Multi-objective computational engineering design optimisation for micro-combustor devices

This thesis describes the development of a multi-objective automated optimisation system to be used for the design optimisation of micro-scale combustion devices. The developed system described within integrates a commercial computational fluid dynamics package and a multi-objective variant of the Tabu Search optimisation algorithm for continuous problems, which is a heuristic optimisation technique that exhibits local search characteristics.

Recent advances in micro-fabrication techniques have resulted in increasing interest from industry and academia to investigate the possibility of replacing the current conventional power supply “battery” with a miniaturised combustion power generation system based on micro-electro-mechanical systems (MEMS) technology. The microcombustor is one of the crucial components of such a power system. The aim is to improve the main micro-scale combustor design characteristics and to satisfy manufacturability considerations from the very beginning of the whole design process.

The main combustor design requirements, challenges and design parameters that influence the device performance at a micro-scale were first defined. Within the optimisation design cycle a robust parameterisation scheme, the geometry and numerical grid representations were implemented. These were achieved by incorporating the knowledge gained from the parametric design study by understanding the design space in depth and identifying issues and their solutions during this design study such as geometry overlapping and mesh refinement. Cont/d.