Full-link intelligent visa making and automatic pricing system and method for project cost

Abstract

The invention discloses a full-link intelligent visa making and automatic pricing system and method for project cost. The system comprises an intelligent acquisition layer which guides field standardized input through a scenario form and synchronously generates an artificial document and a machine-readable pure structured data stream; the data transfer layer is responsible for verification and reliable transmission of structured data; and the intelligent application layer directly analyzes the structured data, completes quota matching and automatic pricing by using an intelligent engine, and generates a settlement file associated with the field evidence chain. According to the method, the core pain points of data non-standardization, flow splitting and evidence disjunction in the engineering cost field are solved from the source, full-flow, automatic and traceable closed-loop management from visa making to pricing auditing is achieved, and efficiency, accuracy and transparency are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the innovative application field of deep integration of computer software engineering, engineering cost management, artificial intelligence, and data science. Specifically, this invention relates to an intelligent management system and method based on a unified data model and standardized protocols, spanning the entire lifecycle of on-site visa processing in engineering construction. This system constructs a complete technical closed loop, from standardized data collection at the on-site data source, to the lossless flow of structured data, and then to intelligent automated pricing at the back end, integrating a complete digital evidence chain throughout the process. This invention aims to fundamentally solve the systemic problems in the engineering cost management industry, such as efficiency bottlenecks, insufficient accuracy, and difficulties in audit traceability caused by inconsistent data formats, numerous system silos, and excessive reliance on manual labor. It is a core technological infrastructure for realizing the digital transformation and intelligent upgrading of engineering project cost management, and can be widely applied to various construction scenarios such as housing construction, municipal transportation, water conservancy and hydropower, and industrial installation. Background Technology

[0002] In current engineering construction project management practices, the creation, review, and subsequent pricing of on-site visas are core aspects of cost control, and also areas prone to disputes and risks. Traditional work models rely on a series of independent software tools and manual operations, resulting in multiple efficiency bottlenecks and data breakpoints, as detailed below.

[0003] 1. Non-standardization and closed nature of data production: Field engineers typically use general office software (such as Microsoft Excel, WPS Spreadsheet) or simple form systems to record visa information. While these tools offer high flexibility, they lack industry-standard constraints, resulting in a wide variety of output document formats. Common problems include: using non-standard units of measurement (such as "one truckload," "one batch"), colloquial descriptions of work conditions (such as "ditching and pipe replacement"), and incomplete information records (lacking key attributes such as construction time and location). This unstructured or semi-structured data is "not directly usable" by the subsequent pricing system and must be manually interpreted and transcribed by cost professionals, becoming the first obstacle in data flow.

[0004] 2. The "Manual Bridging" Model for Cross-System Data Transfer and Its Inherent Defects: Currently, the transfer of visa data from the site to the cost department mainly relies on paper documents or electronic files (such as PDFs and scanned images) sent via email. Cost engineers need to act like "data porters," manually entering information from paper documents or table content from PDFs into professional engineering cost software (such as Glodon and Luban). This process is not only extremely time-consuming (transcription and preliminary pricing of a complex visa can take several hours), but it also inevitably introduces human errors, such as numerical entry errors, misunderstandings of project descriptions, and confusion of units of measurement. More seriously, this "re-entry" model severs the connection between the data and its original context (such as associated drawing numbers, specific construction teams, weather conditions, etc.), causing information to continuously degrade during the transfer process.

[0005] 3. Insufficient intelligence and subjective bias in data consumption (pricing): Even after data is entered into pricing software, the core "quota application" process still heavily relies on the experience and subjective judgment of cost engineers. The national standard quota database contains numerous items and complex rules. Accurately mapping a vague on-site description (e.g., "foundation deepening treatment") to a specific quota item (e.g., "manual excavation of Class III soil, depth within 2m") requires profound professional knowledge and experience. Different cost engineers may choose different quota items and adjustment coefficients, leading to significant differences in pricing results for the same visa item, creating potential contract disputes. While existing pricing software possesses some database query capabilities, it lacks true semantic understanding and intelligent recommendation capabilities, rendering it unsuitable as a reliable automated decision support tool.

