MN
M.C. Naeije
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
48 records found
1
Making launch vehicles fully reusable is a central goal of modern spaceflight, yet recovering the upper stage remains unsolved, because it must dissipate enormous energy during atmospheric re-entry while staying inside a narrow heating and load corridor and reaching a precise terminal state. This thesis develops and validates a closed-loop guidance and control framework for such re-entries. A successive convexification guidance layer solves the constrained re-entry trajectory optimisation problem by convex optimisation, re-planning in a receding horizon, while a three-axis attitude controller tracks its commands on the fully coupled nonlinear dynamics. The framework is then turned into a tool that assesses vehicle configurations and actuator sizing by flying each candidate design through the complete closed loop. Validated on a winged-body vehicle and the Apollo capsule through extensive Monte Carlo campaigns, it quantifies robustness and links early design choices to real mission and propellant outcomes.
...
Making launch vehicles fully reusable is a central goal of modern spaceflight, yet recovering the upper stage remains unsolved, because it must dissipate enormous energy during atmospheric re-entry while staying inside a narrow heating and load corridor and reaching a precise terminal state. This thesis develops and validates a closed-loop guidance and control framework for such re-entries. A successive convexification guidance layer solves the constrained re-entry trajectory optimisation problem by convex optimisation, re-planning in a receding horizon, while a three-axis attitude controller tracks its commands on the fully coupled nonlinear dynamics. The framework is then turned into a tool that assesses vehicle configurations and actuator sizing by flying each candidate design through the complete closed loop. Validated on a winged-body vehicle and the Apollo capsule through extensive Monte Carlo campaigns, it quantifies robustness and links early design choices to real mission and propellant outcomes.
Ariane 6 Engine Bay Recovery
A Trajectory Optimisation for a Semi-Guided Ballistic Re-entry
To improve the cost and sustainability of the Ariane 6, the engine bay of the first stage could be recovered for refurbishment. In this MSc research, the feasibility and cost of this mission profile are assessed from a trajectory point of view, making use of the new reusable Prometheus engine under development. To do this, a 3-DOF TudatPy simulation has been set up, and a guidance algorithm based on Apollo re-entry guidance is implemented into it. Five landing locations were selected to investigate the possibility of crossing the Atlantic during the re-entry following a LEO launch, which was deemed possible if a correction to the initial state is performed after the engine bay is disconnected from the first stage. Following an MC analysis, an MDO was performed to find the set of best identified trajectories to minimise the propellant mass needed for this initial state correction. A trajectory arriving at Cabo Verde was selected as the best option for this mission profile. This reusable mission profile was estimated to be cheaper per launch and over its life cycle and could be more sustainable if the boosters were also made reusable. This is because additional boosters may often be necessary to deal with the significant payload penalty that is incurred, although there are possibilities to work around this penalty. An issue that certainly must be addressed is raising the launch cadence to at least 30 launches per year, as below this number reusability may not be a viable option. The economic viability depends on reaching a sufficient launch cadence as much as on the technical recovery architecture. Furthermore, as this research investigated the trajectory, it must still be seen if the required hardware fits within the engine bay. If these issues can be addressed, the reusable engine bay mission profile may prove worthy of more thorough investigation, to determine if it can be adopted for the Ariane 6 and future vehicles.
...
To improve the cost and sustainability of the Ariane 6, the engine bay of the first stage could be recovered for refurbishment. In this MSc research, the feasibility and cost of this mission profile are assessed from a trajectory point of view, making use of the new reusable Prometheus engine under development. To do this, a 3-DOF TudatPy simulation has been set up, and a guidance algorithm based on Apollo re-entry guidance is implemented into it. Five landing locations were selected to investigate the possibility of crossing the Atlantic during the re-entry following a LEO launch, which was deemed possible if a correction to the initial state is performed after the engine bay is disconnected from the first stage. Following an MC analysis, an MDO was performed to find the set of best identified trajectories to minimise the propellant mass needed for this initial state correction. A trajectory arriving at Cabo Verde was selected as the best option for this mission profile. This reusable mission profile was estimated to be cheaper per launch and over its life cycle and could be more sustainable if the boosters were also made reusable. This is because additional boosters may often be necessary to deal with the significant payload penalty that is incurred, although there are possibilities to work around this penalty. An issue that certainly must be addressed is raising the launch cadence to at least 30 launches per year, as below this number reusability may not be a viable option. The economic viability depends on reaching a sufficient launch cadence as much as on the technical recovery architecture. Furthermore, as this research investigated the trajectory, it must still be seen if the required hardware fits within the engine bay. If these issues can be addressed, the reusable engine bay mission profile may prove worthy of more thorough investigation, to determine if it can be adopted for the Ariane 6 and future vehicles.
From Re-Entry to the Ozone Layer
A One-Dimensional Investigation of Alumina Residence Times in the Middle Atmosphere
The rapid growth of satellite constellations is increasing the frequency of atmospheric re‑entries, raising concerns about the environmental impacts of ablation by‑products such as alumina (Al2O3). Existing studies report widely varying estimates of alumina residence times in the middle atmosphere, ranging from a few years to several decades. This thesis investigates the processes governing alumina particle descent using a one‑dimensional microphysics model that incorporates gravitational sedimentation, eddy diffusion, vertical advection, and Brownian coagulation. Results show that Brownian coagulation produces the largest reduction in residence time for both nano- and micro-scale particle populations, as coagulation shifts particles into larger size ranges where gravitational settling becomes substantially more efficient. Including atmospheric transport and microphysical processes reduces the mean residence time of a nanoscale alumina distribution from approximately 100 years under sedimentation alone to roughly 11 years. These findings highlight the dominant role of atmospheric processes in controlling alumina persistence and provide a computationally efficient framework for future studies of satellite re‑entry by‑products.
...
The rapid growth of satellite constellations is increasing the frequency of atmospheric re‑entries, raising concerns about the environmental impacts of ablation by‑products such as alumina (Al2O3). Existing studies report widely varying estimates of alumina residence times in the middle atmosphere, ranging from a few years to several decades. This thesis investigates the processes governing alumina particle descent using a one‑dimensional microphysics model that incorporates gravitational sedimentation, eddy diffusion, vertical advection, and Brownian coagulation. Results show that Brownian coagulation produces the largest reduction in residence time for both nano- and micro-scale particle populations, as coagulation shifts particles into larger size ranges where gravitational settling becomes substantially more efficient. Including atmospheric transport and microphysical processes reduces the mean residence time of a nanoscale alumina distribution from approximately 100 years under sedimentation alone to roughly 11 years. These findings highlight the dominant role of atmospheric processes in controlling alumina persistence and provide a computationally efficient framework for future studies of satellite re‑entry by‑products.
This thesis addresses the critical challenge of extending the brief payload rendezvous window for Momentum Exchange with Electrodynamic Reboost (MXER) tether systems, a transformative technology for propellantless space transportation. The research systematically conducted a comparative analysis of three distinct actuator configurations using a two-dimensional rigid-body dynamical model, and evaluated a conventional optimal control method against a modern, model-free Reinforcement Learning (RL) algorithm.
The investigation definitively identifies the reeler actuator configuration as the most effective for extending the rendezvous window in an unconstrained dynamic environment. This configuration, which incorporates an intermediate reeling mass, achieved a threefold improvement, extending the uncontrolled rendezvous window of 0.6 seconds to 1.8 seconds. This duration, achieved within specified trajectory tracking tolerances of 10 m for position and 10 m/s for velocity relative to the payload, significantly outperformed both the baseline tip-reeling (0.8 s) and climber (1.0 s) configurations. This superior performance is primarily attributed to the reeler's enhanced control authority over the tether tip's velocity profile, enabling more effective counteraction of the characteristic V-shaped relative velocity curve inherent to rendezvous.
In the unconstrained scenario, both the conventional iterative Linear Quadratic Regulator (iLQR) and the model-free Soft Actor-Critic (SAC) RL agent successfully developed control policies, matching the 1.8-second rendezvous window extension. However, the SAC agent's policy exhibited less smooth, sporadic actuator usage, a trait undesirable in practical applications due to potential structural loads, component wear, and the excitation of unmodelled high-frequency wave dynamics.
The study of constrained control revealed the inherent difficulty of the problem. When realistic operational limits on tether tension, g-loads, and actuator usage were imposed, neither the Augmented-Lagrangian iLQR (AL-iLQR) nor the SAC-based controller could achieve a sustained rendezvous window. The AL-iLQR proved overly conservative, satisfying constraints but failing to exploit the system's full dynamic potential. Conversely, the SAC agent, guided by a simple penalty-based reward function, did not robustly enforce critical constraints, notably violating tension requirements, which would lead to system failure.
Verification and validation studies confirmed the fidelity of the rigid-body model. A variance-based sensitivity analysis highlighted tether length uncertainty as the dominant factor affecting rendezvous accuracy. Additionally, a comprehensive hyperparameter optimisation study for the SAC RL agent identified the learning rate and batch size as highly influential parameters for performance. A brief generalisation test also showed that the RL agent, trained on the reeler configuration, did not successfully generalise to the climber configuration, though its velocity control performance indicated potential for improvement.
Ultimately, this thesis successfully addressed its primary research questions, demonstrating how actuator configuration influences rendezvous window controllability and affirming RL's potential, albeit with current limitations concerning constraint satisfaction and control smoothness. All project goals, from model derivation and iLQR implementation to the deployment and evaluation of the SAC RL algorithm, were addressed, laying foundational groundwork for future advancements in MXER tether control. ...
The investigation definitively identifies the reeler actuator configuration as the most effective for extending the rendezvous window in an unconstrained dynamic environment. This configuration, which incorporates an intermediate reeling mass, achieved a threefold improvement, extending the uncontrolled rendezvous window of 0.6 seconds to 1.8 seconds. This duration, achieved within specified trajectory tracking tolerances of 10 m for position and 10 m/s for velocity relative to the payload, significantly outperformed both the baseline tip-reeling (0.8 s) and climber (1.0 s) configurations. This superior performance is primarily attributed to the reeler's enhanced control authority over the tether tip's velocity profile, enabling more effective counteraction of the characteristic V-shaped relative velocity curve inherent to rendezvous.
In the unconstrained scenario, both the conventional iterative Linear Quadratic Regulator (iLQR) and the model-free Soft Actor-Critic (SAC) RL agent successfully developed control policies, matching the 1.8-second rendezvous window extension. However, the SAC agent's policy exhibited less smooth, sporadic actuator usage, a trait undesirable in practical applications due to potential structural loads, component wear, and the excitation of unmodelled high-frequency wave dynamics.
The study of constrained control revealed the inherent difficulty of the problem. When realistic operational limits on tether tension, g-loads, and actuator usage were imposed, neither the Augmented-Lagrangian iLQR (AL-iLQR) nor the SAC-based controller could achieve a sustained rendezvous window. The AL-iLQR proved overly conservative, satisfying constraints but failing to exploit the system's full dynamic potential. Conversely, the SAC agent, guided by a simple penalty-based reward function, did not robustly enforce critical constraints, notably violating tension requirements, which would lead to system failure.
Verification and validation studies confirmed the fidelity of the rigid-body model. A variance-based sensitivity analysis highlighted tether length uncertainty as the dominant factor affecting rendezvous accuracy. Additionally, a comprehensive hyperparameter optimisation study for the SAC RL agent identified the learning rate and batch size as highly influential parameters for performance. A brief generalisation test also showed that the RL agent, trained on the reeler configuration, did not successfully generalise to the climber configuration, though its velocity control performance indicated potential for improvement.
