电力大系统振荡数据的可靠性挖掘算法

Reliability Mining Algorithm for Power System Oscillation Data

提出一种基于频谱包络特征提取和时频分析的电力大系统振荡数据的可靠性挖掘算法。构建了电力大系统数据传输和存储模型,对电力大系统中的数据进行频谱包络特征提取,以此为数据基础,对电力大系统数据信息库振荡数据的包络矢量实行高斯离散采样,采用决策树分类方法对时频分析后的振荡数据进行初步的数据筛选,采用时频分析和决策树分类方法实现对振荡的可靠性分类挖掘。仿真结果表明,采用该算法进行电力大系统中的振荡数据频谱包络特征提取,能有效挖掘出电力大系统中的振荡数据的分层结构特征和细节,数据挖掘精度较高,抗干扰性能较强。

This paper presents a reliability mining algorithm for power system oscillation data based on spectral envelope feature extraction and time frequency analysis.Construct the power system data transmission and storage model,conduct the spectral envelope features extraction for power system data,based on this basic data,implement discrete Gaussian sampling on power system data repository oscillation data envelope vector,do the preliminary vibration data screening after time-frequency analysis by using decision tree classification method,then achieve the reliability classification mining of oscillation data by using frequency analysis and decision tree classification method.Simulation results show that the algorithm can effectively extract the hierarchical structural features and details of the vibration data in power system with a strong anti-interference performance.