Production project completion commissioning research and judgment method and device based on large model, storage medium and computer equipment

Abstract

The invention discloses a production project completion commissioning research and judgment method and device based on a large model, a storage medium and computer equipment. The method comprises the steps of obtaining production project total data of different power projects from a plurality of production project management systems; the production engineering management system comprises a production management system, a risk control platform, a financial system, a material management system and a document management system. Integrating the production project total data through a pre-trained large model to form a production project data set of each electric power project; the production project data set corresponding to each electric power project is analyzed to determine whether each electric power project has completion and production state abnormity or not, and the completion and production state abnormity comprises false completion, uncompleted first settlement and completion unreported.

Description

Technical Field

[0001] This application relates to the field of power project management technology, and in particular to a method, device, storage medium and computer equipment for evaluating the completion and commissioning of production projects based on a large model. Background Technology

[0002] In the existing production engineering data of the power industry, maintenance operation data is available in the maintenance plan, work order, and technical renovation and overhaul sections of PMS3.0, but the amount of data related to production engineering projects is relatively small. Engineering project information is not a mandatory field between maintenance plans and work orders, so the project information cannot be obtained from the maintenance plans (monthly, weekly, and daily plans).

[0003] Milestone-based project management has been widely adopted in production engineering management systems. However, some project managers fail to adequately implement milestone plans and do not truly control progress according to planned milestones. Currently, the accuracy of determining the completion and commissioning dates of production projects relies heavily on manual methods, involving offline data retrieval from multiple systems and expert verification. This process is inefficient and lacks accuracy. Consequently, problems such as false completion, premature completion, and failure to report completion occur frequently, resulting in inaccurate completion and commissioning dates and posing significant risks to the legal and compliant operation of power companies. Summary of the Invention

[0004] In view of this, embodiments of this application provide a method, apparatus, storage medium, and computer equipment for assessing the completion and commissioning of production projects based on a large model.

[0005] According to one aspect of this application, a method for assessing the completion and commissioning of production projects based on a large-scale model is provided, characterized in that the method includes: The system acquires full production data for different power projects from multiple production engineering management systems, including a production management system, a risk control platform, a financial system, a materials management system, and a document management system. The full dataset of the production projects is integrated by pre-trained large models to form a production project dataset for each power project. The production project dataset corresponding to each power project is analyzed separately to determine whether there are any abnormalities in the completion and commissioning status of each power project. The abnormalities in the completion and commissioning status include...

[0006] Optionally, the full dataset of the production projects is integrated using a pre-trained large model to form a production project dataset for each power project, including: A large-scale model is used to perform semantic analysis on the full data of the production projects, identify and extract key project information for each power project, and integrate the key project information of each power project in chronological order to form a production project dataset for each power project. The production management system stores project information, equipment information, maintenance plans, and work orders for the power projects; the risk control platform stores risk monitoring data, production plans, and control information; the financial system stores information on the use of funds, cost accounting, and settlement for the power projects; the materials management system stores the procurement, supply, and usage of materials required for the power projects; and the document management system stores project plans, project design documents, project progress reports, project quality reports, project acceptance reports, feedback from on-site staff, and feedback from project management personnel for the power projects.

[0007] Optionally, before integrating the full dataset of the production projects using a pre-trained large model to form a production project dataset for each power project, the method further includes: Obtain production project data samples corresponding to multiple power project samples and production project dataset samples constructed based on the production project data samples, and mix the production project data samples corresponding to each power project sample into a training sample set; The general-purpose model is trained in three stages using the training sample set. The first stage involves guiding the model to analyze power projects based on the training sample set, and optimizing its parameters based on the analyzed power project names and the corresponding power project samples to train its ability to identify power projects. The second stage involves guiding the model trained in the first stage to perform data association on the data samples in the training sample set, constructing a data sample set corresponding to each power project name, and optimizing its parameters based on the constructed data sample set and the corresponding power project production dataset samples to train its ability to distinguish between different power project data. The third stage involves guiding the model trained in the second stage to extract key project information corresponding to each power project sample based on the data sample set corresponding to each power project sample and integrate it into production project dataset prediction information. The model's parameters are then optimized based on the production project dataset prediction information and production project dataset samples to train its ability to analyze and integrate key project information from production project data.

[0008] Optionally, the production project dataset corresponding to each power project is analyzed separately to determine whether there are any abnormalities in the completion and commissioning status of each power project, including: If the production project dataset corresponding to the power project contains completed data, the production project dataset is analyzed by a pre-trained fake completion agent to determine whether the power project is fake completed, and the production project dataset is analyzed by a pre-trained incomplete-to-complete agent to determine whether the power project is incomplete-to-complete. If the production project dataset corresponding to the power project does not contain completion data, the production project dataset is analyzed by a pre-trained agent that has not reported completion data to determine whether the power project has any unreported completion data.

[0009] Optionally, the production project dataset is analyzed using a pre-trained fake completion agent to determine whether the power project has fake completions, including: The production project dataset is analyzed by a pre-trained fake completion agent to identify the work progress data of the first engineering project after the completion time of the power project. Based on the type of the power project and the identified work progress data of the first engineering project, it is determined whether the power project is falsely completed. The pre-trained fake completion agent has pre-learned completion standard information for various types of power projects.

