In nuclear facilities, during normal operations in controlled areas, workers could be exposed to
radioactive aerosols (1 μm < dp < 10 μm). One of the airborne contamination sources is particles that are initially seeded on the floor and could be removed by workers while they are walking. During the outage of EDF nuclear facilities, there is a resuspension of some radionuclides in aerosol form (1 μm < dp < 10 μm). Since the number of co-activity will increase in reactors buildings of EDF, it becomes important to understand particle resuspension due to the activity of the operators to reduce their radiation exposure. The purpose of this Ph.D thesis is to quantify the resuspension of particles due to the progress of operators on a contaminated soil. Thus, the approach is to combine an aerodynamic resuspension model with numerical calculations of flow under a shoe, and then to characterize experimentally some input parameters of the model (particle diameter, adhesion forces, shoes motion).
The resuspension model Rock'n'Roll proposed by Reeks and Hall (2001) was chosen because it describes physically the resuspension mechanism and because it is based on the moment of forces applied to a particle. This model requires two input parameters such as friction velocity and adhesion forces distribution applied on each particle.
Regarding the first argument, numerical simulations were carried on using the ANSYS CFX software applied to a safety shoe in motion (digitized by 3D CAO); the mapping of friction velocity shows values of about 1 m.s-1 for an angular average velocity of 200 °.s-1.
As regards the second parameter, AFM (Atomic Force Microscopy) measurements were carried out with alumina and cobalt oxide particles in contact with epoxy surfaces representative of those encountered in EDF power plants. AFM provides the distribution of adhesion forces and reveals a much lower value than what can be calculated theoretically using JKR model (Johnson et al. (1971)). Moreover, this technique, taking into account the surface roughness, shows that adhesion forces decrease while particle diameter increase. Finally, the analysis of AFM measurements gives a correlation linking the distribution of adhesion forces to the particle diameter, replacing the one given by Biasi et al. (2001) originally used in the Rock'n'Roll model and thereby adapt the model to particles and flooring studied in our case.
Coupling, performed in ANSYS CFX software, between the calculations of friction velocity and model of particle resuspension, gives theoretical resuspension rate during shoe motion. This coupling was initially validated by comparison to the experience for the simple case of a rotating plate in a controlled volume. Secondly, experiments at the scale of a ventilated room of 30 m3 were performed by walking on an epoxy coating initially seeded by calibrated particle size (1.1 μm and 3.3 μm). These experiments highlight the parameters influencing the suspension of particles, such as step frequency and particle size.