数据与计算发展前沿 ›› 2026, Vol. 8 ›› Issue (3): 96-109.

doi: 10.11871/jfdc.issn.2096-742X.2026.03.009

• 技术与应用 • 上一篇    下一篇

多源异构数据高质量数据集构建与关联敏感性分析识别技术研究

王迪1(),安冰1,*(),冯函宇1,范梓豪2,李明翰2,茹一伟2   

  1. 1 国家电网有限公司大数据中心， 北京 100052
    2 天津中科智能识别有限公司， 天津 300457
  • 收稿日期:2025-10-16 出版日期:2026-06-20 发布日期:2026-06-18
  • 通讯作者: 安冰
  • 作者简介:王迪，国家电网有限公司大数据中心，硕士，高级工程师，高级技师，国际大电网会议（CRGRE）信息通信专业委员会算力网络工作组成员，长期从事电力网络与数据安全相关工作，牵头完成多项国网公司数据安全规划及顶层设计，牵头或参与编制企业标准10余个、编制电力网络安全相关专著5册。
    本文中负责研究技术路线，提出通过基于扩散模型的MTabGen方法解决数据质量问题，以及负责构建动态图卷积网络DGDCN模型解决关联敏感性分析问题。
    WANG Di, holds a master’s degree and is a senior engineer and senior technician at the Big Data Center of State Grid Corporation of China. She is a member of the Computing-Force Network Task Force under the Information & Communication Committee of CIGRE. She has been long engaged in power-grid and data-security work. She has led several SGCC data-security master plans and top-level designs, headed or participated in drafting more than ten enterprise standards, and co-authored five monographs on power-network security.
    In this paper, she is responsible for the technical roadmap, including proposing the diffusion-model-based MTabGen approach to solve data-quality issues and constructing the dynamic graph convolutional network (DGDCN) model to address data association sensitivity analysis problems.
    E-mail: wangdi1220@aliyun.com|安冰，国家电网有限公司大数据中心，硕士，从事数据安全、规划设计相关工作，主要研究方向包括数据安全与计算机应用技术。
    本文中负责提出本研究的核心思想，课题的总体设计与规划，论文的修改、审阅与最终定稿。承担论文的通讯联络和学术责任。
    AN Bing, holds a master’s degree and works at the Big Data Center of State Grid Corporation of China. She is engaged in data security, planning and design, with main research interests including data security and computer application technology.
    In this paper, she is responsible for proposing the core concept of the research, the overall design and planning of the project, as well as revising, reviewing, and finalizing the manuscript. She serves as the corresponding author, handling correspondence and assuming academic responsibility.
    E-mail: anikab@163.com
  • 基金资助:
    国网大数据中心基于大模型的数据安全风险自动化研判处置关键技术研究项目(SGSJ0000HGJS2500036)

Research on Technology for Construction of High-Quality Multi-Source Heterogeneous Data Sets and Analysis & Identification of Associated Sensitivity

WANG Di1(),AN Bing1,*(),FENG Hanyu1,FAN Zihao2,LI Minghan2,RU Yiwei2   

  1. 1 State Grid Corporation of China Big Data Center, Beijing 100052, China
    2 Tianjin Zhongke Intelligent Recognition Co., Ltd, Tianjin 300457, China
  • Received:2025-10-16 Online:2026-06-20 Published:2026-06-18
  • Contact: AN Bing

RichHTML

4

PDF

49

摘要：

【目的】在数字化时代，多源异构数据呈现爆炸式增长，其蕴含的巨大价值日益凸显。各级电网日均处理超亿条网络访问日志，涵盖数值型、指令类别型、告警文本型等多元数据类型，这些数据广泛分布于调度自动化系统、物联网、新能源并网监测等关键业务场景，蕴含着支撑电网智能决策、设备状态预判、安全风险防控的巨大价值。然而，数据质量缺陷与关联敏感性风险成为制约这些数据价值释放的突出瓶颈。 【方法】为此，本文针对上述两大瓶颈，提出了一套面向数据“质量-安全”的综合技术方案。在高质量数据集构建层面，提出基于扩散模型的MTabGen方法，通过多模态联合优化实现数据缺陷的高精度插补；在数据关联敏感性分析层面，提出采用图卷积神经网络DGDCN构建数据关联图谱，识别敏感关联路径。 【结果】实验验证表明，MTabGen方法在准确率和完整性指标上显著优于传统数据构建方法；图卷积神经网络DGDCN在精确率、召回率和F1值上全面超越传统机器学习方法。

关键词: 多源异构数据, 数据集构建, 数据关联, 数据敏感性

Abstract:

[Background] In the digital era, multi-source heterogeneous data have experienced explosive growth, and its enormous inherent value has become increasingly prominent. Provincial state grid process over 100 million network access logs daily, covering diverse data types such as numerical data, command category data, and alarm text types data. These data are widely distributed in key business scenarios including dispatching automation systems, the Internet of Things, and new energy grid-connected monitoring, containing huge value in supporting intelligent decision-making of power grids, equipment status prediction, and safety risk prevention and control. However, data quality defects and associated sensitivity risks have become prominent bottlenecks restricting the realization of data value. [Methods] To this end, this paper focuses on technologies for constructing high-quality datasets of multi-source heterogeneous data and analyzing and identifying associated sensitivity. In terms of high-quality dataset construction, the MTabGen method based on a diffusion model is proposed, which realizes high-precision imputation of data defects through multi-modal joint optimization. In the aspect of data association sensitivity analysis, the graph convolutional neural network DGDCN is proposed to construct data association graphs and identify sensitive association paths. [Results] Experimental verification shows that the MTabGen method is significantly superior to traditional data construction methods in terms of accuracy and completeness indicators; the graph convolutional neural network DGDCN comprehensively outperforms traditional machine learning methods in precision, recall, and F1-score.

Key words: multi-source heterogeneous data, dataset construction, data association, data sensitivity