Data-driven modelling of compressor stall flutter
File(s)
Author(s)
Rauseo, Marco
Type
Thesis or dissertation
Abstract
Modern aircraft engines need to meet ever more stringent requirements that greatly
increase the complexity of design, which strives for enhanced performance, reduced
operating costs, emissions and noise simultaneously. The drive for performance leads
to the development of thin, lightweight, highly loaded fan and compressor blades which
are increasingly more prone to incur high, sustained vibratory stresses and aeroelastic
problems such as flutter.
The current practice employs preliminary design tools for flutter that are often based
on empiricism or simplified analytical models, requiring extensive use of computational
fluid dynamics to verify aeroelastic stability. As the industry moves to new designs,
fast and accurate prediction tools are needed. In this thesis, data-driven techniques
are employed to model the aeroelastic response of compressor blades.
Machine learning has been applied to a plethora of engineering problems, with particular
success in the field of turbulence modelling. However, conventional, black-box data-
driven methods based on simple input parameters require large databases and are
unable to generalise. In this work a combination of machine learning techniques and
reduced order models is proposed to address both limitations at the same time. Previous
knowledge of flutter is introduced in the physics guided framework by formulating
relevant, steady state input features, and by injecting results from low-fidelity analytical
models.
The models are tested on several unseen cascades and it is found that training on
even a single geometry yields accurate results. The models developed here allow
flutter prediction of fan and compressor flutter stability based on the steady state
flow only without a need for any CPU intensive unsteady simulations. Hence, one can
predict flutter stability of a given blade for different mechanical properties (mode shape,
frequency) at near zero additional cost once the mean flow is known. Moreover, for fan
flutter, the model developed here can be integrated with available analytical models of
intake to analyse the consequences of intake properties, such as length and acoustic
liners location, on the stability of fan blades. The EU goal of climate-neutrality by 2050
requires novel design concepts in aviation which is unachievable without complimentary
novel prediction and design tools. The research presented in this thesis will allow one
to explore the design space for flutter stability based on steady flow only, and hence
offers such an alternative. To the best of the author’s knowledge, no previous research
is available on modelling of compressor stall flutter with data-driven techniques.
increase the complexity of design, which strives for enhanced performance, reduced
operating costs, emissions and noise simultaneously. The drive for performance leads
to the development of thin, lightweight, highly loaded fan and compressor blades which
are increasingly more prone to incur high, sustained vibratory stresses and aeroelastic
problems such as flutter.
The current practice employs preliminary design tools for flutter that are often based
on empiricism or simplified analytical models, requiring extensive use of computational
fluid dynamics to verify aeroelastic stability. As the industry moves to new designs,
fast and accurate prediction tools are needed. In this thesis, data-driven techniques
are employed to model the aeroelastic response of compressor blades.
Machine learning has been applied to a plethora of engineering problems, with particular
success in the field of turbulence modelling. However, conventional, black-box data-
driven methods based on simple input parameters require large databases and are
unable to generalise. In this work a combination of machine learning techniques and
reduced order models is proposed to address both limitations at the same time. Previous
knowledge of flutter is introduced in the physics guided framework by formulating
relevant, steady state input features, and by injecting results from low-fidelity analytical
models.
The models are tested on several unseen cascades and it is found that training on
even a single geometry yields accurate results. The models developed here allow
flutter prediction of fan and compressor flutter stability based on the steady state
flow only without a need for any CPU intensive unsteady simulations. Hence, one can
predict flutter stability of a given blade for different mechanical properties (mode shape,
frequency) at near zero additional cost once the mean flow is known. Moreover, for fan
flutter, the model developed here can be integrated with available analytical models of
intake to analyse the consequences of intake properties, such as length and acoustic
liners location, on the stability of fan blades. The EU goal of climate-neutrality by 2050
requires novel design concepts in aviation which is unachievable without complimentary
novel prediction and design tools. The research presented in this thesis will allow one
to explore the design space for flutter stability based on steady flow only, and hence
offers such an alternative. To the best of the author’s knowledge, no previous research
is available on modelling of compressor stall flutter with data-driven techniques.
Version
Open Access
Date Issued
2023-03
Online Publication Date
2023-10-16T15:09:41Z
Date Awarded
2023-09
Copyright Statement
Creative Commons Attribution NonCommercial Licence
Advisor
Vahdati, Mehdi
Martinez-Botas, Ricardo
Publisher Department
Mechanical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)