module: (I) Implementation of the Data Standardization Submodule: Large-scale EPC projects involve numerous participants, whose systems have different data formats and transmission protocols. This leads to problems such as format incompatibility, data loss, and transmission failures when sharing data across organizations. Existing technologies lack a unified cross-organizational data standard, making efficient collaboration difficult. This step establishes a unified data standard to achieve seamless integration between the systems of all participating parties and this system. Specific technical means are as follows: The data standardization submodule establishes three core standards: data source format standards, data transmission protocol standards, and data interface specification standards.

[0045] The data source format standard stipulates that all data transmitted across organizations should be in JSON format, and clearly defines the field definitions and data types of each type of data: basic cost data includes fields such as project number (string), sub-item number (string), cost value (numerical), and calculation time (timestamp); supplier data includes fields such as supplier number (string), product name (string), specifications (string), and quotation (numerical); change data includes fields such as change number (string), change type (string), associated sub-item (string), and change content (string).

[0046] The data transmission protocol standard uses HTTPS for data transmission, sets the transmission timeout to 30 seconds, the minimum data transmission rate to 1 Mbps, and employs a data fragmentation algorithm to handle large files (files exceeding 100 MB). The fragment size is set to 10 MB, and the fragmentation formula is as follows: , ;in For the original file, For fragmented files, This is the original file size. For the size of the slice, This is the floor function.

[0047] The data interface specification standard stipulates the interface parameters, request methods, and response formats for each participating system to interface with this system. Interface parameters include identity authentication parameters (appID, token) and data query parameters (project number, time range). The request method adopts POST request, and the response format includes status code (200 for success, 400 for parameter error, and 500 for server error), response information, and data content.

[0048] For urban rail transit projects, Supplier A needs to transmit steel reinforcement quotation data to this system, and organize the data into JSON format according to the data source format standard: { "supplierNo":"S2024001", "productName":"HRB400 steel bar", "specification":"25mm", "price":"4200", "timeStamp":"20240510143000" }

[0049] The file is 150MB in size and, according to the data transmission protocol standard, is divided into 15 fragments (each fragment being 10MB) using a fragmentation algorithm. It is transmitted to this system via HTTPS. Supplier A's system sends a POST request through the standard data interface specification, with request parameters including appID="supplierA2024", token="f892d37a6b1e4c988765d231a0bcdef", and projectNo="CRT202401". After verifying the parameters are valid, this system receives the fragmented file, reassembles it, and responds with a status code of 200 and the message "Data transmission successful".

[0050] During the data transmission of BIM models by design firms, core fields are extracted according to a unified format standard to avoid data loss due to format differences.

[0051] Compared to existing technologies, this system offers improved compatibility, higher transmission success rates, and lower data loss rates for cross-organizational data sharing, effectively addressing the issue of inconsistent data formats among multiple participants. Existing publicly available research on cross-organizational engineering project collaborative management only proposes optimization of collaborative processes without addressing the establishment of unified data standards. The uniqueness of this system lies in its ability to achieve seamless integration of systems from all participants through the establishment of multi-dimensional data standards, laying the foundation for cross-organizational collaboration and improving the efficiency and reliability of collaborative data sharing.

[0052] (II) Dynamic Access Control Submodule: In cross-organizational collaboration, different participants have varying access needs to cost data at different project stages. Existing technologies mostly employ static access control, requiring manual adjustments. This leads to issues such as unreasonable access allocation, high risk of data leakage, and delayed access adjustments, making it difficult to balance collaboration efficiency and data security. This step constructs a three-dimensional dynamic access control model to achieve precise access allocation and real-time adjustment. Specific technical methods are as follows: The dynamic access control submodule constructs a three-dimensional access control model based on roles, stages, and data types. It uses an access control matrix algorithm to define the access mapping relationships for each dimension. The access control matrix is ​​represented as follows: in For role dimensions (values ​​1-5, corresponding to owner representative, design engineer, construction project manager, supplier specialist, and auditor). The phase dimension (values ​​1-5, corresponding to the planning phase, design phase, procurement phase, construction phase, and operation and maintenance phase). For data type dimension (values ​​1-4, corresponding to basic cost data, sensitive cost parameters, supplier quotation data, and change approval documents), The values ​​represent permissions: 0 indicates no permission, 1 indicates query permission, 2 indicates modification permission, and 3 indicates approval permission.