Ultimately, this thesis successfully addressed its primary research questions, demonstrating how actuator configuration influences rendezvous window controllability and affirming RL's potential, albeit with current limitations concerning constraint satisfaction and control smoothness. All project goals, from model derivation and iLQR implementation to the deployment and evaluation of the SAC RL algorithm, were addressed, laying foundational groundwork for future advancements in MXER tether control. ...
This thesis addresses the critical challenge of extending the brief payload rendezvous window for Momentum Exchange with Electrodynamic Reboost (MXER) tether systems, a transformative technology for propellantless space transportation. The research systematically conducted a comparative analysis of three distinct actuator configurations using a two-dimensional rigid-body dynamical model, and evaluated a conventional optimal control method against a modern, model-free Reinforcement Learning (RL) algorithm.
The investigation definitively identifies the reeler actuator configuration as the most effective for extending the rendezvous window in an unconstrained dynamic environment. This configuration, which incorporates an intermediate reeling mass, achieved a threefold improvement, extending the uncontrolled rendezvous window of 0.6 seconds to 1.8 seconds. This duration, achieved within specified trajectory tracking tolerances of 10 m for position and 10 m/s for velocity relative to the payload, significantly outperformed both the baseline tip-reeling (0.8 s) and climber (1.0 s) configurations. This superior performance is primarily attributed to the reeler's enhanced control authority over the tether tip's velocity profile, enabling more effective counteraction of the characteristic V-shaped relative velocity curve inherent to rendezvous.
In the unconstrained scenario, both the conventional iterative Linear Quadratic Regulator (iLQR) and the model-free Soft Actor-Critic (SAC) RL agent successfully developed control policies, matching the 1.8-second rendezvous window extension. However, the SAC agent's policy exhibited less smooth, sporadic actuator usage, a trait undesirable in practical applications due to potential structural loads, component wear, and the excitation of unmodelled high-frequency wave dynamics.
The study of constrained control revealed the inherent difficulty of the problem. When realistic operational limits on tether tension, g-loads, and actuator usage were imposed, neither the Augmented-Lagrangian iLQR (AL-iLQR) nor the SAC-based controller could achieve a sustained rendezvous window. The AL-iLQR proved overly conservative, satisfying constraints but failing to exploit the system's full dynamic potential. Conversely, the SAC agent, guided by a simple penalty-based reward function, did not robustly enforce critical constraints, notably violating tension requirements, which would lead to system failure.
Verification and validation studies confirmed the fidelity of the rigid-body model. A variance-based sensitivity analysis highlighted tether length uncertainty as the dominant factor affecting rendezvous accuracy. Additionally, a comprehensive hyperparameter optimisation study for the SAC RL agent identified the learning rate and batch size as highly influential parameters for performance. A brief generalisation test also showed that the RL agent, trained on the reeler configuration, did not successfully generalise to the climber configuration, though its velocity control performance indicated potential for improvement.
Ultimately, this thesis successfully addressed its primary research questions, demonstrating how actuator configuration influences rendezvous window controllability and affirming RL's potential, albeit with current limitations concerning constraint satisfaction and control smoothness. All project goals, from model derivation and iLQR implementation to the deployment and evaluation of the SAC RL algorithm, were addressed, laying foundational groundwork for future advancements in MXER tether control.
The investigation definitively identifies the reeler actuator configuration as the most effective for extending the rendezvous window in an unconstrained dynamic environment. This configuration, which incorporates an intermediate reeling mass, achieved a threefold improvement, extending the uncontrolled rendezvous window of 0.6 seconds to 1.8 seconds. This duration, achieved within specified trajectory tracking tolerances of 10 m for position and 10 m/s for velocity relative to the payload, significantly outperformed both the baseline tip-reeling (0.8 s) and climber (1.0 s) configurations. This superior performance is primarily attributed to the reeler's enhanced control authority over the tether tip's velocity profile, enabling more effective counteraction of the characteristic V-shaped relative velocity curve inherent to rendezvous.
In the unconstrained scenario, both the conventional iterative Linear Quadratic Regulator (iLQR) and the model-free Soft Actor-Critic (SAC) RL agent successfully developed control policies, matching the 1.8-second rendezvous window extension. However, the SAC agent's policy exhibited less smooth, sporadic actuator usage, a trait undesirable in practical applications due to potential structural loads, component wear, and the excitation of unmodelled high-frequency wave dynamics.
The study of constrained control revealed the inherent difficulty of the problem. When realistic operational limits on tether tension, g-loads, and actuator usage were imposed, neither the Augmented-Lagrangian iLQR (AL-iLQR) nor the SAC-based controller could achieve a sustained rendezvous window. The AL-iLQR proved overly conservative, satisfying constraints but failing to exploit the system's full dynamic potential. Conversely, the SAC agent, guided by a simple penalty-based reward function, did not robustly enforce critical constraints, notably violating tension requirements, which would lead to system failure.
Verification and validation studies confirmed the fidelity of the rigid-body model. A variance-based sensitivity analysis highlighted tether length uncertainty as the dominant factor affecting rendezvous accuracy. Additionally, a comprehensive hyperparameter optimisation study for the SAC RL agent identified the learning rate and batch size as highly influential parameters for performance. A brief generalisation test also showed that the RL agent, trained on the reeler configuration, did not successfully generalise to the climber configuration, though its velocity control performance indicated potential for improvement.
Ultimately, this thesis successfully addressed its primary research questions, demonstrating how actuator configuration influences rendezvous window controllability and affirming RL's potential, albeit with current limitations concerning constraint satisfaction and control smoothness. All project goals, from model derivation and iLQR implementation to the deployment and evaluation of the SAC RL algorithm, were addressed, laying foundational groundwork for future advancements in MXER tether control.
Extending Satellite Lifetime in VLEO
Aerodynamic Optimization and Re-entry System Design for a Nano-Satellite
Very Low Earth Orbit (VLEO) has recently emerged as a promising operational regime, offering benefits such as reduced communication latency and improved imaging resolution. However, the elevated atmospheric density at these altitudes generates substantial aerodynamic drag, severely limiting satellite lifetime without active propulsion or drag-reduction measures. In parallel, there is growing cross-sector interest in recovering small satellites or their payloads to enable hardware reuse, reduce mission costs, and expand the scope of in-orbit experimentation.
This thesis addresses the coupled challenge of extending nano-satellite lifetime in VLEO while enabling controlled end-of-life re-entry and intact recovery. Specifically, it focuses on the aerodynamic optimization of satellite geometry for drag reduction during the orbital phase, coupled with the design of a deployable or inflatable re-entry system to ensure both thermal protection and post-reentry retrieval capability.
The heatshield design phase employed a full-factorial grid search over key geometric parameters, applied to both inflatable and deployable concepts. This stage integrated two dedicated models: an entry trajectory analysis to evaluate thermal survivability and stability, and a heatshield mass estimation model to estimate mass and center-of-gravity values. The resulting feasible design space provided the basis for subsequent aerodynamic optimization. For each surviving configuration, a custom Python-based free-molecular panel method using a Cercignani-Lampis-Lord (CLL) gas-surface interaction model was coupled with a multi-objective optimizer to minimize drag while also minimizing satellite length. The optimized designs were then evaluated using an orbital lifetime estimation model under different solar activity levels to quantify performance gains. Verification and validation were performed through cross-tool agreement with ADBSat and DSMC/DS2V for selected cases, and benchmarking against published LOFTID and ADEPT data where applicable.
From the 880 satellite-heatshield configurations evaluated, only 10 (4 inflatable, 6 deployable) met all mass, geometry, thermal, and stability constraints under uncertainty. Lifetime analysis revealed that optimized nosecones could extend orbital lifetime by up to 20%, while variations in solar activity could alter lifetime by up to a factor of three, highlighting the dominant influence of environmental conditions. Although the deployable concept was aerothermally viable, its packed configuration restricted solar array area to the point where even the most favorable geometry produced only 8W of power-an infeasible level for most nano-satellite missions. The final design therefore adopted an inflatable heatshield with the lowest achieved drag, paired with a shuttle-type solar panel layout to improve stability and further reduce drag compared to a feather configuration, at the expense of a small power reduction. Additionally, the need for aerodynamic control during re-entry was identified to reduce the landing footprint to a feasible size.
These results demonstrate the feasibility of integrating aerodynamic optimization with re-entry system design to meet both lifetime extension and safe recovery objectives for nano-satellites in VLEO. While the optimized inflatable configuration achieved significant lifetime gains and robust thermal protection, further work is required to implement aerodynamic control for footprint reduction and to experimentally validate the choice of GSI parameters.
Overall, the work demonstrates a coherent design pathway to reconcile drag minimization in orbit with high-drag requirements at re-entry, and provides a validated, computationally efficient framework for early-phase VLEO mission design with recovery capability. ...
This thesis addresses the coupled challenge of extending nano-satellite lifetime in VLEO while enabling controlled end-of-life re-entry and intact recovery. Specifically, it focuses on the aerodynamic optimization of satellite geometry for drag reduction during the orbital phase, coupled with the design of a deployable or inflatable re-entry system to ensure both thermal protection and post-reentry retrieval capability.
The heatshield design phase employed a full-factorial grid search over key geometric parameters, applied to both inflatable and deployable concepts. This stage integrated two dedicated models: an entry trajectory analysis to evaluate thermal survivability and stability, and a heatshield mass estimation model to estimate mass and center-of-gravity values. The resulting feasible design space provided the basis for subsequent aerodynamic optimization. For each surviving configuration, a custom Python-based free-molecular panel method using a Cercignani-Lampis-Lord (CLL) gas-surface interaction model was coupled with a multi-objective optimizer to minimize drag while also minimizing satellite length. The optimized designs were then evaluated using an orbital lifetime estimation model under different solar activity levels to quantify performance gains. Verification and validation were performed through cross-tool agreement with ADBSat and DSMC/DS2V for selected cases, and benchmarking against published LOFTID and ADEPT data where applicable.
From the 880 satellite-heatshield configurations evaluated, only 10 (4 inflatable, 6 deployable) met all mass, geometry, thermal, and stability constraints under uncertainty. Lifetime analysis revealed that optimized nosecones could extend orbital lifetime by up to 20%, while variations in solar activity could alter lifetime by up to a factor of three, highlighting the dominant influence of environmental conditions. Although the deployable concept was aerothermally viable, its packed configuration restricted solar array area to the point where even the most favorable geometry produced only 8W of power-an infeasible level for most nano-satellite missions. The final design therefore adopted an inflatable heatshield with the lowest achieved drag, paired with a shuttle-type solar panel layout to improve stability and further reduce drag compared to a feather configuration, at the expense of a small power reduction. Additionally, the need for aerodynamic control during re-entry was identified to reduce the landing footprint to a feasible size.
These results demonstrate the feasibility of integrating aerodynamic optimization with re-entry system design to meet both lifetime extension and safe recovery objectives for nano-satellites in VLEO. While the optimized inflatable configuration achieved significant lifetime gains and robust thermal protection, further work is required to implement aerodynamic control for footprint reduction and to experimentally validate the choice of GSI parameters.
Overall, the work demonstrates a coherent design pathway to reconcile drag minimization in orbit with high-drag requirements at re-entry, and provides a validated, computationally efficient framework for early-phase VLEO mission design with recovery capability. ...
Very Low Earth Orbit (VLEO) has recently emerged as a promising operational regime, offering benefits such as reduced communication latency and improved imaging resolution. However, the elevated atmospheric density at these altitudes generates substantial aerodynamic drag, severely limiting satellite lifetime without active propulsion or drag-reduction measures. In parallel, there is growing cross-sector interest in recovering small satellites or their payloads to enable hardware reuse, reduce mission costs, and expand the scope of in-orbit experimentation.