[0010] Optionally, the production project dataset is analyzed using a pre-trained agent that terminates projects before completion to determine whether any power projects are terminated before completion, including: The production project dataset is analyzed by a pre-trained agent that completes projects before completion to identify the work progress data of a second engineering project after the settlement time of the power project. Based on the type of the power project and the identified work progress data of the second engineering project, it is determined whether the power project has been completed before completion. The pre-trained agent that completes projects before completion has pre-learned settlement standard information for various types of power projects.

[0011] Optionally, the production project dataset is analyzed using a pre-trained agent that identifies projects that are not yet reported as completed, including: The production project dataset is analyzed by a pre-trained agent that identifies completed work content of the power projects. Based on the type of the power project and the identified completed work content, it is determined whether there is any unreported completion status for the power project. The pre-trained agent has pre-learned completion standard information and unreported completion rules for various types of power projects. The completion standard information includes work content that meets the completion standard of the corresponding type of power project, and the unreported completion rules include at least the maximum project production and maintenance time threshold for each piece of equipment required for the production of the corresponding type of power project.

[0012] According to another aspect of this application, a device for assessing the completion and commissioning of production projects based on a large model is provided, the device comprising: The data acquisition module is used to acquire full data of production projects for different power projects from multiple production engineering management systems; The data integration module is used to integrate all the data of the production projects through a large model to form a production project dataset for each power project. The anomaly analysis module is used to analyze the production project dataset corresponding to each power project to determine whether there are any anomalies in the completion and commissioning status of each power project.

[0013] Optionally, the data integration module is used for: A large-scale model is used to perform semantic analysis on the full data of the production projects, identify and extract key project information for each power project, and integrate the key project information of each power project in chronological order to form a production project dataset for each power project. The production management system stores project information, equipment information, maintenance plans, and work orders for the power projects; the risk control platform stores risk monitoring data, production plans, and control information; the financial system stores information on the use of funds, cost accounting, and settlement for the power projects; the materials management system stores the procurement, supply, and usage of materials required for the power projects; and the document management system stores project plans, project design documents, project progress reports, project quality reports, project acceptance reports, feedback from on-site staff, and feedback from project management personnel for the power projects.

[0014] Optionally, the data integration module is also used for: Obtain production project data samples corresponding to multiple power project samples and production project dataset samples constructed based on the production project data samples, and mix the production project data samples corresponding to each power project sample into a training sample set; The general-purpose model is trained in three stages using the training sample set. The first stage involves guiding the model to analyze power projects based on the training sample set, and optimizing its parameters based on the analyzed power project names and the corresponding power project samples to train its ability to identify power projects. The second stage involves guiding the model trained in the first stage to perform data association on the data samples in the training sample set, constructing a data sample set corresponding to each power project name, and optimizing its parameters based on the constructed data sample set and the corresponding power project production dataset samples to train its ability to distinguish between different power project data. The third stage involves guiding the model trained in the second stage to extract key project information corresponding to each power project sample based on the data sample set corresponding to each power project sample and integrate it into production project dataset prediction information. The model's parameters are then optimized based on the production project dataset prediction information and production project dataset samples to train its ability to analyze and integrate key project information from production project data.

[0015] Optionally, the anomaly analysis module is used for: If the production project dataset corresponding to the power project contains completed data, the production project dataset is analyzed by a pre-trained fake completion agent to determine whether the power project is fake completed, and the production project dataset is analyzed by a pre-trained incomplete-to-complete agent to determine whether the power project is incomplete-to-complete. If the production project dataset corresponding to the power project does not contain completion data, the production project dataset is analyzed by a pre-trained agent that has not reported completion data to determine whether the power project has any unreported completion data.

[0016] Optionally, the anomaly analysis module is used for: The production project dataset is analyzed by a pre-trained fake completion agent to identify the work progress data of the first engineering project after the completion time of the power project. Based on the type of the power project and the identified work progress data of the first engineering project, it is determined whether the power project is falsely completed. The pre-trained fake completion agent has pre-learned completion standard information for various types of power projects.

[0017] Optionally, the anomaly analysis module is used for: The production project dataset is analyzed by a pre-trained agent that completes projects before completion to identify the work progress data of a second engineering project after the settlement time of the power project. Based on the type of the power project and the identified work progress data of the second engineering project, it is determined whether the power project has been completed before completion. The pre-trained agent that completes projects before completion has pre-learned settlement standard information for various types of power projects.

[0018] Optionally, the anomaly analysis module is used for: The production project dataset is analyzed by a pre-trained agent that identifies completed work content of the power projects. Based on the type of the power project and the identified completed work content, it is determined whether there is any unreported completion status for the power project. The pre-trained agent has pre-learned completion standard information and unreported completion rules for various types of power projects. The completion standard information includes work content that meets the completion standard of the corresponding type of power project, and the unreported completion rules include at least the maximum project production and maintenance time threshold for each piece of equipment required for the production of the corresponding type of power project.

[0019] According to another aspect of this application, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described method for assessing the completion and commissioning of production projects based on a large model.

[0020] According to another aspect of this application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-mentioned method for assessing the completion and commissioning of production projects based on large models.

[0021] By employing the above technical solutions, this application provides a method, apparatus, storage medium, and computer equipment for assessing the completion and commissioning of production projects based on a large-scale model. Addressing the problems in the power industry where poor correlation of production engineering data and inadequate execution of milestone plans lead to inefficient and error-prone judgments of completion and commissioning times, resulting in anomalies such as false completion and significant compliance risks, this method proposes acquiring full-volume data from different power projects across multiple platforms, including production management systems. This data is then integrated using a pre-trained large-scale model to form datasets for each project, and further analyzed to determine if any anomalies exist in the completion and commissioning status. This achieves multi-source data fusion and standardized integration, improving data quality; it can accurately identify anomalies and effectively control risks.

