PARAMETRIC EXPERIMENTS AND MODELING OF A CROSS-FLOW COALESCENCE FILTER

Experiments with a cross-flow precoat coalescence filter using a rotor have been performed in a systematic way, using factor design. Variables like rotor speed, density difference between oil and water, different oil qualities, pore size, drop size, effect of tensides, filter depth, and flow rate have been varied. A physical model has been made taking into account drop capture, drainage, and film rupture. This model is compared with an empirical mathematical model using polynomial approximation for three important experimental parameters. Interception is the most important factor for drop capture. The rotor speed and the density difference also have a significant impact on the filter efficiency. When a drop is captured the drainage rate seems to be high enough to reach the critical film thickness during the residence time for the drop in the filter. This is valid at least for smaller drops in all experiments. The film rupture probability is decreased by surfactants like nonylphenols with variable ethylene-oxide chain length, and especially if these have a hydrophil-lipophil-balance (HLB) value in the range of 12-15.