This thesis addresses the coupled challenge of extending nano-satellite lifetime in VLEO while enabling controlled end-of-life re-entry and intact recovery. Specifically, it focuses on the aerodynamic optimization of satellite geometry for drag reduction during the orbital phase, coupled with the design of a deployable or inflatable re-entry system to ensure both thermal protection and post-reentry retrieval capability.
The heatshield design phase employed a full-factorial grid search over key geometric parameters, applied to both inflatable and deployable concepts. This stage integrated two dedicated models: an entry trajectory analysis to evaluate thermal survivability and stability, and a heatshield mass estimation model to estimate mass and center-of-gravity values. The resulting feasible design space provided the basis for subsequent aerodynamic optimization. For each surviving configuration, a custom Python-based free-molecular panel method using a Cercignani-Lampis-Lord (CLL) gas-surface interaction model was coupled with a multi-objective optimizer to minimize drag while also minimizing satellite length. The optimized designs were then evaluated using an orbital lifetime estimation model under different solar activity levels to quantify performance gains. Verification and validation were performed through cross-tool agreement with ADBSat and DSMC/DS2V for selected cases, and benchmarking against published LOFTID and ADEPT data where applicable.
From the 880 satellite-heatshield configurations evaluated, only 10 (4 inflatable, 6 deployable) met all mass, geometry, thermal, and stability constraints under uncertainty. Lifetime analysis revealed that optimized nosecones could extend orbital lifetime by up to 20%, while variations in solar activity could alter lifetime by up to a factor of three, highlighting the dominant influence of environmental conditions. Although the deployable concept was aerothermally viable, its packed configuration restricted solar array area to the point where even the most favorable geometry produced only 8W of power-an infeasible level for most nano-satellite missions. The final design therefore adopted an inflatable heatshield with the lowest achieved drag, paired with a shuttle-type solar panel layout to improve stability and further reduce drag compared to a feather configuration, at the expense of a small power reduction. Additionally, the need for aerodynamic control during re-entry was identified to reduce the landing footprint to a feasible size.
These results demonstrate the feasibility of integrating aerodynamic optimization with re-entry system design to meet both lifetime extension and safe recovery objectives for nano-satellites in VLEO. While the optimized inflatable configuration achieved significant lifetime gains and robust thermal protection, further work is required to implement aerodynamic control for footprint reduction and to experimentally validate the choice of GSI parameters.
Overall, the work demonstrates a coherent design pathway to reconcile drag minimization in orbit with high-drag requirements at re-entry, and provides a validated, computationally efficient framework for early-phase VLEO mission design with recovery capability.
This thesis addresses the coupled challenge of extending nano-satellite lifetime in VLEO while enabling controlled end-of-life re-entry and intact recovery. Specifically, it focuses on the aerodynamic optimization of satellite geometry for drag reduction during the orbital phase, coupled with the design of a deployable or inflatable re-entry system to ensure both thermal protection and post-reentry retrieval capability.
The heatshield design phase employed a full-factorial grid search over key geometric parameters, applied to both inflatable and deployable concepts. This stage integrated two dedicated models: an entry trajectory analysis to evaluate thermal survivability and stability, and a heatshield mass estimation model to estimate mass and center-of-gravity values. The resulting feasible design space provided the basis for subsequent aerodynamic optimization. For each surviving configuration, a custom Python-based free-molecular panel method using a Cercignani-Lampis-Lord (CLL) gas-surface interaction model was coupled with a multi-objective optimizer to minimize drag while also minimizing satellite length. The optimized designs were then evaluated using an orbital lifetime estimation model under different solar activity levels to quantify performance gains. Verification and validation were performed through cross-tool agreement with ADBSat and DSMC/DS2V for selected cases, and benchmarking against published LOFTID and ADEPT data where applicable.
From the 880 satellite-heatshield configurations evaluated, only 10 (4 inflatable, 6 deployable) met all mass, geometry, thermal, and stability constraints under uncertainty. Lifetime analysis revealed that optimized nosecones could extend orbital lifetime by up to 20%, while variations in solar activity could alter lifetime by up to a factor of three, highlighting the dominant influence of environmental conditions. Although the deployable concept was aerothermally viable, its packed configuration restricted solar array area to the point where even the most favorable geometry produced only 8W of power-an infeasible level for most nano-satellite missions. The final design therefore adopted an inflatable heatshield with the lowest achieved drag, paired with a shuttle-type solar panel layout to improve stability and further reduce drag compared to a feather configuration, at the expense of a small power reduction. Additionally, the need for aerodynamic control during re-entry was identified to reduce the landing footprint to a feasible size.
These results demonstrate the feasibility of integrating aerodynamic optimization with re-entry system design to meet both lifetime extension and safe recovery objectives for nano-satellites in VLEO. While the optimized inflatable configuration achieved significant lifetime gains and robust thermal protection, further work is required to implement aerodynamic control for footprint reduction and to experimentally validate the choice of GSI parameters.
Overall, the work demonstrates a coherent design pathway to reconcile drag minimization in orbit with high-drag requirements at re-entry, and provides a validated, computationally efficient framework for early-phase VLEO mission design with recovery capability.
Bachelor thesis
(2025)
-
P. Kelley, L.Y. Zhang, C. Carrasco Requejo, M. Benedetti, D. Angelopoulos, L. CALDERON, D. Székedi, G. Konstantinou, J. Sarasate Gutierrez, J. Żuchowski, S.A. Abbas, M.C. Naeije, A.S. Secchiaroli, A. Aditya
The Thermodiver project aims to develop a reusable space tug using Air-Breathing Electric Propulsion (ABEP), which collects atmospheric particles from the upper atmosphere to enable in-situ refuelling and extended mission durations. This propulsion method supports a range of missions, including satellite servicing, orbital transfers, and reducing orbital congestion from space debris. A key part of the design process was the use of ARISS (Air-breathing Refuelling Iterative System Solver), a solver developed by the team that encompasses the initial design, sizing, budgeting, and performance assessment of the air-breathing spacecraft. ARISS integrates various simulation tools: Direct Simulation Monte Carlo (DSMC) for intake and aerodynamic performance, the General Mission Analysis Tool (GMAT) for mission planning, and Finite Element Methods (FEM) for thermal and structural analysis. The project produced a detailed 6.2-meter-long, 800 kg spacecraft design with a parabolic intake for efficiency and the bus shielded in its wake. Capable of de-orbiting satellites such as Proba-V and refuelling within five months for continued missions, the system is being further refined in the final stages of development. Ultimately, the Thermodiver offers a sustainable solution to orbital congestion and a versatile platform for affordable in-space transportation and future interplanetary exploration.
...
The Thermodiver project aims to develop a reusable space tug using Air-Breathing Electric Propulsion (ABEP), which collects atmospheric particles from the upper atmosphere to enable in-situ refuelling and extended mission durations. This propulsion method supports a range of missions, including satellite servicing, orbital transfers, and reducing orbital congestion from space debris. A key part of the design process was the use of ARISS (Air-breathing Refuelling Iterative System Solver), a solver developed by the team that encompasses the initial design, sizing, budgeting, and performance assessment of the air-breathing spacecraft. ARISS integrates various simulation tools: Direct Simulation Monte Carlo (DSMC) for intake and aerodynamic performance, the General Mission Analysis Tool (GMAT) for mission planning, and Finite Element Methods (FEM) for thermal and structural analysis. The project produced a detailed 6.2-meter-long, 800 kg spacecraft design with a parabolic intake for efficiency and the bus shielded in its wake. Capable of de-orbiting satellites such as Proba-V and refuelling within five months for continued missions, the system is being further refined in the final stages of development. Ultimately, the Thermodiver offers a sustainable solution to orbital congestion and a versatile platform for affordable in-space transportation and future interplanetary exploration.
Mars entry guidance faces a critical challenge: navigating hypersonic velocities within the thin atmospheric layers (120 – 45 𝑘𝑚 altitude) while balancing conflicting objectives of precision targeting, thermal survival, and structural integrity. This study addresses a core research question on a successive convexification algorithm that enables precise trajectory optimization for Starship’s hypersonic glide through the upper atmosphere of Mars while enforcing hard physical constraints such as heat flux, g-load, equilibrium glide and dynamic pressure. By formulating such successive convexification - based framework, the inherently non-convex entry problem is decomposed into iteratively refined convex sub-problems, enabling computational tractability under Mars’ variable CO₂-rich atmosphere. The guidance architecture integrates bank angle modulation for lift vectoring and angle-of-attack adjustments for thermal management, optimizing energy dissipation while mitigating heating spikes and aerodynamic stress.
Simulations demonstrate that the collocation discretization strategy used ensures trajectory adherence within the entry corridor, achieving terminal positioning errors below 3 𝑘𝑚 at 45 𝑘𝑚 altitude. The algorithm’s robustness is validated under ±10% dispersions in initial velocity (4.3 𝑘𝑚/𝑠) and flight-path angle (−15°) from a parking orbit around the planet, with heat flux, dynamic pressure, and g-load profiles remaining within mission-critical limits. Sensitivity analyses reveal that atmospheric density uncertainties induce predictable deviations compensated by rapid convex optimizations. These results align and improve on previous NASA mission data.
The study bridges theoretical convex optimization with operational reality, demonstrating that modern computational guidance outperforms legacy predictor-corrector methods in handling nonlinear dynamics and path constraints. By extending the convex framework with adaptive trust regions and sequential convex programming, the proposed method reduces terminal errors by 40% compared to state-of-the-art approaches (Mars 2020). This advancement not only enhances Starship’s capability to deliver crewed and cargo payloads to predefined Martian coordinates but also establishes a foundation for integrating the hypersonic glide phase with the subsequent powered descent phases. As humanity strides toward sustained Mars exploration, this work underscores the viability of successive convexification as a paradigm for achieving precise atmospheric glide through the Martian atmosphere. ...
Simulations demonstrate that the collocation discretization strategy used ensures trajectory adherence within the entry corridor, achieving terminal positioning errors below 3 𝑘𝑚 at 45 𝑘𝑚 altitude. The algorithm’s robustness is validated under ±10% dispersions in initial velocity (4.3 𝑘𝑚/𝑠) and flight-path angle (−15°) from a parking orbit around the planet, with heat flux, dynamic pressure, and g-load profiles remaining within mission-critical limits. Sensitivity analyses reveal that atmospheric density uncertainties induce predictable deviations compensated by rapid convex optimizations. These results align and improve on previous NASA mission data.
The study bridges theoretical convex optimization with operational reality, demonstrating that modern computational guidance outperforms legacy predictor-corrector methods in handling nonlinear dynamics and path constraints. By extending the convex framework with adaptive trust regions and sequential convex programming, the proposed method reduces terminal errors by 40% compared to state-of-the-art approaches (Mars 2020). This advancement not only enhances Starship’s capability to deliver crewed and cargo payloads to predefined Martian coordinates but also establishes a foundation for integrating the hypersonic glide phase with the subsequent powered descent phases. As humanity strides toward sustained Mars exploration, this work underscores the viability of successive convexification as a paradigm for achieving precise atmospheric glide through the Martian atmosphere. ...
Mars entry guidance faces a critical challenge: navigating hypersonic velocities within the thin atmospheric layers (120 – 45 𝑘𝑚 altitude) while balancing conflicting objectives of precision targeting, thermal survival, and structural integrity. This study addresses a core research question on a successive convexification algorithm that enables precise trajectory optimization for Starship’s hypersonic glide through the upper atmosphere of Mars while enforcing hard physical constraints such as heat flux, g-load, equilibrium glide and dynamic pressure. By formulating such successive convexification - based framework, the inherently non-convex entry problem is decomposed into iteratively refined convex sub-problems, enabling computational tractability under Mars’ variable CO₂-rich atmosphere. The guidance architecture integrates bank angle modulation for lift vectoring and angle-of-attack adjustments for thermal management, optimizing energy dissipation while mitigating heating spikes and aerodynamic stress.