[0022] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart illustrating a method for assessing the completion and commissioning of a production project based on a large model, as provided in an embodiment of this application, is shown. Figure 2 This paper illustrates a flowchart of another method for assessing the completion and commissioning of production projects based on a large model, as provided in an embodiment of this application. Figure 3 This paper presents a schematic diagram of a production project completion and commissioning assessment device based on a large model, as provided in an embodiment of this application. Detailed Implementation

[0024] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0025] There are some issues with production engineering data in the power industry: maintenance operation data is functional in some modules of PMS3.0, but there is little data related to production engineering projects; project information is not mandatory in maintenance plans and work orders, making it impossible to know the project information from the maintenance plan.

[0026] While milestone-based project management is widely used, some project managers fail to implement it effectively and do not monitor progress according to planned milestones. Currently, determining the completion and commissioning dates of production projects relies on manual labor, involving offline queries of data from multiple systems and manual verification by experts, which is inefficient and difficult to guarantee accuracy. This has led to problems such as false completions, premature closings, and failure to report completions, resulting in inaccurate completion and commissioning dates and posing significant risks to the legal and compliant operation of power companies.

[0027] This embodiment provides a method for evaluating the completion and commissioning of production projects based on a large model, such as... Figure 1 As shown, the method includes: Step 101: Obtain full production data for different power projects from multiple production engineering management systems; among which, the production engineering management system includes production management system, risk control platform, financial system, material management system, and document management system.

[0028] Step 102: Integrate the full data of the production projects using a pre-trained large model to form a production project dataset for each power project.

[0029] Step 103: Analyze the production project dataset corresponding to each power project to determine whether there are any abnormalities in the completion and commissioning status of each power project. The abnormalities in the completion and commissioning status include false completion, completion before completion, and completion without reporting.

[0030] In this embodiment, firstly, full-scale production project data for different power projects is obtained from multiple production engineering management systems. These systems encompass production management systems, risk control platforms, financial systems, material management systems, and document management systems. By acquiring data from these different systems, comprehensive information related to power projects can be collected. Secondly, a pre-trained large-scale model is used to integrate the acquired full-scale production project data, forming a separate production project dataset for each power project. This large-scale model possesses powerful data processing and integration capabilities, effectively integrating data from different systems and with varying formats. It analyzes the production project data from different power projects to identify data belonging to the same power project, making the data more organized and easier to analyze. Finally, the production project dataset for each power project is analyzed to determine if there are any abnormalities in the completion and commissioning status of each project, specifically including false completion, premature completion, and failure to report completion. For example, by comparing the planned project progress with the actual completion status, it is determined whether false completion exists; whether the project was prematurely completed before full completion; and whether the project completion status was reported in a timely and accurate manner to avoid failure to report completion. This application's embodiments acquire full data from multiple production engineering management systems and integrate it with a pre-trained large model. This breaks down information barriers between systems, achieves multi-source data fusion and standardized, orderly integration, and improves data comprehensiveness and accuracy. It can also accurately identify abnormal completion and commissioning statuses such as false completion, premature completion, and failure to report completion, helping power companies to correct and rectify issues in a timely manner and effectively prevent legal, financial, and operational risks. At the same time, it significantly reduces the time and workload of manual data processing, allowing project managers to devote more energy to actual management.

[0031] Example of production project data: Basic project information: Project Name: Energy Efficiency Improvement and Renovation of 10kV Luojiayuntunyan and Other Transformer Areas under State Grid Chongqing Xiushan Company Project Code: 21207123000Q Implementing Unit: Xiushan Power Supply Branch of State Grid Chongqing Electric Power Company Major: Power Distribution Completion and commissioning date: 2024-11-25 Final settlement date: 2024-12-09 Project Closure Date: 2024-12-10 Station and line names: 10kV Meilong Line, 10kV Meifu Line, 10kV Meichang Line, 10kV Gaomeng Line, 10kV Longcui Line, 10kV Longshuai Line, 10kV Meizhai Line 964, 10kV Shizhong Line Current Equipment Status: 1. 10kV Meilong Line Yaguan Tianba Substation Original Transformer S11-200, commissioned on January 10, 1990, manufacturer: Chongqing Zhenglong Electric Power Equipment Co., Ltd. 2. 10kV Meifu Line Shaba Zoujiafen Substation Original Transformer S11-160, commissioned on January 5, 1990, manufacturer: Chongqing Fuling Transformer Factory. 3. 10kV Longcui Line Yuanyang Station Taotai Substation Original Transformer S9-50 / 10, commissioned on July 11, 2004, manufacturer: Chongqing Bolian Transformer Co., Ltd. 4. 10kV Longshuai Line Ganchuan Houlidong Substation Original Transformer S9-125 / 10, commissioned on March 19, 1999, manufacturer: Chongqing Bolian Transformer Co., Ltd.

[0032] Problem identified: 17 transformers, including those in Xiayaguan Tianba, Shaba Zoujiafen, Yuanyangzhan Taotai, and Ganchuan Houlidong, do not meet the latest transformer energy efficiency standards and are in a high-loss state. To reduce transformer losses, these high-loss transformers will be replaced.

