1. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 1 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems David Riley Deimos Space Ltd., UK Davide Bonetti Deimos Space S.L.U., Spain

2. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 2 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Table of Contents Introduction: The Problem Domain The Optimisation Process Tools for Numerical Trade-offs Optimization results Conclusion

3. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 3 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com INTRODUCTION: THE PROBLEM DOMAIN

4. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 4 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com The problem Top-level requirement –Bring a spacecraft safely to rest on the surface of another planet Challenges: Mars –Entry velocity ~ 3 km/s –Very thin atmosphere equivalent to 35 km altitude on Earth –Gravity around 1/3 of Earth’s –Want to maximise the payload delivered to the surface –Subject to constraints in cost, risk, geographical sourcing, development timescale… Image: NASA/JPL

5. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 5 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Entry, Descent and Landing Systems Entry: aerodynamic deceleration using a heatshield Deploy a parachute to slow down faster Retro-rockets for final slowdown and fine control Airbags for touchdown

6. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 6 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Entry, Descent and Landing Systems Some parts are common: –Heatshield for entry –Parachute for deceleration Others vary: –One vs two parachute stages –Whether to have a retro stage –Touchdown on airbags, landing legs, crushables…

7. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 7 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Driving requirements Mass Height loss and verticalisation Thermo-mechanical loads Landing site accuracy Volume Reliability / robustness Cost Development timescale

8. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 8 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Trade-offs Smaller parachute / bigger retros: –Lighter parachute system –Final speed is higher –More retro fuel required Single-stage vs two-stage parachute: –For low final speeds, two-stage generally is lower mass –Two-stage may take more altitude to reach final speed –Single-stage is simpler

9. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 9 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Optimisation problem Problem includes discrete choices, integer values and continuous values Multi-disciplinary approach required –Parachute design, retro system design, airbag design, trajectory, … Can’t split the problem into separate pieces - different parts of the design affect each other –Speed reached under heatshield must be safe for parachute deployment –Each stage must carry the subsequent stages e.g. parachute must slow down the retro fuel and airbags as well as the payload Need the trajectory to be self-consistent –Test through simulation: enough time and altitude to stop before hitting ground

10. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 10 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com High Precision Landers: overview Mission concept –Technology Research Study within the ESA MREP (Mars Robotic Exploration Preparatory) Programme, preliminary to Mars Sample Return –Rover and return vehicle must land near each other –Target is launch in late 2020’s Main study objective –Design an optimum and robust EDL/GNC configuration for MSR lander –Project led by Airbus Defence & Space (Les Mureaux) Main requirements –Achieve landing accuracy of at least 10 km, ideally high precision (3 km) or even pinpoint (100 m) –Mass at entry ~ 800 kg –Use European technology as far as possible Images: NASA/JPL

11. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 11 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Small Mars Landers: overview Mission concept –Technology Research Study within the ESA MREP (Mars Robotic Exploration Preparatory) Programme, preliminary to INSPIRE –Target is launch in 2026 or 2028 Main study objective –Design an optimum and robust EDL/GNC configuration for a Mars Network Science mission –Project led by Deimos Space (Spain) Main requirements –Three identical probes separating from the same carrier and landing in different sites –Each probe mass at entry ~350 kg –Payload: 130 kg, 1150x355 mm (cylinder) –Use European technology as far as possible Image: ESA

12. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 12 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Small Mars Landers: overview One site Global model Challenges of a multi-probe mission –Several studies, only one launched (crashed) –Target entry mass quite different to previous single or multiple landers solutions Wide environment variability –Triple landing site, identical probes –Altitude below 0 MOLA, -15º < Latitude < 30º –Global atmosphere models based on statistics of European Mars Climate Database calls

13. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 13 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com THE OPTIMISATION PROCESS

14. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 14 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Overall Process

15. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 15 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Overall Process

16. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 16 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Preliminary EDL & GNC Trade-offs TRADE-OFFS SUMMARY PhIdTrade-offOptions Arrival A1 Arrival conditions Direct Escape GTO constrained GTO 2022 unconstrained GTO 2024 unconstrained Separation & Coasting SC1 SED Militar/Civil Heritage Beagle-2 Cassini/Hyugens SC2 Entry type Prograde Retrograde SC3 Separation times ESAT dimensions [25:0.5] days before EIP Entry E1 Aeroshell concept 70º cone + sphere 60º cone + sphere IBD Parashield E2 Nose Radius 0.25 diameters 0.45 diameters E3 TPS material SLA-561V Norcoat-Liege AQ60 / Prosial ALESTRASIL PICSIL DO31 / SPA ASTERM E4 Entry Control system Spinning 3-axis RCS damping E5 FPA ESAT dimension [-18:-10.5]º E6 Capsule diameter ESAT dimension [1.4:2] m E7 Capsule backshell angle ESAT dimension [30:50]º Descent and Landing DL1 Parachutes type DGB Ringslot Ringsail Hemisflo Cruciform DL2 Number of parachute stages Single Two stage DL3 Use of retrorockets No retros Retro-assisted parachute phase Fire retros after parachute separation Skycrane DL4 Controllability Fixed Thrust - non pulsed Fixed Thrust – pulsed Throttleable (variable thrust) DL5 Velocity reduction by means of thrust None Only Vertical Only Horizontal Vertical and Horizontal TRADE-OFFS SUMMARY PhIdTrade-offOptions Descent and Landing DL6 Use of lowering system No Yes DL7 Landing system Legs Vented Airbags Non-Vented airbags Crushable (as stand-alone system, not part of legs or airbag system) DL8 Backshell avoidance manoeuvre No Yes DL9 Parachute avoidance manoeuvre / flyaway No Yes GNC G1 Entry Guidance None Hypersonic guidance G2 Descent Guidance None Guided parachute G3 Landing Guidance Apollo derived G-turn None N1 Coasting & Entry Navigation timer+MU timer+IMU+sun sensor N2 Descent & Landing Navigation timer+MU timer+RA timer+IMU+RA timer+IMU+Camera timer+IMU+Camera+RA timer+IMU+Camera+RD C1 Entry Control None Liquid rockets C2 Descent & Landing velocity Control None Solid rockets Liquid rockets C3 Descent & Landing attitude control None Liquid rockets T1 Descent triggering IMU-based correlation Navigated velocity Navigated dynamic pressure T2 Main parachute triggering None Timer T3 RA activation None Timer from mortar T4 Retros triggering None Altitude and velocity T5 Landing triggering None Altitude and velocity

17. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 17 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com 3 Configurations selected for ESAT numerical trade-offs SIMPLE ROBUST TYPE 1 TYPE 2 TYPE3 Sub-systemType Aeroshell70º sphere-cone EntryBallistic Descent2 stages (DGB+Ringsail) Retro-rocketsNone LoweringNone LandingNon vented airbags GNC sensorsSun sensor, RA, IMU Sub-systemType Aeroshell70º sphere-cone EntryBallistic Descent1 stage (DGB) Retro-rocketsSolid (Vertical only) LoweringYes LandingNon vented airbags GNC sensorsSun sensor, RA, IMU Sub-systemType Aeroshell70º sphere-cone EntryBallistic Descent1 stage (DGB) Retro-rocketsSolid (Vertical & Horizontal) LoweringYes LandingNon vented airbags GNC sensorsSun sensor, RA, IMU, VBN MER-LIKE MPF-LIKE Beagle2-LIKE (no lowering)

18. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 18 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Overall Process

19. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 19 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Overall Process ESAT: EDL&GNC Sizing and Analysis Tool

20. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 20 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com TOOLS FOR NUMERICAL TRADE-OFFS

21. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 21 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com ESAT Architecture ESAT relies on a modular and generic infrastructure. Users concentrate efforts on Settings and Wrapper Function to solve a given problem. External module: - multidisciplinary - single discipline EDL&GNC Sizing and Analysis Tool

22. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 22 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com MDO Architecture C: Coasting E: Entry D: Descent L: Landing P/L: Payload Qdyn: Dynamic pressure FS: Frontshield Vt: terminal V NSP: Network Science Probe Different for Type 1, 2 and 3 SIMULATION CORE Overall it is a complex problem: –Multiple phases –Multiple combinations of worst cases (aerodynamics, atmosphere, events…) => robust solutions

23. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 23 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Bi-level MDO Architecture Bi-level surrogate models with internal optimization –Level1: MDO (mission & system objectives) –Level2: Coasting, D&L (nested in level 1) Achieve efficient optimizations PESDO (TESSELLA)

24. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 24 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com OPTIMIZATION RESULTS

25. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 25 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Problem 1: Probe separation timing 3 probes, want to optimise the separation events ESAT wrapper for Small Mars Landers Coasting and separation overview => simplified ESAT example =>

26. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 26 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Problem 1: Probe separation timing Look at results for sites 1 and 3 Fuel mass for retargeting EIP FPA dispersion Pareto Front: Dominant solutions

27. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 27 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Problem 2: Minimisation of EDLS Mass EDL&GNC Robust Optimization Problem –7 variable dimensions covering EDL, GNC and environment aspects TypePhaseDescription ConstraintLLanding Accuracy EMaximum heat flux at nose EMaximum heat load at nose EMaximum dynamic pressure E/DMaximum load factor E/DMaximum dynamic pressure at parachute deployment E/DMinimum altitude at parachute deployment E/D/LPercentage of non feasible simulations CTotal fuel mass for coasting phase E/D/LConsistency of volumes and sizes FixedE/D/LESAT EDL/GNC configuration InputsLPayload mass LPayload diameter LPayload height ECapsule nose radius / Capsule diameter ECapsule front cone angle ECapsule shoulder radius / Capsule diameter ECapsule back plate diameter/capsule diameter ETPS density (Norcoat Liege) Variable ESAT Dimensions E/DCapsule diameter ECapsule back cone angle EFlight Path Angle at EIP E/DDispersion on Mach number at parachute triggering DDispersion on Altitude at Powered Descent Initiation D/LDispersion on Altitude at Free Fall start LWinds scale factor Minimize system mass Maximize altitude at parachute dep.

28. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 28 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Problem 2: Minimisation of EDLS Mass Configurations selection process: Query and Filtering –More than 100 performances have been managed in 7 dimensions with ESAT –2D slices / N-dim data mining –Statistics of 50000 different samples

29. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 29 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Pareto Frontier: optimum solutions, = chosen one –Predicted and validated Pareto frontiers match with very good accuracy Problem 2: Minimisation of EDLS Mass TYPE 2 (mass includes system margins)

30. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 30 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Optimum configurations comparison Overall summary

31. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 31 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Configurations selected Type 1&2: Mass budget and Sizes Type 1: BC ~ 73 kg/m 2 Type 2: BC ~ 79 kg/m 2 (mass includes system margins) P/L = 32.2% of Total P/L = 29.6% of Total ElementType 1Type 2Units Payload volume 0.369 m3m3 Capsule diameter 2.0382.045 m Capsule front cone angle 70.00 deg Capsule back cone angle 44.29044.809 deg Capsule back cover diameter 1.0191.022 m Internal volume 1.3831.381 m3m3 Sensors volume 0.0030.002 m3m3 Parachute diameter13.36613.573m Parachute volume (stowed) 0.0740.076 m3m3 RCS volume 0.012 m3m3 RCS diameter 0.124 m Airbags volume (stowed) 0.0480.129 m3m3 Airbags thickness (stowed) 0.0140.035 m DL volume 0.1340.216 m3m3 Ballistic coefficient 72.878.7 Kg/m 2 kg Type 1: winds < 21.3 m/s Type 2: Winds < 13 m/s

32. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 32 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Configurations selected Type 1&2: GNC solutions Type 1: Modes –Events (blue) –Sensors (grey) –State vector (red) FunctionPhaseGNC solutionEquipments GUIDANCE EDLN/A NAVIGATION Inertial: Coasting & Entry Ballistic mode Inertial kinematic Sun sensor + IMU Relative: Descent and Landing Hybridized with lateral velocity estimation RA + IMU + VBN CONTROL EntryN/A Velocity control: Descent and Landing Vertical velocity control/reduction Lateral velocity compensation/control Solid rockets for vertical/lateral control (RAD+TIRS) Attitude control: Descent and Landing N/A Only for Type 1 configuration

33. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 33 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com CONCLUSIONS

34. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 34 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Conclusions: Mars Landers Identified optimal landing site sequences, configurations and component trade-offs for the Small Mars Landers project Robustness is critical for a network Mars landers mission Higher GNC complexity is the price of adding flexibility to site selection –Vision based navigation –Lateral control ESAT allows the System, Mission and GNC engineers to perform high- fidelity EDL-GNC architecture trade-offs relaying on high-fidelity and end- to-end approach. It increases the reliability of the selected design solutions with a reduced number of iteration loops (number and extent)

35. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 35 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Conclusions: Design Optimisation Tools We have found the approach adds a lot to EDLS optimisation Need to choose an appropriate level of optimisation based on the level of fidelity of the model –No point fine-tuning something that exists in the model, not in reality Surrogate modelling tools like ESAT are valuable when the models are complicated, slow to run Multi-disciplinary approach is vital – we’ve got as far as we can with tuning each component separately, choosing handover conditions by guesswork Mathematical optimisation is effectively “yet another discipline” –Important to add appropriate support not just extra complication ESAT is a useable front-end to a sophisticated optimisation approach –Allows the user to focus on creating the “wrapper function”

36. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 36 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Acknowledgements Deimos Space (Spain) –Gabriele De Zaiacomo, Irene Pontijas Fuentes, Rodrigo Haya Ramos, Gabriele Bellei, Jordi Freixa Mallol Airbus Defence and Space –Timothee Verwaerde, Aurélien Pisseloup, Cédric Renault European Space Agency –Eric Bornschlegl, Kelly Geelen, Alvaro Martinez Barrio, Thomas Voirin

37. Multi-Disciplinary Optimisation for Planetary Entry, Descent and Landing Systems - 37 - 2 nd UK Workshop on Optimisation in Space Engineering Cambridge, 19 Mar 2014 © 2014 Deimos Space UK Ltd. - www.deimos-space.com Thank you!