Simulations demonstrate that the collocation discretization strategy used ensures trajectory adherence within the entry corridor, achieving terminal positioning errors below 3 𝑘𝑚 at 45 𝑘𝑚 altitude. The algorithm’s robustness is validated under ±10% dispersions in initial velocity (4.3 𝑘𝑚/𝑠) and flight-path angle (−15°) from a parking orbit around the planet, with heat flux, dynamic pressure, and g-load profiles remaining within mission-critical limits. Sensitivity analyses reveal that atmospheric density uncertainties induce predictable deviations compensated by rapid convex optimizations. These results align and improve on previous NASA mission data.
The study bridges theoretical convex optimization with operational reality, demonstrating that modern computational guidance outperforms legacy predictor-corrector methods in handling nonlinear dynamics and path constraints. By extending the convex framework with adaptive trust regions and sequential convex programming, the proposed method reduces terminal errors by 40% compared to state-of-the-art approaches (Mars 2020). This advancement not only enhances Starship’s capability to deliver crewed and cargo payloads to predefined Martian coordinates but also establishes a foundation for integrating the hypersonic glide phase with the subsequent powered descent phases. As humanity strides toward sustained Mars exploration, this work underscores the viability of successive convexification as a paradigm for achieving precise atmospheric glide through the Martian atmosphere.
Simulations demonstrate that the collocation discretization strategy used ensures trajectory adherence within the entry corridor, achieving terminal positioning errors below 3 𝑘𝑚 at 45 𝑘𝑚 altitude. The algorithm’s robustness is validated under ±10% dispersions in initial velocity (4.3 𝑘𝑚/𝑠) and flight-path angle (−15°) from a parking orbit around the planet, with heat flux, dynamic pressure, and g-load profiles remaining within mission-critical limits. Sensitivity analyses reveal that atmospheric density uncertainties induce predictable deviations compensated by rapid convex optimizations. These results align and improve on previous NASA mission data.
The study bridges theoretical convex optimization with operational reality, demonstrating that modern computational guidance outperforms legacy predictor-corrector methods in handling nonlinear dynamics and path constraints. By extending the convex framework with adaptive trust regions and sequential convex programming, the proposed method reduces terminal errors by 40% compared to state-of-the-art approaches (Mars 2020). This advancement not only enhances Starship’s capability to deliver crewed and cargo payloads to predefined Martian coordinates but also establishes a foundation for integrating the hypersonic glide phase with the subsequent powered descent phases. As humanity strides toward sustained Mars exploration, this work underscores the viability of successive convexification as a paradigm for achieving precise atmospheric glide through the Martian atmosphere.
Electron Rocket Launch and Recovery Performance Analysis
Development of Simulink Simulation Tools for Ascent and Descent Trajectories with Parafoil and Unguided Circular Parachute Recovery Mechanisms
Rocket Lab has recovered the first stage of its dedicated small satellite launcher Electron twice by means of splashdown. The ability to reuse the entire first stage can reduce production cost at the current launch frequency and enable a higher launch rate. To lower the launch cost further, Rocket Lab is developing Mid-Air Recovery system for the first stage of the Electron. The research objective of this thesis is to develop a simulation tool for analyzing the launch performance of the Electron rocket and its first-stage recovery, involving mid-air recovery using an air-ram parachute and splashdown recovery using an unguided circular parachute. For the ascent phase, a two-stage rocket Simulink model is developed that implements quaternion representation of six-degrees-of freedom (6-DOF) equations of motion for custom variable mass in the Earth-centered Earth-fixed (ECEF) reference frame to simulate ascent trajectories. Another Simulink model is created to simulate the ballistic trajectory of the first stage after separation using quaternion representation of 6-DOF equations of motion for constant mass in the ECEF reference frame, particularly when first stage recovery is not implemented. It was found that he additional mass, which is required to transform the Electron’s first stage into a reusable one is 100 kg, with the parafoil accounting for 74 kg, the ballute 9 kg, and the remaining mass attributed to modifications such as the parachute attachment and release mechanism and for the circular parachute 71.5 kg with 22.8 kg for canopy, 35.3 kg for suspension lines and 13.4 kg actuators. The parafoil provides superior control, whereas the circular parachute is lighter and more straightforward in design. Consequently, if refurbishment costs are identical for both, the circular parachute emerges as the more advantageous choice.
...
Rocket Lab has recovered the first stage of its dedicated small satellite launcher Electron twice by means of splashdown. The ability to reuse the entire first stage can reduce production cost at the current launch frequency and enable a higher launch rate. To lower the launch cost further, Rocket Lab is developing Mid-Air Recovery system for the first stage of the Electron. The research objective of this thesis is to develop a simulation tool for analyzing the launch performance of the Electron rocket and its first-stage recovery, involving mid-air recovery using an air-ram parachute and splashdown recovery using an unguided circular parachute. For the ascent phase, a two-stage rocket Simulink model is developed that implements quaternion representation of six-degrees-of freedom (6-DOF) equations of motion for custom variable mass in the Earth-centered Earth-fixed (ECEF) reference frame to simulate ascent trajectories. Another Simulink model is created to simulate the ballistic trajectory of the first stage after separation using quaternion representation of 6-DOF equations of motion for constant mass in the ECEF reference frame, particularly when first stage recovery is not implemented. It was found that he additional mass, which is required to transform the Electron’s first stage into a reusable one is 100 kg, with the parafoil accounting for 74 kg, the ballute 9 kg, and the remaining mass attributed to modifications such as the parachute attachment and release mechanism and for the circular parachute 71.5 kg with 22.8 kg for canopy, 35.3 kg for suspension lines and 13.4 kg actuators. The parafoil provides superior control, whereas the circular parachute is lighter and more straightforward in design. Consequently, if refurbishment costs are identical for both, the circular parachute emerges as the more advantageous choice.
Reinforcement Learning for Hypersonic Glide Vehicle Trajectory Optimization
A Soft Actor-Critic Approach
Hypersonic glide vehicle trajectory optimization requires generating complete reachable footprints for mission planning under strict path constraints. Current methods face limitations including equilibrium glide assumptions, fixed angle of attack profiles and incomplete footprint coverage requiring re-optimization for each target direction. This work presents the first reinforcement learning approach for complete global footprint generation with dual control authority (bank angle and angle of attack) on a rotating spherical Earth model including direct target point guidance and no-fly zone avoidance. The trained policy generates footprints for any location on Earth while incorporating full three-degree-of-freedom dynamics including Coriolis effects and control rate limitations. All trajectories satisfy operational constraints including dynamic pressure, g-load and temperature limits. The policy learns purely from the objective to maximize range in every direction without pre-designed control profiles. Based on the Soft Actor-Critic algorithm, optimal control strategies are learned directly from environment interaction. Large-scale Monte Carlo validation with 100 million randomly sampled trajectories confirms the learned policy discovered the boundaries of the reachable domain, with the reinforcement learning footprint exceeding the Monte Carlo footprint by 5.1% in footprint area. The framework provides precision target guidance to arbitrary global coordinates and allows for no-fly zone avoidance.
...
Hypersonic glide vehicle trajectory optimization requires generating complete reachable footprints for mission planning under strict path constraints. Current methods face limitations including equilibrium glide assumptions, fixed angle of attack profiles and incomplete footprint coverage requiring re-optimization for each target direction. This work presents the first reinforcement learning approach for complete global footprint generation with dual control authority (bank angle and angle of attack) on a rotating spherical Earth model including direct target point guidance and no-fly zone avoidance. The trained policy generates footprints for any location on Earth while incorporating full three-degree-of-freedom dynamics including Coriolis effects and control rate limitations. All trajectories satisfy operational constraints including dynamic pressure, g-load and temperature limits. The policy learns purely from the objective to maximize range in every direction without pre-designed control profiles. Based on the Soft Actor-Critic algorithm, optimal control strategies are learned directly from environment interaction. Large-scale Monte Carlo validation with 100 million randomly sampled trajectories confirms the learned policy discovered the boundaries of the reachable domain, with the reinforcement learning footprint exceeding the Monte Carlo footprint by 5.1% in footprint area. The framework provides precision target guidance to arbitrary global coordinates and allows for no-fly zone avoidance.
With the unprecedented surge in demand for launching Low Earth Orbit (LEO) satellites in the coming years, exploring more sustainable and cost-effective launch methods like the space tether concept has become essential. This study delves into the feasibility of configuring a rotating orbital space tether as an alternative upper stage for LEO launch systems. An orbital space tether, combined with a reusable rocket stage, is proposed for launching 1000 kg payloads to a reference 550 km circular orbit. A problem of excess velocity, identified in the literature for the pickup and toss of payloads to LEO using a rotating tether, is addressed using a unique two-part ballast and a two-part tether system. An integrated optimization framework employing Particle Swarm Optimization is employed to determine an optimal tether configuration. The proposed system is compared to state-of-the-art launch systems using a weighted dry mass metric, indicating significant potential to reduce launch costs. The study suggests that space tether-based upper stages offer a promising alternative to conventional rocket-based systems.
...
With the unprecedented surge in demand for launching Low Earth Orbit (LEO) satellites in the coming years, exploring more sustainable and cost-effective launch methods like the space tether concept has become essential. This study delves into the feasibility of configuring a rotating orbital space tether as an alternative upper stage for LEO launch systems. An orbital space tether, combined with a reusable rocket stage, is proposed for launching 1000 kg payloads to a reference 550 km circular orbit. A problem of excess velocity, identified in the literature for the pickup and toss of payloads to LEO using a rotating tether, is addressed using a unique two-part ballast and a two-part tether system. An integrated optimization framework employing Particle Swarm Optimization is employed to determine an optimal tether configuration. The proposed system is compared to state-of-the-art launch systems using a weighted dry mass metric, indicating significant potential to reduce launch costs. The study suggests that space tether-based upper stages offer a promising alternative to conventional rocket-based systems.
The increasing number of rocket launches has intensified concerns about the environmental effects of spaceflight. Quantifying rocket emissions at different altitudes and locations within the atmosphere is essential to assess these impacts. This requires realistic launcher models, but much of the publicly available technical data is incomplete or inconsistent. This research aimed to develop a semi-automated framework for launcher remodelling that can handle uncertain or missing input parameters while producing reliable and efficient results.
The framework was developed at the German Aerospace Center (DLR) and integrates a flexible input data structure, statistical estimation methods, and trajectory optimisation tools. Three statistical techniques — Monte Carlo, Latin Hypercube, and Approximate Bayesian Computation — were evaluated to estimate unknown parameters and define their valid ranges. Their performance was tested using several expendable launch vehicles with liquid propulsion.
Results showed that Latin Hypercube sampling achieved the best balance between accuracy and computational cost. When applied to real rockets, the framework produced configurations with payload estimates within 2% of reference values, even when up to five input parameters were uncertain.
Overall, the developed framework demonstrates that launcher remodelling can be automated while effectively handling uncertainty. It facilitates the generation of realistic launcher models and supports ongoing efforts to quantify the environmental impact of rocket emissions ...
The framework was developed at the German Aerospace Center (DLR) and integrates a flexible input data structure, statistical estimation methods, and trajectory optimisation tools. Three statistical techniques — Monte Carlo, Latin Hypercube, and Approximate Bayesian Computation — were evaluated to estimate unknown parameters and define their valid ranges. Their performance was tested using several expendable launch vehicles with liquid propulsion.
