This research work presents methods and techniques for multi-fidelity global optimisation of low-thrust trajectories. In the early stages of the definition of a space mission, tools that can provide a fast and preliminary estimation of the cost of low-thrust transfers are required; a more accurate optimisation process of the trajectories is left for subsequent phases. Therefore, models of different levels of fidelity are needed, based on the current phase of the design and on the desired accuracy. An efficient global optimisation algorithm has then to be used in conjunction with these models, in order to identify the global optimal solution to a given problem. The development of multi-fidelity methods, of an efficient global optimisation algorithm and their application for the solution of low-thrust global optimisation problems, are addressed in this thesis. The lower fidelity models consist of analytical laws for the cost of low-thrust transfers. Higher fidelity innovative laws for transfers between Earth's orbits have been derived. Moreover, a set of analytical equations for the motion of the spacecraft subject to low-thrust acceleration and orbital perturbations is presented. These models are used in conjunction with a novel adaptive multi-population global optimisation algorithm, validated using several test functions and real world problems. To allow for the use of the global solver with higher fidelity, and therefore computationally more expensive models, the use of surrogate model for low-thrust transfer is proposed. Various applications are presented where these tools and methods are successfully applied, and that represent an original scientific contribution. Missions have been designed to deorbit objects from Low Earth Orbit and deploy a constellation in Medium Earth Orbit. The optimisation of a transfer from Geostationary Transfer Orbit to Geosynchronous Orbit is also presented. Interplanetary applications include missions to visit the asteroids of the inner solar system and of the main belt.
|Date of Award||17 Sep 2018|
- University Of Strathclyde
|Sponsors||Astrium Ltd & University of Strathclyde|
|Supervisor||Massimiliano Vasile (Supervisor) & Malcolm Macdonald (Supervisor)|