基于ACO-LS-SVM的漏磁信号二维轮廓重构  

2-D PIPELINE DEFECT RECONSTRUCTION FROM MAGNETIC FLUX LEAKAGE SIGNALSBASED ON ACO-LS-SVM

作　　者：纪凤珠[1] 孙世宇[1] 苑希超[1] 王瑾[1] 左宪章[1] 

机构地区：[1]军械工程学院电气工程系,河北石家庄050003

出　　处：《西南石油大学学报（自然科学版）》2011年第6期165-169,215-216,共5页Journal of Southwest Petroleum University(Science & Technology Edition)

基　　金：河北省自然科学基金项目"漏磁缺陷三维成像关键问题研究"(E2008001258)

摘　　要：漏磁检测技术被广泛应用于铁磁材料的无损评估中,由漏磁信号描述缺陷的几何特征一直是漏磁检测的难点。为此提出应用LS-SVM对缺陷轮廓进行重构的新方法,利用蚁群算法优化LS-SVM及核函数的参数,并采用剪枝算法改善LS-SVM的稀疏性。支持向量机输入采用漏磁信号Bx、By分量的特征融合信号,输出是缺陷轮廓数据,建立了由缺陷的漏磁信号到缺陷二维轮廓的映射关系。实现了人工裂纹缺陷二维轮廓的重构,并与BP神经网络、GA-LS-SVM和PSO-LS-SVM等3种方法重构效果进行了比较。结果表明:该方法速度快、精度高。Nondestructive evaluation of ferromagnetic material is widely used in the magnetic flux leakage(MFL)techniques,and it is a nodus to describe the characters of defects from MFL inspection signals. A novel method forthe reconstitution of 2-D profiles is presented based on least squares support vector machines(LS-SVM)technique,Ant Colony Optimization(ACO)is adopted to optimize the model parameter of LS-SVM,and the pruning arithmeticis applied to improve the sparseness. The input data sets of SVM is feature fusion signals of Bx and By whichis the component of MFL signals,and output data sets is 2-D profiles parameter,the mapping relationship fromMFL signals to 2-D profiles of defects is established. The reconstitution of 2-D profiles of artificial crack defectsin the magnetic flux leakage testing is implemented by this algorithm. After comparing with the reconstitution performanceof BP network,GA-LS-SVM and PSO-LS-SVM,we reach the conclusion that LS-SVM possesses quickspeed、high accuracy and very good generalization ability,and it is a good way for the quantization of the MFLtesting.

关 键 词：漏磁检测 LS-SVM ACO 特征融合 轮廓重构 

分 类 号：TE838[石油与天然气工程—油气储运工程]