Skip to Main content Skip to Navigation

Modélisation d'une population d'aérosols multi-sources et recherche des contributions de chaque source à l'échelle urbaine avec le modèle de dispersion CHIMERE

Abstract : The objective of this thesis is the development and validation of a numerical size resolved and externally mixed model of the particle dynamics. In order to trace several chemical compositions for each size class, a new approach is presented in which the particle chemical composition is itself discretized according to the mass fraction of one or several of its components (e.g. soot, sulfate). This approach aims to improve the simulation of the particle population evolution at local scale and to emphasize chemical compositions which are specific to some sources. In atmosphere, particles interacts essentially between themselves and gaseous pollutants through coagulation and condensation/evaporation. The first part of this thesis is dedicated to the model development for the coagulation process, which happens to be the most complex to model with our external mixing approach. First, coagulation equations in external mixing were set up and discretized with an arbitrary number of size and chemical composition classes. Several numerical simulations were then performed with this model according to the same case 6 study, using two, three and four chemical components. We checked each time that the simulation results in external mixing agreed well with those of the case study internal mixing. The results of these simulations are useful to understand how coagulation mix particles and produces, from monocomposed ones, bicomposed and tricomposed particles. Given the growing complexity of such a model, the numerical implementation has been carried out with carefullness and algorithms have been optimized. The extension of this approach to condensation/evaporation is the next development step of this model, the theoretical basis are adressed in appendix. Size resolved particle measurements (SMPS) do exist nowadays, but truely suitable data to validate the external mixing model still lack, that is to say measurements which would quantitatively distinguish several chemical compositions per size class. That is why, in the second part of this thesis, we considered the protocol of a chamber experiment allowing to highlight the mixing by coagulation of two particle populations with distinct compositions and to bring validating data for the model developed. Two series of experiments were conducted, the first one with the CESAM large volume chamber and the second, with the small reactor of INERIS. The first serie underlined the homocoagulation of each polydispersed distribution taken separately and to a lower extent, the heterocoagulation of the two distributions of different kinds (NaBr et KBr) between themselves. The second serie showed the possibility to observe simultaneously two monodispersed distributions of particles with different compositions (CaSO4 et KBr), which was required in this case. Finally, measurement results happened to be insufficiant to produce validating data for the model, because of the great deviation of polydispersed distributions in the large volume chamber and because of the dominating wall losses in the small reactor. However, some microscope electronic analysis showed evidences of particles produced from coagulation between both kind of particles. Further to these experiments, we come back to the planed protocol and propose some improvements
Document type :
Complete list of metadata

Cited literature [65 references]  Display  Hide  Download
Contributor : ABES STAR :  Contact
Submitted on : Monday, April 1, 2013 - 11:22:10 AM
Last modification on : Tuesday, January 28, 2020 - 3:25:10 AM
Long-term archiving on: : Tuesday, July 2, 2013 - 2:33:10 AM


Version validated by the jury (STAR)


  • HAL Id : pastel-00806520, version 1



Hilel Dergaoui. Modélisation d'une population d'aérosols multi-sources et recherche des contributions de chaque source à l'échelle urbaine avec le modèle de dispersion CHIMERE. Autre [q-bio.OT]. Université Paris-Est, 2012. Français. ⟨NNT : 2012PEST1123⟩. ⟨pastel-00806520⟩



Record views


Files downloads