自动化分割识别与数据校验方法在核电电子文件归档审查中的应用研究

Research on the Automatic Segmentation Recognition and Data Checking Method in Nuclear Electronic File Archiving

摘要: 为有效规范管控海量设计文件的质量，中广核开发了过程管理系统，对文件全生命周期通过系统进行管控，同时将文件质量控制提前到文件生成阶段，实现核电文档管理一体化、电子化和自动化，该系统建立起全新的项目文档前端控制模式，确保项目文档有效受控与准确传递。但该系统的“移交归档”电子文件审查环节，即核查结构化文档元数据与电子文件实体一致性，仍需投入大量人力。为解决该问题，笔者将人工智能(AI)技术引入系统，并依托系统进行二次开发，成功解决这一难题。本文将从企业知识管理、软件及文档管理信息化的交叉技术领域出发，阐述人工智能(AI)技术如何在文档管理领域创新应用，提高文档流转的规范性与效率。

Abstract: In order to effectively regulate and control the quality of mass design documents, CGN developed a process management system, which not only controlled the whole life cycle of documents, but advanced the document quality control to the stage of file generation as well, and realized the integration and automation of nuclear power document management. The system established a new project document front-end control mode to ensure the effective control and accurate transmission of the project documents. However, the electronic document review process in the system, which is to verify the consistency between structured document metadata and electronic document entities, still requires a large amount of manpower. To solve the problem, we will introduce AI technology into the system for secondary development. This paper starts from the enterprise knowledge management and the intersection of software and document management informatization, and elaborates how AI technology innovates in the field of document management and improves the stand-ardization efficiency of document flow.