Multi-disciplinary Design Optimisation Strategy in Multi-stage Launch Vehicle Conceptual Design

1. Multi-disciplinary Design Optimisation Strategy in Multi-stage Launch Vehicle Conceptual Design

2. Introduction to Launch Vehicle (LV) conceptual design process • Literature survey on Multi-disciplinary Design Optimisation(MDO) in LV design • Proposed research work • Preliminary work done

3. Introduction to LV Conceptual Design Process Design of Launch Vehicle Making appropriate comprises to achieve balance among many coupled objectives. Objectives High performance Safety Simple operation Low cost Conceptual design To reveal trends and allow relative comparison among alternatives, while design flexibility exists.

4. Introduction to LV Conceptual Design Process Outcome of conceptual design • Number of stages • Type of stage/propellant • Mass split of stages • Thrust levels and propulsive system details • External layout

5. Mission Requirements Propulsion Options & design Layout & Surface geometry Vehicle sizing Weight & C.G Structural, Control & Thermal Analyses Trajectory Analyses Vehicle Configuration Dimensions Steering rate history Aerodynamic Analysis Launch vehicle conceptual design process

6. Resize Vehicle Vehicle sizing Vehicle performance Determine Performance Introduction to LV Conceptual Design Process Determining the optimum configuration • The evaluation of the interaction between the vehicle systems • The impact of this system upon the vehicles ability to perform the desired mission.

7. Introduction to LV Conceptual Design Process Optimum values of design parameters Vehicle performance will be carried out to examine the value of each parameters by fixing the values of remaining parameters. “one-variable-at-a-time approach”

8. Introduction to LV Conceptual Design Process Conceptual design of Advanced Manned Launch System Ref: Stanley, D.O., Talay, A.T., Lepsch, R.A., Morris, W.D., Kathy, E.W. “Conceptual Design Of A Fully Reusable Manned Launch System”. Journal of Spacecraft And Rockets, Vol 29, No.4, pp 529-537, July-August, 1992 • Reference vehicle geometry is chosen after all discipline analyses were carried out. • After finalising the reference vehicle, a series of parametric trade studies were performed to determine the major vehicle parameters.

9. Introduction to LV Conceptual Design Process

10. Layout & Surface geometry Vehicle Concept options Aerodynamic analysis Mission Requirements Trajectory analysis Structural, Control, thermal Propulsion analyses Propulsion option Configuration, Weights and sizing Technology options Cost Analysis Operational analysis Operational options Rethink/modify requirements and options

11. Introduction to LV Conceptual Design Process The limitations associated with conceptual design are • Conceptual design is carried out with low fidelity models • The relationships among design and the conceptual design parameters are often not well modeled or understood. • In ‘one-variable-at-a-time’ approach, the impact of simultaneously considering all variables is not considered – Result in near optimum configuration.

12. Introduction to LV Conceptual Design Process • These limitations result in probably inefficient final design - Leaving room for significant improvements in performance and reduction in life costs.

13. Introduction to LV Conceptual Design Process To improve results during conceptual design: • Improvement of disciplinary analysis, modelling and tools that capture, with sufficient fidelity, the major relationships among design variables and system objectives. • The development of methods for coordinating the engineering analysis and optimising the total launch vehicle system. (iii) ‘All-at-same-time’ approach is to be adopted.

14. Introduction to LV Conceptual Design Process • All these can be achieved by design of all systems together should be iteratively refined together, with sufficient fidelity models, by MDO scheme. • This was not practical earlier because of high computational expenditure associated with numerical prediction methods. • Now, with availability of various methods and computational capabilities an MDO based conceptual design can be made.

15. Introduction to LV Conceptual Design Process MDO based conceptual design will allow system engineers to systematically explore the vast trade space and consider many more configurations during the conceptual design phase before converging on the final design.

16. Literature Review on MDO Works Related to Launch Vehicle Design

17. Literature Review on MDO Works on LV Design Performance optimisation of launch Vehicle • System design – Vehicle characteristic and parameters like number of stages engine sizing. • Trajectory optimisation - Control vector that optimises the performance for the chosen configuration. • Ideally, design of the vehicle and propulsion system and trajectory shaping should be iteratively refined together by a coupled MDO scheme to obtain solution.