[0053] Permission allocation uses an automatic matching algorithm. Based on the user's login role and the current project stage, it extracts the corresponding permission value from the permission matrix and automatically assigns operation permissions. When the project stage changes or the user role is adjusted, a real-time permission update algorithm is used to complete the permission adjustment within 3 seconds. The update formula is as follows: ;in For the new character identifier, This marks the beginning of a new phase. For permission change matrix, This is a matrix XOR operation.

[0054] To ensure access security, an access auditing algorithm is used to record all users' access operation logs, including the operator, operation time, operation type, and access data type. The audit logs are retained for 3 years. When an unauthorized operation is detected, an alarm mechanism is automatically triggered and the user's access is frozen.

[0055] Example: In the procurement phase of an urban rail transit project, the supplier specialist (role) Log in to the system; current project stage. (Procurement phase), based on the permission matrix, basic cost data ( Permission values (Query permissions), sensitive cost parameters ( Permission values (No permission), Supplier quotation data ( Permission values (Modify permissions), Change approval documents ( Permission values (Query Permissions) The system automatically assigns corresponding operation permissions to this user, allowing them to only query basic cost data and change approval documents, modify their own supplier quotation data, and not access sensitive cost parameters.

[0056] When a project moves from the procurement phase to the construction phase ( (From 3 to 4) The real-time permission update algorithm automatically adjusts the permission matrix. The supplier specialist's permission value for supplier quotation data changes from 2 to 1, retaining only query permissions and no longer able to modify quotation data. The permission adjustment is completed within 2 seconds. During the audit, a construction project manager was detected attempting to access sensitive cost parameters (permission value 0). The system automatically triggered an alarm and froze their permissions. Furthermore, compared to the static permission allocation of existing technologies, this step achieves precise and dynamic permission control, improves permission adjustment efficiency, reduces the risk of data leakage, and balances cross-organizational collaboration efficiency with data security. Existing public documents use a single-dimensional permission allocation based on roles, without addressing the dimension division of stages and data types. The uniqueness of this system lies in constructing a three-dimensional permission model, combined with automatic matching and real-time update algorithms, to achieve dynamic adaptation of permissions. This solves the problems of insufficient flexibility and low security in traditional permission management, providing security guarantees for cross-organizational collaboration.

[0057] IV. PLM Supply Chain Real-Time Data Coupling Module: (I) Implementation of the Real-Time Data Acquisition Submodule: Civil engineering costs are significantly affected by the supply chain. Fluctuations in raw material prices, supplier inventory, and logistics costs directly lead to cost changes. Existing technologies often rely on periodic manual updates of supply chain data, which suffers from significant data lag and untimely updates, resulting in a disconnect between cost assessments and actual scenarios, and a high risk of cost overruns. This step utilizes real-time data acquisition technology to achieve dynamic acquisition of supply chain data. Specific technical methods are as follows: The real-time data acquisition submodule uses API interface technology to connect to three types of supply chain data sources: building materials e-commerce platforms (such as a building materials information platform), supplier ERP systems (such as SAP systems), and logistics tracking systems (such as a logistics tracking platform).

[0058] The system connects to building materials e-commerce platforms using a REST API interface, with a data collection frequency of once per hour. The core data collected includes raw material names, specifications, real-time prices, and price fluctuation ranges. The system connects to supplier ERP systems using a WebService API interface, with a data collection frequency of once every 2 hours. The collected data includes supplier inventory quantities, capacity utilization rates, and delivery cycles. The system connects to logistics tracking systems using an HTTP API interface, with a data collection frequency of once every 6 hours. The collected data includes transportation routes, transportation distances, transportation costs, and estimated delivery times.

[0059] To ensure the stability of data acquisition, a breakpoint resumption algorithm is employed. When data transmission is interrupted, the interruption point is recorded, and acquisition resumes from the breakpoint upon reconnection, preventing data duplication or loss. A data verification algorithm is used to verify the integrity of the acquired data; the verification formula is as follows: ;in For checksum, For the first The values ​​of each data field, The number of data fields is specified. If the checksum calculated by the receiver matches that of the sender, the data is considered complete; otherwise, a retransmission is requested.

