基于深度学习算法的大型飞机电缆故障识别  

Cable fault identification of large aircraft based on deep learning algorithm

作　　者：林伟 罗群[1] 陈龑斌 Lin Wei;Luo Qun;Chen Yanbin(Aviation Industry Xi'an Aircraft Industry(group)Co.,Ltd.,Shaanxi Xi'an,710089,China)

机构地区：[1]航空工业西安飞机工业(集团)有限责任公司,陕西西安710089

出　　处：《机械设计与制造工程》2022年第1期62-66,共5页Machine Design and Manufacturing Engineering

摘　　要：为了提高大型飞机电缆故障识别效果,提出了基于深度学习算法的大型飞机电缆故障识别方法。首先利用S变换方法获取大型飞机电缆信号的复时频矩阵,利用熵和奇异值分解理论提取复时频矩阵的特征向量,将特征向量作为深度学习网络的输入、大型飞机电缆故障作为输出;然后采用随机梯度下降法更新深度学习网络的权重参数、偏置参数,从而建立大型飞机电缆故障精准识别模型;最后对大型飞机电缆故障识别实例进行了分析,分析结果表明,在含有白噪声的情况下,深度学习算法的电缆故障识别精度仍高于99%,识别误差控制在有效范围内,具有较高的实际应用价值。In order to improve the effect of cable fault identification for large aircraft,a cable fault identification method based on deep learning algorithm is proposed.Firstly,the complex time-frequency matrix of large aircraft cable signal is obtained by S-transform method,and the eigenvector of complex time-frequency matrix is extracted by using entropy and singular value decomposition theory.The eigenvector is used as the input of deep learning network,and the cable fault of large aircraft is taken as the output.The weight parameters and bias parameters of the deep learning network are updated by using stochastic gradient descent method.Finally,an example of cable fault identification for large aircraft is carried out.The results show that in the case of white noise,the cable fault recognition accuracy of deep learning algorithm is still higher than 99%,and the identification error is controlled within the effective range,which has high practical application value.

关 键 词：深度学习算法 大型飞机 电缆故障识别 输入层 隐含层 复时频矩阵 特征向量 

分 类 号：TM247[一般工业技术—材料科学与工程]