Results showed that Latin Hypercube sampling achieved the best balance between accuracy and computational cost. When applied to real rockets, the framework produced configurations with payload estimates within 2% of reference values, even when up to five input parameters were uncertain.
Overall, the developed framework demonstrates that launcher remodelling can be automated while effectively handling uncertainty. It facilitates the generation of realistic launcher models and supports ongoing efforts to quantify the environmental impact of rocket emissions ...
The increasing number of rocket launches has intensified concerns about the environmental effects of spaceflight. Quantifying rocket emissions at different altitudes and locations within the atmosphere is essential to assess these impacts. This requires realistic launcher models, but much of the publicly available technical data is incomplete or inconsistent. This research aimed to develop a semi-automated framework for launcher remodelling that can handle uncertain or missing input parameters while producing reliable and efficient results.
The framework was developed at the German Aerospace Center (DLR) and integrates a flexible input data structure, statistical estimation methods, and trajectory optimisation tools. Three statistical techniques — Monte Carlo, Latin Hypercube, and Approximate Bayesian Computation — were evaluated to estimate unknown parameters and define their valid ranges. Their performance was tested using several expendable launch vehicles with liquid propulsion.
Results showed that Latin Hypercube sampling achieved the best balance between accuracy and computational cost. When applied to real rockets, the framework produced configurations with payload estimates within 2% of reference values, even when up to five input parameters were uncertain.
Overall, the developed framework demonstrates that launcher remodelling can be automated while effectively handling uncertainty. It facilitates the generation of realistic launcher models and supports ongoing efforts to quantify the environmental impact of rocket emissions
The framework was developed at the German Aerospace Center (DLR) and integrates a flexible input data structure, statistical estimation methods, and trajectory optimisation tools. Three statistical techniques — Monte Carlo, Latin Hypercube, and Approximate Bayesian Computation — were evaluated to estimate unknown parameters and define their valid ranges. Their performance was tested using several expendable launch vehicles with liquid propulsion.
Results showed that Latin Hypercube sampling achieved the best balance between accuracy and computational cost. When applied to real rockets, the framework produced configurations with payload estimates within 2% of reference values, even when up to five input parameters were uncertain.
Overall, the developed framework demonstrates that launcher remodelling can be automated while effectively handling uncertainty. It facilitates the generation of realistic launcher models and supports ongoing efforts to quantify the environmental impact of rocket emissions
As climate change continues to drive an increase in extreme weather events, the need for reliable predictions and observational data has never been more critical. Oceans, playing a vital role in regulating the Earth’s climate, are central to understanding these changes. Specifically, accurately modeling ocean wave dynamics, i.e., how waves are generated, evolve, and interact with oceanic processes such as currents, helps track ocean circulation and predict future variations. Thanks to Earth Observation satellites, continuous global observations have been available over the past few decades. Satellite altimeters, active sensors utilizing radar’s ranging capabilities, have emerged as pioneers in space oceanography. These instruments measure critical geophysical parameters such as sea surface height, significant wave height and near-surface wind speed along satellite tracks. Recognizing the immense value of these measurements for climate studies and operational activities, continuous technological innovations are essential for optimizing the performance and use of these instruments. One major breakthrough in satellite altimetrywas the incorporation of Synthetic Aperture Radar (SAR) technology in 2010, which enhanced spatial resolution from around 7 km, provided by Low-ResolutionMode sensors, to about 300m. In 2017, the fully-focused coherent processing of pulse echoes was implemented, a concept widely used in SAR imaging, enabling meter-scale resolution. This improvement led to significant benefits in near-coast applications, improving the quality of geophysical parameters by reducing signal contamination from surrounding land features. Additionally, for the first time, offnadir signals, previously considered a nuisance, were exploited to map narrow inland water bodies and detect sea-ice leads and floes. Recognizing this imaging potential led to investigating its capabilities over open oceans. Existing challenges in conventional, or unfocused, SAR altimetry relate to the accuracy of significant wave height estimates, especially when long waves, known as swells, dominate the sea surface. Swell waves, with wavelengths exceeding 150 meters, are often too long to be fully captured within the SAR altimeter’s footprint, leading to noisy, multi-peaked waveforms. Recognizing the interference of swell signals in SAR responses, combined with the high-resolution data provided by fully-focused processing, this dissertation first investigated the feasibility of transforming what was previously considered a nuisance into valuable information about the sea surface. To achieve this, the identification of swell-induced power variations in off-nadir altimeter’s signals, representing the so-called trailing edge of the returned echo, was first confirmed. These patterns were analyzed to compute a fully-focused modulation spectrumderived from altimetry. The modulation spectrumis a commonly used Level-2 product provided by satellites designed to measure the wave field, such as Sentinel-1 and CFOSAT, and allows for the estimation of swell characteristics, including wavelength, direction and wave height. The proposed method involved normalizing the signal intensity and re-projecting the range bins to cross-track ground locations, followed by spectral analysis akin to side-looking SAR systems. The analysis revealed that fully-focused altimetrymodulation spectra display power in all four quadrants due to the inherent 180-degree SAR ambiguity, plus two additional ambiguities caused by inseparable signals received from both sides of the radar footprint. The study also identified the main modulation mechanisms, using as reference numerical and analytical models. Range bunching was found to be a dominant mechanism alongside velocity bunching, with their relative strength highly dependent on the wave propagation angle. Fully-focused altimetry modulation spectra, derived from Cryosat-2, were evaluated through comparisons with buoy-derived directional wave spectra, showing good agreement. Furthermore, applying the proposed technique to Sentinel-6A data demonstrated that exploiting its full-beamfootprint, which is partially truncated onboard for data volume efficiency, improves swell retrieval. This is particularly true for waves propagating in or near the cross-track direction, due to its extended observational window and higher resolution compared to the operational truncated data. Yet, the development of a method to invert modulation-derived spectra to real ocean wave spectra is necessary to reliably use these instruments as a new source for providing operational global swell observations. The dissertation further explored the limitations of SAR altimeters in ocean wave imaging, focusing on resolution loss. This was addressed by estimating the azimuth cutoff wavelength, which serves as a proxy for the shortest detectable waves across different sea states and wave directions. The method used to estimate this parameter involved a SAR imaging technique applied in the spatial domain, minimizing residuals between the along-track autocorrelation function of fully-focused SAR radargrams, representing successive waveforms, and a fitted Gaussian function. Sentinel-6A data were then used to evaluate the method’s performance through comparisons with model-derived values globally. The analysis revealed that the method performs well under the majority of sea states but tends to underestimate values in extreme wind wave conditions. Furthermore, sensitivity to swell presence was observed, leading to pronounced over estimations, with the magnitude of these errors influenced by the swell direction. To mitigate these errors, an alternative approach was developed in the wave number domain. Results revealed an improvement in the correlation between azimuth cutoff estimates and model-derived values by 10%. Given the strong relationship between resolution loss and sea state conditions, the azimuth cutoff was further used to derive a new sea-state parameter: the variance of wave orbital velocities. Wave orbital velocity statistics offer valuable insights into wave climate by isolating wave components associated with developing seas. Comparisons between modeled and estimated wave orbital velocity variances showed similar sensitivities to swell presence and high sea states, suggesting further refinement of the proposed methods. Despite these challenges, the ability to extract these two additional parameters from the radar signal is valuable for identifying sensors capabilities and providing a new geophysical parameter for oceanographic studies. Lastly, the dissertation assessed the impact of wave-current interactions on wave products derived from both models and satellites, focusing on the Agulhas Current region, one of the most dynamic ocean environments. In situ wave measurements, collected during the One Ocean Expedition in 2023, in which the author participated, served as a reference for this study. The study first examined ocean current products. A clear underestimation of surface current velocities exceeding 0.5 m/s was found for both the Mercator operational model and the altimetry-derived Globcurrent product, with Mercator showing greater variability. Next, wave products, both with and without these ocean current products included in their modeling, were validated. The ECMWF reanalysis v5, known as ERA5, consistently underestimated wave heights above 2.5 m, whereas the MFWAM, which is the Global Ocean Wave Analysis and Forecast system from Meteo-France, showed good agreement with in situ data. This discrepancy was attributed to the lack of ocean current forcing in ERA5, underscoring the need for refinement in areas dominated by currents. Customized MFWAM simulations, including and excluding current data, further supported this finding. MFWAM forced with Globcurrent aligned most closely with drifter measurements, outperforming the operational product that uses Mercator currents. Comparisons between satellite altimeter observations and drifters also showed good agreement in significant wave height, with clear evidence of current-induced wave height variations along satellite tracks. Additionally, a multi-mission analysis of swellinduced modulation spectra from Sentinel-1, CFOSAT and SAR altimeters demonstrated alignment with in situ data and between them, highlighting the potential for synergistic use of these instruments in operational oceanography and climate studies.
...