18. Literature Review on MDO works on LV design MDO approaches in LV design To optimise vehicle performance is collect all elements of the trajectory control vector and system design variables in one vector of optimisation parameters to be manipulated by an appropriate algorithm. This approach has been applied successfully to ascent mission of Rocket powered single-stage-to-orbit. (I) Iterative loop MDO strategy (ii) Sequential compatibility constraint solution (iii) Collaborative Optimization

19. Literature Review on MDO works on LV design References: • Braun, R. D., Powell, R. W., Lepsch, R. A.. Stanley, D. 0., and Kroo, 1. M., "Comparison of Two Multidisciplinary Optimization Strategies for Launch-Vehicle Design," Journal of Spacecraft and Rockets,Vol. 32, No. 3, 1995,pp.404-410. • Braun, R.D. and Moore., “Collaborative approach to launch vehicle design” Journal of Spacecraft and Rockets, Vol. 34, No.4, pp 478-485, July-August,1997.

20. Iterative-loop solution strategy Optimizer Minimize J=dry weight Design variables(40) Subject to inflight and terminal constraints Initial guess at GLOW, Sref Base diameter Landed weight propulsion GLOW=GLOWc Sref=Srefc Landed wt =Landed wtc base diameter =base diameterc Inflight & Terminal Constraints Trajectory N0 Delta =(GLOWc-GLOW)2 +(Srefc-Sref) 2 +(Landed wtc-Landed wt) 2 +(base diameterc- base diameter) 2 Is Delta small Weights & Sizing GLOWc, Srefc Base diameterc Landed weightc Dry weight Yes Done

21. Sequential compatibility-constraint solution Optimizer Minimize J=dry weight Design variables(40) Subject to inflight and terminal constraints Trajectory propulsion Inflight & terminal constraints Weights & Sizing Compatibility constraints GLOWc-GLOW =0 Srefc-Sref = 0 Landed wtc-Landed wt= 0 base diameterc- base diameter= 0 GLOWc Srefc Landed wtc base diameterc Dry weight

22. Literature Review on MDO works on LV design Advantages of sequential compatibility constraint approach: i) being 3-4 times more computationally efficient ii) providing greater flexibility in the way in which consistency is maintained across disciplinary boundaries, and iii) a smoother design space. Disadvantage: The compatibility constraint approach is in situations terminates without reaching the solution - Because multidisciplinary feasibility is only guaranteed at a solution in this approach, the design information could be invalid.

23. Literature Review on MDO works on LV design Collaboration optimization • A problem is decomposed into subproblems along domain-specific boundaries. • Through subspace optimization, each group is given control over its own set of local design variables and is charged with satisfying its own domain-specific constraints. • The objective of each subproblem is to reach agreement with the other groups on values of the interdisciplinary variables. • A system-level optimizer is employed to orchestrate this interdisciplinary compatibility process while minimizing the overall objective

24. Literature Review on MDO works on LV design Collaborative optimization architecture for launch vehicle design

25. Literature Review on MDO works on LV design

26. Literature Review on MDO works on LV design Ref: Tsuchiya, T. and Mori. T. “Multidisciplinary Design Optimization to future space transportation vehicle”. AIAA 2002-5171. • MDO method to choose the best among the seven typical concepts of RLV. • The design variables are representing geometry and shape of vehicles, flight performance parameters • Similar to Sequential Compatibility Constraint Solution.

27. Literature Review on MDO works on LV design Ref:Hillesheimer, M., Schotlle, U. M. and Messerschmid, E., "Optimiza­tion of Two-Stage Reusable Space Transportation Systems with Rocket and Airbreathing Propulsion Concepts," International Astronautical Federation Paper 92-O863, Sept. 1992 Though these MDO architectures has been applied successfully to the ascent mission of single stage vehicle, it has shown poor convergence properties even for less complex mission examples of an expendable multistage rocket launches, when major system design parameters such as the mass split of stages or engine sizing were included to optimize trajectory control and vehicle parameters simultaneously

28. Literature Review on MDO works on LV design Proposed another approach that avoids this difficulty is a multistep sequential optimization procedure. • Consists of a performance optimization cycle (inner loop) and a vehicle design cycle (outer loop). • Inner loop uses the data of the latter to determine the control functions and major system parameters yielding the optimum performance - responds to varying vehicle size needs as long as the departure from the preset design (outer loop) remains small.

