Evolutionary computing strategies for robust design with application to shipboard engineering plant design
Abstract
In this work, a novel evolutionary optimization technique for robust design is proposed. This technique considers perturbations to the parameters during the design process such that the final design is expected to be robust, i.e., to satisfy all design specifications when the parameters are slightly perturbed. A unique feature of the proposed evolutionary algorithm is that it allows users to specify the degree of robustness of the final design prior to the optimization. Hence, the degree of robustness is specified as part of the design as opposed to being computed after the design is complete. This new methodology is motivated by and demonstrated on a problem concerning the energy storage allocation on a warship. Energy storage units are considered to increase system performance in the event of battle damage. An optimal energy storage unit allocation is desired due to the increase in cost and mass of the warships. In the course of solving the problem, a new method to represent power system for early design studies is also set forth. The proposed evolutionary algorithm is also tested on an electromagnet design problem.
Degree
Ph.D.
Advisors
Sudhoff, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.