Focused Averaging for Low-Thrust Primer Vector Theory with Eclipsing
摘要
Focused averaging is introduced to improve the computational efficiency of many-revolution, low-thrust spacecraft trajectory dynamics. The new method applies to smoothed models for bang-bang control and eclipsing constraints, using judicious spacing of quadrature points to efficiently average the state and costate dynamics equations. Unlike existing models, the averaged Hamiltonian is constant and is introduced as a validation and comparison tool. A numerical sensitivity is identified that arises in any averaged dynamics model with the typical conical shadow model, and is fixed with a novel redefinition of the eclipsing function. The unaveraged and averaged dynamics models are directly compared with an optimal 44-revolution minimum fuel transfer. Additionally, the capabilities of focused averaging are demonstrated with an optimal 911-revolution LEO to GEO minimum fuel transfer. These examples incorporate eclipsing in an optimal manner, generic two-body perturbations, and variational equations for use in targeting and optimization.