Industrial ultrasonic tomography (UT) possesses unique advantages in visualizing multiphase medium and has received broad attention in the past decades. Facing small bubble distribution imaging, the current independent-transducer-based UT faces bottleneck problems of limited projection number and large projection diameter. To image the cross-sectional bubble distribution, an original concept of ultrasonic phase array tomography (UPAT) is proposed by introducing the ultrasound phase array (UPA) transducers and novel designed sensing strategy into UT. The UPAT emits a wide and directional acoustic beam toward the receivers, on which each element forms a narrow projection to detect small bubbles on its path. Hence, the UPAT forms a novel sensing strategy with shifting and multiangled acoustic beam scanning of the whole cross section, as well as the precise measurement on a high amount of narrow projections. To mathematically model the UPAT, the acoustic field distribution and beam steering algorithm are designed for beam scanning with shifted apertures, while the forward/inverse solutions are proposed for tomographic imaging. In numerical simulations, the eight-transducer UPAT is with much lower imaging error than the state-of-art 32-transducer transmissive/reflective UT with relative error reduced by 65.66% on average. In lab-scale experiments, the reconstructed bubble size, shape, and distribution are with good accuracy, boundary preservation, and image purity.

中文翻译：

---

超声相控阵断层扫描技术用于双相介质分布成像的合成孔径光束扫描

工业超声断层扫描（UT）在可视化多相介质方面具有独特的优势，并且在过去的几十年中受到了广泛的关注。面对小气泡分布成像，当前基于独立换能器的UT面临投影数量有限和投影直径较大的瓶颈问题。为了对横截面气泡分布进行成像，通过将超声相阵列（UPA）换能器和新颖设计的传感策略引入UT，提出了超声相阵列层析成像（UPAT）的原始概念。UPAT向接收器发射一个宽广的定向声束，在每个接收器上，每个元素形成一个狭窄的投影，以检测其路径上的小气泡。因此，UPAT通过对整个横截面进行位移和多角度声束扫描，形成了一种新颖的传感策略，以及对大量狭窄投影的精确测量。为了对UPAT进行数学建模，设计了声场分布和波束控制算法，用于位移孔径的波束扫描，同时提出了层析成像的前向/后向解决方案。在数值模拟中，八传感器UPAT的成像误差比现有技术的32传感器透射/反射UT低得多，相对误差平均降低了65.66％。在实验室规模的实验中，重构后的气泡大小，形状和分布具有良好的精度，边界保留和图像纯度。同时提出了用于层析成像的正向/反向解决方案。在数值模拟中，八传感器UPAT的成像误差比现有技术的32传感器透射/反射UT低得多，相对误差平均降低了65.66％。在实验室规模的实验中，重构后的气泡大小，形状和分布具有良好的精度，边界保留和图像纯度。同时提出了用于层析成像的正/反解决方案。在数值模拟中，八传感器UPAT的成像误差比现有技术的32传感器透射/反射UT低得多，相对误差平均降低了65.66％。在实验室规模的实验中，重构后的气泡大小，形状和分布具有良好的精度，边界保留和图像纯度。