Pal Soma, Lin Fu-Sung, Hsieh Ching-Chuan, Liu Ya-Han, Lu Chen-Yuan, Du Shan-Wen, Huang Chih-Hsien
IEEE Trans Ultrason Ferroelectr Freq Control. 2021 Apr;68(4):1370-1379. doi: 10.1109/TUFFC.2020.3030541. Epub 2021 Mar 26.
Acoustic pyrometer is expected to be a useful noninvasive method for monitoring gas temperature distribution inside a steel-making furnace. On the superficial layer above the burden of a blast furnace, most of the high-temperature gas is concentrated near the center, and tracking the position of the hotspot is critical for productivity. However, most of the existing acoustic temperature distribution reconstruction algorithms are developed with relatively uniform temperature distribution environments. Besides, their capabilities of tracking the pinnacle of temperature distribution in the region of interest (ROI) are rarely discussed. In this research, a reconstruction method of acoustic temperature tomography dedicated for highly centralized gas temperature distribution is proposed and demonstrated. The key metrics include the reproducibility of 2-D temperature distribution, the sensitivity of hotspot shift, and the accuracy of point-to-point (P2P)/peak temperature. To optimize the result of each metric, previous approaches of acoustic temperature tomography have been first evaluated. Then, the investigation of effects from the shape and size of meshes is proceeded to improve the performance. After that, a novel method to address convergence issues while using the iterative method is introduced. Consequently, the reconstruction method proposed in this article could effectively visualize the temperature map while hotspot moves to different locations. It could also sense the occurrence of a hotspot (2.56% of ROI) traveled from center to 1% of ROI's diameter. Moreover, a competitive accuracy with 5.89% and 1.46% error at P2P root-mean-square (rms) and peak temperature is achieved, respectively. Finally, a practical acoustic 2-D pyrometer consisted of 12 ultrasonic transducers arranged in a circular pattern with a 1-m width of ROI successfully detected the shift of a hotspot when the displacement of a heater reaches 5 cm.
声学高温计有望成为一种用于监测炼钢炉内气体温度分布的有用的非侵入性方法。在高炉炉料上方的表层，大部分高温气体集中在中心附近，追踪热点位置对生产率至关重要。然而，现有的大多数声学温度分布重建算法是在温度分布相对均匀的环境下开发的。此外，很少讨论它们在感兴趣区域（ROI）追踪温度分布峰值的能力。在本研究中，提出并论证了一种专门用于高度集中气体温度分布的声学温度层析成像重建方法。关键指标包括二维温度分布的可重复性、热点偏移的灵敏度以及点对点（P2P）/峰值温度的准确性。为了优化每个指标的结果，首先评估了声学温度层析成像的先前方法。然后，研究网格形状和大小的影响以提高性能。之后，引入了一种在使用迭代方法时解决收敛问题的新方法。因此，本文提出的重建方法在热点移动到不同位置时能够有效地可视化温度图。它还能够检测到热点（ROI的2.56%）从中心移动到ROI直径的1%处的情况。此外，在P2P均方根（rms）和峰值温度方面分别实现了具有5.89%和1.46%误差的有竞争力的精度。最后，一个由12个超声换能器以圆形模式排列且ROI宽度为1米组成的实用声学二维高温计，当加热器位移达到5厘米时成功检测到了热点的移动。
