Development of a novel system for progressive freeze concentration process

• Main Author: Jusoh, Mazura
• Format: Thesis
• Language: English
• Published: 2010
• Subjects: TP Chemical technology
• Online Access: http://eprints.utm.my/id/eprint/18754/16/MazuraJusohPFKKKSA2010.pdf; http://eprints.utm.my/id/eprint/18754/

Freeze concentration (FC) was investigated focusing in terms of assessment towards benefiting industrial wastewater treatment. A new more productive crystalliser was aimed to be designed as the main component in a progressive freeze concentration (PFC) process, a more economical version of FC which forms ice crystals as a layer or a block on cooled surface. Subsequent analysis on its performance, a process optimisation and heat transfer study were intended following completion of design. As a result, a helical structured copper crystalliser was successfully developed and fabricated named coil crystalliser, designed in such a way to provide high surface area per coolant volume, easy scale up and practical. In the subsequent performance analysis carried out using glucose solution as simulated industrial wastewater treatment, it was found that the effective partition constant (K) was satisfactorily low at high circulation flowrates, low initial concentration and intermediate coolant temperature. Low coolant temperature of -10 °C was observed to cause high ice front growth rate which subsequently promote high solute entrapment in the ice layer formed, affecting its purity. High circulation flowrates of 1000 ml/min on the other hand resulted in low solute inclusion in solid phase thus giving out high ice purity. In terms of volume reduction, the highest achieved was 76.65 % and coolant temperature seems to influence the most where low temperatures resulted in high solution volume reduction as a consequence of high ice growth rate. A process optimisation employing Response Surface Methodology (RSM) in Statistica software to yield the optimum conditions to produce the best K, ice purity (IP) and solution volume reduction (�V) generated three regression models, which have been proven adequate by good R-squared, residual and Analysis of Variance (ANOVA) analysis. Interactions between process variables according to the model agree well with the fundamental theory of FC and finally optimum conditions for each response have been identified. The best K predicted was 0.17, 78.5 % for �V and 0.05 mg/ml for IP. In the subsequent heat transfer study standard lines for overall heat transfer coefficient (Uo) were plotted. The Uo lines then facilitate in generating a model to predict ice crystal mass produced, originated from a heat balance analysis. Error analyses have proven the model’s reliability with Rsquared of 0.997 and Absolute Average Relative Deviation (AARD) of 10.6% between experimental and model data, with the highest predicted mass of 973.2g