29. Literature Review on MDO works on LV design Multistepsequentialprocedure

30. Literature Review on MDO works on LV design • Otherwise, a vehicle redesign including system modifications and reevaluation of the aerodynamic coefficients (which are held constant in the inner optimization cycle) is performed in separate computations in the outer iteration loop. • The latter requires manual interaction and is supported by graphic interface tools. • This scheme outlined above is applied to enhance the performance of a reusable rocket launcher which is part of Ariane X family.

31. Literature Review on MDO works on LV design Two design software based on the schemes similar to multistage sequential optimization process. FASTPASS (Flexible analysis for synthesis trajectory and performance for advanced space systems) developed by Lockheed Martin Astronautics and SWORD (Strategic Weapon Optimisation for rapid Design) developed by Lockheed Missile design and space Co. for solid motor missile. References Szedula, J.A., FASTPASS: A Tool For Launch Vehicle Synthesis, AIAA-96-4051-CP, 1996. Hempel, P. R., Moeller C. P., and Stuntz L. M., “Missile Design Optimization Experience And Developments”, AIAA-94-4344,1994-CP

32. Literature Review on MDO works on LV design Ref:Rahn, M. and Schottle, U. M., "Decomposition Algorithm for Performance Optimization of a Launch Vehicle," Journal of Spacecraft and Rockets, Vol. 33, No. 2, 1996, pp. 214--221. Though Multistep sequential scheme was able to solve the optimization problem of a two-stage, winged rocket launch vehicle designed for vertical takeoff, severe convergence problems were encountered when it was applied to the more complex mission of an airbreathing launch vehicles. These difficulties were attributed in part to different performance sensitivities of the various flight phases, controls, and major system design parameters, and to scaling problems.

33. Literature Review on MDO works on LV design Proposes adecomposition approach to solve the overall optimization problem of a Reentry launch system. • Partitioning the trajectory into subarcs such that each mission segment can be optimized independently. • These subproblems constitute the first level of optimization. • A second-level controller is then used to optimize the entire mission. • Hence, a two-level optimization procedure results, with the master-level algorithm optimally coordinating the solution of the subproblems.

34. Literature Review on MDO works on LV design

35. Literature Review on MDO works on LV design Master Problem: Maximize upper-stage payload mass Independent variables: Staging Mach number Longitude at staging Load factor at pull-up Time interval for pull-up Subproblem 1: Minimize: Booster stage ascent propellant Subject to: Staging Mach no. (master contr.) Staging longitude(master contr.) Latitude at staging heading staging Independent variables: flight heading after take-off supersonic cruise flight length bank angle control parameter determines the length of the turn flight Subproblem 1: Minimize: Booster stage flyback propellant Subject to: Max flight acceleration Max dynamic pressure End head towards landing site Independent variables: Angle of attack control Bank angle control Parameter determines the length of the turn flight. Subproblem 1: Minimize: Orbiter ascent propellant Subject to: Max long. Flight acceleration Perigee velocity Perigee altitude Perigee path angle Independent variables: Angle of attack control

36. Literature Review on MDO works on LV design MDO methods may be divided into three groups: i) Parameters methods based on design of experiments (DOE) techniques ii) Gradient or Calculus based methods iii) Stochastic methods such as genetic algorithm and simulated annealing. Parametric methods as well as gradient based methods are applicable at conceptual design phase.

37. Literature Review on MDO works on LV design Ref: Stanley, D. O., Unal, R., and Joyner, C. R., "Application of Taguchi Methods to Dual Mixture Ratio Propulsion System Optimization for SSTO Vehicles," Journal of Spacecraft and Rockets, Vol. 29, No. 4, 1992, pp. 453-459. • Taguchi design method to determine the thrust levels of a variety of engine and vehicle parameter for single-stage-to-orbit vehicle. • This study considers five design parameters.

