Browsing by Author "Spratt, Alexander"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Recovery of organic solvents from dewaxed oil mixtures using organic solvent nanofiltration technology
    (Stellenbosch : Stellenbosch University, 2019-04) Spratt, Alexander; Burger, A. J.; Van der Gryp, Percy; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.
    ENGLISH ABSTRACT: Solvents are valuable chemicals, which are conventionally recovered through energy intensive processes such as distillation. In lube oil dewaxing processes, four times more solvent relative to solute is required. Solvents, such as toluene and methyl ethyl ketone (MEK), are commonly used in the dewaxing process. Due to high-energy costs, potential alternatives to recover these solvents are investigated. An alternative to solvent recovery through distillation is the use of Organic Solvent Nanofiltration (OSN) technology, which incorporates nanofiltration membranes designed to separate solvent-oil mixtures. The Max DeWax operation using OSN technology demonstrated successful recovery of solvent, while providing lower energy usage in solvent recovery. The research documented in this thesis focused on solvent recovery, membrane performance, transport modelling and techno-economic evaluation. The main aim of the project was to investigate the viability of OSN separation as an alternative method to conventional process separation. OSN viability evaluation was accomplished through demonstrating the recovery of solvent-oil mixtures using novel membranes. Experimental investigations performed in this study focused on recovery of four commercially available solvents used in lube-oil dewaxing processes, namely toluene, methyl ethyl ketone (MEK), methyl-isobutyl ketone (MIBK), di-chloro-methane (DCM) and the solute species which represented long-chain paraffin solutions, n-hexadecane (C16H34). Duramem™150, Duramem™200 and Puramem™280 membranes were used for the recovery process. Operating parameters, which include pressure, solute feed concentration, solvent type and membrane type, were varied and the effects thereof on membrane performance were investigated. The investigation also focused on describing the mass transfer through membranes using transport models, such as solution-diffusion and pore-flow transport models, using Matlab R2013a. Modelling of the mass transfer of MEK and toluene, well-known commercial solvents used in lube-oil dewaxing and processing, was done according to two pore-flow models (PF-1, PF-2) and two solution-diffusion models (SD-1, SD-2) using Duramem™150, Duramem™200 and Puramem™280 membranes. Furthermore, the permeability of hexadecane was regressed to fit the experimental data. The viability of OSN operations in comparison to distillation operations was determined through a preliminary techno-economic evaluation using simulation software (Aspen Plus V8.8) to describe the mass and energy and provide supporting data for use in cost evaluation. Energy consumption, equipment performance as well as operating and capital costs were investigated. The main contributions made by this study are threefold: (i) to demonstrate the successful recovery of solvent-oil mixtures using novel membranes through experimentation, (ii) to describe the transport through membranes using transport models and (iii) to investigate the economic feasibility of OSN systems, using simulation software. i)Recovery of solvents from oil mixtures This study found that the recovery of MEK from solute was the most successful while providing high membrane fluxes and high rejections over the rest, followed by DCM. Overall, MEK and n-hexadecane, at feed concentrations above 20 wt/wt% separated using Duramem™150 membranes, provided >90% rejection, while permeating at fluxes of approximately 12 L.mˉ².hrˉ¹. Membrane performance and solvent behaviour of permeating species were found to be affected mainly by applied pressure, chemical properties that describe polarity such as di-electric constant and dipole moment as well as properties such as molar volume, viscosity and solubility parameters. ii)OSN modelling and simulation The transport of pure solvent and binary solvent-solute mixtures was described using transport models based on literature. Using MEK and toluene, the two-parameter pore-flow model (PF-2)and the classical solution-diffusion model (SD-1) provided relatively good predictions forthe transport through the polar stable membranes such as Duramem™150 and Duramem™200, but poor predictive models for non-polar stable membranes such as Puramem™280. The PF-2 model and SD-1 model provided Pearson coefficients of >0.988 and >0.996, respectively, for the Duramem™ series membranes. The SD-1 model was further improved after regressing the estimated permeability parameter of hexadecane which provided an optimized Pearson coefficient of 0.9995. iii)OSN solvent recovery For both OSN and distillation systems, while ignoring the cost of raw material, It was found that the energy required to recover a ton of MEK solvent by OSN (i.e. 2.5 kWh.tonsolvent_product-1)is approximately 50 times less than that of distillation (135 kWh.tonsolvent_product-1) with energy recovery incorporated, which also results in a lower carbon footprint. However, by using theNelson-Farrar cost index, it was found that the capital costs for OSN ($0.85 million) in 2015 were approximately ~25% of the capital cost for distillation (i.e. $3.36 million). The operating costs, while ignoring the cost of raw material and having a total operating feed capacity of 1 ton.hr-1, were approximately $0.075 million.yr-1 for OSN operation and $0.155 million.yr-1 for distillation operation with a recycling stream and heat integration, while providing solute rejections as high as 97%. Total operating costs of OSN are less than half the amount required for distillation with heat integration, where energy and maintenance costs differ significantly between the two processes.