As the clean fuels research has become a significant area of research, many studies were proposed for modelling and optimizing desulfurization process. However, most of the conducted research studies have formulated only a single objective function to be achieved which is maximizing the removal of sulfur compounds from fuel products. In reality, maximizing the removal percentage of sulfur compounds has expensive cost either economically (total annual cost) or environmentally (SO2 emissions). Hence, such process optimization requires considering more than one objective, simultaneously, even if the desired objectives may conflict or compete with each other. This kind of problem is identified as a multiobjective optimization (MOO) problem. In practical uses, MOO can help communities balance productivity, profitability and environmental damage factors and serve as a guide for real-world decision making. Usually, MOO application results in a set of optimal solutions for the undertaken problem. However, practitioners and designers are looking to learn the principles representing these solutions as optimal. Innovization (innovation through optimization) has been recognized as a new method to learn and deduce common patterns from the obtained optimal solutions in order to make a feasible solution which is extremely useful for practical problem In this research project, we aim to handle real-life optimization problem, namely hydro-desulfurization, as a MOO problem taking into account competition between three important desired objectives functions total annual cost, SO2 emissions, and sulfur content. The problem involves a competition among the specified objectives, i.e. maximizing the reduction of the sulfur concentration in fuel products increases inevitably total annual cost and SO2 emissions. To find the Pareto optimal solutions of the understudy problem, we aim to investigate a Pareto front approach with Non-dominated Sorting Genetic Algorithm-II (NSGA-II) as an evolutionary MOO algorithm. For the sake of the optimal design of the process, the obtained optimal solutions will be analysed against each other to discover interesting principles commonly embedded in most obtained solutions. These shared features among high-performance solutions will provide valuable insights about what makes a solution optimal in particular problem. Such innovization analysis would help designers to pinpoint and manipulate the variable(s) that lead to optimum designs. The project work will start with further updating survey on existing MOO approaches proposed for desulfurization process in oil refinery. The survey may go further to explore the oil-upgrading processes in order to investigate the utilization of MOO approaches in similar applications. In addition, a part of this research is to collect data samples for desulfurization of oil using nanomaterials of MoCo (molybdenum cobalt) catalysts, and study their characteristics and significance for the MOO approach. Furthermore, to validate and evaluate the resultant optimal solutions, more real experiments will be conducted. The long-term objective of this work is to emphasis the utilization of the evolutionary computation algorithms to optimize the complex and nonlinear problems in the petroleum industry.
|Effective start/end date
|15/04/19 → 15/04/22
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.