As climate change continues to drive an increase in extreme weather events, the need for reliable predictions and observational data has never been more critical. Oceans, playing a vital role in regulating the Earth’s climate, are central to understanding these changes. Specifically, accurately modeling ocean wave dynamics, i.e., how waves are generated, evolve, and interact with oceanic processes such as currents, helps track ocean circulation and predict future variations. Thanks to Earth Observation satellites, continuous global observations have been available over the past few decades. Satellite altimeters, active sensors utilizing radar’s ranging capabilities, have emerged as pioneers in space oceanography. These instruments measure critical geophysical parameters such as sea surface height, significant wave height and near-surface wind speed along satellite tracks. Recognizing the immense value of these measurements for climate studies and operational activities, continuous technological innovations are essential for optimizing the performance and use of these instruments. One major breakthrough in satellite altimetrywas the incorporation of Synthetic Aperture Radar (SAR) technology in 2010, which enhanced spatial resolution from around 7 km, provided by Low-ResolutionMode sensors, to about 300m. In 2017, the fully-focused coherent processing of pulse echoes was implemented, a concept widely used in SAR imaging, enabling meter-scale resolution. This improvement led to significant benefits in near-coast applications, improving the quality of geophysical parameters by reducing signal contamination from surrounding land features. Additionally, for the first time, offnadir signals, previously considered a nuisance, were exploited to map narrow inland water bodies and detect sea-ice leads and floes. Recognizing this imaging potential led to investigating its capabilities over open oceans. Existing challenges in conventional, or unfocused, SAR altimetry relate to the accuracy of significant wave height estimates, especially when long waves, known as swells, dominate the sea surface. Swell waves, with wavelengths exceeding 150 meters, are often too long to be fully captured within the SAR altimeter’s footprint, leading to noisy, multi-peaked waveforms. Recognizing the interference of swell signals in SAR responses, combined with the high-resolution data provided by fully-focused processing, this dissertation first investigated the feasibility of transforming what was previously considered a nuisance into valuable information about the sea surface. To achieve this, the identification of swell-induced power variations in off-nadir altimeter’s signals, representing the so-called trailing edge of the returned echo, was first confirmed. These patterns were analyzed to compute a fully-focused modulation spectrumderived from altimetry. The modulation spectrumis a commonly used Level-2 product provided by satellites designed to measure the wave field, such as Sentinel-1 and CFOSAT, and allows for the estimation of swell characteristics, including wavelength, direction and wave height. The proposed method involved normalizing the signal intensity and re-projecting the range bins to cross-track ground locations, followed by spectral analysis akin to side-looking SAR systems. The analysis revealed that fully-focused altimetrymodulation spectra display power in all four quadrants due to the inherent 180-degree SAR ambiguity, plus two additional ambiguities caused by inseparable signals received from both sides of the radar footprint. The study also identified the main modulation mechanisms, using as reference numerical and analytical models. Range bunching was found to be a dominant mechanism alongside velocity bunching, with their relative strength highly dependent on the wave propagation angle. Fully-focused altimetry modulation spectra, derived from Cryosat-2, were evaluated through comparisons with buoy-derived directional wave spectra, showing good agreement. Furthermore, applying the proposed technique to Sentinel-6A data demonstrated that exploiting its full-beamfootprint, which is partially truncated onboard for data volume efficiency, improves swell retrieval. This is particularly true for waves propagating in or near the cross-track direction, due to its extended observational window and higher resolution compared to the operational truncated data. Yet, the development of a method to invert modulation-derived spectra to real ocean wave spectra is necessary to reliably use these instruments as a new source for providing operational global swell observations. The dissertation further explored the limitations of SAR altimeters in ocean wave imaging, focusing on resolution loss. This was addressed by estimating the azimuth cutoff wavelength, which serves as a proxy for the shortest detectable waves across different sea states and wave directions. The method used to estimate this parameter involved a SAR imaging technique applied in the spatial domain, minimizing residuals between the along-track autocorrelation function of fully-focused SAR radargrams, representing successive waveforms, and a fitted Gaussian function. Sentinel-6A data were then used to evaluate the method’s performance through comparisons with model-derived values globally. The analysis revealed that the method performs well under the majority of sea states but tends to underestimate values in extreme wind wave conditions. Furthermore, sensitivity to swell presence was observed, leading to pronounced over estimations, with the magnitude of these errors influenced by the swell direction. To mitigate these errors, an alternative approach was developed in the wave number domain. Results revealed an improvement in the correlation between azimuth cutoff estimates and model-derived values by 10%. Given the strong relationship between resolution loss and sea state conditions, the azimuth cutoff was further used to derive a new sea-state parameter: the variance of wave orbital velocities. Wave orbital velocity statistics offer valuable insights into wave climate by isolating wave components associated with developing seas. Comparisons between modeled and estimated wave orbital velocity variances showed similar sensitivities to swell presence and high sea states, suggesting further refinement of the proposed methods. Despite these challenges, the ability to extract these two additional parameters from the radar signal is valuable for identifying sensors capabilities and providing a new geophysical parameter for oceanographic studies. Lastly, the dissertation assessed the impact of wave-current interactions on wave products derived from both models and satellites, focusing on the Agulhas Current region, one of the most dynamic ocean environments. In situ wave measurements, collected during the One Ocean Expedition in 2023, in which the author participated, served as a reference for this study. The study first examined ocean current products. A clear underestimation of surface current velocities exceeding 0.5 m/s was found for both the Mercator operational model and the altimetry-derived Globcurrent product, with Mercator showing greater variability. Next, wave products, both with and without these ocean current products included in their modeling, were validated. The ECMWF reanalysis v5, known as ERA5, consistently underestimated wave heights above 2.5 m, whereas the MFWAM, which is the Global Ocean Wave Analysis and Forecast system from Meteo-France, showed good agreement with in situ data. This discrepancy was attributed to the lack of ocean current forcing in ERA5, underscoring the need for refinement in areas dominated by currents. Customized MFWAM simulations, including and excluding current data, further supported this finding. MFWAM forced with Globcurrent aligned most closely with drifter measurements, outperforming the operational product that uses Mercator currents. Comparisons between satellite altimeter observations and drifters also showed good agreement in significant wave height, with clear evidence of current-induced wave height variations along satellite tracks. Additionally, a multi-mission analysis of swellinduced modulation spectra from Sentinel-1, CFOSAT and SAR altimeters demonstrated alignment with in situ data and between them, highlighting the potential for synergistic use of these instruments in operational oceanography and climate studies.
This thesis addresses the optimization of on-demand satellite servicing in Sun-Synchronous Orbits through dynamic routing and refueling strategies. Traditional satellite operations, limited by finite onboard fuel, lead to premature mission termination. This research focuses on optimizing a refueling infrastructure comprising fuel stations, a service satellite, and client satellites, specifically for SunSynchronous Orbits. The algorithm is designed to efficiently manage fuel resources by determining the most effective servicing routes to minimize mission time and fuel consumption. A multi-objective optimization framework is developed to model these operations, utilizing a custom algorithm that refines transfer trajectories through high-precision numerical simulations that account for perturbations. Robustness analysis demonstrates that the algorithm can effectively handle uncertainties, consistently converging to the same optimal solutions across varying conditions. The developed algorithm efficiently identifies optimal servicing paths and is scalable to more complex missions involving multiple satellites and refueling stations.
...
This thesis addresses the optimization of on-demand satellite servicing in Sun-Synchronous Orbits through dynamic routing and refueling strategies. Traditional satellite operations, limited by finite onboard fuel, lead to premature mission termination. This research focuses on optimizing a refueling infrastructure comprising fuel stations, a service satellite, and client satellites, specifically for SunSynchronous Orbits. The algorithm is designed to efficiently manage fuel resources by determining the most effective servicing routes to minimize mission time and fuel consumption. A multi-objective optimization framework is developed to model these operations, utilizing a custom algorithm that refines transfer trajectories through high-precision numerical simulations that account for perturbations. Robustness analysis demonstrates that the algorithm can effectively handle uncertainties, consistently converging to the same optimal solutions across varying conditions. The developed algorithm efficiently identifies optimal servicing paths and is scalable to more complex missions involving multiple satellites and refueling stations.
Breaking Barriers
Revolutionizing communication with laser tech aboard high-altitude pseudo-satellite (haps) platforms
Bachelor thesis
(2024)
-
T.A. Stomp, A.H.A. Groen, C. Margaritis, T.A.F. Veerman, M.J.C. Oortwijn, A.M.G. Poirier, J.A. Lans, M. Rancaño Martinez, J.A.J. Lapidaire, M. Ebrahimi, M.C. Naeije, T.C. Darikwa, J.A.W. Poland
Final Report Design Synthesis Exercise Spring 2024. AE3200: Design Synthesis Exercise. DSE Group 22
...
Final Report Design Synthesis Exercise Spring 2024. AE3200: Design Synthesis Exercise. DSE Group 22
Analyzing the Impact of Earth-Sun Distance Variations on Global Temperature
A Comparison of Simplified Solar System Models
This thesis investigates how variations in the Earth-Sun distance influence global temperatures, by comparing a simplified model of the solar system with an existing paper from V.V. Zharkova, claiming that increasing temperatures can be explained naturally. Over a 5000-year period, numerical simulations including planetary gravitational influences, solar inertial motion, and Milankovitch cycles, this study looks at distance variations and Earth hemispheric differences in solar intensity due to albedo differences, to asses this statement. The result shows that while orbital mechanics influence the global temperature, Their role is minimal. It should see a slight decrease in temperature, and thus V.V. Zharkova’s research does not represent the actual situation. This offers valuable insight into the relationship between the Earth's orbital mechanics and climate. However, further research into the accuracy of the model is required.
...
This thesis investigates how variations in the Earth-Sun distance influence global temperatures, by comparing a simplified model of the solar system with an existing paper from V.V. Zharkova, claiming that increasing temperatures can be explained naturally. Over a 5000-year period, numerical simulations including planetary gravitational influences, solar inertial motion, and Milankovitch cycles, this study looks at distance variations and Earth hemispheric differences in solar intensity due to albedo differences, to asses this statement. The result shows that while orbital mechanics influence the global temperature, Their role is minimal. It should see a slight decrease in temperature, and thus V.V. Zharkova’s research does not represent the actual situation. This offers valuable insight into the relationship between the Earth's orbital mechanics and climate. However, further research into the accuracy of the model is required.
The design of multi-target rendezvous trajectories, which see a spacecraft approaching a sequence of objects in orbit as efficiently (by some metric) as possible, is a challenging problem of critical importance for Active Debris Removal (ADR), On-Orbit Servicing (OOS) and cis-Lunar logistics more widely. This thesis investigates two primary challenges in space Vehicle Routing Problems (VRPs): the application of Neural Combinatorial Optimization (NCO) methods for ADR missions and the integration of verifiable trajectory optimization techniques for OTV payload deployment.
The first research focus assesses the efficacy of NCO methods in designing multi-target rendezvous trajectories for ADR missions. An Attention-based routing policy, comprising a Graph Attention Network and a Pointer Network, was developed and trained using Reinforcement Learning (RL) algorithms, including REINFORCE, Advantage Actor-Critic (A2C), and Proximal Policy Optimization (PPO). Through hyperparameter analysis utilizing ANOVA, embedding dimension and the number of encoder layers were identified as critical factors influencing model performance. The trained policy was evaluated on scenarios involving 10, 30, and 50 transfers based on the Iridium 33 debris cloud. In missions with 10 transfers, the NCO policy achieved a mean optimality gap of 32%, outperforming the Dynamic RAAN Walk (DRW) heuristic in both mission cost and runtime. However, performance degraded in more complex scenarios with 30 and 50 transfers, indicating limited generalization beyond the training conditions. Grid search hyperparameter optimization revealed that while model performance improves with increased complexity, gains are marginal, and larger training datasets enhance convergence speed with only slight improvements in final performance. These findings demonstrate that NCO methods are effective for ADR missions with a limited number of targets but face scalability and generalization challenges in more complex scenarios.
The second research focus involves the design and optimization of multi-rendezvous trajectories for the UARX Space OSSIE mission using a modular framework that integrates Heuristic Combinatorial Optimization (HCO) with Sequential Convex Programming (SCP). This framework successfully determined optimal target sequences and generated near fuel-optimal trajectories for OSSIE, a translational and mass-dynamic payload delivery platform.
An Attention-based routing policy trained with RL was integrated into the combinatorial optimization process, enhancing the efficiency of mission planning. Applied to the OSSIE mission, the framework effectively explored the mission design space, optimizing 5000 mission scenarios and affirming the vehicle’s capability to fulfill advertised services. The modularity of the framework ensures adaptability to mission-specific constraints and facilitates future extensions, such as the incorporation of low-thrust propulsion profiles. Overall, this thesis confirms that NCO methods are applicable and effective for specific instances of space VRPs, particularly in optimizing ADR missions with a limited number of targets and in near-static mission scenarios where RAAN convergence is not required. The integration of verifiable trajectory optimization techniques with advanced routing policies presents a viable approach for efficient and adaptable mission planning. However, scalability and generalization remain challenges that necessitate further research. Recommendations include refining NCO model architectures to enhance scalability and generalization, exploring hybrid approaches that combine NCO with traditional heuristics, and developing automated machine learning frameworks to optimize model performance and robustness.
The project successfully achieved its primary objectives: developing and implementing heuristic and neural combinatorial optimization solvers for space VRPs, designing a modular trajectory optimization framework, and conducting comprehensive mission analyses for the OSSIE OTV. In doing so it has increased the mission design capabilities for space logistics missions at SENER Aerospace & Defence, as well as provided a strong foundation for future research and development aimed at addressing the increasing complexities of space operations. ...
