Machine-learning based TMP prediction in MF process |
---|
학술지명 International Environmental Engineering Conference
저자 김지혜,이경혁,임재림
발표일 2017-11-15
|
Membrane filtration processes such as microfiltration (MF) and ultrafiltration (UF), physically separate the particles, organic matters and colloidal species by permitting the water through the membrane by pressure. Due to their high product water quality, operation convenience, and less footprints, membrane filtration technologies are widely used in water treatment as well as pre-treatments for seawater desalination processes2). However, several drawbacks are existed; relatively high equipment costs, short membrane life-time, and vulnerability to membrane fouling which leads to decrease in permeate flux. According to the accumulation of foulants, transmembrane pressure (TMP) is being increased1). When the TMP reaches a certain level, cleaning-in-place (CIP) was needed. Fouling is inevitably occurred in the membrane processes, therefore, determination of the CIP timing highly influences the operation costs. |