Design optimization of hemodialyzer membrane modelled in chemical reaction module using statistical algorithms and ANOVA testing

Document Type

Article

Publication Title

South African Journal of Chemical Engineering

Abstract

High performing hemodialyzers membranes such as high flux membranes, high cut-off membranes and medium cut-off membranes are always at research interest due to their better efficiency than conventional membranes. These membranes provide greater toxin clearance, however retention of essential solutes in a preferable way are still under study. This paper aims at the design of high performing membrane to study the role of its parameters in solute removal and its capability of holding back important molecules. One of the most effective design of experiments (DOE) tool, namely Taguchi Algorithm was used for the formation of fractional factorial design of parameters. The simulation results were benchmarked with that of experimental data from literature and with manufacturers data sheets. Once the benchmarking was done, the error quantification and significance of each design were analysed using statistical method, Analysis of Variance (ANOVA) testing. The most relevant parameters that helped in better clearance in these membranes were thus identified and substantial conclusions were drawn which can be used in the future for designing optimal dialyzer designs. Results shows that clinically used dialyzer membranes such as RevaclearMax and FxCorDiax series on modelling using COMSOL Multiphysics with a blood flow rate of 400 ml/min and dialysate flow of 500 ml/min showed better urea clearance rate of above 300 indicating that the membranes thus designed was superior to the conventional high flux membranes that have a clearance rate of 297 of less for the exact same functional, geometrical, and parametric conditions. The model replication and thus validation of the design helped in understanding the influence of various parameters in toxin clearance. These parameters can be further investigated, and optimal models can be delivered with more of clinical examinations and trials.

First Page

417

Last Page

424

DOI

10.1016/j.sajce.2024.03.007

Publication Date

4-1-2024

This document is currently not available here.

Share

COinS