摘要
状态识别是保证光电检测系统正常工作的技术基础,当前光电检测系统状态识别方法对噪声信息敏感,无法获得高精度的光电检测系统状态识别结果,为了提高光电检测系统状态识别正确率,提出了回声状态网络的光电检测系统状态识别方法。首先采集光电检测系统状态信号,采用小波分析去除光电检测系统振动信号的噪声,抑制噪的干扰,并提取光电检测系统状态振动号故障特征,然后将特征向量作为输入,光电检测系统状态作为输出,通过回声状态网络行训练建立光电检测系统状态识别的分类器,最后与其它检测方法进行了对比测试。结果表明,本方法的光电检测系统状态识别正确率均值超过95%,而且减少了光电检测系统状态识别时间,实际应用价值更高。
state recognition is the technical basis to ensure the regular operation of the photoelectric detection system. The current state recognition methods of the photoelectric detection system are sensitive to noise information and cannot obtain high-precision state recognition results of the photoelectric detection system. To improve the state recognition accuracy of the photoelectric detection system,a state recognition method of a photoelectric detection system based on an echo state network is proposed. Firstly,the state signal of the photoelectric detection system is collected,and wavelet analysis is used to remove the noise of the vibration signal of the photoelectric detection system to suppress noise interference and extract fault characteristics of the vibration signal of the photoelectric detection system. Then,the feature vector is used as the input vector,and the state of the photoelectric detection system is taken as output.Through the training of echo state network,the state recognition of photoelectric detection system is established. At last,it is compared with other detection methods. The results show that the state recognition accuracy of the photoelectric detection system is more than 95%,and the time of the state recognition of the photoelectric detection system is reduced,and the practical application value is higher.
作者
鲁明珠
刚建华
LU Mingzhu;GANG Jianhua(Department of Mechanical and Electrical Engineering,Cangzhou Normal University,Cangzhou Hebei 061001,China)
出处
《激光杂志》
CAS
北大核心
2021年第5期143-146,共4页
Laser Journal
基金
河北省科技厅项目(No.18211844)
河北省“三三三人才工程”资助项目(No.A202001101)
沧州师范学院检测技术与自动化装置科研创新团队成果(No.cxtdl1905)。
关键词
光电检测系统
工作状态
回声状态网络
噪声抑制
信号采集
期望输出
photoelectric detection system
working state
echo state network
noise suppression
signal acquisition
expected output
