CFD simulation is a very useful tool when it comes to studying the resuspension and transfer of particles deposited in a cleanroom.
Different numerical techniques can be implemented depending on the conditions and /en/controlled-environment/requirements.
Cleanroom ventilation requirements can be better defined when it is possible to calculate the balance between contamination sources and the supply of clean air.
For this purpose, CFD simulation is a valuable tool, but it requires the most accurate possible information about the sources that will actually be present (people, equipment, materials).
The resuspension of particles deposited on surfaces is an important parameter, since these surfaces can become potential sources depending on their nature. The operating life of the cleanroom (localized contamination, reduced operating mode) and the trajectory of particles from primary sources (sedimentation, impaction) must also be considered.
A particle deposited on a surface can be released when sufficient energy is applied to overcome the adhesion forces holding it in place:
- Electrostatic forces (when opposite charges are present);
- Van der Waals forces (interaction between the molecules of the particle and those of the surface);
- Capillary forces (when a moisture film is present on the surface of the materials).
The influence of these forces varies depending on the particle characteristics (size, density, shape, electrical charge) and environmental conditions.
If a large quantity of particles is present on the surface, additional layers may form. In this case, interparticle cohesive forces must also be considered, together with a weakening of the surface forces as the distance from the surface increases.
In a cleanroom, particles can often be considered to be isolated from one another. If this is not the case, the cohesive forces binding the particles together must be taken into account, as they may form groups with more or less complex geometries.
Likewise, particle shape and the roughness of the deposition surfaces are important characteristics when assessing detachment. It is difficult, however, to incorporate this level of complexity into a CFD model at a scale larger than one meter.
Particles are treated as a second fluid, without distinguishing the trajectory of an individual particle. This is therefore a statistical approach that makes it possible to simulate particle concentration while taking into account particle size and density.
Indeed, particles subject to gravity and inertia may either rebound from or remain attached to surfaces.
Depending on the requirements, external forces can be added through specific functions, either integrated into the model or developed specifically (electrostatic forces, surface adhesion, elasticity during impacts).
In this Eulerian approach, the resuspension process requires the implementation of a resuspension-rate model, followed by the treatment of particle dispersion through interaction with air movements.
This method has, for example, been applied to study particle resuspension caused by a person walking through a cleanroom.
This approach is suitable for studies requiring particle quantification:
- Transport balance from a source, dispersion and contamination;
- Particle resuspension using dedicated models;
- Operation of dynamic filtration processes and sensors;
- Positioning of sensors for particle monitoring;
- Efficiency of localized air-cleaning processes.
The Lagrangian approach, on the other hand, enables individual particle tracking, at the expense of increased computational time. This results in limitations on the number and size of particles that can be processed.
Indeed, calculating particle interactions such as collisions requires fine spatial and temporal discretization, particularly when dealing with small particles.
Among Lagrangian approaches, the DEM (Discrete Element Method) is used to model systems of interacting particles or discrete elements (granulates, particles, molecules).
Particle motion is determined by calculating the forces acting between each pair of particles. These particle pairs generate elastic forces, modeled using linear springs, as well as damping forces that dissipate energy between the particles.
The mesh associated with this method is less demanding than that required for DNS (Direct Numerical Simulation), while still allowing individual particles to be modeled.
Another approach is Direct Numerical Simulation (DNS), which is reserved for microscopic-scale studies and requires an extremely fine mesh.
Therefore, the following questions should be considered when selecting one of these methods:
- Are interparticle collisions significant for the application?
- Is the domain large compared with the particle size?
- Is the event duration long compared with the computational time step?
This approach is therefore suitable for more complex studies in which interactions between particles must be taken into account.
- Behavior of a powder bed or particle agglomerate (accidental release, flow, conveyors, drying, cohesion);
- Mixing of particles in a liquid (suspension, dissolution);
- Coarse filtration.
Multiphysics calculations can also be performed, including interactions with moving objects and flexible materials such as fabrics.
It is very difficult to model particles other than as simple spheres.
A realistic representation of complex particle shapes introduces numerous difficulties related to geometry creation, their influence on fluid behavior (aerodynamics in air), and collision calculations.
These constraints mean that spherical particles are used in most cases.
The simulation of non-spherical particles with regular or random shapes and different shape factors is mainly reserved for aerosol research (high-density particle clouds with agglomeration), filtration studies (particle–fiber interactions), and human exposure studies (penetration into the respiratory tract).
It should also be noted that porous particles can be modeled by modifying the material density in order to reproduce the same aerodynamic behavior as an actual porous particle.
Thus, by knowing the characteristics of the particles and the available simulation tools, it is possible to model situations in which information about particle dispersion and transport is required.