Browsing by Author "Fraser, Duncan"

Now showing 1 - 6 of 6
  • Thumbnail Image
    Item
    Open Access
    Assessment of pre-treatment technologies for bio-ethanol production using multi-objective optimisation
    (2015) Crimes, Joanne; Isafiade, Adeniyi Jide; Fraser, Duncan
    South Africa’s liquid fuels have a large carbon footprint due to coal- to -liquid fuels however, this could be reduced by blending bio-ethanol in the fuel. It is estimated that 3.3 Mt/year of sugarcane bagasse, a non-food biomass, could be available for biofuel production in South Africa if steam generation from bagasse at the sugarcane mills was more efficient (Lynd et al., 2003). Bagasse comprises lignocellulose which does not contain free sugars, but requires pre-treatment so as to promote access to polysaccharides for hydrolys is to sugars prior to fermentation to ethanol. Lignin present in bagasse prevents access to cellulose, thus lignin is often solubilised in a basic solution prior to hydrolysis. A variety of methods exist for pre-treating bagasse which require different raw materials and have different operating conditions, and thus have different costs and environmental impacts associated with them. In order to determine an optimal pre-treatment network of sugarcane bagasse for the production of bio-ethanol, a systematic procedure which considers economics and environmental impact as objectives should be employed. This thesis uses a systematic approach to develop mixed integer non-linear programs (MINLPs) of pre- treatment options for sugarcane bagasse. The superstructure of pre-treatment options is aimed at embedding the key pre-treatment alternatives, and the optimisation of each of these alternatives is performed using GAMS (General Algebraic Modelling System). The superstructure incorporates the following pre-treatment options: acid pre-treatment and steam explosion (acid-catalysed and un- catalysed), and both acid and enzymatic hydrolysis. The use of delignification using sodium hydroxide prior to hydrolysis was investigated. The benefits of producing methane from the xylose-rich liquid leaving the pre-treatment unit were also included. The superstructure, which embeds the aforementioned pre-treatment options, was developed using insights obtained from detailed modelling and simulation of some key aspects of individual unit operations involved in possible pre-treatment flowsheets. The acid pre-treatment unit was developed in Matlab using reaction kinetic data to generate 13 sets of black box data at differing acid weight percentages and temperatures. The two steam explosion methods and the enzymatic hydrolysis unit, used black box data obtained from Aspen Plus simulations from CTBE (Brazilian Bio-ethanol Science and Technology Laboratory ) (Bonomi, Dayan, Jesus, Cunha, & Mantelatto, 2011). Kinetic equations describing the acid hydrolysis of cellulose were included directly in the GAMS model for acid hydrolysis. Linear relationships describing the solubilisation of solid components with sodium hydroxide weight percentage during delignification were used in the delignification model. The superstructure was decomposed into fixed flowsheets, which involved all possible combinations of these models. The optimal pre-treatment flowsheet was s then chosen based on both economic and environmental objectives by evaluating the solution space. It was found that recycling of sodium hydroxide is needed for profitability in the delignification flowsheets. A recycle cost of 25% of the total annual sodium hydroxide cost with no recycling was used in the flowsheets although the recovery process could possibly be more efficient. However, adding delignification reduced the profitability of all flowsheets except steam explosion with enzymatic hydrolysis. Acid-catalysed steam explosion with acid hydrolysis was one of the most profitable flowsheets and had the lowest environmental impact, however the glucose flowrate produced by this flowsheet was low. Acid-catalysed steam explosion followed by enzymatic hydrolysis produces more glucose and was more profitable however, the environmental impact of this method may be very large due to the use of enzymes. Enzymes (excluding transportation) can contribute significantly to environmental impact if the production method is energy intensive and the e energy production method is carbon-intensive method . More research into the environmental impact of enzymes should be conducted to determine which hydrolysis method should be chosen.
  • Thumbnail Image
    Item
    Open Access
    The development of a model for high pressure propene oligomerisation over H-ZSM-5
    (1996) Sealy, Sarah Jane; O'Connor, Cyril; Fraser, Duncan
    The primary aim of this thesis is to develop a fundamental, mathematical model for high pressure propene oligomerisation over H-ZSM-5. To this end, this introductory chapter first revises the characteristics and applications of catalysts which have been shown to be active for oligomerisation. This is further split into acid catalysts, heterogeneous nickel catalysts, other transition metal catalysts and finally homogeneous metal catalysts. The next section covers the various mechanisms of oligomerisation on different catalyst types since an understanding of these mechanisms is an important step in the modelling process. Following this, the modelling of oligomerisation reactions is reviewed with an additional section on other complex reaction models. Throughout this chapter heterogeneous acid catalysts are given the most emphasis, especially H-ZSM-5, since this thesis is concerned with modelling oligomerisation over acid catalysts. The final sections of this chapter cover the objectives of this study and the approach to studying this complex reaction in detail.
  • Thumbnail Image
    Item
    Open Access
    The effect of selected academic development programmes on the academic performance of academic development students at a South African university : an empirical study.
    (2012) Smith, Leonard Cowper; Case, Jenni; Fraser, Duncan; Van Walbeek, Corné; Ranchhod, Vimal
    The case studies that make up this thesis cover the three largest academic development programmes at the University of Cape Town. A variety of statistical methods are used to estimate the effect of educational interventions in selected first- and second-year academic development courses on the academic performance of academic development students in these courses and through to graduation, relative to mainstream students. In general, research in this area in South Africa and internationally has been characterised by small sample sizes and a lack of statistical rigour. Few studies control for the range of independent variables that can affect students’ academic performance, in addition to the academic development programme or course, and the great majority ignore the sampleselection problem that arises in the selection of students for academic development and mainstream programmes. The theoretical rationale underpinning this thesis is informed by the postpositivist and evidence-based approaches to empirical investigation. Demographic, academic and other data for some 9000 students for the years 1999?2005 was obtained from the university’s data base and academic departments. Statistical techniques including multivariate analysis and propensity score matching are used in an attempt to finesse the problems associated with the use of non-experimental data as students are selected into different courses and programmes.
