| Popis: |
Liquid rocket engine program management involves decision-making with multiple, usually conflicting, and correlated objectives like performance/weight, demonstrated reliability, initial operational capability, and affordability that include program risks and epistemic uncertainty. This paper presents a methodology to combine all objectives into a single fitness function with inclusion of customer preferences for aspiration levels for each objective as well as uncertainties for attaining these levels. Performance and mass are combined in a single performance figure of exergy loss, calculated in a thermodynamic cycle design model including epistemic uncertainty propagation covering both reliability based (RBDO) and robust design and optimization (RDO) for system sizing. The other objectives are estimated by dedicated models which include programmatic aspects like test strategy and technology readiness level. The final multiobjective optimization problem is reduced to a single objective function by combining truncated utility functions for which the shape is determined by the utility probability duality. The technology readiness level is expressed with a knowledge factor that links physical measures from cycle design with programmatic aspects. Finally, a satisficing solution is searched that exceeds all of the aspiration levels (targets) of each objective, rather than optimizing, leading to a fast solution answering all customer demands. |
