System architecture selection in a multi - disciplinary system design optimization framework

In today's automotive business environment, responsibility for module and sub-system design and development is increasingly driven into the supply chain. Simultaneously, the pressure to reduce product cost is maintained if not increased. The engineering organizations of the higher tier suppliers are saddled with the task of managing a product development portfolio ranging from largely mature products and sub-systems to new, and risky, innovations, all on a shoe-string budget. The primary problem the work in this thesis attempts to resolve is the R&D investment decision for those projects deemed to be high risk and high potential. The decision to invest in a particular highrisk technology is generally based on an economic analysis, often based on far too optimistic expectations of product performance. Multi-disciplinary System Design Optimization (MSDO) has been used for decades in the Aeroand Astro-nautical industry to design and evaluate system concepts that cannot be built and tested for practical reasons. This thesis attempts to extend MSDO to architectural concept evaluation -in an automotive industry environmentwith the specific intent to provide a more rigorous and quantitative technical analysis of promising technologies and the system level performance impact they may have. With a more rigorous technical analysis, quality of the economic analysis can be improved, enhancing R&D investment decisions. A representative test case of a Plasmatron enabled internal combustion engine is used throughout this thesis. The novel technology is a fuel-reforming device invented at the MIT Plasma Science and Fusion Center. The device can reform fuel on-board a vehicle and provide hydrogen rich gas on demand. This hydrogen rich gas can be used as a fuel additive to significantly enhance combustion of a spark ignited, homogeneous burn internal combustion engine. Neither the performance of the fuel reformer, nor the impact on engine (and vehicle) efficiency is well understood. Within the MSDO framework, a set of Matlab models part physical and part empirical have been developed to describe the various aspects of the system of interest. System simulations then focus on the trade-off between the primary benefit of the technology -vehicle fuel efficiency improvementand the primary constraints, emissions regulations and system cost impact.