Modelling of particulate contamination redistribution inside purged cavities

Abstract

Sophisticated instruments on-board most of satellite payloads, especially scientific ones, are sensitive to particulate contamination since their performance can be affected in terms of straylight or surface obscuration. The surface cleanliness levels have to be minimized on ground. However, even if the deposition rate is controlled in cleanroom, the hardware is continuously exposed to particles and fibers, moving in the airflow and settling according to their various sizes and matter compositions. Thus, depending on the initial surface cleanliness level, the redistribution of particulate contamination can be a concern during integration and tests activities until launch, when venting and/or vibrations occur which results in more or less random fallout of particles. Previous studies identified different steps for the transfer mechanism such as detachment, transport and interaction with the surface (with or without bounce). Therefore, it is paramount to understand the physical phenomena of particle resuspension that happens either inside purged cavities of spacecraft or under fairing in order to be able to model them as for molecular contamination. After performing a state of the art of the existing detachment models, a Computational Fluid Dynamics software, scSTREAM, has been chosen among the most relevant ones for a feasibility study to simulate particles redistribution. This tool was particularly interesting due to the possibility to add specific functions for computation. For the purpose, the Rock’n’Roll (R’n’R) model has been implemented to consider particle resuspension. In parallel, some simple test configurations have been defined and a mock-up with several adjacent cavities has been used to compare theory and experiment when a purging under dry nitrogen was applied in the vicinity of a pre-contaminated surface. The relocation of FluoDust particles, as the source of contamination, deposited on a specific substrate has been evaluated thanks to many collector plates disseminated in one or two cavities and the levels quantified. The impact of different test conditions (flow rate, material of the substrate, presence of leakage, initial contamination level…) has been investigated. This paper presents the comparative results that confirmed the redistribution of the particulate contamination with similar trajectories but also some discrepancies due to the limitations of the R’n’R model and depending on experimental parameters such as initial distribution of particles, adherence force to the substrate…