Cite this article: Wu, J., Chen, L., Zhou, J., Wu, X., Gao, X., Gréhan, G. and Cen, K. (2017). Particle Size Distribution of Soot from a Laminar/Diffusion Flame.
Aerosol Air Qual. Res.
17: 2095-2109. https://doi.org/10.4209/aaqr.2017.06.0216
An explicit solution was proposed for acquiring soot particle size distribution.
Critical time intervals for the inversions with the solution were found.
Influences of experimental conditions on the inversions were evaluated.
Experiments on a laminar diffusion flame were performed to validate the solution.
A practical method is proposed to determine the inflame soot particle size distribution via an explicit solution derived from time-resolved laser-induced incandescence (TiRe-LII). Three appropriate time intervals are selected from the TiRe-LII decay signal. The equivalent mean sizes as well as relative ratios of number densities for three classes of monodisperse particles are determined with the mono-exponential fit to each interval, which allows an explicit solution for the particle size distribution. Simulations show that inversed log-normal distributions from the explicit solution are coincident with the input parameters in terms of trend, and there exist critical time intervals where inversed results are most insensitive to the variation of interval length. The error of inversion is critically dependent on the geometric standard deviation but weakly dependent on the count median particle diameter. Influences of experimental conditions on the inversed error are additionally evaluated. The results show that flame temperature has a significant impact on the error of inversion. Thus, a database of the error as a function of flame temperature at a fixed aggregate size is established. The error database allowed the inversed results from experimental TiRe-LII signals to be readily corrected at various flame locations by interpolation. The corrected inversed log-normal distributions were consistent in trend with those determined from the established non-linear regression method, and the modeling LII signals reproduced agree with the experimental data. The small deviation of the results potentially stemmed from the statistical noise contained in recorded LII signals.