基于TLGWO-SVM的光纤Bragg光栅传感器故障诊断
Fault Diagnosis of Fiber Bragg Grating Sensor Based on TLGWO-SVM
李锦 1陈静 2肖纯 2刘晖 3刘岩4
作者信息
- 1. 武汉理工大学自动化学院,武汉 430070;国家能源氢能及氨氢融合新能源技术重点实验室,佛山仙湖实验室,佛山 528200
- 2. 国家能源氢能及氨氢融合新能源技术重点实验室,佛山仙湖实验室,佛山 528200;武汉理工大学自动化学院,武汉 430070
- 3. 武汉理工大学土木工程与建筑学院,武汉 430070;武汉理工大学三亚科教创新园,三亚 572025
- 4. 武汉理工大学自动化学院,武汉 430070
- 折叠
摘要
为了快速准确地识别光纤Bragg光栅(FBG)传感器故障,提出一种基于教与学算法改进灰狼算法优化支持向量机的故障诊断方法.该方法首先利用经验模态分解将信号分解为若干个本征模态函数,并计算其对应的多尺度近似熵,从而得到故障特征向量;其次,构建支持向量机故障诊断模型,利用教与学改进的灰狼算法对支持向量机的参数进行迭代寻优;最后,将FBG传感器的原始测量信号输入故障诊断模型,可以有效区分传感器正常状态和啁啾故障状态,并与广泛使用的算法进行了对比分析.结果表明,该方法能够更精准、更迅速地对FBG传感器的故障进行识别,具有一定的工程参考价值.
Abstract
In order to quickly and accurately identify the faults of fiber Bragg grating(FBG)sensor,this paper proposes a fault diagnosis method based on the teaching and learning algorithm to improve the grey wolf algorithm to optimize the support vector machine.Firstly,the method uses empirical mode decomposition to decompose the signal into a number of intrinsic mode functions and calculate its corresponding multi-scale approximate entropy,so as to obtain the fault eigenvec-tors.Secondly,the support vector machine fault diagnosis model is constructed,and the parameters of the support vector machine are iteratively optimized by grey wolf algorithm improved by teaching and learning algorithm.Finally,by input-ting the original measurement signal of FBG sensor into the fault diagnosis model,the normal state and chirped fault state of the sensor can be distinguished effectively,and the comparison with the widely used algorithm is analyzed.The results show that this method can identify the fault of FBG sensor more accurately and quickly,and has certain engineering refer-ence value.
关键词
光纤Bragg光栅传感器/故障诊断/经验模态分解/近似熵/灰狼算法/支持向量机Key words
fiber Bragg grating sensor/fault diagnosis/empirical mode decomposition/approximate entropy/grey wolf algorithm/support vector machine引用本文复制引用
基金项目
海南省自然科学基金(522CXTD517)
海南省自然科学基金(522RC879)
先进能源科学与技术广东省实验室佛山分中心(佛山仙湖实验室)开放基金(XHD2020-003)
出版年
2024
NETL
NSTL
万方数据