In hazardous industrial activities such as in nuclear facilities, High Efficiency Particulate Air filters (HEPA filters) are essential to ensure the containment of airborne contamination. Most of the filters used in ventilation networks are pleated, in order to offer a larger surface of filtration. For industrial risks likely to lead to an important release of particles (e.g. fire), predicting the evolution of the pressure drop of pleated filters is very important, in order to anticipate any dysfunction, failure or breaking of these devices. Pressure drop variations are linked to airflow rate variations and to clogging process of the medium by airborne particles.
Thus, the airflow pattern in a pleat channel is essential for optimizing the filter design and enhancing its lifetime. Particles are transported by the airflow and deposited at the filter surface; hence, the geometry of the dust cake (shape and location) is partially determined knowing the velocity streamlines.
The present paper focuses on the characterization of airflows in a clean HEPA filter. The difficulty to perform fine measurement on a real scale filter led us to develop an experimental device, consisting in the reproduction of a single pleat, identical to a real pleat constituting industrial filters. The small dimension of the pleat makes the velocity measurement difficult to establish. That is why μ-PIV method has been adapted to measure the velocity field inside the filter for different filtration velocities at the first moments of the experiment, in order to avoid the impact of clogging by particles used to seed the flow. These particles are DEHS droplets 0.01 < St < 0.05. In the future, these well-characterized airflows will be the basis for CFD computation of particle transport and deposition inside the pleats. Ultimately, the aim is to develop or upgrade physical models predicting the pressure drop evolution of pleated filters, during clogging process in accidental situations.