摘要
采用自行研制的模块组合式漏磁传感器实现了微细裂纹的定量化研究。漏磁传感器核心器件为AMR磁敏元件,基于漏磁信号切向分量,提取与人工预制裂纹尺寸相关的一组特征量;利用三次样条法拟合实验数据,构建虚拟实验样本,采用GA-BP神经网络对样本库进行训练,并定量识别微细裂纹。研究表明,文中所提技术思路能有效实现微细裂纹的定量化评估。
To carry quantitative research on micro cracks,magnetic sensor of module combination was developed. Core component of the sensor is AMR,then a set of features related to artificial crack size can be extracted based on the tangential component of magnetic signals. Fitting the experimental data through the three spline method,virtual experimental samples were constructed,samples were trained by using GA- BP neural network,and micro cracks were identified quantitatively. Final results show that the proposed approach can achieve quantitative assessment of the micro crack effectively.
出处
《仪表技术与传感器》
CSCD
北大核心
2016年第3期1-3,共3页
Instrument Technique and Sensor
基金
国家自然科学基金资助项目(51275048)
关键词
漏磁检测
模块组合式
阵列传感器
微细裂纹
定量识别
magnetic flux leakage detection
module combination
sensor array
micro crack
quantitative identification
