Improving Gas Distribution Tomography: Evaluation of Algorithms and Fan-Beam Measurement Geometries Indoors

Abstract

Gas distribution mapping (GDM) is widely used to evaluate the efficiency of environmental control systems and locate pollution sources. In this study, the optimized fan-beam measurement geometries and tomography algorithms were proposed for mapping the gas distributions indoors accurately. The airflow and gas distributions in two different types of rooms were predicted using the validated computational fluid dynamics (CFD) method. Then, the performances of five algorithms, including iterative and Tikhonov regularization methods, were quantitatively compared, and the measurement geometries (i.e., number of optical paths, grid resolution, and laser emitter amount) were investigated. The effect of measurement noise on reconstruction accuracy was also analyzed. The results have shown that the simultaneous algebraic reconstruction technique (SART) and the maximum likelihood expectation maximization (MLEM) algorithms were better than the other algorithms, and their averaged relative root mean square errors (RRMSEs) were at least 39% lower with the iteration residual of 0.0015. The optimum grid resolution and the number of optical paths were <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$10\times20$ </tex-math></inline-formula> grid cells and 200 paths, respectively. In addition, the sensitivity of the SART and MLEM algorithms to noise was almost the same, and the addition of the laser emitter could improve the reconstruction accuracy of most cases in this study.