PDF《FRFT-KPCA》

第17 卷第8期2013 年8月电机与控制学报Electric Machines and Control Vol.

17 No.

8 Aug.

2013 基于FRFT-KPCA的模拟电路非线性故障特征提取 孙靖杰1,赵建军1,王汉昌1,乔凤兰1 (1. 海军航空工程学院青岛校区 航空军械火控系, 山东 青岛 266041;

2. 海军航空工程学院 兵器科学与技术系, 山东 青岛 264001) 摘要:针对模拟电路受非线性及元件容差影响而导致响应信号在时域和频域都出现耦合, 造成 故障特征提取困难的问题, 结合分数阶傅里叶变换和核主成分分析理论提出一种非线性故障特征 提取方法.利用分数阶傅里叶变换对耦合信号进行预处理, 采用粒子群优化算法寻找最优分数 阶p, 实现耦合信号在分数阶域最大程度的分离.采用核主成分分析对非线性特征进行维数压缩, 实现故障特征提取.实验结果表明, 在时域或频域相互耦合的信号经分数阶傅里叶变换后, 在分 数阶域上耦合程度明显减弱, 核主成分分析能够有效处理信号中的非线性信息, 特征提取效果要 优于其他线性特征提取方法.经过分数阶傅里叶变换和核主成分分析相结合的方法所提取的故 障特征使故障模式具有更好的可分性. 关键词:分数阶傅里叶变换;

核主成分分析;

模拟电路;

类内类间散布矩阵;

特征提取 中图分类号:TP

306 文献标志码:A 文章编号:1007C449X(2013)08C0001C10 Nonlinear fault features extraction for analog circuit based on FRFT-KPCA SUN Jing-jie1 , ZHAO Jian-jun2 , WANG Hang-chang1 , QIAO Feng-lan1 (1. Department of Aeronautical Armament and Fire Control, Naval Aeronautical and Astronautical University Qingdao Branch, Qingdao 266041, China;

2. Department of Ordnance Science and Technology, Naval Aeronautical and Astronantical University, Yantai 264001, China) Abstract: Aiming at the dif?culty of fault features extraction since response signals of analog circuit cou- pling in both time domain and frequency domain under the effect of nonlinear and tolerance, a nonlinear fault features extraction method based on fractional Fourier transform (FRFT) and kernel principal com- ponent analysis (KPCA) was proposed. FRFT was used as decoupling pretreatment for signals, which used particle swarm optimization (PSO) to seek the optimal fractional order p to separate signals great- ly in fractional domain. KPCA was applied to compress the dimension of nonlinear features to extract fault features. The experiment results show that after FRFT, signals coupling in time domain of frequen- cy domain can decouple obviously in fractional domain. Compared with other linear feature extraction methods, KPCA can obtain better extraction effect since it can analysis nonlinear information of sig- nals. So the obtain features extracted by the method based on FRFT-KPCA can enhance the divisible characteristic of different modes. Key words: fractional fourier transform;

kernel principal component analysis;

analog circuit;

within-class and among-class scatter matrix;

feature extraction 收收收稿 稿 稿日 日 日期 期期: : :2012C06C01 基基基金 金 金项 项 项目 目目: : :国家自然科学基金项目(60802088);

教育部新世纪优秀人才支持计划项目(NCET-05-0912) 作作作者 者 者简 简 简介 介介: : :孙靖杰(1983―), 女, 博士, 讲师, 研究方向为故障诊断、状态监测;

赵建军(1965―), 男, 博士, 教授, 博士生导师, 研究方向为测控技术、复杂系统故障诊断、参数标校;