Modern wind energy industry has embraced larger turbine solutions with ultra-long
blades that has reached a 100-150 meter scale for maximising the amount of power
extraction and increasing the levelized cost of energy (LCOE) of the wind turbines.
One key driving factor of this technological advancement is the necessity for in-depth
understandings of the blade aeroelastic behaviour and its structural responses of the
composite blades, particularly under complex operational conditions such as those
experienced by floating offshore wind turbines (FOWTs). However, analysing the
aeroelastic responses of such composite structures can encounter significant
computational challenges, such as the difficulties in capturing global-local coupled
behaviours and anisotropic material effects due to the excessive computational
expenses.
This thesis is structured in two main parts: (1) the development of a high-fidelity
aeroelastic analysis framework for composite FOWT blades via fluid–structure
interaction (FSI); and (2) multi-objective structural optimization of composite blades
using surrogate-assisted algorithms.
In the first part, an FSI framework aiming for the structural response investigations of
composite wind turbine blade on FOWT is developed. This work is a further
development based on previous developed FSI framework established by Yuanchuan
Liu (2018) who integrated the open-source computational fluid dynamics (CFD) code
OpenFOAM and the multibody dynamics (MBD) method MBDyn for the flexible
structure aeroelastic predictions. In parallel, the fully-resolved finite element analysis
(FEA) for the composite blade is conducted using the commercial package Abaqus
CAE, where the blade aeroelastics being resolved in FSI can be applied explicitly for
the field recovery for stress inspections on the blade.
This FSI framework is used for FOWT blade aeroelastic investigations with
consideration of anisotropic composite material properties. To reasonably account for
the influence of platform motions, a prescribed sinusoidal motion function resembling
the FOWT platform motions under a regular wave condition is applied, allowing a
realistic reproduction of the dynamics on the FOWT blades. The present FSI framework performs more computationally efficient than existing FSI strategies by
reducing nearly 25% of core-hours of computing resources while offering detailed
multi-hierarchy composite structural insights into the non-uniform stress behaviour
across the blade under dynamic loading conditions.
In the second part of this work, we presented a further extension based on the
established FSI framework for blade structural optimizations, aiming to achieve higher
strength-weight ratio blade designs to support the upscaling trend of the future wind
turbine blades. A nondominated sorting genetic algorithm II (NSGA-II), is integrated
with a machine learning (ML) based artificial neural network (ANN) surrogate model
for approximating the objective outputs (i.e. blade weight and max. Von Mises stress).
This approach streamlined the conventional FEA approach so that a significant
reduction in computational expenses is achieved. A notable challenge of the
distribution drifting issue of the surrogate model is identified and addressed,
improving the generalisability and predictive accuracy of the ANN during iterative
optimization. The framework demonstrates its robustness and effectiveness in highdimensional design spaces, achieving substantial blade weight reduction without
compromising structural integrity.
This work systematically introduced a numerical FSI analysing and optimization
workflow by taking the advantages of CFD, MBD and FEA, for the FSI investigations
for the flexible composite structures. The novelty of this work is that we provide a
general-purpose FSI-driven surrogate-assist structural optimization framework for
designing flexible composite structures with higher strength-to-weight performance.
The proposed framework in this work is also capable of handling applications beyond
wind turbines to other complex systems that are prone to interactive environments
between the fluid and structure physical fields accompany with different forms of
dynamic motions, offering detailed insights in aero- or hydroelastics terms and has a
great potential in the light-weight designs for composite structures.
| Date of Award | 21 Nov 2025 |
|---|
| Original language | English |
|---|
| Awarding Institution | - University Of Strathclyde
|
|---|
| Sponsors | University of Strathclyde |
|---|
| Supervisor | Qing Xiao (Supervisor) & Liu Yang (Supervisor) |
|---|