[0060] Example: During the construction phase of an urban rail transit project, the real-time data acquisition submodule connects to a building materials information platform via a REST API interface to collect real-time prices of concrete, steel bars, and waterproofing materials hourly. At 10:00 AM on May 10, 2024, the real-time price of C40 concrete was collected as 580 yuan / cubic meter, an increase of 10 yuan / cubic meter compared to the previous hour, representing a price fluctuation of 1.75%. The submodule also connects to supplier B's SAP system via a WebService API interface to collect data on the inventory quantity of HRB400 Φ25mm steel bars as 500 tons, with a capacity utilization rate of 85% and a delivery cycle of 7 days. Finally, the submodule connects to SF Express's logistics tracking system via an HTTP API interface to collect data on the transportation route of this batch of steel bars: steel mill A to the construction site, a transportation distance of 300 kilometers, a transportation cost of 20 yuan / ton, and an estimated arrival time of May 17, 2024.

[0061] During a data collection process, a network outage caused a disruption in logistics data transmission. The breakpoint resume algorithm recorded the interruption point as the transportation cost field. After the connection was restored, only the transportation cost and subsequent fields were re-collected, avoiding a full data retransmission. During data verification, the sender calculated a checksum of 125, and the receiver's calculation result was consistent, thus determining that the data was complete.

[0062] Compared to the periodic manual updates of existing technologies, this step achieves real-time collection and automatic updates of supply chain data, reducing data lag, improving collection efficiency, and effectively solving the problem of disconnect between cost assessment and supply chain status. Existing publicly available documents only collect raw material price data and do not cover multi-dimensional supply chain data such as inventory and logistics. The uniqueness of this system lies in its multi-API interface to connect to full-dimensional supply chain data sources, combined with breakpoint resume and data verification algorithms to ensure data quality, providing comprehensive and timely data support for real-time cost correction.

[0063] (II) Data Coupling Processing Submodule: Existing technologies only treat supply chain data as an independent reference factor, failing to deeply integrate it with PLM's full lifecycle cost data. This results in supply chain data being unable to effectively empower cost assessment, and cost adjustments lacking accurate basis. This step employs a data coupling algorithm to integrate real-time supply chain data with cost data in the PLM data resource pool, achieving dynamic cost correction. Specific technical methods are as follows: The data coupling processing submodule uses a weighted fusion algorithm to fuse multi-source supply chain data with cost data. The fusion formula is as follows: ;in For the merged data, For raw material price data, For supplier inventory data, For logistics cost data, Let be the weighting coefficient, satisfying The weighting coefficients were determined using the analytic hierarchy process (AHP). Raw material price data had a weight of 0.6, supplier inventory data had a weight of 0.2, and logistics cost data had a weight of 0.2.

[0064] Based on the fused data, the initial cost value is updated in real time using the cost correction model in claim 1. The model formula is as follows: ;in for Value is created through constant adjustments. To create value from the beginning This is the material price impact coefficient (value 0.7). for Real-time material prices The initial material price, The logistics cost impact coefficient (value 0.3) for Real-time logistics costs This represents the initial logistics costs.

[0065] When the fluctuation of supply chain data exceeds the preset threshold (±5%), the emergency cost correction mechanism is triggered, shortening the correction cycle from once per hour to once every 15 minutes, ensuring that the cost is updated in a timely manner to keep up with data fluctuations.

[0066] Example: The initial cost of steel reinforcement procurement for a specific sub-project of an urban rail transit project. 10,000 yuan, initial material price Yuan / ton, initial logistics cost Yuan / ton, material price impact coefficient Logistics cost impact coefficient . Real-time supply chain data collected at 10:00 AM on May 10, 2024: Real-time material prices Yuan / ton, real-time logistics cost The data for the unit price per ton fluctuated by 5% and 33.3% respectively, both exceeding the preset thresholds and triggering an emergency correction mechanism. First, a weighted fusion algorithm was used to calculate the fused data: (here) Take the standardized value of the inventory quantity of 500 tons. Then substitute it into the cost correction model to calculate the corrected cost value: The revised cost estimate is 10,000 yuan. This revised cost estimate is synchronized to the PLM data resource pool, allowing all stakeholders to view it in real time through a cross-organizational collaboration platform. The construction team can then adjust its procurement plan based on the revised cost estimate to avoid cost overruns.