[0033] Equipment Information: 1. 71C1B4C3-7302-4B29-9CDF-BC803779D2A8-52677 Gaoxiu Jujiaping Pole-Mounted Transformer. 2. CEF2EC4E-8B2C-417C-AD7B-443C695164B7-77784 Mangdong Meishatai Pole-Mounted Transformer. Maintenance plan and work order information: Line Name: 10kV Meichang Line Work tasks: Adjusting the tap pitch of the Huaxing transformer at pole #15-1 on the 10kV Meichang line; adjusting the tap pitch of the Pingyang Huangnidong transformer at pole #44-10-4 on the Houqi branch line of the Changhou branch of the 10kV Meichang line. Planned working time: 2024-11-17 Risk control platform weekly plan, daily management, and work tickets: Weekly Work Plan Title: Inspection of Communication Power Supply and Battery Charge / Discharge Test at 35kV Zhonghe Substation Weekly planned substation / line: 35kV Zhonghe Substation Weekly plan start and end times: 2024-08-28 09:00 - 2024-08-28 23:00 Zhou's main business unit: Xiushan Power Supply Branch of State Grid Chongqing Electric Power Company Weekly work plan: Inspection of communication power supply and battery charge / discharge test at 35kV neutral station. Daily Work Plan Title: Inspection of Communication Power Supply and Battery Charging / Discharging Test at 35kV Zhonghe Substation Daily planned substation / line: 35kV Zhonghe Substation Daily plan start and end times: 2024-08-28 09:00 - 2024-08-28 23:00 Daily planning main business unit: Xiushan Power Supply Branch of State Grid Chongqing Electric Power Company Daily planned tasks: Inspection of communication power supply and battery charge / discharge test at 35kV Zhonghe substation. Work order: Name of the substation and equipment busy status, work task information, planned work time, and workplace name.

[0034] By applying the technical solution of this embodiment, this paper addresses the problems in the power industry, such as poor correlation of production engineering data, inadequate execution of milestone plans leading to inefficient and error-prone judgment of completion and commissioning times, resulting in anomalies such as false completion and significant compliance risks. The solution proposes acquiring full-volume data from different power projects from multiple platforms, including production management systems, integrating this data into pre-trained large-scale models to form datasets for each project, and then analyzing and judging whether there are any anomalies in the completion and commissioning status. This achieves multi-source data fusion and standardized integration, improving data quality; it can accurately identify anomalies and effectively prevent and control risks.

[0035] Optionally, in this embodiment, the entire production project data is integrated using a pre-trained large model to form a production project dataset for each power project. This includes: performing semantic analysis on the entire production project data using the large model to identify and extract key project information for each power project, and integrating the key project information for each power project in chronological order to form a production project dataset for each power project. The production management system stores project information, equipment information, maintenance plans, and work orders for the power projects; the risk control platform stores risk monitoring data, production plans, and control information; the financial system stores information on the use of funds, cost accounting, and settlement for the power projects; the materials management system stores the procurement, supply, and usage of materials required for the power projects; and the document management system stores project plans, project design documents, project progress reports, project quality reports, project acceptance reports, feedback information from on-site staff, and feedback information from project management personnel for the power projects.

[0036] In this embodiment, a pre-trained large-scale model is used for deep processing of full-volume production data from different power projects obtained from a production management system (which stores key information such as project information, equipment information, maintenance plans, and work orders for power projects), a risk control platform (which stores risk monitoring data, production plans, and control information), a financial system (which records information on the use of funds, cost accounting, and settlement for power projects), a materials management system (which covers the procurement, supply, and usage of materials required for power projects), and a document management system (which stores project plans, project design documents, project progress reports, project quality reports, project acceptance reports, and feedback information from on-site and project management personnel). The large-scale model first performs semantic analysis on this full-volume data, leveraging its powerful language understanding capabilities to accurately identify and extract key project information for each power project. Subsequently, this key information is systematically integrated according to temporal logic, ultimately forming an independent and complete production project dataset for each power project. This approach not only breaks down data barriers between systems and achieves the fusion and unification of multi-source heterogeneous data, providing a comprehensive and accurate data foundation for subsequent analysis, but also allows for a clearer presentation of the project's progress based on the time-series integrated data, helping to assess the project's status and promptly identify potential problems.

[0037] Optionally, before integrating the full data of the production projects through a pre-trained large model to form a production project dataset for each power project, the method further includes: acquiring production project data samples corresponding to multiple power project samples and production project dataset samples constructed based on the production project data samples, and mixing the production project data samples corresponding to each power project sample into a training sample set; using the training sample set to train the general large model in three stages; wherein, the first stage of training includes guiding the general large model to perform power project analysis based on the training sample set, and optimizing the parameters of the general large model based on the analyzed power project names and the power project names corresponding to the power project samples, so as to train the general large model's ability to identify power projects; The second-stage training includes guiding the general-purpose large model trained in the first stage to perform data association on the data samples in the training sample set, constructing a data sample set corresponding to each power project name, and optimizing the parameters of the general-purpose large model based on the constructed data sample set and the corresponding power project production project dataset samples to train the general-purpose large model's ability to distinguish data from different power projects. The third-stage training includes guiding the general-purpose large model trained in the second stage to extract the key project information corresponding to each power project sample based on the data sample set corresponding to each power project sample and integrate it into the production project dataset prediction information, and optimizing the parameters of the general-purpose large model based on the production project dataset prediction information and the production project dataset samples to train the general-purpose large model's ability to analyze and integrate the key project information of the production project data.

