Event Detail

Event Type: 
Applied Mathematics and Computation Seminar
Date/Time: 
Friday, November 9, 2012 - 04:00
Location: 
GLK 113

Speaker Info

Institution: 
OSU Mechanical Engineering
Abstract: 

Designing complex systems for mission-critical applications, such as aerospace or military missions, requires a design process with a focus upon maximizing the probability of meeting a set of design specifications. These design specifications typically come from the contracting agency, and proof of compliance to design specifications is required before investing in physical prototypes. Typically the design process for these systems consists of filtering and refining an initial set of conceptual designs to produce a final set of detailed designs most capable of meeting specifications. In this talk, a framework is presented for a multi-stage design process in which a Probabilistic Design Certification with respect to a set of requirements is utilized as a metric to sequentially filter designs from the abstract conceptual phase through the detailed design phase. This framework utilizes methods for uncertainty quantification (UQ); the methods are organized within a framework to match the UQ method with the model fidelity and data type available at each stage of the design process. The key element of this approach is the consideration of multiple requirements and the ability to determine the joint probability of meeting the complete set of requirements. The challenges faced in implementing such a process are the non-linearly of the governing equations and potentially large number of decision variables, the need to accurately represent the tail of the distribution for reliability calculations, and the desire to validate a given UQ method for a given class of problems.