摘要
针对目前电能质量扰动识别的实时性和识别率不高的问题,研制了一种基于DSP与FPGA的实时电能质量监测系统。该监测系统采用FPGA控制高精度的8通道16位同步AD转换器AD7606完成电能质量信号的同步采集,DSP(TMS320DM642)从FPGA内部的FIFO存储器读取采集的电能质量数据,然后对其进行双密度双树离散小波变换(DD-DTDWT),并计算其小波系数Shannon熵,最后用训练得到的最小二乘支持向量机分类树对扰动信号进行分类识别。实验测试结果表明:该电能质量监测系统能够连续实时对8路电能质量进行准确监测,对8种常见扰动信号的平均识别率可以达到99.1%。
A real time power quality monitoring system based on DSP and FPGA is developed, which is capable of enhancing the recognition rate of power quality disturbances. The monitoring system adopts 8-channel 16-bit AD converter AD7606 controlled by FPGA to acquire the power quality signals synchronously. Then DSP (TMS320DM642) reads the power quality data from the FIFO storage in FPGA and then treats them with double-density dual-tree discrete wavelet transform (DD-DT DWT), and calculates the wavelet coefficient Shannon entropy. At last, a binary tree of SVM is adopted to classify the power quality disturbances. The experimental results indicate that the developed instrument can continuously monitor power quality of eight channels respectively, and the average recognition rate for eight kinds of common disturbance signals can achieve 99.1%.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2012年第12期125-129,共5页
Power System Protection and Control
基金
国家自然科学基金项目(50177139)
重庆市科技攻关项目(CSTC2011AC3066)~~
关键词
电能质量
双密度双树离散小波变换
小波系数Shannon熵
支持向量机
识别率
power quality
double-density dual-tree discrete wavelet transform (DD-DT DWT)
wavelet coefficient Shannon entropy
support vector machine (SVM)
recognition rate
作者简介
王平(1976-),男,博士,副教授,主要研究方向为电力系统信号处理、电能质量分析,E-mail:cqu_dq@163.com;高阳(1987-),男,硕士研究生,主要研究方向为电力系统分析、电路与系统;王林泓(1974-),女,博士,副教授,主要研究方向为电力系统信号获取与处理及模式识别。