摘要
提出了一种基于遗传算法优化支持向量机的故障诊断模型。它利用遗传算法对支持向量机同时对传统的时域特征参量子集和核参数同时优化,以达到选择最优的设备故障主导特征参数组合的目的,实现对机器不同类型故障的识别。对齿轮故障诊断的结果表明它有效提高了多分类支持向量机的故障分类准确性。
A new classification model based on genetic programming ( GA ) and support vector machine ( SVM ) for machine fault diagnosis is proposed. The model adopts the hybrid GA-SVM strategy which simultaneously performs the optimization of conventional time domain fea- tures parameters subset and core parameters for achieving the goal of selecting the optimal feature parameters combination related to fault of mechanical equipment, and realizing the recognition of different kinds of faults of machines. Experiments of gears fault detection shows the classification ability of multi-class support vector machine is improved after feature extraction and selection.
出处
《计算机应用与软件》
CSCD
2009年第1期62-63,共2页
Computer Applications and Software
基金
中国科学院创新基金资助项目(200417009)
关键词
支持向量机
遗传算法
特征选择
故障诊断
Support vector machine Genetic algorithm Feature selection Fault detection