Design and Optimization of Ultrafiltration Membrane Setup for Wastewater Treatment and Reuse

Abstract: With the advances in the membrane technology, there is an ongoing quest to determine the best optimal configuration for an adopted treatment as well as it’s polishing to achieve cumulative sustainability for the treatment process. Henceforth, this thesis report is an evaluation to devise a membrane filtration process for investigating the possibility of treating pre-sedimented municipal wastewater with ceramic ultrafiltration, optimizing the membrane as a pre-treatment for reverse osmosis as an overall strategy for recovering wastewater. Methods and various technological trends pertaining to membrane filtration of municipal wastewater were researched and documented, Interestingly the five most influential factors governing the membrane performance are identified: 1) Back pulsing Frequency 2) VRF 3) Run Time 4) Cross-Flow Rate 5) Trans Membrane Pressure (TMP). To get a thorough and holistic overview of parametric influence design of experiment (DOE) is devised to find the influence of above-given factors on outcoming responses as COD Reduction (%), Membrane Flux and Turbidity reduction (%). 16+3 DOE factorial tests are executed at Hammarby Sjöstadsverk, Joint Research Facility of IVL Swedish Environmental Research Institute & KTH Royal Institute of Technology on pilot plant WASLA incorporating an ATECH GmBh 20kDa, Type 7/6 Ultrafiltration membrane module where Factorial experiments resulted in a maximum value of flux of 274 LMH, 88.75% reduction of COD and 99.94% reduction of Turbidity. Moreover, response values obtained from the Results of factorial experiments are fed in MODDE, generating a model using PLS Regression, The model summary presented predictivity and reproducibility trends w.r.t responses used. Furthermore, COD resulted in the worst fit followed by Turbidity, and the best fit was observed for Membrane Flux where model fit represented the ability to predict the respective parameter. Optimization tool is utilised to simulate a case scenario where the Membrane flux response is maximized to a high value of 300 LMH and correspondingly 211.885 LMH value is recorded, Furthermore factor influence is identified to be TMP> VRF> Cross Flow >BP Frequency >Runtime. Overall COD reductions are found out to be heavily influenced by the varying incoming feed therefore it is hard to analyze their interactions and predict their subsequent reduction behavior. Back pulsing overall was found out to be another non-influential factor colluding with results throughout the experimental duration with very little or no effect on the permeate water quality.