Optimal sizing of an embedded electrical system with an approach for limiting the search space

Source abstract

Purpose The purpose of this paper is to describe an optimization methodology based on a mixed (analytical‐numerical) design model and evolutionary algorithms. Design/methodology/approach In this paper, optimal sizing is used to design the new architecture of an embedded electrical system. An approach is presented for defining the specifications and a method for modeling an embedded electrical network. This will enable the optimization problem to be set. Findings This paper shows that the proposed methodology is useful for sizing an embedded electrical system. It is illustrated by its application to the sizing of an aircraft electrical power channel (auto‐transformer‐rectifier unit type). Originality/value An original approach is proposed for limiting the search space before using the evolutionary algorithms. The advantage of this approach is to increase convergence speed of the evolutionary algorithms.