38. Literature Review on MDO works on LV design Ref: Stanley, D. 0., Engelund. W. C., Lepsch. R. A., McMillin, M. L.Wt K. E.. Powell. R. W., Guinta. A. A., and Unal, R. "Rocket-Powered Single Stage Vehicle Configuration Selection and Design," Journal of Spacecraft and Rockets,Vol. 31, No. 5, 1994. pp. 792-798; also AIAA Paper93-Feb. 1993. • The configuration selection for rocket powered single stage vehicle configuration using RSM. • Five configuration parameters considered for study. • RSM was used to determine the minimum dry weight entry vehicle to meet constraints on performance.

39. Literature Review on MDO works on LV design Ref: Olds, J., and Walberg, G., ”Multidisciplinary Design of a Rocket-Based Combined-Cycle SSTO Launch Vehicle using Taguchi Methods” , AIAA 93-1096, Feb,1993. Taguchi method was used to evaluate the effects of changing 8 design variables (2 of which were discrete) in an "all at the same time" approach. Design variables pertained to both the vehicle geometry (cone half-angle, engine cowl wrap around angle) and trajectory parameters (dynamic pressure limits, heating rate limits, and airbreathing mode to rocket mode transition Mach number).

40. Literature Review on MDO works on LV design Ref: Anderson, m., Burkhalter J., and Jenkins R “Multidisciplinary Intelligence Systems Approach To Solid Rocket Motor Design, Part I: Single And Dual Goal Optimization. AIAA 2001-3599, July, 2001. Investigated the potential of using a multidisciplinary genetic algorithm approach to the design of a solid rocket motor propulsion system as a component within overall missile system. Aerodynamics and trajectory performance disciplines were considered in this study

41. Literature Review on MDO works on LV design A survey on literature reveals that MDO works related to conceptual design, that is, simultaneous optimization of system and trajectory are limited to • Enhancement of an existing reference vehicle system • Selecting one among canididate configurations • Subsystem optimization with respect to vehicle performance.

42. Literature Review on MDO works on LV design • This may be attributed to the focused effort on the Advanced Manned Launch System (AMLS) activity since 1988. Two vehicles, single stage and two stages were used for this AMLS mission and all further design studies are to optimize the performance of these configuration. • Also, other recently developed vehicles are designed by evolution strategy.

43. Proposed Research Work • An MDO strategy with following capability would be useful in developing a new vehicle. • That is, given the range of realizable mass fraction and specific impulse, the scheme should be able to decide number of stages, mass and propellant fraction and iterate this vehicle, propulsion system and trajectory shaping and give optimum configuration and trajectory that meets the specification.

44. Proposed Research Work • This would be useful when no propulsion system or technological constraints are identified and the initial trade space is being defined. • This scheme may come up with a design which is non- intuitive and much better than traditional design technique. Development of such scheme is the aim of present research effort.

45. Preliminary Work Done

46. Preliminary Work Done Aim: To demonstrate the effect of bringing ‘Mass estimation discipline’ into conceptual design Problem considered: Choose a configuration with two-stage-to-orbit vehicle to inject 20t payload at 400km circular orbit. Assumptions: V loss Structural factors (1, 2 ) Specific Impulse

47. Preliminary Work Done

48. Preliminary Work Done Orbit Specifications Payload Choice of propulsion Isp1, Isp2 ms1,mp1 ms2,mp2, mpf LOW Vtotal 1, 2 Ideal velocity calculations Initialize V1 Assumptions V loss Structural factors (1, 2 ) Optimum LOW & Configuration Is LOW minimum Vary V1 Yes No

49. Preliminary Work Done Dy. Pressure Load factor Area ratios Fineness ratios mp1 mp2, mpf Sizing of tanks Weight estimation ms1e,ms2e

50. Dy. Pressure Load factor Area ratios Fineness ratios Preliminary Work Done Orbit Specifications Payload Choice of propulsion Isp1, Isp2 ms1,mp1 ms2,mp2, mpf LOW Vtotal 1, 2 Ideal velocity calculations Initialize V1 Sizing of tanks ms1e,ms2e Assumptions V loss Structural factors (1, 2 ) Is ms1= ms1e ms2= ms2e Weight estimation No Vary 1, 2 Yes LOW Optimum LOW & Configuration Is LOW minimum Vary V1 Yes No