一种变压器健康状态的改进免疫识别方法  

Improved immune recognition method of transformer health

作　　者：程宏波 刘嘉 康嘉斌 王勋 CHENG Hong-bo;LIU Jia;KANG Jia-bin;WANG Xun(School of Electrical and Automatic Engineering,East China Jiaotong University,Nanchang 330013,China)

机构地区：[1]华东交通大学电气与自动化工程学院,江西南昌330013

出　　处：《电力科学与技术学报》2018年第2期141-147,共7页Journal of Electric Power Science And Technology

基　　金：国家自然科学基金(51267004);江西省重点研发计划项目(20161BBH80033);江西省博士后科研择优资助项目(2016KY36)

摘　　要：变压器是电力系统的核心设备,对变压器的状态进行准确识别有利于掌握变压器的健康状况,对实施变压器的状态检修具有重要的指导意义。在分析变压器特征指标与运行状态之间联系的基础上,将变压器的健康状态划分成4个不同的区域。借鉴免疫系统的变异和克隆选择原理,构建变压器健康状态的免疫识别方法,并针对状态空间中检测器的冗余低效问题,引入清除和吞并机制,实现对变压器状态的自适应免疫识别。仿真对比表明,改进策略后的免疫识别算法具有更好的实时性和适应性,进一步提高了变压器状态识别的速度和灵敏度。Transformer is the core equipment of power system.It is of great significance to implement the transformer condition maintenance via recognizing transformer states that reflect transformer health.In this case,the health state is divided into four different areas on the basis of analyzing the relationship between the characteristic indexes and the operating state of transformer.The variation and clonal selection principle of immunity theory is referred to construct an immune recognition method for transformer health state.Besides,the clearance and annexation mechanism of transformer state recognition strategies is invoked for the redundant and inefficient problems of detectors in the state space.A simulation is included to show that the improved immune recognition algorithm has better real-time and adaptive performance.The recognition speed and sensitivity of transformer state is well improved.

关 键 词：变压器 状态识别 免疫识别算法 检测策略 自适应 

分 类 号：TM417[电气工程—电器]