机组振动状态时序识别系统的判别函数分类性能分析  

Classification Quality Analysis for Discriminate Function of Turbogenerator Vibration State Time Series Identification System

作　　者：程懋华[1] 高亹[1] 

机构地区：[1]东南大学火电机组振动国家工程研究中心,南京210096

出　　处：《汽轮机技术》2003年第1期46-48,共3页Turbine Technology

基　　金：国家重点基础研究发展规划项目(G19998020300)。

摘　　要：对基于自回归模型机组振动状态时序识别系统常用的5种判别函数的判别性能,从理论上进行了定性分析,同时定量地比较了它们对试验摩擦振动状态的分类概率。指出了这5种判别函数各自的特点,对它们分类能力进行了评价。其评价结果为选择机组振动状态时序识别系统的判别函数提供了依据。The paper qualitatively analyzes five common discriminate functions of turbogenerator vibration state time series identification system based on AR model. The paper also compares their classification probabilities to experiment rub vibration states quantitatively. The paper gives out their features and evaluates their classification quality. These results provide a base for choosing proper discriminate function of turbogenerator vibration state time series identification system.

关 键 词：振动状态 时序识别系统 判别函数 自回归模型 

分 类 号：TH113.1[机械工程—机械设计及理论]