  • Thumbnail Image
    Item
    Open Access
    Lion Battery Museum
    (2012) Fraser, Duncan; Noero, Jo
    The idea of space lies at the root of all architecture; architecture is about spatial experience. This thesis is an investigation around the ideas of architectural spatial experience how we experience space and how to construct rich spatial experiences in architecture. The design project revolves around re-imagining the Lion Battery site on the slopes of Signal Hill. The design tests out the theoretical explorations laid out in this document and then incorporates tradtional and parametric approaches in the resolution of the design. The aim of this project is to explore experiential architecture.
  • Thumbnail Image
    Item
    Open Access
    Optimisation of South Africa's biomass to bio-ethanol supply chain network
    (2016) Mutenure, Mildred; Isafiade, Adeniyi Jide; Fraser, Duncan
    This study is about the optimisation of bio-ethanol supply chains for economic and environmental objectives, using a mathematical programming approach. A superstructure presented as a Mixed Integer Linear Programme (MILP) model that adequately captures the key variables in South Africa's bio-ethanol supply chain network is developed. The MILP model accounts for food demand, geographical distribution of biomass cultivation areas and biomass diversity, feedstock, product and by-product distribution, product demand and tax subsidies. The study focuses on the use of sugarcane, bagasse and crop waste from maize, wheat, barley and sorghum in the production of bio-ethanol. In the supply chain, one processing technology for ethanol production is considered and one mode of transportation for both feedstock and products is considered. A detailed profitability analysis of the optimised MILP model is also provided. To account for the environmental impact of the supply chain, the model is integrated with life cycle analysis through multi-objective optimisation. The ε- constraint method is used to solve the multi-objective optimisation problem and Pareto analysis is done to check the trade-offs between the economic and environmental objectives, which is measured mainly by greenhouse gas emissions. In addition to greenhouse gas emissions, other impact categories namely eutrophication, human toxicity, acidification and global warming potential were also considered. Bio-ethanol production has been a subject of many studies. It is a renewable and potentially environment-friendly product, which after blending with petrol can be used as a fuel in the transport sector. The use of bio-fuels has the potential to relieve pressure on fossil-based fuels, and achieve a reduction in the emissions of greenhouse gases. The use of bio-fuel results in net savings in carbon dioxide gas emissions as plants absorb the carbon dioxide released during bio-fuel production during biomass cultivation. The bio-fuel industry worldwide, however, faces many challenges, which compromises its economic viability and commercialisation, especially where lignocellulosic biomass is to be used in bio-fuel production. These challenges include the uncertainty or discontinuous availability of biomass, fluctuations in market prices, high logistics and high maintenance costs of the processing equipment. The high logistics costs arise from the low density of the feedstock and from distribution of the feedstock, which is usually scattered over a wide area thereby making the process energy intensive. To overcome these challenges, an optimised supply chain network is required.
  • Thumbnail Image
    Item
    Open Access
    Processing cost analysis of the African biofuels industry with special reference to capital cost estimation techniques
    (2008) Amigun, Bamikole; Von Blottnitz, Harro; Fraser, Duncan
    Access to energy, in the form of electricity and fuels, is a necessary condition for development. There are several reasons for biofuels to be considered important in many African countries. They include energy security, environmental concerns, foreign exchange savings and socio-economic opportunities for the rural population. Biofuels such as biogas, biodiesel and bioethanol may be easier to commercialise than other alternatives to crude-oil derived fuels, considering performance, infrastructure and other factors. Biofuels are in use in a number of developing countries (including some African ones for example, Mauritius, South Africa, Kenya), and have been commercialised in several OECD countries, as well as Brazil and China. A good understanding of the production cost of biofuels, and the availability of robust and indigenous cost estimation models is essential to their eventual commercialization. However, available process engineering cost estimation relationships and factors are based on plant costs from developed countries, and thus have limited applicability and unknown accuracy when applied to African installations. The need to develop indigenous cost prediction relationships, which is central to economic feasibility studies, is driven not only by the limitations in terms of current data bases and methodologies for the generation of such. There is also a requirement for a more systematic presentation of cost data in equation forms, which will ensure easier and more rapid use of the data in numerical and economic models, and in preliminary design and plant optimisation in a time and cost effective manner, providing decision-makers with key information in the early design stages of a project. It is these shortcomings and challenges that this dissertation attempts to address, through an analysis of the economic input factors, and the development of more robust, indigenous cost estimation relationships for both capital and operating costs for the biofuels process industry in Africa. The conceptual approach developed within this thesis addresses the current data gaps and deficiencies through analyses of establishment and operating costs of existing biofuels plants both on the continent and elsewhere. It aims to determine which factors most influence the production cost, and then proceed to modify known cost estimation tools for both capital and operating costs specifically for African biomass-to-biofuel conversion plants as a function of plant size, feedstock, location, exchange rates, and other site-specific variables. Shortcomings in the use of existing cost estimation models are addressed with the aid of a literature study, supported by the analysis of African biofuels plant establishment (biogas and bio-ethanol) and operating (bio-ethanol) costs‡. Plant establishment costs are analysed at two different levels of detail, corresponding to the concept development and pre-feasibility phases in the project planning cycle.