[0067] Compared to the separate application of supply chain data and cost data in existing technologies, this step achieves deep coupling and real-time linkage between the two types of data. This improves the accuracy of cost correction, reduces the risk of cost overruns, and solves the problem of traditional cost assessment lagging behind supply chain fluctuations. No publicly available technology couples PLM lifecycle data with real-time supply chain data. The uniqueness of this system lies in its use of a weighted fusion algorithm to achieve multi-source data fusion, combined with a cost correction model to dynamically update cost value. This enables cost assessment to respond to supply chain changes in real time, improving the real-time nature and accuracy of cost control.

[0068] V. Dynamic Optimization Module for PDM Cost Estimation Process: (I) Change Identification Submodule: Large-scale EPC projects involve numerous change scenarios throughout their entire lifecycle, such as design optimization, changes in geological conditions, and supply chain anomalies. Existing technologies mostly rely on manual reporting to identify changes, which suffers from problems such as identification lag, high omission rates, and inaccurate judgment of change types, resulting in the inability to timely control the impact of changes on costs. This step constructs an automatic change identification mechanism to achieve rapid identification and accurate classification of changes. Specific technical means are as follows: The change identification submodule uses a multi-source data fusion identification algorithm to integrate three types of change trigger source data: change application data submitted across organizational collaboration platforms, abnormal data from the PLM supply chain real-time data coupling module, and historical change data from the PLM data resource pool.

[0069] Keywords from change request data were extracted using a Natural Language Processing (TF-IDF) algorithm. These keywords included design changes, geological changes, material replacements, and process adjustments. The change type was determined based on the keyword matching degree, calculated using the following formula: ; in For keyword matching degree, For the set of keywords in the change application data, For the preset set of keywords for change types, Keywords The weight value is used to determine the corresponding change type if the matching degree is higher than 0.6.

[0070] Anomaly detection algorithms are used to identify abnormal changes in the supply chain. When the fluctuation range of supply chain data exceeds ±10%, it is determined to be an abnormal change in the supply chain. Comparative analysis algorithms are used to identify design changes. The current design data is compared with the historical version data. If the difference is higher than 5%, it is determined to be a design change.

[0071] Example: During the construction phase of an urban rail transit project, the design unit submitted a change request through a cross-organizational collaboration platform. The request changed the structural form of the station transfer passage from rectangular to circular to improve structural stability. The change identification submodule used the TF-IDF algorithm to extract the keywords of the request: transfer passage, structural form, rectangle, circle, and structural stability. The matching degree with the preset set of design change keywords was calculated to be 0.75, which is higher than the threshold of 0.6, thus it was judged as a design change. At the same time, the PLM supply chain real-time data coupling module monitored the real-time price of waterproof materials as 120 yuan / square meter, a 20% increase compared to the initial price of 100 yuan / square meter, with a fluctuation exceeding 10%, thus it was judged as an abnormal supply chain change. By comparing the current construction drawings with the original drawings through a comparative analysis algorithm, it was found that the burial depth of a certain section of the tunnel was adjusted from 15 meters to 18 meters, a difference of 20%, thus it was judged as a design change. All three types of changes were quickly identified and classified, and synchronized to the subsequent stages of the PDM cost process dynamic optimization module.

[0072] Compared to existing manual change identification technologies, this step achieves automatic identification and accurate classification of changes, improving identification efficiency, reducing omission rates, and increasing the accuracy of change type judgment. It solves the problems of lagging change identification and untimely control in traditional methods. Existing publicly available documents only identify changes through single change requests, without involving multi-source data fusion for identification. The uniqueness of this system lies in integrating multiple types of change trigger source data and employing natural language processing and anomaly detection algorithms to achieve change identification, improving the comprehensiveness and accuracy of change identification and laying the foundation for subsequent change impact quantification.

