Refine
Year of publication
- 2009 (1)
Institute
Has Fulltext
- no (2)
Language
- English (2)
Document Type
- Article (1)
- Conference Proceeding (1)
Keywords
- Hybrid Propellants (1)
- Reusable Rocket Engines (1)
- Spacecraft (1)
Is part of the Bibliography
- no (2)
The recently proposed NASA and ESA missions to Saturn and Jupiter pose difficult tasks to mission designers because chemical propulsion scenarios are not capable of transferring heavy spacecraft into the outer solar system without the use of gravity assists. Thus our developed mission scenario based on the joint NASA/ESA Titan Saturn System Mission baselines solar electric propulsion to improve mission flexibility and transfer time. For the calculation of near-globally optimal low-thrust trajectories, we have used a method called Evolutionary Neurocontrol, which is implemented in the low-thrust trajectory optimization software InTrance. The studied solar electric propulsion scenario covers trajectory optimization of the interplanetary transfer including variations of the spacecraft's thrust level, the thrust unit's specific impulse and the solar power generator power level. Additionally developed software extensions enabled trajectory optimization with launcher-provided hyperbolic excess energy, a complex solar power generator model and a variable specific impulse ion engine model. For the investigated mission scenario, Evolutionary Neurocontrol yields good optimization results, which also hold valid for the more elaborate spacecraft models. Compared to Cassini/Huygens, the best found solutions have faster transfer times and a higher mission flexibility in general.
Our objective function value for the given mission is J = 1.2 with a secondary performance index T = 1,271 d. To find the optimal sequence of flyby bodies to solve the given problem, we have used an adapted version of InTrance [1-3]. Because InTrance is a purely global trajectory optimization program based on Artificial Neural Networks (ANN) and Evolutionary Algorithms (EA), its local optimization capabilities are rather poor. Therefore, its accuracy is typically limited to a final distance ΔR of about 10 5 km and a final relative velocity ΔV of about 100 m/s (to the respective target asteroid). Nevertheless, for single leg optimization it is possible to optimize with the accuracy requirements needed to solve the given problem. For trajectory integration, we have used a maximal integration error of 10 -10 . We have used the following stepwise approach: 1. We let InTrance find a promising rendezvous candidate out of 27 manually target asteroids 2. The first rendezvous to the determined target asteroid was optimized to fulfill the given accuracy requirements of 1,000 km in distance and 1 m/s relative velocity. 3. The steps 1 and 2 were again carried out for another target asteroid. 4. A flyby to the first found body was optimized and another rendezvous to another candidate out of the list of remaining targets. The transfer with the shortest flight time was chosen. 5. The steps 3 and 4 were repeated until no more rendezvous or flybys could be achieved. We are absolutely aware of the fact that this approach can by principle not yield a globally optimal trajectory as the key to success is the generation of a promising sequence of rendezvous and flybys, but further adaptation of our tools to the given problem was not possible in the given time.