Emanuela Gaglio

emanuela.gaglio-ssm[at]unina.it

PhD in: SPACE
Ciclo: XXXVII

Supervisor(s): Raffaele Savino, Riccardo Bevilacqua

  • Scarica il CV

  • Title: Aerothermodynamic and optimal control assessment of a deployable system for atmospheric re-entry.

    Atmospheric re-entry is a crucial part of any space mission with recovery and landing requirements, for instance human re-entry from Low Earth Orbit or scientific space experiments demanding the recovery of samples for further analysis. Controlled atmospheric entry is also critical for deep space mission that involve the exploration of the surface of celestial bodies in the solar system. The extremely high specific energies involved, the reactive nature of the gasses and the prohibitive temperatures experienced during this phase require a proper design of the vehicle, with particular attention to those areas most affected by the detached shock wave phenomena. Re-entry technologies have been developed originally for large systems like the first manned spacecraft and later also for relatively small capsules able to return to Earth samples from Low Earth Orbit or other planets or asteroids. With the recent interest in micro and nanosatellites, including CubeSats, different solutions have been proposed to offer de-orbit and recovery capabilities after Low Earth Orbit scientific or technologic missions.

    In this framework, deployable re-entry systems and related algorithms represent research topics of increasing interest among scientists and engineers working in the development of Entry, Descent, and Landing (EDL) technologies. The rationale is to reduce the ballistic coefficient without exceeding the volumetric constraints of current and future space launchers fairings. At the same time, a deployable system, i.e., a surface increasing system, may raise the probability of collisions in low Earth orbit, thus its design needs to minimize the compound risk by substantially reducing the time to re-enter or enter. In the last decades, several projects have been taking shape all over the world, from the United States [1-2] to Italy [3], both combining an umbrella-like heat shield with a proper number of flaps embedded at the extremities to ensure aerodynamic control during the re-entry phase. Meanwhile, inflatable re-entry systems represent an alternative lightweight return technology to provide mass and cost savings compared to conventional fixed heat shields and parachute systems. They are folded into a small volume by employing a flexible film and inflated in orbit to change their aerodynamic shape to have a large diameter. One of their main advantages lies in the possibility of achieving lower ballistic coefficients compared to rigid shell systems, resulting in a reduction of heath fluxes as well as payload deceleration at a higher altitude in the atmosphere [4-5].

    Another essential aspect when dealing with re-entry systems is the identification of a proper Guidance, Navigation and Control (GNC) system to ensure precision landing in the SPOUA (South Pacific Ocean Uninhabited Area) or a safe, targeted, and timely recovery. Among all techniques, alternative control mechanisms such as drag modulation (DM) can be employed instead of the de-orbit burn to go from orbit to the re-entry interface. DM is based on the effective drag force modulation on the vehicle, to achieve the necessary reduction in orbital energy. In literature, several works presented a method to target the atmospheric re-entry point by controlling aerodynamic drag through ballistic coefficient modulation [6-7]. However, none of these works provides for optimality as well as instantaneous feedback, resulting in large errors and/or numerical instabilities under certain conditions.  In this framework, optimization is extremely important in re-entry problems to minimize time to come down when crosswind surface area increases, keeping accuracy under check. Recently, researchers are focusing their attention on Deep Learning techniques to overcome the issues referred to above. Training Deep Neural Networks as optimal controllers during probe landing is proving to be a promising technique [8-10].

    In this framework, the current research project seeks obtain two fundamental objectives:

