Multistate Design Approach to Analysis of Twin-Engine Aircraft Performance Robustness

Abstract

DOI: 10.2514/1.C031338 This article introduces an integrated approach for the early-stage multistate design and analysis of aircraft requiringrobustperformanceacrossarangeofoff-nominalsystemstates.Itisappliedtothecasestudyofanexisting twin-engineaircraft,enabledbythedevelopmentofamultimodalperformancemodelcalibratedwith flight-testdata collected at Edwards Air Force Base Flight Test Center. The methodology includes elements for determining responsesinaircraftexpectedperformanceandavailabilitytochangesinstaticdesignvariables(geometry),dynamic design variables (control gains), and component failure rates, which are the three driving input categories affecting performance robustness. This is accomplished using Markov chain analysis within the design loop to stochastically modelstatetransitionsbasedonfailureratesaswellasgainoptimizationthroughsimulationtoensurecontrollability at each design point. Results showed unsafe performance in 15% of the King Air geometry-state responses under investigation.Althoughmanyoccurredinthefullyfailedstateandwereexpected, fiveoccurredinpartiallydegraded states where the majority of geometries were able to meet performance requirements. Furthermore, this behavior clearly exhibited itself in the resulting design sensitivities, confirming that such an approach allows designers to identifyelementsthatmightdrivesystemlossthroughanalysisofperformancechangesacrosssystemstatesandtheir respective response to changes in design variables.