[0038] In this embodiment, to ensure that the pre-trained large model can accurately and efficiently integrate all the production project data to form a high-quality production project dataset for each power project, the following large model training process needs to be executed before integrating the data using the large model. First, production project data samples corresponding to multiple power project samples are collected. These samples cover various information throughout the entire lifecycle of a power project, from planning to implementation to acceptance. Simultaneously, production project dataset samples are constructed based on these production project data samples to provide a reference standard for subsequent training. Next, the production project data samples corresponding to each power project sample are mixed to form a rich and diverse training sample set, which fully reflects the diversity and complexity of power project data. Then, the general large model is trained in three phases using the training sample set. In the first phase of training, the general large model is guided to analyze power projects based on the training sample set. By comparing the analyzed power project names with the actual names corresponding to the power project samples, the large model parameters are optimized and adjusted. This process aims to train the general large model's ability to accurately identify power projects, enabling it to precisely locate specific projects from massive amounts of data, laying the foundation for subsequent processing. The second phase of training is used to improve the large model's ability to distinguish between different power project data. The first phase of training guides the general-purpose large model to perform deep correlation analysis on the data samples in the training sample set, constructing a data sample set corresponding to each power project name. Then, the constructed data sample set is compared with the corresponding power project production project dataset samples, and the large model parameters are optimized based on the differences. Through this training, the large model can clearly distinguish data from different power projects, avoiding data confusion and ensuring the accuracy and relevance of data integration. For example, project information is not mandatory in maintenance plans and work orders; therefore, the project information cannot be determined from the maintenance plans (monthly, weekly, and daily plans). The large model can analyze the equipment information required by the project and the maintenance equipment information in the maintenance plan, identifying the equipment information required by the project from the maintenance equipment information. This allows it to associate maintenance equipment information without project information with the corresponding project, thus obtaining the full data of the power projects. The third phase of training focuses on cultivating the general-purpose large model's ability to analyze and integrate key information from production project data. The second phase of training guides the general-purpose large model to extract key information from the data sample sets corresponding to each power project sample, and integrates this information into the production project dataset prediction information. The predicted information is then compared with the production project dataset samples, and the parameters of the large model are optimized based on the comparison results. After this stage of training, the large model can accurately grasp the core points of the power project data and integrate them in an orderly manner to form a complete and accurate production project dataset. Through these three stages of systematic training, the general-purpose large model possesses powerful capabilities in power project identification, data differentiation, and key information analysis and integration, enabling it to efficiently and accurately integrate all production project data.

[0039] Optionally, in this embodiment of the application, the production project dataset corresponding to each power project is analyzed to determine whether there are any abnormalities in the completion and commissioning status of each power project, including: If the production project dataset corresponding to the power project contains completed data, the production project dataset is analyzed by a pre-trained fake completion agent to determine whether the power project is fake completed, and the production project dataset is analyzed by a pre-trained incomplete-to-complete agent to determine whether the power project is incomplete-to-complete. If the production project dataset corresponding to the power project does not contain completion data, the production project dataset is analyzed by a pre-trained agent that has not reported completion data to determine whether the power project has any unreported completion data.

[0040] In this embodiment, such as Figure 2 As shown, when the production project dataset corresponding to a power project contains completion data, it indicates that the project has entered the completion stage at the data level. Therefore, two pre-trained agents are used for in-depth analysis. Firstly, a pre-trained fake completion agent is used. This agent, trained on a large amount of fake completion case data, possesses the ability to accurately identify fake completion characteristics. It comprehensively scans and analyzes the production project dataset, searching for clues related to fake completion from multiple dimensions of data such as project progress, quality acceptance, and fund settlement, thereby determining whether the project has a fake completion issue. Secondly, a pre-trained incomplete-to-complete agent is also used. This agent focuses on identifying incomplete-to-complete anomalies. It analyzes the completion status of various tasks, resource input, and overall project progress in the dataset, and determines whether the project has been prematurely completed without completing all work based on the typical characteristics and judgment criteria of incomplete-to-complete completion. If the production project dataset corresponding to a power project does not contain completion data, it means that the project has not yet shown as completed in the data records. However, this does not preclude the possibility that a project may have been completed but not reported in a timely manner, i.e., a "completed but not reported" anomaly. To address this, a pre-trained intelligent agent is used for analysis. This agent, trained on a wealth of case data on unreported completions, can deeply analyze project progress information, personnel feedback, and material usage within the dataset. By comparing this data with normal project progress and reporting procedures, it can determine whether the project has completed all work but failed to report completion information according to the prescribed procedures and timelines.

[0041] In this embodiment of the application, optionally, the production project dataset is analyzed by a pre-trained fake completion agent to determine whether the power project has been falsely completed. This includes: analyzing the production project dataset by the pre-trained fake completion agent to identify the work progress data of the first engineering project after the completion time of the power project; determining whether the power project has been falsely completed based on the type of the power project and the identified work progress data of the first engineering project; wherein the pre-trained fake completion agent has pre-learned completion standard information for various types of power projects.

