Decoupling pyrolysis and combustion of organic powders to determine the laminar flame speed

Abstract

Determining the laminar flame speed of dusts is far from straightforward. A strong dependency on the experimental setup and the data treatment's high complexity makes it a true challenge. This work compares three complementary experimental setups to measure the laminar flame speed of organic dust (here, cellulose): a modified Hartmann tube, a 20 L sphere, and a micro-fluidized bed (MFB) burner. The first two consider the flame propagation phenomenon in its globality, which means that numerous steps are involved simultaneously (particle heating, pyrolysis, oxidation, radiative transfer, flame stretching), while the third one decouples pyrolysis and combustion, to focus mainly on the oxidation rate. An MFB was conceived to generate pyrolysis products and burn them in a laminar flame. Unstretched flame velocities determined with the first two setups were consistent and equal to 22.0 and 26.6 cm ∙ s⁻¹, respectively. Using Silvestrini's equation, values ranging between 14.0 and 33.4 cm ∙ s⁻¹ were obtained according to the dust concentration. With the MFB burner, the flame speed was much higher (135–155 cm ∙ s⁻¹), due to the higher temperature of the fresh mixture and the fact that only the oxidation of the pyrolysis gases is considered. A numerical simulation (Chemkin) confirmed these results since the range 135 to 231 cm ∙ s⁻¹ was obtained for equivalence ratios of 0.6 and 1.2, respectively. The discrepancy between the laminar flame speed determined in the sphere or in the tube and that obtained in the MFB highlights the significant influence of particle heating and pyrolysis during a dust explosion.