    • The aerothermodynamic assessment of a Low Earth Orbit (LEO) deployable re-entry system, also with reference to possible scenarios involving small satellites. Investigation of aerodynamic behavior in relevant flight conditions of its trajectory, which will correspond to different flight regimes, from hypersonic to transonic to subsonic. Computational Fluid Dynamic (CFD) software will be employed to investigate the flow field evolution around the system, also allowing studies on aerodynamic stability. In addition, engineering tools such as Direct Simulation Monte Carlo (DSMC), will be used to carry out probabilistic analyzes at altitudes characterized by rarefied flow regimes [11]. To conclude, a preliminary study on the thermal protection system will be eventually performed to ensure thermal loads compatible with the structure. These aspects will be investigated under the supervision of Professor Raffaele Savino (University of Naples “Federico II”), with the collaboration of Professor Rodrigo Cassineli (Universidad Técnica Santa Maria, Santiago del Chile).
    • The investigation of an optimal control system. The aim is to take advantage of Deep Neural Networks (DNN) to ensure a precise landing. Attention will be focused on both deorbiting, from initial orbit to re-entry interface, and a final phase from the re-entry interface to the landing point. First, a class of open loop optimal re-entry trajectories, based on the sole modulation of the drag force, will be generated, and will constitute the training dataset for Deep Neural Networks. In this regard, different initial conditions and cost functions will be employed, e.g., error on longitude and latitude, crosswind surface area and/or descent time, etc. The OpenOCL Matlab tool will be employed to generate the open loop optimal trajectories. Concerning the second part (DNN), it will be addressed through Python programming language and analyzing different architectures to find the most suitable one. This work will be carried out under the supervision of Professor Riccardo Bevilacqua (Embry Riddle Aeronautical University).

    REFERENCES

    1. Reddish, B. J., Nikaido, B. E., D’Souza, S. N., Hawke, V. M., Hays, Z. B., & Kang, H. S. Pterodactyl: Aerodynamic and Aerothermal Modeling for a Symmetric Deployable Earth Entry Vehicle with Flaps. Rep.
    2. Johnson, B. J., Rocca-Bejar, D., Lu, P., Nikaido, B., Hays, Z. B., D’Souza, S., & Sostaric, R. R. (2020). Pterodactyl: Development and Performance of Guidance Algorithms for a Mechanically Deployed Entry Vehicle. In AIAA SciTech 2020 Forum (p. 1011).
    3. Fedele, A., Carannante, S., Grassi, M., & Savino, R. (2021). Aerodynamic control system for a deployable re-entry capsule. Acta Astronautica, 181, 707-716.
    4. Marraffa, L., Vennemann, D., Anschuetz, U., Walther, S., Stelter, C. S., Pitchkhadze, K. M., & Finchenko, V. S. (2003, April). IRDT-Inflatable Re-entry and Descent Technology. In Hot Structures and Thermal Protection Systems for Space Vehicles (Vol. 521, p. 19).
    5. Beck, R. A., White, S., Arnold, J., Fan, W., Stackpoole, M., Agrawal, P., & Coughlin, S. (2011, May). Overview of initial development of flexible ablators for hypersonic inflatable aerodynamic decelerators. In 21st AIAA Aerodynamic Decelerator Systems Technology Conference and Seminar (p. 2511).
    6. Virgili, J., Roberts, P. C., & Hara, N. C. (2015). Atmospheric interface reentry point targeting using aerodynamic drag control. Journal of Guidance, Control, and Dynamics, 38(3), 403413.
    7. Fedele, A., Omar, S., Cantoni, S., Savino, R., & Bevilacqua, R. (2021). Precise re-entry and landing of propellantless spacecraft. Advances in Space Research, 68(11), 4336-4358.
    8. Sánchez-Sánchez, C., & Izzo, D. (2018). Real-time optimal control via deep neural networks: study on landing problems. Journal of Guidance, Control, and Dynamics, 41(5), 1122-1135.
    9. Gaudet, B., Linares, R., & Furfaro, R. (2020). Deep reinforcement learning for six degreeof-freedom planetary landing. Advances in Space Research, 65(7), 1723-1741.
    10. Furfaro, R., Bloise, I., Orlandelli, M., Di Lizia, P., Topputo, F., & Linares, R. (2018). A recurrent deep architecture for quasi-optimal feedback guidance in planetary landing. In IAA SciTech Forum on Space Flight Mechanics and Space Structures and Materials (pp. 1-24).
    11. Carná, S. R., & Bevilacqua, R. (2019). High fidelity model for the atmospheric re-entry of CubeSats equipped with the drag de-orbit device. Acta Astronautica, 156, 134-156.