[0073] (II) Impact Quantification Submodule: Existing technologies for quantifying the impact of changes mostly rely on empirical estimation methods, lacking scientific data support. This leads to inaccurate judgments of the scope of impact and large deviations in the calculation of the magnitude of impact, failing to provide a reliable basis for cost adjustments. This step employs mathematical modeling and algorithm analysis to achieve accurate quantification of the impact of changes on costs. Specific technical methods are as follows: The impact quantification submodule uses a change correlation analysis algorithm to identify affected sub-projects. Based on the sub-project correlation data in the PLM data resource pool, it constructs a correlation network model, where nodes represent sub-projects and edges represent correlations (such as process correlations and material correlations). The shortest path algorithm is used to calculate the correlation degree between changes and each sub-project. The correlation degree formula is: ; in For relevance, To change the shortest path length between nodes and sub-project nodes, The maximum path length is used; items with a correlation coefficient higher than 0.3 are considered affected sub-projects. The change impact quantification formula is used to calculate the total cost change: ;in This represents the total cost change. For the first The value created by each affected sub-project For the first The impact weight of changes to each sub-item project (determined by the analytic hierarchy process, with a process-related weight of 0.6 and a material-related weight of 0.4). For the first The degree of change coefficient for each sub-item project (0.1-0.3 for design changes, 0.05-0.2 for abnormal supply chain changes, and 0.2-0.4 for geological changes).

[0074] Example: Design changes to an urban rail transit project, including adjustments to the structural form of transfer passages, affect the quantitative sub-module. First, a network model of the inter-item engineering relationships is constructed. The change node is the construction of the transfer passage structure. Using the shortest path algorithm, the affected inter-items are calculated to include the reinforcement, concrete, formwork, and waterproofing works of the transfer passage, with correlation degrees of 0.9, 0.85, 0.7, and 0.6 respectively, all exceeding the threshold of 0.3. The construction value of each affected inter-item is then calculated. The changes in the amounts of 200,000 yuan, 300,000 yuan, 100,000 yuan, and 150,000 yuan are weighted accordingly. All are 0.6 (process-related), change degree coefficient All are 0.2 (design change). Substitute into the quantitative formula to calculate the total cost change: The calculation results show that the design change increased the total cost by 90,000 yuan, providing a precise basis for the generation of subsequent cost adjustment plans.

[0075] Compared to empirical estimations in existing technologies, this step achieves precise quantification of the impact of changes, improves the accuracy of impact scope identification, reduces the deviation in impact magnitude calculation, and solves the problems of unscientific and inaccurate traditional change impact assessments. Among existing publicly available technologies, none have used mathematical models to quantify the impact of changes on costs. The uniqueness of this system lies in identifying affected sub-projects through a relational network model and calculating the total cost change using quantitative formulas, thus providing scientific data support for change impact assessments and improving the rationality and accuracy of cost adjustments.

[0076] (III) Solution Generation Submodule: Existing technologies often employ a single cost adjustment solution after changes occur, lacking multi-solution comparison and optimization. Furthermore, the solution approval process is cumbersome and inefficient, hindering the rapid implementation of cost adjustments and impacting project progress. This step utilizes intelligent algorithms to generate multiple optimized solutions and integrates them with cross-organizational collaborative processes for rapid approval. Specific technical methods are as follows: The scheme generation submodule uses an improved genetic algorithm to generate multiple cost adjustment schemes, with the optimization objective being to minimize the total cost change. The constraints include the feasibility of construction technology, the stability of material supply, and the compliance of quality standards.

[0077] The implementation steps of the improved genetic algorithm are as follows: 1. Encoding: A binary encoding method is used, with each chromosome representing a cost adjustment scheme. The gene position corresponds to the cost adjustment ratio of the sub-project (0.8-1.2). 2. Population initialization: Generate 50 initial chromosomes, with a population size of 50. 3. Fitness function: ,in The total cost change is represented by the fitness value, indicating a better solution. 4. Selection: A roulette wheel selection algorithm is used to select the parent chromosome; the selection probability is proportional to the fitness value. 5. Crossover: A single-point crossover operator is used, with a crossover probability of 0.8. 6. Mutation: A bit mutation operator is used, with a mutation probability of 0.05. 7. Iteration Termination: The iteration terminates when the number of iterations reaches 100 generations or the fitness value converges (the change rate is less than 1% for 10 consecutive generations), outputting the 3 solutions with the highest fitness values.

[0078] The scheme approval process adopts a cross-organizational collaborative approval mechanism, which is processed through the collaborative platform of the PDM cross-organizational collaborative management module. It follows the process of "the initiator of the change submits the scheme - the relevant parties review it - the auditor reviews it - the owner confirms it". The review opinions are confirmed by electronic signature, and the approval results are fed back to the scheme generation sub-module in real time.

