摘要
针对传统电力设备故障预测方法在对关联维特征提取时,存在故障信息冗余、误差较大的缺点,提出基于改进关联规则特征分析方法的电力设备故障预测算法。采用平均互信息方法和虚假最近邻点方法进行电力设备故障信息相空间重构,在高维相空间中,将电力设备的故障信号模拟为一个非线性时间序列波形,构建故障信号关联规则指向性特征约束函数,提取故障信号关联维特征,通过关联规则指导实现故障预测改进。仿真结果表明,该算法在进行电力设备故障预测时,能有效反应电力设备故障信号的关联内部特征信息,实现对电力设备故障类别的诊断,对提高电力设备故障类别诊断的准确率有现实意义。
The correlation dimension feature extraction in the traditional fault prediction algorithm for power equipment is prone to failure information redundancy and large errors. To address these shortcomings,this paper proposes a new electric power equipment fault prediction algorithm based on the improved association rule feature analysis method. The average mutual information method and false nearest adjacent points method are used to reconstruct the fault information phase space for power equipment. In the high dimensional phase space,the fault signal is simulated as a nonlinear time series,and the directional feature constraint function of the fault signal association rules is constructed to extract the fault signal characteristics of correlation dimensions,and improvement of fault prediction is realized through the association rule guidance. The simulation results show that the algorithm adopted to improve the power equipment failure prediction can effectively reflect the associated internal feature information of the fault signal of power equipment,and realize the diagnosis of power equipment fault category, and it is of realistic significance to enhance the accuracy in power equipment failure diagnosis.
出处
《电网与清洁能源》
北大核心
2015年第10期83-88,共6页
Power System and Clean Energy
基金
内蒙古教育厅项目(NJZC14387)~~
关键词
电力设备
故障信号
改进关联规则
power equipment
fault signal
improved association rules