[0006] 4. The dual disconnect between the physical and logical nature of business data and process evidence: Project settlement audits emphasize the integrity of the "chain of evidence." In traditional management models, process evidence such as on-site photos, videos, and meeting minutes are typically stored in independent file management systems, cloud drives, or even personal mobile phones, while pricing results are confined within cost estimation software. During audits, auditors need to manually search for corresponding supporting materials in a vast sea of ​​independent files, comparing them to the settlement statement—a tedious process prone to omissions. This physical and logical separation of data and evidence not only significantly increases audit costs and time but also reduces the credibility and transparency of project payment, exposing the construction company, contractor, and auditor to higher compliance risks.

[0007] 5. Existing technical improvement attempts and their limitations: The industry has recognized the above problems and some technical improvement solutions have emerged, but none of them have addressed the root cause: (1) Publication No. CN120634593A (Intelligent matching method for bill of quantities) proposes to adopt a two-stage process of "coarse screening-fine sorting" and semantic similarity calculation to improve the matching accuracy of bill of quantities item descriptions. However, its processing object is still existing, non-standardized text with varying quality. The algorithm's efficiency is limited by the standardization of the input text. Moreover, this solution is only an isolated "matching algorithm" and is not integrated with the front-end data production and back-end pricing process, so it cannot solve the problems of where the data comes from and how to ensure its quality. (2) Publication No. CN120850976A (Data entry method for engineering cost software) connects Excel table fields with software database fields through an intelligent mapping algorithm, which improves the data import efficiency. However, its essence is to perform "post-event remedial" processing on the generated unstructured / semi-structured files, which fails to standardize the data production format from the source and does not involve the association of process evidence. (3) Although the public document CN121146363A (Power Engineering Cost Review System) applies natural language processing technology, it focuses on the post-review and analysis of cost documents and belongs to the supervision link, rather than the in-process production and real-time pricing link of the construction process. There is an essential difference between the technical goal and the business scenario. (4) Many other solutions (such as CN120851509A, etc.) focus on using BIM and GIS technology to automatically calculate the quantity of work or manage materials. They belong to the technical path based on geometric models. They are different from solving the standardization of visa descriptions and process automation based on business forms. The two can complement each other but cannot replace each other.

[0008] In summary, the fundamental problem with existing technical architectures lies in the lack of a unified, end-to-end, machine-readable data standard and flow protocol to seamlessly connect on-site physical events (visa) with their digital cost representation (pricing). This invention presents a systematic solution to this fundamental pain point. Summary of the Invention

[0009] Purpose of the invention: The fundamental purpose of this invention is to break down the traditional barriers between visa creation and quota pricing in engineering cost management and to build a new work paradigm that is full-link, intelligent, and digitally integrated. Specific objectives include: (1) Achieving one-time data entry and full-process reuse: Through the forced guidance and structured design of the front-end system, high-quality, standardized, machine-readable data is generated at the source of visa creation, directly serving the back-end automated pricing and completely eliminating manual secondary entry. (2) Improving the intelligence and objectivity of quota matching: Using natural language processing (NLP) and machine learning technology, an intelligent matching engine is built to reduce the dependence of quota application on personal experience and improve the consistency, accuracy, and efficiency of pricing results. (3) Establishing a dynamically linked digital evidence chain: Visa items are strongly bound to multimedia evidence, time, and space information of the construction process to achieve visualization and one-click traceability of pricing data and process evidence, thereby enhancing the auditability of engineering management. (4) Provide a flexible and configurable system architecture: Design an open data interface and modular system so that it can be adapted to quota libraries and pricing rules of different regions and professions, and can be integrated with the project's existing OA, BIM and project management platforms.

[0010] Technical Solution: To achieve the above-mentioned objectives, this invention proposes a full-chain intelligent visa creation and automatic pricing system for engineering cost estimation. This system is not a simple patchwork of multiple independent software programs, but rather an organic whole based on a unified data model and communication protocol. Its core architecture consists of three layers: a visa creation subsystem (front-end), a data flow layer (middle platform), and a quota management subsystem (back-end). The overall architecture and data flow diagram of the system are shown below. Figure 1 As shown.

