Uncertainty Quantification and Management on Aircraft Weight Estimation

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)


Weight is a key element in aircraft design, having a major influence on its performance and being a common factor to all disciplines involved in the decision making process: aerodynamics, structural sizing, materials, loads, geometry, cost, manufacturing, etc. To ensure an optimal trade-off is achieved, alongside a smooth convergence to the desired final aircraft weight, it is essential to be able to model the aircraft weight estimation process throughout the design, including assessment of uncertainty, sensitivity and risk.
Weight estimation processes and uncertainty analysis are well established bodies of literature; and uncertainty in the aircraft design process has been a topic much explored in recent years, both in academia and in industry. Applications of uncertainty quantification in aeroelasticity, including uncertainty in aircraft design features and structural sizing, aircraft life-cycle cost, aircraft’s environmental impacts, aeroelastic stability and aerodynamic characteristics, have been studied independently and for the purposes of robust and reliability-based design optimisations. On the other hand, by analysing the industrial processes and methodologies, the need to integrate empirical evidence, from physics-based models and statistical evidence, when assessing weight uncertainty at project milestones is identified. The unification of UQ and weight estimation in aircraft design into a framework that can deliver exhaustive, meaningful and innovative technical information, not only for the purposes of optimisation but also for risk and project management assessments, was lacking; The present work is set out to close this gap. Historical aircraft data is explored in order to trace patterns and trends on weight data, for the objective of concluding on its drivers, causes and sensitivities and ultimately help predict its fluctuations. The aircraft data proves scarce to constitute statistical evidence and the lack of standardisation in reporting is appointed. For these reasons, a framework has been developed that emulates the weight convergence corridor for an aircraft wing. It combines a traditional wing-box sizing method for primary weight with alternative methods for secondary weight. The alternative methods mimic the different phases of design in the aircraft development cycle. Maturity of design translates to the status of the information available, which translates to accuracy in the weight estimation method in use. This process incorporates uncertainty in the form of modelling the desired input parameters as Probability Density Functions (PDFs). The uncertain input space may include wing and engine planform geometry, wing-box material properties, load cases, general aircraft weights and fuselage dimensions. Design features and aircraft components are correlated and therefore an underlying dependency grid prevails. Combining the PDFs on the grid propagates the uncertainty towards an ultimate distribution of the total wing weight. The methodology is demonstrated on a representative commercial jet airliner wing. This work investigates the use of the framework developed for wing weight estimation by quantifying design sensitivities impact on wing weight. This includes the effect of uncertainties on: the aerodynamic loads, stress and sizing parameters; the material properties, specifically the stress allowable, of the wingbox; the stress formulae and activation of stress constraints; an external mass such as the engine properties; and the secondary structure weight. The results are combined into a weight convergence corridor, where the final wing weight can be compared against a target weight with a reliability requirement. As design features become fixed and information matures, uncertainty in the weight convergence corridor decreases, and that translates to a narrowing uncertainty band in the corridor. Uncertainty is inherently present in our world, and engineering is no exception. It is crucial that UQ becomes an integral part of all engineered systems; only then can they become more reliable.
Date of Award13 May 2020
Original languageEnglish
Awarding Institution
  • The University of Bristol
SponsorsAirbus Operations Ltd.
SupervisorJonathan E Cooper (Supervisor)


  • Uncertainty Quantification
  • Aircraft Weight
  • Structural Sizing
  • Sensitivity Analysis
  • Weight Engineering

Cite this