Assessment of pre-treatment technologies for bio-ethanol production using multi-objective optimisation

Includes bibliograpical references.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