[0042] In this embodiment, the agent first extracts the "completion time" field from the project dataset and uses this time as a base point to scan subsequent time-series data. For example, if the completion time of a power transmission project is December 5, 2025, the agent focuses on analyzing the records of engineering activities such as equipment installation, commissioning, and quality inspection generated between December 6 and December 31, identifying "first engineering project work progress data" as an anomaly clue. Based on the completion standards of various types of power projects injected during the pre-training phase (such as thermal power projects needing to complete 72 hours of full-load trial operation, and new energy power plants needing to pass grid connection acceptance), the agent semantically matches the identified subsequent work data with the completion standards of the corresponding project type. For example, if a wind power project has a "wind turbine blade replacement" record after completion, and the standard for this type of project requires that such work should be completed before acceptance, an anomaly warning is triggered. Furthermore, the agent will also simultaneously associate and identify data from dimensions such as the financial system (e.g., large payments for completed projects), the materials system (e.g., main material requisition not terminated), and the document system (e.g., missing acceptance reports). If at least one dimension of the data points to "the project has not reached a substantial completion status", it will ultimately be judged as a false completion.

[0043] In this embodiment of the application, optionally, the production project dataset is analyzed by a pre-trained agent for "unfinished and settled" to determine whether the power project has unfinished and settled status. This includes: analyzing the production project dataset by the pre-trained agent for "unfinished and settled" to identify the work progress data of a second engineering project after the settlement time of the power project; determining whether the power project has unfinished and settled status based on the type of the power project and the identified work progress data of the second engineering project; wherein the pre-trained agent for "unfinished and settled" has pre-learned settlement standard information for various types of power projects.

[0044] In this embodiment, the agent first precisely extracts the "settlement time" field from the production project dataset. This time point marks the point at which the project is deemed completed in the financial process. Based on this, the agent focuses on scanning engineering activity records generated after the settlement time to identify "secondary engineering project work progress data." For example, if the settlement time for a power distribution project is November 15, 2025, but there are still "cable laying" records on November 20 and "equipment commissioning" applications on November 25, these work data with timestamps later than the settlement time are marked as potential anomalies. The agent has built-in settlement standard information for various types of power projects (such as power transmission and transformation projects requiring completion of all equipment installation and commissioning and passing acceptance before settlement, and distributed photovoltaic projects requiring grid connection and stable operation before settlement). For the identified post-settlement work data, the intelligent agent dynamically matches it with the project type: If the work belongs to the core stage of the project, such as a "tower erection" record after the settlement of a transmission line project, and the standard for this type of project requires that such work be completed before settlement, it is directly judged as incomplete work settled first. If the work belongs to the auxiliary or rectification stage, such as a "fire protection facility reinstallation" record after the settlement of a substation project, further analysis is needed to determine whether it is rectification work caused by the failure of the previous acceptance. If the rectification work was not completed before settlement and was not clearly stated at the time of settlement, it is still considered as incomplete work settled first. In addition, the intelligent agent simultaneously correlates the following dimensions of data for judgment: Financial system: Check whether there are still project payment records after settlement (such as payment for unfinished projects); Materials system: Analyze the main material requisition status after settlement (such as the requisition of key equipment after settlement); Document system: Verify whether the acceptance report, settlement report and other documents are complete (such as the settlement report not mentioning unfinished work); Personnel system: Verify whether key personnel are still involved in the project after settlement (such as the project manager still submitting work logs after settlement). If at least one dimension of data points to "the project not reaching a substantial completion status at the time of settlement," it is confirmed as settlement before completion. Scenario 1: A wind power project's settlement date is October 1, 2025, but a "wind turbine blade replacement" record appears on October 5. After identification by the intelligent agent, combined with the wind power project settlement standard (requiring all wind turbines to pass 240 hours of trial operation), it is determined that: if the blade replacement was caused by a trial operation failure, and the trial operation was not completed before settlement, it is determined as settlement before completion. Scenario 2: A distribution network renovation project's settlement date is September 20, 2025, but a "meter installation" record appears on September 25. After analysis by the intelligent agent, it is found that: this work falls within the scope of the project contract, but it was not completed before settlement due to the supplier's delayed delivery, and this was not clearly stated in the settlement report; therefore, it is determined as settlement before completion.

[0045] In this embodiment of the application, optionally, the production project dataset is analyzed by a pre-trained agent for unreported completion to determine whether the power project has unreported completion status. This includes: analyzing the production project dataset by the pre-trained agent for unreported completion status to identify the completed engineering content of the power project; determining whether the power project has unreported completion status based on the type of the power project and the identified completed engineering content; wherein the pre-trained agent for unreported completion status has pre-learned completion standard information and unreported completion rules for various types of power projects; the completion standard information includes engineering content that conforms to the completion standard of the corresponding type of power project; and the unreported completion rules include at least the maximum project production maintenance time threshold for each piece of equipment required for the production of the corresponding type of power project.

[0046] In this embodiment, the agent first performs a full scan of the production project dataset, extracting explicit descriptions of "completed project content" (such as "cable laying completed" and "transformer installation accepted") and implicit clues (such as "72 hours of trouble-free operation" records in the equipment operation log and "main material requisition terminated" signals in the material system) through natural language processing technology. Simultaneously, by combining the timestamps in the project progress report, the actual completion time of each project content is determined, constructing a time-series completion map of the entire project lifecycle. Based on the completion standard information of various types of power projects injected during the pre-training phase (such as overhead line projects requiring completion of the entire process of "conductor erection - insulator installation - grounding resistance testing" and passing acceptance), the agent dynamically matches the identified completed project content: Forward matching: If the project type is "substation project," it verifies whether it includes standard project content such as "main transformer installation - commissioning - 72-hour trial operation"; Reverse verification: If a distribution project only completes "pole erection" but not "conductor erection," it is determined that it has not reached a substantial completion state. In addition, the "maximum project production maintenance time threshold" in the "completion without reporting" rule can be combined (e.g., the maximum maintenance time for a certain piece of equipment during production shall not exceed 15 days to ensure that the equipment is in normal working condition). If the equipment has completed maintenance but has not carried out the next maintenance after exceeding the threshold, a completion without reporting warning will be triggered.

