Methodology to investigate instantaneous and local transmembrane pressure within Rotating and Vibrating Filtration (RVF) module

Abstract Dynamic filtration exhibits high performances by generating wall shear stress (tangential to membrane) and pressure stress (normal to membrane) by the mechanical movements, such as rotating, oscillating or vibrating systems. Rotating and Vibrating Filtration (RVF) module includes rotating flat blades impellers which generate high and fluctuating shear stress and pressure at the membrane surface. To understand performances in turbulent regime and to optimise the operating conditions, global parameters (power consumption, pressure drop) and driving forces (mean, instantaneous and local pressure at the membrane surface) were characterised. For global approach, friction and mixing power in the RVF module were described by semi-empirical correlations. Euler number correlations were integrated based on feeding and mixing conditions. The balance between nominal power and thermal dissipation was reported. On the other hand, the mechanical power calculated with the empirical correlation of local shear stress was underestimated. For semi-local and local approaches, the local pressure at the membrane surface was measured with a specially designed and instrumented porous substrate. Mean radial pressure and core velocity coefficients were quantified versus flowrate and mixing rate. The core velocity coefficient decreases with mixing rate and radius up to a plateau value close to 0.6. For fluctuating component, pressure oscillation and its amplitude were treated by statistical analysis, probability distribution function and Fast Fourier transform. These methods show similar results with maximum fluctuating intensity between 15 and 30 Hz, which increase with radius. The maximum value can be obtained at the outer edge of the impeller with a relative standard deviation of over 25%. It indicates that the influence of pressure fluctuations should be carefully considered to enhance filtration performances. Pressure fluctuation distributions were accurately modelled by the convolution of periodic (sinusoid wave) and random (normal) functions. The area of intensive fluctuation was identified, in which periodic component accounts for 60% up to 97% of total energy input.