[0011] First layer: Intelligent data collection layer (front-end visa processing subsystem), such as Figure 2As shown. This layer is deployed at the construction site or project management department and is the source of data and "digital tool" for the whole chain. Its design philosophy is "guidance is better than input, structure is better than text". The core modules include: (1) Contextualized intelligent form engine: The system dynamically generates structured electronic forms according to the project type (such as civil engineering, installation, decoration) and visa category (such as design change, on-site negotiation, concealed works). The form fields are pre-bound to the standard terminology library. For example, the "work content" field provides a drop-down list with search function. The content comes from the description of the work content in the national standard quota. For the input of engineering quantity, the system integrates the automatic unit conversion function. When the user inputs "3 cubic meters" or "3m³", the system internally stores it as a standardized value and unit code. At the same time, the rationality of the data is verified by logical rules. For example, when the description is "pouring concrete", the system will prompt the associated "concrete strength grade" required field. (2) Multimodal evidence integration module: It supports direct photo and video upload and can automatically read the Exif information (shooting time, GPS coordinates) of the photo from the mobile device as metadata. The "evidence marking" function is innovatively introduced. Users can annotate, mark with arrows, and add text annotations (such as "This is the changed part") on the uploaded images. These markings are stored together with the images and strongly associated with specific visa entries, forming a reinforced evidence package with contextual information. (3) Dual-channel data output engine (core innovation): Channel A: Human-readable document generator. Generates a beautifully formatted PDF visa form that meets the company's stamping requirements according to a preset template, for use in the traditional signing, stamping, and archiving process; Channel B: Machine-readable data serializer (key): Serializes all information entered by the user in the form (including structured entries, IDs of associated evidence, and metadata) into a lightweight, unambiguous structured data format. This format completely strips away visual presentation and contains only pure data and semantic tags, such as Figure 3 As shown.

[0012] The second layer: Data flow layer (middle platform service and governance platform) This layer is the "central nervous system" and "data bus" of the system, responsible for receiving, verifying, transforming, routing and storing data. Its main components include: (1) Unified API gateway: provides a secure and standardized API endpoint, receives structured data packets from the front-end system, and performs identity authentication, permission verification and anti-replay attack protection. (2) Data cleaning and enhancement engine: performs logical verification on the received data (such as whether the quantity of work is a non-negative number), and can automatically supplement information according to rules (such as automatically associating the contract number and the current information price version according to the project ID). (3) Real-time message queue and data warehouse: uses message queues (such as Kafka, RabbitMQ) to ensure that data is transmitted asynchronously and reliably to the back-end pricing system. At the same time, all standardized visa data is stored in a time series database or data lake to form a complete and queryable project change data asset for big data analysis and historical data mining.

[0013] The third layer: Intelligent application layer (backend intelligent quota pricing subsystem). This layer is the value realization end of the entire chain, deployed in the cost department or company cost center. Its core mission is to automatically complete the transformation from data to cost documents, such as... Figure 4(1) Dedicated high-fidelity parser: Since the input data is in a preset structured format (B channel output), this module does not require complex OCR or natural language processing technology for text recognition. It directly acts as a "data consumer", parsing JSON / XML data packets and extracting key fields such as description.std_term (standard term) and quantity.value (work quantity) in milliseconds, achieving zero-error data import. (2) Intelligent matching engine based on deep semantic understanding: 2.1 The core of this engine is a trained quota semantic model. It not only performs keyword matching, but also understands the semantics of the work content. For example, if the current end description is "installation of galvanized steel pipe with a diameter of 50 mm", the engine can understand that "diameter of 50 mm" is equivalent to "DN50" and "galvanized steel pipe" is the material, thus accurately matching the sub-item "indoor galvanized steel pipe installation (threaded connection) with a nominal diameter of less than 50 mm" in the quota library. 2.2 The engine has a built-in recommendation algorithm. When the matching degree reaches the threshold (such as 95%), it will automatically apply the algorithm. When the matching degree is within the threshold range (such as 80%-95%), it will provide 2-3 most likely options for users to quickly select and display the matching degree comparison and basis. 2.3 It supports rule configuration and can adapt to the differences in quota libraries in different regions and the special pricing habits of enterprises. (3) Automated pricing and dynamic book generation module: 3.1 After the matching is completed, the system will automatically connect to the official or enterprise private material and machine price information platform to obtain real-time or current information price. 3.2 According to the project's pre-configured fee template (management fee, profit, tax, regulatory fee, etc. rate), it will automatically complete the calculation of comprehensive unit price and total price. 3.3 Finally, the system can generate a complete settlement draft containing a cost summary table, bill of quantities, pricing details table and embedded evidence links with one click according to the prescribed report format. Users can review the system results on the graphical interface and make necessary batch adjustments (such as uniformly adjusting the earthwork transportation distance coefficient) or individual corrections.

