A mesoscale model for the relationship between efficiency and internal liquid distribution of droplet mist filters

  • Author:

    H.E. Kolb, A.K. Watzek, V. Zaghini Francesconi, J. Meyer, A. Dittler, G. Kasper

  • Source:

    Journal of Aerosol Science, 2018, 123, 219-230, https://doi.org/10.1016/j.jaerosci.2018.05.013

  • Building on work by Kampa et al. (2014) about the liquid distribution in multi-layer oil mist filters, a mesoscale empirical model is proposed for their efficiency. It decomposes the total penetration into a product of independent ‘building blocks’ that are aligned with the mesoscale liquid distribution in steady state, essentially in the form of an oil film and a channel region. The film (typically one very thin region per sandwich) is associated with a film penetration Pfilm and the channel region (consisting typically of multiple layers) with a channel penetration Pchannel. The first layer of wettable media also represents a separate contribution P1st layer.

    A method is described to derive these unit penetrations from a set of measurements. Experiments are then carried out with filters composed of conventional glass microfiber media, both wettable and non-wettable to compressor oil, in which the fractional penetration of the entire sandwich was measured by SMPS between about 30 and 700 nm. It is shown that the film penetration is substantially reduced for particles above about 100 nm in comparison to that of the dry media due to enhanced inertial deposition, while the channel penetration is generally higher across the board, probably because of a loss of active fiber in the partially saturated media. The last part of the paper is dedicated to practical consequences of this efficiency model.

    It is shown that the fractional efficiency for an arbitrary number of (identical) layers can be predicted quite accurately from the unit penetrations. Additional measurements of the total downstream concentration as a function of time are used to discuss and explain changes from start-up until steady state.