[0047] For example, a photovoltaic project has completed all engineering work, including "support installation, module laying, and grid connection commissioning," and passed a 72-hour trial run, but failed to report completion information within 3 working days. After identification by the intelligent agent, based on the photovoltaic project completion standards and equipment maintenance thresholds, it is determined that the project was completed but not reported.

[0048] Furthermore, as Figure 1 In terms of specific implementation, this application provides a device for assessing the completion and commissioning of production projects based on a large model, such as... Figure 3 As shown, the device includes: The data acquisition module is used to acquire full data of production projects for different power projects from multiple production engineering management systems; The data integration module is used to integrate all the data of the production projects through a large model to form a production project dataset for each power project. The anomaly analysis module is used to analyze the production project dataset corresponding to each power project to determine whether there are any anomalies in the completion and commissioning status of each power project.

[0049] Optionally, the data integration module is used for: A large-scale model is used to perform semantic analysis on the full data of the production projects, identify and extract key project information for each power project, and integrate the key project information of each power project in chronological order to form a production project dataset for each power project. The production management system stores project information, equipment information, maintenance plans, and work orders for the power projects; the risk control platform stores risk monitoring data, production plans, and control information; the financial system stores information on the use of funds, cost accounting, and settlement for the power projects; the materials management system stores the procurement, supply, and usage of materials required for the power projects; and the document management system stores project plans, project design documents, project progress reports, project quality reports, project acceptance reports, feedback from on-site staff, and feedback from project management personnel for the power projects.

[0050] Optionally, the data integration module is also used for: Obtain production project data samples corresponding to multiple power project samples and production project dataset samples constructed based on the production project data samples, and mix the production project data samples corresponding to each power project sample into a training sample set; The general-purpose model is trained in three stages using the training sample set. The first stage involves guiding the model to analyze power projects based on the training sample set, and optimizing its parameters based on the analyzed power project names and the corresponding power project samples to train its ability to identify power projects. The second stage involves guiding the model trained in the first stage to perform data association on the data samples in the training sample set, constructing a data sample set corresponding to each power project name, and optimizing its parameters based on the constructed data sample set and the corresponding power project production dataset samples to train its ability to distinguish between different power project data. The third stage involves guiding the model trained in the second stage to extract key project information corresponding to each power project sample based on the data sample set corresponding to each power project sample and integrate it into production project dataset prediction information. The model's parameters are then optimized based on the production project dataset prediction information and production project dataset samples to train its ability to analyze and integrate key project information from production project data.

[0051] Optionally, the anomaly analysis module is used for: If the production project dataset corresponding to the power project contains completed data, the production project dataset is analyzed by a pre-trained fake completion agent to determine whether the power project is fake completed, and the production project dataset is analyzed by a pre-trained incomplete-to-complete agent to determine whether the power project is incomplete-to-complete. If the production project dataset corresponding to the power project does not contain completion data, the production project dataset is analyzed by a pre-trained agent that has not reported completion data to determine whether the power project has any unreported completion data.

[0052] Optionally, the anomaly analysis module is used for: The production project dataset is analyzed by a pre-trained fake completion agent to identify the work progress data of the first engineering project after the completion time of the power project. Based on the type of the power project and the identified work progress data of the first engineering project, it is determined whether the power project is falsely completed. The pre-trained fake completion agent has pre-learned completion standard information for various types of power projects.

[0053] Optionally, the anomaly analysis module is used for: The production project dataset is analyzed by a pre-trained agent that completes projects before completion to identify the work progress data of a second engineering project after the settlement time of the power project. Based on the type of the power project and the identified work progress data of the second engineering project, it is determined whether the power project has been completed before completion. The pre-trained agent that completes projects before completion has pre-learned settlement standard information for various types of power projects.

[0054] Optionally, the anomaly analysis module is used for: The production project dataset is analyzed by a pre-trained agent that identifies completed work content of the power projects. Based on the type of the power project and the identified completed work content, it is determined whether there is any unreported completion status for the power project. The pre-trained agent has pre-learned completion standard information and unreported completion rules for various types of power projects. The completion standard information includes work content that meets the completion standard of the corresponding type of power project, and the unreported completion rules include at least the maximum project production and maintenance time threshold for each piece of equipment required for the production of the corresponding type of power project.

[0055] It should be noted that other corresponding descriptions of the functional units involved in the production project completion and commissioning assessment device based on a large model provided in this application embodiment can be found in the following references. Figures 1 to 2 The corresponding descriptions in the method will not be repeated here.

[0056] This application also provides a computer device, specifically a personal computer, server, network device, etc. The computer device includes a bus, processor, memory, and communication interface, and may also include input / output interfaces and a display device. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores location information. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the various method embodiments.

[0057] Those skilled in the art will understand that the structure of the computer device described above is only a partial structure related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. A specific computer device may include more or fewer components, or combine certain components, or have different component arrangements.