The first research focus assesses the efficacy of NCO methods in designing multi-target rendezvous trajectories for ADR missions. An Attention-based routing policy, comprising a Graph Attention Network and a Pointer Network, was developed and trained using Reinforcement Learning (RL) algorithms, including REINFORCE, Advantage Actor-Critic (A2C), and Proximal Policy Optimization (PPO). Through hyperparameter analysis utilizing ANOVA, embedding dimension and the number of encoder layers were identified as critical factors influencing model performance. The trained policy was evaluated on scenarios involving 10, 30, and 50 transfers based on the Iridium 33 debris cloud. In missions with 10 transfers, the NCO policy achieved a mean optimality gap of 32%, outperforming the Dynamic RAAN Walk (DRW) heuristic in both mission cost and runtime. However, performance degraded in more complex scenarios with 30 and 50 transfers, indicating limited generalization beyond the training conditions. Grid search hyperparameter optimization revealed that while model performance improves with increased complexity, gains are marginal, and larger training datasets enhance convergence speed with only slight improvements in final performance. These findings demonstrate that NCO methods are effective for ADR missions with a limited number of targets but face scalability and generalization challenges in more complex scenarios.
The second research focus involves the design and optimization of multi-rendezvous trajectories for the UARX Space OSSIE mission using a modular framework that integrates Heuristic Combinatorial Optimization (HCO) with Sequential Convex Programming (SCP). This framework successfully determined optimal target sequences and generated near fuel-optimal trajectories for OSSIE, a translational and mass-dynamic payload delivery platform.
An Attention-based routing policy trained with RL was integrated into the combinatorial optimization process, enhancing the efficiency of mission planning. Applied to the OSSIE mission, the framework effectively explored the mission design space, optimizing 5000 mission scenarios and affirming the vehicle’s capability to fulfill advertised services. The modularity of the framework ensures adaptability to mission-specific constraints and facilitates future extensions, such as the incorporation of low-thrust propulsion profiles. Overall, this thesis confirms that NCO methods are applicable and effective for specific instances of space VRPs, particularly in optimizing ADR missions with a limited number of targets and in near-static mission scenarios where RAAN convergence is not required. The integration of verifiable trajectory optimization techniques with advanced routing policies presents a viable approach for efficient and adaptable mission planning. However, scalability and generalization remain challenges that necessitate further research. Recommendations include refining NCO model architectures to enhance scalability and generalization, exploring hybrid approaches that combine NCO with traditional heuristics, and developing automated machine learning frameworks to optimize model performance and robustness.
The project successfully achieved its primary objectives: developing and implementing heuristic and neural combinatorial optimization solvers for space VRPs, designing a modular trajectory optimization framework, and conducting comprehensive mission analyses for the OSSIE OTV. In doing so it has increased the mission design capabilities for space logistics missions at SENER Aerospace & Defence, as well as provided a strong foundation for future research and development aimed at addressing the increasing complexities of space operations. ...
The design of multi-target rendezvous trajectories, which see a spacecraft approaching a sequence of objects in orbit as efficiently (by some metric) as possible, is a challenging problem of critical importance for Active Debris Removal (ADR), On-Orbit Servicing (OOS) and cis-Lunar logistics more widely. This thesis investigates two primary challenges in space Vehicle Routing Problems (VRPs): the application of Neural Combinatorial Optimization (NCO) methods for ADR missions and the integration of verifiable trajectory optimization techniques for OTV payload deployment.
The first research focus assesses the efficacy of NCO methods in designing multi-target rendezvous trajectories for ADR missions. An Attention-based routing policy, comprising a Graph Attention Network and a Pointer Network, was developed and trained using Reinforcement Learning (RL) algorithms, including REINFORCE, Advantage Actor-Critic (A2C), and Proximal Policy Optimization (PPO). Through hyperparameter analysis utilizing ANOVA, embedding dimension and the number of encoder layers were identified as critical factors influencing model performance. The trained policy was evaluated on scenarios involving 10, 30, and 50 transfers based on the Iridium 33 debris cloud. In missions with 10 transfers, the NCO policy achieved a mean optimality gap of 32%, outperforming the Dynamic RAAN Walk (DRW) heuristic in both mission cost and runtime. However, performance degraded in more complex scenarios with 30 and 50 transfers, indicating limited generalization beyond the training conditions. Grid search hyperparameter optimization revealed that while model performance improves with increased complexity, gains are marginal, and larger training datasets enhance convergence speed with only slight improvements in final performance. These findings demonstrate that NCO methods are effective for ADR missions with a limited number of targets but face scalability and generalization challenges in more complex scenarios.
The second research focus involves the design and optimization of multi-rendezvous trajectories for the UARX Space OSSIE mission using a modular framework that integrates Heuristic Combinatorial Optimization (HCO) with Sequential Convex Programming (SCP). This framework successfully determined optimal target sequences and generated near fuel-optimal trajectories for OSSIE, a translational and mass-dynamic payload delivery platform.
An Attention-based routing policy trained with RL was integrated into the combinatorial optimization process, enhancing the efficiency of mission planning. Applied to the OSSIE mission, the framework effectively explored the mission design space, optimizing 5000 mission scenarios and affirming the vehicle’s capability to fulfill advertised services. The modularity of the framework ensures adaptability to mission-specific constraints and facilitates future extensions, such as the incorporation of low-thrust propulsion profiles. Overall, this thesis confirms that NCO methods are applicable and effective for specific instances of space VRPs, particularly in optimizing ADR missions with a limited number of targets and in near-static mission scenarios where RAAN convergence is not required. The integration of verifiable trajectory optimization techniques with advanced routing policies presents a viable approach for efficient and adaptable mission planning. However, scalability and generalization remain challenges that necessitate further research. Recommendations include refining NCO model architectures to enhance scalability and generalization, exploring hybrid approaches that combine NCO with traditional heuristics, and developing automated machine learning frameworks to optimize model performance and robustness.
The project successfully achieved its primary objectives: developing and implementing heuristic and neural combinatorial optimization solvers for space VRPs, designing a modular trajectory optimization framework, and conducting comprehensive mission analyses for the OSSIE OTV. In doing so it has increased the mission design capabilities for space logistics missions at SENER Aerospace & Defence, as well as provided a strong foundation for future research and development aimed at addressing the increasing complexities of space operations.
The first research focus assesses the efficacy of NCO methods in designing multi-target rendezvous trajectories for ADR missions. An Attention-based routing policy, comprising a Graph Attention Network and a Pointer Network, was developed and trained using Reinforcement Learning (RL) algorithms, including REINFORCE, Advantage Actor-Critic (A2C), and Proximal Policy Optimization (PPO). Through hyperparameter analysis utilizing ANOVA, embedding dimension and the number of encoder layers were identified as critical factors influencing model performance. The trained policy was evaluated on scenarios involving 10, 30, and 50 transfers based on the Iridium 33 debris cloud. In missions with 10 transfers, the NCO policy achieved a mean optimality gap of 32%, outperforming the Dynamic RAAN Walk (DRW) heuristic in both mission cost and runtime. However, performance degraded in more complex scenarios with 30 and 50 transfers, indicating limited generalization beyond the training conditions. Grid search hyperparameter optimization revealed that while model performance improves with increased complexity, gains are marginal, and larger training datasets enhance convergence speed with only slight improvements in final performance. These findings demonstrate that NCO methods are effective for ADR missions with a limited number of targets but face scalability and generalization challenges in more complex scenarios.
The second research focus involves the design and optimization of multi-rendezvous trajectories for the UARX Space OSSIE mission using a modular framework that integrates Heuristic Combinatorial Optimization (HCO) with Sequential Convex Programming (SCP). This framework successfully determined optimal target sequences and generated near fuel-optimal trajectories for OSSIE, a translational and mass-dynamic payload delivery platform.
An Attention-based routing policy trained with RL was integrated into the combinatorial optimization process, enhancing the efficiency of mission planning. Applied to the OSSIE mission, the framework effectively explored the mission design space, optimizing 5000 mission scenarios and affirming the vehicle’s capability to fulfill advertised services. The modularity of the framework ensures adaptability to mission-specific constraints and facilitates future extensions, such as the incorporation of low-thrust propulsion profiles. Overall, this thesis confirms that NCO methods are applicable and effective for specific instances of space VRPs, particularly in optimizing ADR missions with a limited number of targets and in near-static mission scenarios where RAAN convergence is not required. The integration of verifiable trajectory optimization techniques with advanced routing policies presents a viable approach for efficient and adaptable mission planning. However, scalability and generalization remain challenges that necessitate further research. Recommendations include refining NCO model architectures to enhance scalability and generalization, exploring hybrid approaches that combine NCO with traditional heuristics, and developing automated machine learning frameworks to optimize model performance and robustness.
The project successfully achieved its primary objectives: developing and implementing heuristic and neural combinatorial optimization solvers for space VRPs, designing a modular trajectory optimization framework, and conducting comprehensive mission analyses for the OSSIE OTV. In doing so it has increased the mission design capabilities for space logistics missions at SENER Aerospace & Defence, as well as provided a strong foundation for future research and development aimed at addressing the increasing complexities of space operations.
Hybrid Propelled Access to Space
Feasibility Study on a Hybrid Propulsion Concept for the Mk-III Spaceplane
In pursuit of cost-effective and sustainable space access, Dawn Aerospace is pioneering innovative launch systems. Their focus lies on a semi-reusable, two-stage-to-orbit launch vehicle, designed for horizontal take-off and landing. While current launch vehicles rely solely on rocket propulsion, the potential of airbreathing engines on the first stage is explored. Airbreathing engines offer increased fuel efficiency and utilize the incoming airflow as oxidiser.
This study delves into the feasibility of integrating airbreathing propulsion into Dawn Aerospace's Mk-III vehicle, with a goal to reduce gross take-off mass while meeting mission requirements. Airbreathing engine types are evaluated, with ramjets and turbine engines emerging as primary candidates. Detailed design processes, including vehicle modeling, trajectory analysis, and optimization, are employed.
The study reveals the potential of the ramjet concept which could have a lower gross take-off mass compared to fully rocket powered designs. Nevertheless, the application requires an improved vehicle design to obtain a feasible design. ...
This study delves into the feasibility of integrating airbreathing propulsion into Dawn Aerospace's Mk-III vehicle, with a goal to reduce gross take-off mass while meeting mission requirements. Airbreathing engine types are evaluated, with ramjets and turbine engines emerging as primary candidates. Detailed design processes, including vehicle modeling, trajectory analysis, and optimization, are employed.
The study reveals the potential of the ramjet concept which could have a lower gross take-off mass compared to fully rocket powered designs. Nevertheless, the application requires an improved vehicle design to obtain a feasible design. ...
In pursuit of cost-effective and sustainable space access, Dawn Aerospace is pioneering innovative launch systems. Their focus lies on a semi-reusable, two-stage-to-orbit launch vehicle, designed for horizontal take-off and landing. While current launch vehicles rely solely on rocket propulsion, the potential of airbreathing engines on the first stage is explored. Airbreathing engines offer increased fuel efficiency and utilize the incoming airflow as oxidiser.
This study delves into the feasibility of integrating airbreathing propulsion into Dawn Aerospace's Mk-III vehicle, with a goal to reduce gross take-off mass while meeting mission requirements. Airbreathing engine types are evaluated, with ramjets and turbine engines emerging as primary candidates. Detailed design processes, including vehicle modeling, trajectory analysis, and optimization, are employed.
The study reveals the potential of the ramjet concept which could have a lower gross take-off mass compared to fully rocket powered designs. Nevertheless, the application requires an improved vehicle design to obtain a feasible design.
This study delves into the feasibility of integrating airbreathing propulsion into Dawn Aerospace's Mk-III vehicle, with a goal to reduce gross take-off mass while meeting mission requirements. Airbreathing engine types are evaluated, with ramjets and turbine engines emerging as primary candidates. Detailed design processes, including vehicle modeling, trajectory analysis, and optimization, are employed.
The study reveals the potential of the ramjet concept which could have a lower gross take-off mass compared to fully rocket powered designs. Nevertheless, the application requires an improved vehicle design to obtain a feasible design.
