A probabilistic framework for single-sensor acoustic emission source localization in thin metallic plates
A probabilistic framework for single-sensor acoustic emission source localization in thin metallic plates
复制标题
DOI：
10.1088/1361-665x/aa78de
复制
发表时间：
2017-09-01
影响因子：
4.1
通讯作者：
Salamone, Salvatore
中科院分区：
文献类型：
--
作者：
Ebrahimkhanlou, Arvin;Salamone, Salvatore
跟踪边缘反射声发射（AE）波可以定位其源。具体来说，在有界各向同性板结构中，只能使用一个传感器来执行这些源定位。本文的主要目标是开发一个三步概率框架来量化与这种单传感器定位相关的不确定性。根据这个框架，概率的方法是第一次使用的AE源和传感器之间的直接距离估计。然后，一个分析模型被用来重建的基础上的源到传感器的距离估计和他们的第一个到达的边缘反射AE信号的包络。最后，概率重建的包络和记录的AE信号之间的相关性被用来估计AE源的位置的置信度轮廓。为了验证所提出的框架，Hsu-Nielsen铅笔芯断裂（PLB）测试上进行的表面以及铝板的边缘。定位结果表明，估计的置信度轮廓包围的实际源位置。此外，该框架的性能进行了测试，在噪声环境中模拟两个虚拟换能器和任意波发生器。结果表明，在低噪声环境中，置信度轮廓的形状和大小取决于源及其位置。然而，在高噪声环境中，置信度轮廓的大小随着噪声基底单调增加。这样的概率结果表明，建议的概率框架，从而可以提供更全面的信息，AE源的位置。
Tracking edge-reflected acoustic emission (AE) waves can allow the localization of their sources. Specifically, in bounded isotropic plate structures, only one sensor may be used to perform these source localizations. The primary goal of this paper is to develop a three-step probabilistic framework to quantify the uncertainties associated with such single-sensor localizations. According to this framework, a probabilistic approach is first used to estimate the direct distances between AE sources and the sensor. Then, an analytical model is used to reconstruct the envelope of edge-reflected AE signals based on the source-to-sensor distance estimations and their first arrivals. Finally, the correlation between the probabilistically reconstructed envelopes and recorded AE signals are used to estimate confidence contours for the location of AE sources. To validate the proposed framework, Hsu-Nielsen pencil lead break (PLB) tests were performed on the surface as well as the edges of an aluminum plate. The localization results show that the estimated confidence contours surround the actual source locations. In addition, the performance of the framework was tested in a noisy environment simulated by two dummy transducers and an arbitrary wave generator. The results show that in low-noise environments, the shape and size of the confidence contours depend on the sources and their locations. However, at highly noisy environments, the size of the confidence contours monotonically increases with the noise floor. Such probabilistic results suggest that the proposed probabilistic framework could thus provide more comprehensive information regarding the location of AE sources.