摘要
在导弹试验过程中,及时进行故障诊断是保障试验安全的关键,故障诊断的方法一般是采用分析提取到的特征数据,来预判导弹飞行的预期效果和发生故障的概率。在分析过程中需要处理来至多只传感器的测量数据,而其特征向量维数较多,一般的数据处理方法不能满足要求。因此,提出了1种利用小波包分析辅以人工神经网络分析处理的故障诊断方法,实现了对导弹飞行状态的故障预判和特性预测。
In the course of the missile tests,timely diagnosis is the key to protect the safety test,fault diagnosis method is generally used to extract the characteristics of data to predict the expected missile flight performance and the probability of failure.Need to be addressed during the analysis to only the sensor measurements up to,and its feature vector dimension of more general data processing methods can not meet the requirements.Therefore,presents a wavelet packet analysis combined with artificial neural network analysis and processing of fault diagnosis method to realize the failure of the missile flight characteristics of pre-judgment and prediction.
出处
《电子测量技术》
2011年第4期100-102,共3页
Electronic Measurement Technology
关键词
小波
神经网络
导弹
故障诊断
wavelet
neural network
missile
fault diagnosis