[0079] Example: A design change to an urban rail transit project, involving adjustments to the transfer passage structure, resulted in a total cost change of 90,000 yuan. The scheme generation submodule used an improved genetic algorithm to generate three cost adjustment schemes: Option 1: Adjust the cost of steel reinforcement works by 30,000 yuan, concrete works by 40,000 yuan, formwork works by 10,000 yuan, and waterproofing works by 10,000 yuan, for a total change of 90,000 yuan. Option 2: Replace ordinary steel bars with high-strength steel bars. This will increase the cost of steel reinforcement by 20,000 yuan, concrete by 30,000 yuan, formwork by 15,000 yuan, and waterproofing by 25,000 yuan, for a total change of 90,000 yuan. At the same time, it will improve the durability of the structure. Option 3: Optimize the construction process. The cost of steel reinforcement will increase by 25,000 yuan, concrete engineering by 35,000 yuan, formwork engineering by 5,000 yuan, and waterproofing engineering by 25,000 yuan, with a total change cost of 90,000 yuan. The construction period will be shortened by 5 days.

[0080] The three plans were approved through a cross-organizational collaborative platform: the construction party reviewed and found that the process optimization of Plan 3 was more in line with the construction schedule requirements, the supplier reported that the supply of high-strength steel bars for Plan 2 was sufficient, the auditor confirmed that all three plans complied with the cost specifications, the owner finally selected Plan 3, the approval process was completed within 24 hours, and the plan generation submodule fed the final plan back to the PLM data resource pool to complete the closed-loop update of cost data.

[0081] Compared to the single-solution approach and cumbersome approval processes of existing technologies, this step enables the intelligent generation and efficient approval of multiple optimized solutions. This enhances the rationality and feasibility of the solutions, improves approval efficiency, and solves the problems of traditional change management solutions being singular and slow to implement. Existing publicly available documents only generate single change solutions and do not involve intelligent optimization algorithms or cross-organizational collaborative approval. The uniqueness of this system lies in generating multiple optimized solutions through an improved genetic algorithm, combined with cross-organizational collaborative processes to achieve rapid approval. This makes cost adjustment solutions more scientific, implementation more efficient, and improves the flexibility and timeliness of change management.

[0082] VI. Example of a complete system integration and operation process: The system integration and operation process of urban rail transit EPC projects from the planning stage to the operation and maintenance stage is as follows: During the planning phase, the PLM full lifecycle cost data closed-loop management module collects historical cost data and policy standard data from similar projects through the data acquisition submodule. The data analysis submodule uses a multiple linear regression algorithm to predict the overall initial project cost to be 5 billion yuan. The feedback optimization submodule generates data acquisition optimization instructions to increase the frequency of collecting key material price data. The PDM cross-organizational collaborative management module establishes cross-organizational data standards. Owners and design units share planning scheme data through a collaborative platform, and the dynamic access control submodule assigns query and approval permissions for basic cost data to owner representatives.

[0083] During the design phase, the design unit uploads BIM model data through a standardized interface. The PLM data storage submodule stores the data in partitions according to the design phase. The data analysis submodule discovers the correlation between concrete strength and steel reinforcement usage through association rule mining, optimizing the initial cost model. The PDM cost process dynamic optimization module identifies a design change (adjustment to the transfer passage structure). The impact quantification submodule calculates a total cost change of 90,000 yuan. The scheme generation submodule generates three optimization schemes. After cross-organizational collaborative approval, Scheme 3 is selected, and the cost value in the PLM data resource pool is updated.

[0084] During the procurement phase, the PLM supply chain real-time data coupling module collects real-time prices of materials such as steel bars and concrete every hour. The data coupling processing submodule uses a cost correction model to adjust the cost value during the procurement phase in real time. When the price of steel bars increases by 20%, an emergency correction mechanism is triggered, and the corrected cost value is synchronized to the collaboration platform, allowing the construction party to adjust its procurement plan. The PDM cross-organizational collaborative management module assigns supplier specialists the authority to modify quotation data. Suppliers submit quotations through the collaboration platform, and the dynamic access control submodule ensures that sensitive cost parameters are not leaked.