[0014] Collaborative approach: The full-link data flow and business collaboration approach includes the following 5 steps: (1) In the project initiation phase, configure the project-specific quota library version, fee standard, material price source and visa approval process in the middle platform system. (2) On-site guided data collection and evidence solidification. Construction workers use the front-end APP or Web terminal to enter visas under the guidance of standard forms, and take photos, upload and mark related evidence in real time. After submission, the data is packaged and sent to the middle platform. (3) Structured data synchronization and status tracking. After receiving the data, the middle platform immediately updates its status to "pending pricing" and notifies the back-end pricing system and related cost engineers through message push. (4) Intelligent pricing and manual review. Cost engineers view the "pending" tasks in the back-end system. The system has automatically completed parsing, matching and preliminary pricing. Engineers focus on rationality review and approval process follow-up, improving efficiency by more than 90%. (5) Output of results and audit support. When the final settlement statement is output, each pricing item contains a clickable hyperlink or QR code. Auditors can scan a QR code to directly access all the original structured data, on-site photos, videos, and process records corresponding to that item, achieving "what you see is what you audit".

[0015] Beneficial effects: Compared with the existing fragmented technical solutions, the full-link intelligent system provided by this invention brings multi-dimensional and disruptive improvements: (1) Geometric growth in production efficiency: The average visa processing cycle (including data entry, pricing, and verification) required in the traditional model is shortened to 10-30 minutes, with purely mechanical data processing work almost reduced to zero. In a large project, it is expected to save the cost team more than 70% of man-hours, allowing professionals to focus on high-value work such as cost analysis and risk management. (2) Fundamental change in data quality and pricing accuracy: Through source governance (terminology standardization) and process-free (automated flow), human transcription errors, descriptive ambiguities, and quota misapplication are systematically eliminated. Tested in standard construction projects, the system can reduce the error rate of pricing data from about 5-15% in the traditional model to less than 1%, significantly reducing contract disputes and rework caused by calculation errors. (3) A trustworthy and transparent digital audit infrastructure has been built: For the first time in the field of engineering cost, the integration of "business data" and "process evidence" at the digital native level has been realized. Each final cost figure comes with a complete and tamper-proof "data gene" and "evidence fingerprint", which makes engineering settlement move from "black box calculation" to "white box traceability", greatly enhancing the mutual trust among all parties involved in the construction and meeting the increasingly strict digital audit supervision requirements in the future. (4) Promote the accumulation of industry knowledge and intelligent upgrading: During operation, the system continuously accumulates matching relationship data of "standard terminology-quota" and manual review feedback data. These data can feed back into the algorithm model of the intelligent matching engine to optimize the system, forming a positive cycle of becoming smarter the more it is used. At the same time, the structured project data accumulated throughout the entire chain provides unprecedented high-quality data raw materials for the construction of enterprise cost database, benchmarking analysis and artificial intelligence prediction model. (5) It has excellent scalability and ecosystem compatibility: The system adopts a microservice architecture and open API design. Its front end can be embedded into existing mobile inspection APPs, and its back end pricing engine can be provided as a service to third-party project management software. Attached Figure Description

[0016] Figure 1 Description: This section illustrates the overall system architecture of the present invention, which consists of a front-end visa creation subsystem and a back-end quota management subsystem. The two systems are connected through a structured data stream (dedicated API interface), with the back-end interacting with the quota database to ultimately generate the project settlement statement.

[0017] Figure 2Note: This is a flowchart of the front-end subsystem data processing. The process begins with "Start Input". Data is entered according to the input method (standard terminology or free text). The free text is automatically mapped and converted into structured data. Then, the export mode (archive or machine recognition) is selected to generate a visually appealing document or structured data for transmission, respectively.