[0058] In one embodiment, a computer-readable storage medium is provided, which may be non-volatile or volatile, having stored thereon a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0059] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0060] It should be noted that the user personal information involved in the embodiments of this application is all authorized (with the knowledge and consent) by the relevant parties or fully authorized by all parties, and the executing entity can obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with the relevant laws and regulations of the relevant countries and regions, and do not violate public order and good morals. It should be noted that if any software tools or components other than those of this company appear in the embodiments of this application, they are merely illustrative examples and do not represent actual use.

[0061] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0062] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0063] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for evaluating the completion and commissioning of production projects based on a large-scale model, characterized in that, The method includes: The system acquires full production data for different power projects from multiple production engineering management systems, including a production management system, a risk control platform, a financial system, a materials management system, and a document management system. The full dataset of the production projects is integrated by pre-trained large models to form a production project dataset for each power project. The production project datasets corresponding to each power project are analyzed separately to determine whether there are any abnormalities in the completion and commissioning status of each power project. The abnormalities in the completion and commissioning status include false completion, completion before completion, and failure to report completion.

2. The method according to claim 1, characterized in that, By integrating the full dataset of the production projects using a pre-trained large model, a production project dataset for each power project is formed, including: A large-scale model is used to perform semantic analysis on the full data of the production projects, identify and extract key project information for each power project, and integrate the key project information of each power project in chronological order to form a production project dataset for each power project. The production management system stores project information, equipment information, maintenance plans, and work orders for the power projects; the risk control platform stores risk monitoring data, production plans, and control information; the financial system stores information on the use of funds, cost accounting, and settlement for the power projects; the materials management system stores the procurement, supply, and usage of materials required for the power projects; and the document management system stores project plans, project design documents, project progress reports, project quality reports, project acceptance reports, feedback from on-site staff, and feedback from project management personnel for the power projects.

3. The method according to claim 2, characterized in that, Before integrating the full dataset of the production projects using a pre-trained large model to form a production project dataset for each power project, the method further includes: Obtain production project data samples corresponding to multiple power project samples and production project dataset samples constructed based on the production project data samples, and mix the production project data samples corresponding to each power project sample into a training sample set; The general-purpose model is trained in three stages using the training sample set. The first stage involves guiding the model to analyze power projects based on the training sample set, and optimizing its parameters based on the analyzed power project names and the corresponding power project samples to train its ability to identify power projects. The second stage involves guiding the model trained in the first stage to perform data association on the data samples in the training sample set, constructing a data sample set corresponding to each power project name, and optimizing its parameters based on the constructed data sample set and the corresponding power project production dataset samples to train its ability to distinguish between different power project data. The third stage involves guiding the model trained in the second stage to extract key project information corresponding to each power project sample based on the data sample set corresponding to each power project sample and integrate it into production project dataset prediction information. The model's parameters are then optimized based on the production project dataset prediction information and production project dataset samples to train its ability to analyze and integrate key project information from production project data.

4. The method according to claim 3, characterized in that, The production project datasets for each power project were analyzed separately to determine whether any power projects exhibited abnormalities in their completion and commissioning status, including: If the production project dataset corresponding to the power project contains completed data, the production project dataset is analyzed by a pre-trained fake completion agent to determine whether the power project is fake completed, and the production project dataset is analyzed by a pre-trained incomplete-to-complete agent to determine whether the power project is incomplete-to-complete. If the production project dataset corresponding to the power project does not contain completion data, the production project dataset is analyzed by a pre-trained agent that has not reported completion data to determine whether the power project has any unreported completion data.

5. The method according to claim 4, characterized in that, The production project dataset is analyzed using a pre-trained fake completion agent to determine whether any power projects are falsely completed, including: The production project dataset is analyzed by a pre-trained fake completion agent to identify the work progress data of the first engineering project after the completion time of the power project. Based on the type of the power project and the identified work progress data of the first engineering project, it is determined whether the power project is falsely completed. The pre-trained fake completion agent has pre-learned completion standard information for various types of power projects.

6. The method according to claim 4, characterized in that, The production project dataset is analyzed using a pre-trained agent that terminates projects before completion to determine whether any power projects are terminated before completion, including: The production project dataset is analyzed by a pre-trained agent that completes projects before completion to identify the work progress data of a second engineering project after the settlement time of the power project. Based on the type of the power project and the identified work progress data of the second engineering project, it is determined whether the power project has been completed before completion. The pre-trained agent that completes projects before completion has pre-learned settlement standard information for various types of power projects.

7. The method according to claim 4, characterized in that, The production project dataset is analyzed using a pre-trained agent that identifies projects that are not yet reported as completed, to determine whether any such projects exist. The production project dataset is analyzed by a pre-trained agent that identifies completed work content of the power projects. Based on the type of the power project and the identified completed work content, it is determined whether there is any unreported completion status for the power project. The pre-trained agent has pre-learned completion standard information and unreported completion rules for various types of power projects. The completion standard information includes work content that meets the completion standard of the corresponding type of power project, and the unreported completion rules include at least the maximum project production and maintenance time threshold for each piece of equipment required for the production of the corresponding type of power project.

8. A device for evaluating the completion and commissioning of production projects based on a large model, characterized in that, The device includes: The data acquisition module is used to acquire full data of production projects for different power projects from multiple production engineering management systems; The data integration module is used to integrate all the data of the production projects through a large model to form a production project dataset for each power project. The anomaly analysis module is used to analyze the production project dataset corresponding to each power project to determine whether there are any anomalies in the completion and commissioning status of each power project.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

10. A computer device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

Images