[0085] During the construction phase, the PLM full lifecycle cost data closed-loop management module continuously collects construction progress data and cost accounting data. The data analysis submodule continuously updates the cost prediction model, reducing the prediction error to 8%. The PLM supply chain real-time data coupling module connects to the logistics tracking system to update the material transportation status in real time, ensuring timely cost adjustments. The PDM cost process dynamic optimization module identifies abnormal changes in the supply chain (such as an increase in waterproof material prices), quickly generates cost adjustment plans, and completes approval to avoid cost overruns.

[0086] During the operation and maintenance phase, the PLM data resource pool continuously collects operation and maintenance energy consumption and repair cost data. The data analysis submodule analyzes the patterns of operation and maintenance cost changes, and the feedback optimization submodule generates operation and maintenance resource configuration optimization instructions to reduce operation and maintenance costs. The data traceability submodule records cost data changes throughout the entire lifecycle, and the auditing unit can query the traceability chain through the collaborative platform to ensure the compliance of cost data.

[0087] Throughout the project lifecycle, the four modules are deeply coupled and operate collaboratively, enabling full lifecycle, cross-organizational, real-time, and dynamic cost control. This reduces the total cost of the project throughout its lifecycle, improves the accuracy of cost assessment, enhances cross-organizational collaboration efficiency, and accelerates change response speed, effectively addressing various pain points of traditional cost control.

[0088] In summary, the unique technical approach of this system lies in breaking through the limitations of existing technologies that rely on single-module applications or simple layering. It constructs an architecture system with deep coupling of four major modules: PLM full lifecycle data management, PDM cross-organizational collaboration, PLM supply chain real-time coupling, and PDM process dynamic optimization, forming a complete closed loop for cost control.

[0089] This system adds a new module for cross-organizational collaborative management and control and real-time supply chain data coupling, enabling multi-participant collaboration and real-time data empowerment, solving the problem that single-enterprise internal management cannot cover cross-organizational scenarios; this system integrates PLM full lifecycle data support and supply chain real-time data linkage, enabling change processing to have a more accurate data foundation, solving the problem of lack of full-dimensional data support for change optimization; this system realizes the integrated application of technologies such as unified data standards, dynamic access control, data coupling correction, and intelligent solution generation, solving the technical challenges of cross-organizational collaboration and deep data integration.

[0090] This system achieves high-quality collection of multi-source heterogeneous data through multi-interface adaptation and data cleaning technology, adopts a distributed hybrid storage architecture to achieve efficient management of large-scale data, combines machine learning algorithms to achieve in-depth data mining and accurate prediction, constructs a three-dimensional dynamic permission model to ensure cross-organizational collaborative security, achieves real-time linkage between supply chain data and cost data through data coupling algorithms, and adopts intelligent optimization algorithms to achieve accurate quantification of the impact of changes and optimization of solutions, forming a series of non-obvious technical solutions.

[0091] These technical solutions not only overcome the shortcomings of traditional cost assessment systems, such as data fragmentation, inefficient collaboration, insufficient real-time performance, and delayed change response, but also achieve full lifecycle coverage of cost control, cross-organizational collaboration, and real-time dynamic optimization, resulting in significant technical effects. They meet the cost control needs of large-scale EPC projects and promote the development of civil engineering cost assessment technology.

Claims

1. A civil engineering cost assessment and optimization system, characterized in that: It includes a PLM full lifecycle cost data closed-loop management module, a PDM cross-organizational collaborative control module, a PDM cost process dynamic optimization module, and a PLM supply chain real-time data coupling module; The PLM full lifecycle cost data closed-loop management module builds a unified data resource pool covering the entire stages of project planning, design, procurement, construction, and operation and maintenance. It realizes the full-process flow of data collection, storage, analysis, feedback, and optimization through data closed-loop algorithms. The PDM cross-organizational collaborative management module establishes a unified data standard and dynamic access control model across organizations, enabling seamless integration between multiple stakeholders and the data resource pool. The PLM supply chain real-time data coupling module captures real-time supply chain data through API interfaces and updates the creation value in the data resource pool in real time based on the cost correction model. The PDM cost estimation process dynamic optimization module identifies project changes and quantifies their impact, generates cost adjustment plans, and after approval through cross-organizational collaborative processes, feeds them back to the data resource pool, completing the closed-loop update of cost data; the expression for the cost correction model is: ,in for Value is created through constant adjustments. To create value from the beginning The coefficient representing the influence of material prices. for Real-time material prices The initial material price, The impact coefficient of logistics costs. for Real-time logistics costs This represents the initial logistics costs.

