摘要
本文主要利用超声相控阵技术进行了裂纹方向识别研究。首先,对线性超声相控阵探头采集的全矩阵数据进行了全聚焦成像,确定缺陷的位置。然后将线性阵列划分为若干子阵列,研究了缺陷位置处子阵列的散射系数分布,从中提取出缺陷的方向信息。在此基础上,研究了子阵列参数选择(子阵列包含晶片数及相邻子阵列间隔晶片数)及探头位置对裂纹方向识别的影响。通过对散射系数分布图中提取的3个特征指标,角度测量误差、角度分辨率及相对脊带宽度,进行主成分分析,评价了子阵列参数设置(如子阵列包含晶片数及相邻子阵列间隔晶片数)对裂纹方向识别的影响,优化出最佳的检测位置及子阵列参数设置。仿真和检测实验结果表明,当相邻子阵列间隔晶片数为1,包含晶片数为11个时,可以利用超声散射系数分布进行裂纹方向的准确测量,测量误差小于2%。
Ultrasonic phased array technique was applied to characterize the orientation angle of the crack-like defects in simulation and experiment. The total focusing image was generated first using the full matrix capture data obtained from the linear ultrasound array to determine the location of the defect. Then, scattering coefficient distribution was deduced to extract the information on the defect orientation by dividing the linear array into a certain number of subarrays. Besides, the influence of subarrays parameters setting and the inspection location of phased array transducer was studied. Principal component analysis （PCA） was used to evaluate the influence of subarrays parameters setting on the crack-like defects orientation characterization, such as element number in one subarray and element number between adjacent subarrays, and the best parameters were obtained by evaluation indicators extracted from scattering coefficient distribution, namely measuring error, angular resolution and relative width of the ridge. Results in simulation and experiment show that the crack orientation angle can be measured accurately in the conditions of optimal subarrays parameters setting, 11 elements in every subarray and 1 element between adjacent subarrays, and the error is less than 2% by the approach of crack-like defects orientation characterization based on scattering coefficient distribution.
出处
《声学学报》
EI
CSCD
北大核心
2016年第1期13-24,共12页
Acta Acustica
基金
国家自然科学基金(51075012
11272017)
北京市自然科学基金(1122005)
质检公益项目(201310156)资助
关键词
散射系数
方向识别
裂纹
超声
角度测量误差
线性阵列
缺陷位置
参数设置
Angular distribution
Characterization
Cracks
Defects
Principal component analysis
Ultrasonic applications
Ultrasonic transducers
