How to validate numerical results on primary particle diameter to experimental data from laminar sooting flames

Combustion generated soot particles have harmful effects on human health and our environment. An important aspect is to accurately determine the surface area of the particle population, which can be estimated from the particle size distribution (PSD) and morphology. Experimental investigations showed that large particles are aggregates constituted of several small primary particles [1]. Therefore, the determination of the primary particle size distribution (PPSD) is essential for the characterization of soot population. On the one side, sectional methods can be used to numerically predict the particle population of sooting flames. However, most of models assumes that large particles are spherical for all sections [2] or aggregates constituted of primary particles of identical size for all sections [3-5]. These strong assumptions can affect the results’ quality and the validity of the models themselves. On the other side, Time Resolved Laser Induced Incandescence (TiRe-LII) is a powerful, nonintrusive experimental method, which exploits the fact that the temporal decay of the LII signal is related to the primary particle diameter dp. Information on the PPSD can then be derived once the PPSD shape is presumed [6]. The general approach is to assume log-normal distribution, but Transmission Electron Microscopy measurements showed that this assumption may be not always valid [7]. In this context, the comparison of numerical results on the PPSD with experiments is extremely complex due to the strong assumptions underlying the numerical models and the fact that TiRE-LII technique does not measure directly the PPSD. In this work, we propose a new way to compare numerical to experimental data on PPSD. First, we improved our existing CFD code to obtain the mean size of the primary particles for each section, based on what proposed in [4]. Second, the TiRe-LII signal is reconstructed from the numerical PPSD and compared directly to the measured signal [8] to avoid any potential errors due to a presumed PPSD shape. This approach is applied to the investigation of an ethylene laminar-coflow diffusion flame [9], which is a target of the ISF workshop [10], and potential sources of errors are discussed.