Critical analysis of uncertainties during particle filtration

Using the law of propagation of uncertainties we show how equipment- and measurement-related uncertainties contribute to the overall combined standard uncertainties (CSU) in filter permeability and in modelling the results for polystyrene latex microspheres filtration through a borosilicate glass filter at various injection velocities. Standard uncertainties in dynamic viscosity and volumetric flowrate of microspheres suspension have the greatest influence on the overall CSU in filter permeability which excellently agrees with results obtained from Monte Carlo simulations. Two model parameters "maximum critical retention concentration" and "minimum injection velocity" and their uncertainties were calculated by fitting two quadratic mathematical models to the experimental data using a weighted least squares approximation. Uncertainty in the internal cake porosity has the highest impact on modelling uncertainties in critical retention concentration. The model with the internal cake porosity reproduces experimental "critical retention concentration vs velocity"-data better than the second model which contains the total electrostatic force whose value and uncertainty have not been reliably calculated due to the lack of experimental dielectric data.