Bachelor thesis
(2023)
-
L. Ramos Archilla, R. Bereś, A. Chumsena, T.P. Dammerman, R.J. van Dijk, Bo Lee, O. Miller, M.A. Ploeg, E.J.C. Watson, D.M. van Wolffelaar, M.C. Naeije, H.S. Saluja, O. Stroosma
Closed-loop guidance for micro launchers
Improvement to robustness and failure tolerance
The startup Rocket Factory Augsburg (RFA) is developing the RFA ONE orbital launch vehicle. A significant hurdle for its closed-loop guidance is the need to fully deplete the third stage for insertion during the first flight, which results in lengthy burn-arcs and complexity not managed properly by traditional guidance algorithms. Additionally, new testing methods introduce uncertainty and component failure risks, demanding an adaptable guidance system.This thesis therefore aims to enhance RFA's closed-loop guidance system to ensure optimal orbit insertion in all feasible scenarios. This was achieved by redesigning the closed-loop guidance system to make it more robust for extended thrust-arcs.Furthermore an in-flight estimation algorithm for the key performance parameters was implemented. Finally the guidance can now fall back to alternate target orbits in case of insufficient performance margins. It was shown that the new guidance is able to successfully complete the maiden flight with the large amount of expected uncertainty.
...
The startup Rocket Factory Augsburg (RFA) is developing the RFA ONE orbital launch vehicle. A significant hurdle for its closed-loop guidance is the need to fully deplete the third stage for insertion during the first flight, which results in lengthy burn-arcs and complexity not managed properly by traditional guidance algorithms. Additionally, new testing methods introduce uncertainty and component failure risks, demanding an adaptable guidance system.This thesis therefore aims to enhance RFA's closed-loop guidance system to ensure optimal orbit insertion in all feasible scenarios. This was achieved by redesigning the closed-loop guidance system to make it more robust for extended thrust-arcs.Furthermore an in-flight estimation algorithm for the key performance parameters was implemented. Finally the guidance can now fall back to alternate target orbits in case of insufficient performance margins. It was shown that the new guidance is able to successfully complete the maiden flight with the large amount of expected uncertainty.
RLV development can be considered as the modern step towards mission design due to financial and strategic decisions. In the past, reusability has been addressed however the level of maturity of the technology, both in terms of hardware and software was not yet reached. There are several aspects to developing a RLV, and these can be categorized into optimization of the LV, optimization of the trajectory, and cost analysis. TO be able to determine the feasibility of the mission, it is not just necessary to develop a suitable configuration, but to also determine the physical feasibility of the trajectory. Several methods exist of which convex optimization is selected. This class of algorithms have risen in popularity in the regime of powered descent guidance, and present a desirable trade-off between performance and computational cost. An already existing algorithm, DESCENDO, for a two-staged vehicle CALLISTO purposed for a mission to a geo-synchronous orbit, is taken as reference. The algorithm is rewritten in YALMIP allowing it to perform more efficiently by saving computation time through creation of multiple controllers based on a discretized burn schedule. A potential candidate for reusability in the future is selected, which is a VEGA variant, considered as a two-staged vehicle with set requirements on the mission and configuration. Through closed-loop simulations, the feasibility of RTLS for a particular mission of this VEGA variant can be studied. The disciplines involved in the study include the launch vehicle optimization, engine sizing, preliminary ascent & descent, and 3-DoF simulations. Previous research at TU Delft on RLV has included the work Rozenmeijer, Vandamme, Van Kesteren, Miranda, and Contant, graduate students of the TU Delft Aerospace Engineering faculty. The work relied on the usage of the TUDAT C++ software environment and based its feasibility or reusability of operations through a cost-analysis. A shift in direction is taken away from cost-analysis to examine at a greater detail the physical feasibility of the trajectory for a nominal candidate RLV. This is done by examining the influence of simulator to guidance algorithm dynamics and guidance algorithm parameters. Moreover, a nominal payload class between 100 and 500 kg is selected to determine the configuration of the vehicle ideal for this mission. To be able to determine feasibility of RTLS, three metrics are considered, which are the final landing velocity, final landing position, and maximumdynamic pressure. The study performs higher fidelity analysis only on the return phase, and as such the starting conditions for descent are determined through a preliminary design process by considering a drag-less ascent. This returns a starting altitude of around 26-30 km, with similar values for starting downrange position, and varying conditions of initial mass and velocity. For the preliminary descent, it is found that the metric of dynamic pressure does not reach more than around 60% of the limit imposed by the VEGA-C, which is similar for other VEGA variants. This coincides with research done with the CALLISTO vehicle. All in all, the best cases for these metrics and one included as an overall best case where candidates for the 400 kg payload class. The selection criteria for best cases of the preliminary descent involved the dynamic pressure, final velocity, and required propellant mass for descent. Moreover, the vehicle optimization results showed that this contained the most variation of vehicle characteristics, and as such it was deemed as a desirable class to work with for its flexibility in design. The best case burnt mass result was selected as the best case velocity required extensive propellant mass to burn for only a less than 7 m/s difference in result, which would not be indicative of what the convex algorithm could achieve due to dynamics involved in the preliminary experiment. This nominal candidate is then tested for various variations of controller tuning parameters combinations, burn schedules, and simulator/guidance frequencies. The results showed a clear desirable region for the final time of just between 300 and 310 seconds for return, favouring shorter burns. The solution envelope for the burn schedule showed gaps in zones of feasibility as well as optimality, suggesting some performance issues with the algorithm due to it failing to find solutions. Nevertheless this envelope is well defined and several feasible solutions existed. The influence of parameter tuning and simulator frequency was studied. It was determined that no set of guidance parameters could give an advantage over the other, but that some values did favour feasibility more. This is somewhat in contrast to the selection of frequencies, as despite the fact there was also a large difference for higher frequency ratios between the Q3 to Q4 and Q0 to Q2, there is a noticeably trend that higher ratios are favoured. Moreover, a local optimal ratio of frequency of 10-1 was also found, and being the same ratio used for the other experiments as well as the CALLISTO study, provides more evidence that this effect is intended. The uncertainties studied are for the initial state variations, errors in reading of position and velocities, and process time delays. It was noted that almost all the errors in the initial state variations shared similar distributions for ranges of values of the metrics. The overall majority returned feasible as well as the large minority of this returned optimal. Of little to no significance was the processing time delay, of which the overwhelming majority returned optimal results, and the rest where outliers whose process time factor where beyond the 3σ limit imposed in the creation of normal random variables. The largest errors arose from the real-time uncertainty in the velocities and position, modelled after pseudo-range errors. Although results showed a high density in the feasible and optimal regions for metrics of final time and position, there was also a high density past the feasible regions. It can be considered that the feasibility of RTLS operations for such a VEGA vehicle is restricted by such errors as expected, but nevertheless results are promising in what can be the main steps to lead to an error analysis study by the use of state estimation techniques and testing various modifications made to the SOCP problem to improve performance.
...
RLV development can be considered as the modern step towards mission design due to financial and strategic decisions. In the past, reusability has been addressed however the level of maturity of the technology, both in terms of hardware and software was not yet reached. There are several aspects to developing a RLV, and these can be categorized into optimization of the LV, optimization of the trajectory, and cost analysis. TO be able to determine the feasibility of the mission, it is not just necessary to develop a suitable configuration, but to also determine the physical feasibility of the trajectory. Several methods exist of which convex optimization is selected. This class of algorithms have risen in popularity in the regime of powered descent guidance, and present a desirable trade-off between performance and computational cost. An already existing algorithm, DESCENDO, for a two-staged vehicle CALLISTO purposed for a mission to a geo-synchronous orbit, is taken as reference. The algorithm is rewritten in YALMIP allowing it to perform more efficiently by saving computation time through creation of multiple controllers based on a discretized burn schedule. A potential candidate for reusability in the future is selected, which is a VEGA variant, considered as a two-staged vehicle with set requirements on the mission and configuration. Through closed-loop simulations, the feasibility of RTLS for a particular mission of this VEGA variant can be studied. The disciplines involved in the study include the launch vehicle optimization, engine sizing, preliminary ascent & descent, and 3-DoF simulations. Previous research at TU Delft on RLV has included the work Rozenmeijer, Vandamme, Van Kesteren, Miranda, and Contant, graduate students of the TU Delft Aerospace Engineering faculty. The work relied on the usage of the TUDAT C++ software environment and based its feasibility or reusability of operations through a cost-analysis. A shift in direction is taken away from cost-analysis to examine at a greater detail the physical feasibility of the trajectory for a nominal candidate RLV. This is done by examining the influence of simulator to guidance algorithm dynamics and guidance algorithm parameters. Moreover, a nominal payload class between 100 and 500 kg is selected to determine the configuration of the vehicle ideal for this mission. To be able to determine feasibility of RTLS, three metrics are considered, which are the final landing velocity, final landing position, and maximumdynamic pressure. The study performs higher fidelity analysis only on the return phase, and as such the starting conditions for descent are determined through a preliminary design process by considering a drag-less ascent. This returns a starting altitude of around 26-30 km, with similar values for starting downrange position, and varying conditions of initial mass and velocity. For the preliminary descent, it is found that the metric of dynamic pressure does not reach more than around 60% of the limit imposed by the VEGA-C, which is similar for other VEGA variants. This coincides with research done with the CALLISTO vehicle. All in all, the best cases for these metrics and one included as an overall best case where candidates for the 400 kg payload class. The selection criteria for best cases of the preliminary descent involved the dynamic pressure, final velocity, and required propellant mass for descent. Moreover, the vehicle optimization results showed that this contained the most variation of vehicle characteristics, and as such it was deemed as a desirable class to work with for its flexibility in design. The best case burnt mass result was selected as the best case velocity required extensive propellant mass to burn for only a less than 7 m/s difference in result, which would not be indicative of what the convex algorithm could achieve due to dynamics involved in the preliminary experiment. This nominal candidate is then tested for various variations of controller tuning parameters combinations, burn schedules, and simulator/guidance frequencies. The results showed a clear desirable region for the final time of just between 300 and 310 seconds for return, favouring shorter burns. The solution envelope for the burn schedule showed gaps in zones of feasibility as well as optimality, suggesting some performance issues with the algorithm due to it failing to find solutions. Nevertheless this envelope is well defined and several feasible solutions existed. The influence of parameter tuning and simulator frequency was studied. It was determined that no set of guidance parameters could give an advantage over the other, but that some values did favour feasibility more. This is somewhat in contrast to the selection of frequencies, as despite the fact there was also a large difference for higher frequency ratios between the Q3 to Q4 and Q0 to Q2, there is a noticeably trend that higher ratios are favoured. Moreover, a local optimal ratio of frequency of 10-1 was also found, and being the same ratio used for the other experiments as well as the CALLISTO study, provides more evidence that this effect is intended. The uncertainties studied are for the initial state variations, errors in reading of position and velocities, and process time delays. It was noted that almost all the errors in the initial state variations shared similar distributions for ranges of values of the metrics. The overall majority returned feasible as well as the large minority of this returned optimal. Of little to no significance was the processing time delay, of which the overwhelming majority returned optimal results, and the rest where outliers whose process time factor where beyond the 3σ limit imposed in the creation of normal random variables. The largest errors arose from the real-time uncertainty in the velocities and position, modelled after pseudo-range errors. Although results showed a high density in the feasible and optimal regions for metrics of final time and position, there was also a high density past the feasible regions. It can be considered that the feasibility of RTLS operations for such a VEGA vehicle is restricted by such errors as expected, but nevertheless results are promising in what can be the main steps to lead to an error analysis study by the use of state estimation techniques and testing various modifications made to the SOCP problem to improve performance.