Abstract

The program DDMS (Digital Design Manufacturing and Services) at Airbus aims to increase the CRL (Capability Readiness Level) of various modelling simulation features, including MDAO (Multi-Disciplinary Analysis and Optimisation) capabilities. Additional objectives of the work described in this paper were to: • show the value-added of the methods and tools on a practical Space System use-case • assess the infrastructure and tools required to deliver the capability on realistically-sized scenarios • develop engineering skills in Airbus’s Space Systems business line • evolve the way of working from: - focusing the design around a single point to answer the customer needs, - to delivering inside the boundary of the design space, a behavioural model able to grasp the 3V challenge: Velocity, Variety, Volume and to propose a set of solutions in adequacy to the customer needs. During the early design stage of spacecraft development, the design freedom is high whereas knowledge about the validity of the design is very low. With traditional spacecraft process development only a small number of iterations can be performed, due to a lack of robust and automated processes covering several disciplines. Thanks to automated workflows using DoE (Design of Experiments) or optimization, hundreds if not thousands of designs can be studied and the design space can be thoroughly evaluated. The objective of this paper is to describe the application of MDAO methods and tools to an industrial use case, namely phase 0 study of a low earth orbit space infrastructure. The complexity of the modelling and implementation, discontinuities in the design space and the wide range of time-scales involved need not to be underestimated when moving from research purpose to industrial deployment. The modelled disciplines are: orbit dynamics, cell illumination, electrical power subsystem (including energy storage and solar array sizing), electric propulsion, power budget, mass budget. Coupling between the various discipline models is performed and, for some disciplines, surrogate based models are created thanks to DoE and model training techniques based on artificial intelligence. The global workflow/dataflow is then executed with multi criteria optimization techniques.