改进的子空间匹配追踪导波信号识别方法

针对子空间匹配追踪计算复杂的缺点,提出一种改进的子空间匹配追踪（MSMP）方法.采用线调频小波函数作为匹配原子,选用微分进化算法（DEA）实现改进的子空间匹配追踪方法.利用29kHz t（0,1）导波对含缺陷的铝管进行检测实验,采用MSMP对检测信号进行匹配分解与重构.将匹配结果与基于微分进化算法的匹配追踪（MP）及基于t算子的进化规划算法（tEP）的正交匹配追踪（OMP）所得结果进行比较,并比较了基于DEA的MSMP和MP,基于tEP的OMP匹配所得参数.结果发现：重构所得信号质量明显提高,基于DEA的MSMP和MP方法匹配所得参数均能比较准确地反映缺陷位置以及激励信号的中心频率,基于DEA的MSMP匹配所得的参数更加准确且耗时更短,改进的方法可有效识别管道导波无损检测信号并定位缺陷.

Guided wave signal recognition by modified subspace matching pursuit method

The modified subspace matching pursuit(MSMP) was proposed to simplify the complex computation of the subspace matching pursuit.The differential evolution algorithm(DEA) was applied to the MSMP by choosing chirplet function as match atoms.An aluminum pipe with holes was tested by 29 kHz t(0,1) guided wave and the measured signal was decomposed and reconstructed by MSMP.The quality of the processed signal was improved obviously,and the defect echo can be identified easily.The matched result was compared to the matched result from matching pursuit(MP) with DEA and orthogonal matching pursuit(OMP) with evolutionary programming using mutations based on the t probability distribution(tEP).The matched parameters get by MSMP,MP with DEA,and OMP with tEP were compared and analyzed.It is found that the quality of the matched waveform from MSMP and MP are better than that from OMP.The defect locations and the center frequency of the excitation signal are more exactly get from MSMP than from MP.The computation time by MSMP is shorter than the one by MP and OMP.Therefore,MSMP is a useful signal recognition and defect location approach for pipes guided wave non-destructive testing.