一种基于傅里叶描述子的轴心轨迹自动识别方法  

A METHOD BASED ON FOURIER DESCRIPTORS TO RECOGNIZE SHAFT ORBIT

作　　者：付波[1] 周建中[1] 陈文清[1] 余炳辉[1] 

机构地区：[1]华中科技大学水电与数字化工程学院,湖北省武汉市430074

出　　处：《电力系统自动化》2004年第12期40-44,50,共6页Automation of Electric Power Systems

基　　金：国家跨世纪优秀人才基金资助项目(2003274)

摘　　要：傅里叶描述子是分析和识别物体形状的重要方法之一。文中详细讨论了傅里叶描述子的原理、不变性和实际应用方法;运用人工神经网络分析方法,以傅里叶描述子作为特征向量,对汽轮发电机组轴心轨迹的自动识别进行了探讨;并对3类典型的轴心轨迹的自动识别进行了仿真试验,结果表明该方法是有效、可行的。Fourier Descriptors (FDs) is an important method used in shape analysis and recognition. A new method based on FDs is presented to recognize the shaft orbits of turbine rotor. The principle, invariants and application of FDs are discussed in detail. The method of Artificial Neural Network combined with FDs is used to identify the shaft orbit of turbine rotor. Simulation results of automatic identification for three typical axis orbits are provided. It is proved that the method is effective and practicable.

关 键 词：傅里叶描述子 人工神经网络 汽轮发电机 轴心轨迹 识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]