中文摘要：

笔者系统地介绍了基于Measurement Studio软件平台研制的一种新型交流局部放电宽带检测与分析仪。其主要包括利用NI5112高速示波采集卡和前置信号处理单元实现的采样率为100MS/s、模拟宽带为10kHz～40MHz的局部放电脉冲宽带检测技术,和基于脉冲波形时频特征的脉冲群实用快速分类技术以及基于PRPD和幅值统计分布谱图的放电识别技术。其工作原理为：首先,被同步交流电压信号触发的检测系统,单周期（20ms）或多周期采集来自宽带传感器的脉冲信号,并利用移动提取窗实现脉冲波形及其对应相位信息提取;其次,利用脉冲群实用快速分类技术将脉冲群分为若干个由相同波形特征脉冲组成的子类脉冲群,从而实现多局放分类或随机干扰脉冲群剔除;最后,利用统计算子对各个子类脉冲群对应的PRPD和幅值统计分布谱图进行放电指纹提取,共38个放电指纹形成识别样本文件。220kV电压等级的气体绝缘组合电器（GIS）的母线尖刺和筒壁尖刺放电的试验结果分析表明,研制的检测系统具有抗干扰能力强、放电识别准确,并能够解决和处理多局放源问题。

英文摘要：

This paper systematically introduced a novel wide-band detection and analysis device for AC partial discharge（PD） pulse signals,which was developed based on measurement studio software platform.It contains some practical techniques of PD detection and analysis,while the PD pulse signals extraction resorts to a high speed digitizer with a sampling rate up to 100 MS/s and a bandwidth of 10 kHz to 40 MHz.The PD pulse sequence classification was realized by the fast grouping technique based on pulse waveshape characteristics in time and frequency domain,and the PD identification worked on the fingerprints deduced from several histograms of phase-resolved partial discharge（PRPD） and magnitude statistical distributions（MSD）.The PD data processing of the system were splited into three steps.Firstly,signal detection channel samples synchronization PD data with a length of one AC period（20 ms） or several AC periods after triggered by the AC voltage signals from a divider.Then pulse waveshapes and their corresponding phase information were extracted by the moving extraction window（MEW） with a simple algorithm.Secondly,the PD pulse sequence was separated into sub-groups using the pulse fast grouping technique,which results each one of the sub-groups was relevant to a well defined PD phenomenon or electrical noise.And lastly,the PD identification was made based on 38 fingerprints deduced from the PRPD and MSD of each sub-group.GIS experimental results showed that the developed AC PD measurement system has good performance in anti-interference,PD classification and PD identification,and can solve and deal with the multi-PD source problem.