改进S变换自适应算法与电能质量检测及扰动信号特征提取方法研究

项目名称： 改进S变换自适应算法与电能质量检测及扰动信号特征提取方法研究

项目编号： No.51277058

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 电工技术

项目作者： 唐求

作者单位： 湖南大学

项目金额： 75万元

中文摘要： 电能质量问题越来越引起重视，其指标检测与扰动信号特征提取是发现电能质量问题并进行有效治理的前提。本项目研究电能质量扰动信号时频分析改进算法，建立电能质量指标统一检测与扰动信号特征提取新方法，具体包括：①以S变换为基础，构建适宜电能质量参数分析的新型优化窗函数，通过加窗插值改进FFT和卷积运算创建改进S变换算法；②探索改进S变换数学特性，确定频谱最佳探索函数，实现离散频谱自适应分割，形成改进S变换自适应算法；③分析各种扰动信号的频谱特性、持续时间、幅值变化等特征信息，确定能准确反映各种扰动信号特征的时域、频域特征量，简化电能质量各项指标的检测方法，建立统一的基于改进S变换自适应算法的电能质量指标检测与扰动信号特征提取方法；④构建基于嵌入式系统的电能质量监测实验平台，实验验证理论研究成果的正确性。项目研究的关键技术和方法可应用于电能质量分析、数字信号处理、模式识别和故障诊断等领域。

中文关键词： 电能质量扰动；Kaiser自卷积窗；包络极值算法；改进S变换；时频分析

英文摘要： Power quality (PQ) issues cause more and more attentions. The parameters measurement and disturbances feature extraction are the fundament of the detection and improvement of PQ issues. The aim of this project is to study the improved time-frequency analysis method of the PQ disturbance signals, and present the innovation method of detection of unique PQ parameters and extraction of PQ disturbances feature. The major works include: ① the construction of the new optimization window based on the S transform which is suitable for the PQ parameter analysis, and the improved S transform based on the interpolation FFT and convolution operation; ② the establish of the improved adaptive S transform by investigating the mathematic features of the improved S transform and by optimizing spectral analysis function to perform the adaptive spectral segmentation; ③ the analysis of the characteristics of the PQ disturbance signals including the spectral, continuous time, the changing of the amplitude, etc., the confirmation of the features of the PQ disturbance signals in frequency and time domains, the simplification of the measurement method of PQ parameters, and the establish of the unique improved adaptive S transform for PQ measurement and disturbances feature extraction; ④the design of the PQ measurement platform based on

英文关键词： power quality disturbance；the self-convolution Kaiser window；envelope extremal algorithm；improved S-transform；time-frequency analysis