CN-122022708-A - Method and device for judging completion of production project based on large model, storage medium and computer equipment

CN122022708ACN 122022708 ACN122022708 ACN 122022708ACN-122022708-A

Abstract

The application discloses a large-model-based production project completion production research judgment method, a large-model-based production project completion research judgment device, a large-model-based production project completion research judgment storage medium and computer equipment, wherein the large-model-based production project completion research judgment method comprises the steps of acquiring 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 wind control platform, a financial system, a material management system and a document management system, wherein the production project total data are integrated through a pre-trained large model to form a production project data set of each electric power project, and the production project data sets corresponding to each electric power project are respectively analyzed to determine whether an completion production state abnormality exists in each electric power project or not, wherein the completion production state abnormality comprises false completion, incomplete completion and incomplete completion.

Inventors

CHEN JIE
YOU CHUN
Hou Rongjun

Assignees

国网重庆市电力公司经济技术研究院

Dates

Publication Date: 20260512
Application Date: 20251209

Claims (10)

1. The method for judging completion and production of the production project based on the large model is characterized by comprising the following steps of: Acquiring production project total data of different power projects from a plurality of production project management systems, wherein the production project management systems comprise a production management system, a wind control platform, a financial system, a material management system and a document management system; integrating the full production project data through a pre-trained large model to form a production project data set of each electric power project; and analyzing the production project data set corresponding to each electric power project respectively to determine whether each electric power project has abnormal completion production state, wherein the abnormal completion production state comprises false completion, incomplete completion and complete non-report.
2. The method of claim 1, wherein integrating the production item total data by a pre-trained large model forms a production item dataset for each power item, comprising: The production project total data are subjected to semantic analysis through a large model, project key information of each electric power project is identified and extracted, the project key information of each electric power project is integrated according to time sequence to form a production project data set of each electric power project, the production management system stores project information, equipment information, maintenance plans and work tickets of the electric power project, the wind control platform stores risk monitoring data, production plans and management information, the financial system stores fund use information, cost accounting information and settlement information of the electric power project, the material management system stores purchase, supply and use conditions of materials required by the electric power project, and the document management system stores project plans, project design files, project progress reports, project quality reports, project acceptance reports, on-site staff feedback information and project manager feedback information of the electric power project.
3. The method of claim 2, wherein prior to integrating the production project volume data by a pre-trained large model to form a production project data set for each power project, the method further comprises: Acquiring production project data samples corresponding to a plurality of power project samples and a production project data set sample constructed according to the production project data samples, and mixing the production project data samples corresponding to the power project samples into a training sample set; The method comprises the steps of training a general large model by utilizing a training sample set, carrying out three stages of training on the general large model, wherein one stage of training comprises the steps of guiding the general large model to carry out electric project analysis based on the training sample set, carrying out parameter optimization on the general large model based on the analyzed electric project names and the electric project names corresponding to the electric project samples so as to train the recognition capability of the general large model on the electric projects, the two stages of training comprises the steps of guiding the general large model after one stage of training to carry out data association on data samples in the training sample set, constructing a data sample set corresponding to each electric project name, carrying out parameter optimization on the general large model based on the constructed data sample set and production project data set samples of corresponding electric projects so as to train the data distinguishing capability of the general large model on different electric projects, and the three stages of training comprises the steps of guiding the general large model after two stages of training to extract project key information corresponding to each electric project sample based on the data sample set corresponding to the production project data set prediction information and integrating the project key information into the production project data set prediction information, carrying out parameter optimization on the general large model based on the production project data set prediction information and the production project data set sample so as to train the general large model to carry out parameter optimization on the project key information and the project data and the integration capability.
4. A method according to claim 3, wherein analyzing the production item data set for each power item to determine whether there is an as built production status anomaly for each power item, respectively, comprises: If the production item data set corresponding to the electric power item contains finishing data, analyzing the production item data set by a pre-trained false finishing agent to determine whether the electric power item has false finishing, and analyzing the production item data set by a pre-trained outstanding finishing agent to determine whether the electric power item has outstanding finishing; And if the production project data set corresponding to the electric power project does not contain the finishing data, analyzing the production project data set by a pre-trained finishing unreported agent to determine whether the electric power project has finishing unreported.
5. The method of claim 4, wherein analyzing the production item data set by a pre-trained false completion agent to determine whether a false completion exists for the power item comprises: And analyzing the production project data set through a pre-trained false finishing agent to identify first project work development data after the finishing time of the power project, and determining whether the power project has false finishing according to the type of the power project and the identified first project work development data, wherein the pre-trained false finishing agent learns finishing standard information of various types of power projects in advance.
6. The method of claim 4, wherein analyzing the production item data set by a pre-trained pre-harvest agent to determine whether the power item has a pre-harvest comprises: And analyzing the production project data set through a pre-trained outstanding settlement agent to identify second project work development data after the settlement time of the electric power project, and determining whether the electric power project has outstanding settlement according to the type of the electric power project and the identified second project work development data, wherein the pre-trained outstanding settlement agent learns settlement standard information of various electric power projects in advance.
7. The method of claim 4, wherein analyzing the production item data set by a pre-trained complete unreported agent to determine whether the power item has a complete unreported comprises: And analyzing the production project data set through a pre-trained finished unreported agent to identify finished engineering content of the electric power project, and determining whether the electric power project has finished unreported according to the type of the electric power project and the identified finished engineering content, wherein the pre-trained finished unreported agent learns finishing standard information and finishing unreported rules of various types of electric power projects in advance, the finishing standard information comprises the finishing standard engineering content of the electric power project conforming to the corresponding type, and the finishing unreported rules at least comprise maximum project production and maintenance time thresholds of all equipment required by the production of the electric power project of the corresponding type.
8. A large model-based production project completion commissioning research and judgment device, characterized in that the device comprises: The data acquisition module is used for acquiring the production project total data of different power projects from a plurality of production project management systems; the data integration module is used for integrating the production project total data through a large model to form a production project data set of each electric power project; And the abnormality analysis module is used for respectively analyzing the production project data sets corresponding to each electric power project to determine whether the completion production state abnormality exists in each electric power project.
9. A storage medium having stored thereon a computer program, which when executed by a processor, implements the method of any 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 the processor implements the method of any one of claims 1 to 7 when executing the computer program.

