Conceptual Spacecraft Design Using a Genetic Algorithm Trade Selection Process

Abstract

Spacecraft design is a highly coupled problem. The design of the spacecraft must balance payload objectives and orbital design against cost and schedule guidelines. In commmon practice, a limited set of spacecraft design alternatives are considered and the design objectives are often not formally stated. With this approach there is no guarantee that a system-level focus will be taken, and feasibility rather than optimality is commonly all that is achieved. With support from NASA, the author has investigated better methods for performing the earliest stage of spacecraft design (including cost analysis) to ensure that a systems perspective is taken. This paper describes a tool for conceptual spacecraft design, the spacecraft concept optimization and utility tool (SCOUT), which uses a set of design-estima ting relationships and cost-estimating relationships coupled with genetic algorithm optimization. A retrospective comparison between SCOUT results and the actual design of the Near Earth Asteroid Rendezvous spacecraft is made. This approach helps to enhance the understanding of the feasible design space, as well as uncovering interesting design points that merit further investigation.