摘要
不同于传统的模拟电路故障诊断思路,提出了一种针对复杂电子设备某型雷达接收机的故障方法;首先,对设备各性能参数监测数据进行标准化处理,将各特征变量的属性值转换为[0,1]区间内的数据序列,该序列可视为某一信号采样序列,对其进行时域分析以实现复杂电子设备的状态特征提取;其次,建立其HMM模型,将提取出的特征向量输入到各个故障状态HMM并进行训练,从而设计出复杂电子设备运行状态分类器;最后以某型雷达接收机电子设备为例进行故障诊断,各故障状态的平均识别精度为93.33%,说明了该方法的可行性和有效性。
Unlike the traditional fault diagnosis of analog circuit, a novel health evaluation method is proposed for complexity electronic e quipment radar receiver. Firstly, monitoring data of performance parameters is processed and eigenvectors＇ attributes are transformed into data sequence with [0, 1] interval. The achieved sequence can be viewed as a signal＇s samples. Performance feature extraction for complexi ty electronic equipment is carried out by analyzing the achieved samples in time domain. Then, HMM model of complexity electronic equip ment is established. The resulted eigenvector is inputted to HMM for training, and running states classification model of complexity electron ic equipment is constructed. And finally an example of the radar receiver is given. The average diagnosis rate of all faults is 93. 33%. The result has showed the feasibility and availability of the proposed method.
出处
《计算机测量与控制》
北大核心
2013年第11期3011-3013,共3页
Computer Measurement &Control
关键词
复杂电子设备
故障诊断
特征提取
隐马尔可夫模型
complexity electronic equipment
fault diagnosis
feature extraction
hidden markov models