2. The civil engineering cost assessment and optimization system as described in claim 1, characterized in that: The PLM full lifecycle cost data closed-loop management module includes a data acquisition submodule, a data storage submodule, a data analysis submodule, and a feedback optimization submodule; The data acquisition submodule collects design drawing data, construction plan data, historical cost data, and supplier basic information through a standardized interface, and then transmits them to the data storage submodule after format conversion. The data storage submodule adopts a distributed storage architecture, classifying and storing data according to project stage and data type; the data analysis submodule uses multivariate statistical analysis algorithms to extract features and mine patterns from the stored data. The feedback optimization submodule transforms the analysis results into cost optimization instructions, synchronizes them to the data resource pool, and drives other modules to perform adjustments.

3. The civil engineering cost assessment and optimization system as described in claim 1, characterized in that: The PDM cross-organizational collaborative management module includes a data standardization submodule and a dynamic access control submodule. The data standardization submodule defines a common data source format and data transmission protocol for multiple participants, enabling seamless integration between the PLM data resource pool and the systems of owners, design units, construction parties, and suppliers. The dynamic access control submodule constructs a data access, modification, and approval permission allocation mechanism based on project stage and user identity, ensuring the information security of the data resource pool.

4. The civil engineering cost assessment and optimization system as described in claim 1, characterized in that: The PLM supply chain real-time data coupling module includes a real-time data acquisition submodule and a data coupling processing submodule. The real-time data acquisition submodule connects to building materials e-commerce platforms, supplier ERP systems, and logistics tracking systems through API interfaces to capture raw material prices, supplier inventory status, logistics transportation costs, and production capacity fluctuation data in real time. The data coupling processing submodule inputs real-time data into the cost correction model, calculates the real-time cost value, synchronizes it to the PLM data resource pool, and updates the cost assessment results.

5. The civil engineering cost evaluation and optimization system as described in claim 1, characterized in that: The PDM cost estimation process dynamic optimization module includes a change identification submodule, an impact quantification submodule, and a scheme generation submodule. The change identification submodule receives design change requests through a cross-organizational collaboration platform and, in conjunction with the anomaly monitoring results of the PLM supply chain real-time data coupling module, identifies the change type and trigger source. The impact quantification submodule, based on the change correlation analysis algorithm, calculates the scope and magnitude of the impact of changes on the cost. The quantification formula is as follows: ,in This represents the total cost change. For the first The value created by each related sub-project For the first The impact weight of changes to each sub-item of the project. For the first The change factor for each sub-project; the scheme generation sub-module uses an improved genetic algorithm to generate multiple cost adjustment schemes, which are then submitted to the cross-organizational collaboration platform to initiate the approval process.

6. The civil engineering cost assessment and optimization system as described in claim 2, characterized in that: The data acquisition submodule is equipped with a data cleaning unit, which uses an outlier detection algorithm to remove invalid data and a data completion algorithm to fill in missing data. The data analysis submodule is equipped with a data fusion unit, which uses a weighted average fusion algorithm to integrate multi-source basic data and a feature extraction algorithm to extract key cost influencing factors.

7. The civil engineering cost evaluation and optimization system as described in claim 3, characterized in that: The dynamic access control submodule constructs a three-dimensional access model of roles, stages, and data types. Roles include owner representatives, design engineers, construction project managers, supplier specialists, and auditors. Stages include planning, design, procurement, construction, and operation and maintenance. Data types include basic cost data, sensitive cost parameters, supplier quotation data, and change approval documents. The three-dimensional permission model assigns operation permissions for corresponding data types based on the matching relationship between roles and stages.

8. The civil engineering cost evaluation and optimization system as described in claim 1, characterized in that: The PLM full lifecycle cost data closed-loop management module also includes a data traceability sub-module. The data traceability sub-module assigns a unique identifier to each piece of cost data, records the data collection source, modification records, flow path, and associated approval information, forming a full lifecycle data traceability chain. The cost data details of any node can be queried through the identifier.

Images