Uncertainty Quantification in Aeroelasticity Using Reduced Order Models
Computational aeroelasticity to predict instabilities, such as flutter, incurs in high computing times and lacks reliability due to several sources of uncertainty. While the impact of uncertainties in the structural parameters are well studied, little has been done on assessing the influence of aerodynamic uncertainties on the system aeroelastic behaviour. In a computational setting, turbulence models based on Reynolds-averaged Navier-Stokes (RANS) equations remain the workhorse in the computation of high Reynolds-number wall-bounded flows. While these methods have proved their utility in an industrial setting, their deficiencies in modelling complex flows are well-documented. Even the most sophisticated turbulence models invoke radically simplifying assumptions about the structure of the underlying turbulence. Although a model is based on physically and mathematically sound ideas, the model formulation typically devolves into the calibration of many free parameters, called closure coefficients, which balance the turbulence equations. The numerical values of the closure coefficients in current turbulence models are chosen by using a combination of heuristic and empirical decision making using a small set of canonical problems. Therefore, the central question posed in this research project is: Can we quantify the uncertainty and sensitivity of commonly-used RANS models on the system aeroelastic behaviour?
The aim is to explore novel computational strategies to perform uncertainty quantification of large-scale aeroelastic problems with large numbers of uncertain variables with minimal computing costs. However, propagation of the uncertainties is generally computational intractable due to the expensive fluid component, whereas the structural part is generally cheap. In this research project, we will alleviate this problem through the use of: 1) a machine learning approach which drives an adaptive design of experiments algorithm1 to investigate the propagation of aleatory and epistemic uncertainties to aerodynamics output quantities of interest; and 2) a reduced-order model that propagates efficiently the above-mentioned uncertainties. Armed with this reduced-order model, we will perform uncertainty quantification to estimate the system aeroelastic behaviour.
The successful applicant will have a background in physics, engineering or applied mathematics. Experience with programming is essential.
 Da Ronch A et al, “Adaptive design of experiments for efficient and accurate estimation of aerodynamic loads”, Aircraft Engineering and Aerospace Technology, 2017; 89(4): 558-569
If you wish to discuss any details of the project informally, please contact Andrea Da Ronch, Aerodynamics and Flight Mechanics research group, Email: [Email Address Removed].
This 3 year studentship covers home-rate tuition fees and provides an annual tax-free stipend at the standard EPSRC rate, which is £14,777 for 2018/19.
The funding is only available to UK citizens or EU citizens who have been resident in the UK for at least 3 years prior to the start of the studentship and not mainly for the purpose of receiving full-time education. For further guidance on funding, please contact [Email Address Removed]
How good is research at University of Southampton in General Engineering?
FTE Category A staff submitted: 192.23
Research output data provided by the Research Excellence Framework (REF)
Click here to see the results for all UK universities