The performance of progressive freeze concentration for water purification using rotating cylindrical crystalliser with anti-supercooling holes
Progressive freeze concentration (PFC) is a freeze concentration method which forms ice crystals as a layer or a block on a cooled surface, which can be applied in purification of seawater to obtain pure water in ice form and leave the impurity behind. The aim of this study is to design a rotating c...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2016
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/84032/1/FarahHanimAbPFChE2016.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Progressive freeze concentration (PFC) is a freeze concentration method which forms ice crystals as a layer or a block on a cooled surface, which can be applied in purification of seawater to obtain pure water in ice form and leave the impurity behind. The aim of this study is to design a rotating cylindrical crystalliser with anti-supercooling holes in order to prevent supercooling phenomenon as well as to improve the solution movement in the crystalliser to increase productivity of the PFC system. Prevention of supercooling is important as it will affect the purity of ice produced. The performance analysis was carried out by using saline solution as simulated seawater, and was evaluated by the value of effective partition constant (K), desalination rate (Rd), efficiency (E %) and solute recovery (Rs). The system performance was found to be at its best at 300 rpm of rotation speed, four hours of rotation time, coolant temperature at -8 °C and 35 g/L of initial concentration with K value, E%, Rd and Rs of 0.376, 62.37%, 35.71% and 0.672, respectively. Low K value, high E%, high Rd and high Rs represent the best performance due to higher purity of ice crystal produced. Response surface methodology (RSM) in STATISTICA software was employed for optimisation process to obtain the optimum conditions in producing the best K value and Rd. Due to the limitation of optimization process by RSM offered by STATISTICA, a hybrid Artificial Neural Network and Genetic Algorithm in MATLAB was implemented for multiple response optimisations, where the best K value and Rd predicted were 0.26 and 49%, respectively. A mathematical heat transfer model in predicting ice crystal growth at different coolant temperature was successfully developed with an Average Absolute Relative Deviation of 5.56% and R2 value of 0.873. This newly designed crystalliser was found to be capable of producing ice crystals with high efficiency and productivity. |
|---|