基于集成学习的电力设备运行全流程数据共享方法  

Integrated Learning Based Data Sharing Method for the Whole Process of Power Equipment Operation

作　　者：刘维嘉 LIU Weijia(Information Center of Guizhou Power Grid Co.,Ltd.,Guiyang 550000,Guizhou,China)

机构地区：[1]贵州电网有限责任公司信息中心,贵州贵阳550000

出　　处：《电力大数据》2023年第3期37-43,共7页Power Systems and Big Data

摘　　要：电力系统设备的检修计划将逐步由传统的定期检修转变为基于设备运行状态的检修,从而提高检修任务针对性,避免造成人力物力的浪费。基于设备运行状态安排检修计划的关键在于及时准确把握设备状态,随着相关检测技术的不断发展,在线监测、离线试验、人工巡视、无人机巡视等多种状态检测手段目前广泛应用于设备的状态检测之中,并积累了大量的历史数据。电力设备运行数据共享过程中存在数据安全系数过低问题,为此,本文引入集成学习技术,开展对电力设备运行全流程数据共享方法设计研究。通过构建电力设备运行全流程数据共享框架,根据集成学习提取电力设备运行全流程特征,通过半诚实模型和恶意模型实现电力设备运行全流程数据共享。实验证明,新的共享方法与区块链共享方法相比,能够有效扩大共享数据在电力设备运行全流程中的覆盖范围,并提高数据在共享时的安全性。The maintenance plan for power system equipment is gradually shifting from traditional periodic maintenance to maintenance based on equipment operational status.This shift aims to improve the targeted nature of maintenance tasks and avoid unnecessary manpower and resource wastage.The key to scheduling maintenance plans based on equipment operational status lies in promptly and accurately assessing the equipment's condition.With the continuous development of relevant detection technologies,various status detection methods such as online monitoring,offline testing,manual inspections,and drone inspections are widely employed and have accumulated a large amount of historical data.However,there is a low data security coefficient in the process of sharing power equipment operational data.To address this issue,this paper introduces ensemble learning technology to design a method for sharing comprehensive operational data of power equipment.By constructing a framework for sharing comprehensive operational data of power equipment and utilizing ensemble learning,the paper extracts comprehensive operational features of power equipment and realizes data sharing throughout the entire operational process using semi-honest models and malicious models.The experimental results demonstrate that the proposed sharing method,compared to blockchain-based sharing methods,effectively expands the coverage of shared data in the comprehensive operational process of power equipment and improves the security of data sharing.

关 键 词：集成学习 半诚实模型 恶意模型 全流程数据 电力设备 

分 类 号：TM732[电气工程—电力系统及自动化]