Description

Method and device for judging completion of production project based on large model, storage medium and computer equipment Technical Field The application relates to the technical field of power project management, in particular to a method, a device, a storage medium and computer equipment for judging completion and production of a production project based on a large model. Background In the existing production engineering data of the power industry, overhaul operation data is available in PMS3.0 overhaul plans, work tickets and technical improvement overhaul plate functions, but the data quantity associated with production engineering projects is small. Since the project information is unnecessary between the maintenance schedule and the work ticket, the project information cannot be known from the maintenance schedule (month, week, day schedule). The milestone plan management mode is widely applied to a production engineering management system, but part project management staff cannot execute the milestone plan in place, the progress is not really mastered according to a planning node, the real accuracy judgment of the completion time of the production engineering at the present stage is mostly dependent on manpower, the data among a plurality of systems are queried offline and expert judgment is added for manual verification, the efficiency is low, and the accuracy cannot be guaranteed. Therefore, problems such as false completion, incomplete report and the like frequently occur, and the completion production date is unrealistic and inaccurate, so that great risks are brought to legal compliance management of the electric company. Disclosure of Invention In view of this, the embodiment of the application provides a method, a device, a storage medium and computer equipment for judging completion and production of a production project based on a large model. According to one aspect of the present application, there is provided a method for judging completion of a production project based on a large model, the method comprising: Acquiring production project total data of different power projects from a plurality of production project management systems, wherein the production project management systems comprise a production management system, a wind control platform, a financial system, a material management system and a document management system; integrating the full production project data through a pre-trained large model to form a production project data set of each electric power project; And analyzing the production project data set corresponding to each electric power project respectively to determine whether each electric power project has abnormal completion production state, wherein the abnormal completion production state comprises. Optionally, integrating the production project volume data by a pre-trained large model to form a production project data set for each power project, comprising: The production project total data are subjected to semantic analysis through a large model, project key information of each electric power project is identified and extracted, the project key information of each electric power project is integrated according to time sequence to form a production project data set of each electric power project, the production management system stores project information, equipment information, maintenance plans and work tickets of the electric power project, the wind control platform stores risk monitoring data, production plans and management information, the financial system stores fund use information, cost accounting information and settlement information of the electric power project, the material management system stores purchase, supply and use conditions of materials required by the electric power project, and the document management system stores project plans, project design files, project progress reports, project quality reports, project acceptance reports, on-site staff feedback information and project manager feedback information of the electric power project. Optionally, before integrating the production project volume data by the pre-trained large model to form a production project data set for each power project, the method further comprises: Acquiring production project data samples corresponding to a plurality of power project samples and a production project data set sample constructed according to the production project data samples, and mixing the production project data samples corresponding to the power project samples into a training sample set; The method comprises the steps of training a general large model by utilizing a training sample set, carrying out three stages of training on the general large model, wherein one stage of training comprises the steps of guiding the general large model to carry out electric project analysis based on the training sample set, carrying out parameter optimization on the general large model based on the analyzed el