[0018] Figure 3 Explanation: This illustrates the data file structure exported in machine recognition mode, including a file header, structured data body, and format specifications. The data body records items such as work content and quantities in key-value pairs, using plain text format for easy automated processing by the backend.

[0019] Figure 4 Description: This demonstrates the back-end subsystem processing and pricing workflow. Starting with receiving data from the front end, information is extracted via an AI intelligent recognition interface, entering the intelligent quota matching stage (including fuzzy and exact matching). After manual confirmation of the sub-items, the automated pricing core is activated to generate a preliminary settlement statement. Finally, after review and fine-tuning, the official result is output.

Claims

1. A fully intelligent visa creation and automatic pricing system for engineering cost estimation, characterized in that, It comprises an intelligent acquisition layer, a data flow layer, and an intelligent application layer. The intelligent acquisition layer guides users at the engineering site to input visa information based on a preset standard terminology database, and simultaneously generates a first-format human-readable document and a second-format purely structured machine-readable data stream. The data flow layer receives, verifies, and forwards the purely structured machine-readable data stream. The intelligent application layer parses the purely structured machine-readable data stream, performs intelligent quota matching and automatic pricing based on the standardized information within it, and outputs a settlement document. The pricing items in the settlement document are bound to the original multimedia evidence collected by the intelligent acquisition layer through embedded association identifiers.

2. A fully intelligent visa creation and automatic pricing system for engineering cost estimation, characterized in that: It comprises an intelligent acquisition layer, a data flow layer, and an intelligent application layer. The intelligent acquisition layer guides users at the engineering site to input visa information based on a preset standard terminology database, and simultaneously generates a first-format human-readable document and a second-format purely structured machine-readable data stream. The data flow layer receives, verifies, and forwards the purely structured machine-readable data stream. The intelligent application layer parses the purely structured machine-readable data stream, performs intelligent quota matching and automatic pricing based on the standardized information within it, and outputs a settlement document. The pricing items in the settlement document are bound to the original multimedia evidence collected by the intelligent acquisition layer through embedded association identifiers.

3. The system according to claim 2, characterized in that, The intelligent acquisition layer includes a dual-channel data output engine, which is used to generate, in parallel, a first-format document and a second-format data stream with different physical formats but related logical content from the same input data.

4. The system according to claim 1, characterized in that, The intelligent data collection layer also includes a contextualized intelligent form engine, which dynamically loads form templates based on visa type and associates the input controls in the form with the standard terminology database to constrain or guide the user's input.

5. The system according to claim 1, characterized in that, The intelligent application layer includes a dedicated parser configured to directly parse the second-format data stream and extract standardized terminology fields, quantity values, and units without requiring optical character recognition or free text parsing steps.

6. The system according to claim 5, characterized in that, The intelligent application layer also includes an intelligent matching engine, which receives standardized terminology fields extracted by the dedicated parser as input, searches for matching items in the quota library by calculating semantic similarity, and outputs the matching results and confidence scores.

7. The system according to claim 1, characterized in that, The association identifier is a unique resource identifier for multimedia evidence stored in the data transfer layer or intelligent acquisition layer, and this identifier is written into the second format data stream and the final generated settlement file.

8. A method for end-to-end data flow and automatic pricing based on the system described in any one of claims 1-7, characterized in that, Includes the following steps: S1. On-site, standardized information entry and evidence association are completed through the intelligent acquisition layer to generate a first-format document and a second-format pure structured data stream; S2. The second format pure structured data stream is sent to the intelligent application layer through the data stream transfer layer; S3. In the intelligent application layer, the data stream is parsed, quota matching and cost calculation are automatically completed, and a settlement document draft containing evidence association identifiers is generated; S4. Based on the association identifier in the aforementioned settlement document draft, it is possible to reverse query and retrieve the corresponding original input information and multimedia evidence.

9. The method according to claim 8, characterized in that, The generation of the second-format pure structured data stream in step S1 specifically involves serializing the form input content into an independent self-describing data packet. The semantics of this data packet are defined by its internal key-value pairs and do not depend on a specific file reader.

10. The method according to claim 8, characterized in that, The automatic quota matching described in step S3 is based directly on the description field extracted from the second format data stream and standardized by the front-end standard terminology library, without the need for preprocessing of non-standardized text.

Images