On the application of generative adversarial networks for nonlinear modal analysis

Tsialiamanis, George; Champneys, Max D.; Dervilis, Nikolaos; Wagg, DJ; Worden, Keith

doi:10.1016/j.ymssp.2021.108473

Cited by 18 publications

(16 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…linear embedding Dervilis et al (2019), generative adversarial networks Tsialiamanis et al (2022) and variational autoencoders Simpson et al (2023). Through these NNMs, parallels may also be drawn between autoencoders and structural dynamics, similarly to the relationship between PCA and linear modes.…”

Section: Figurementioning

confidence: 99%

Reduced order modeling of non-linear monopile dynamics via an AE-LSTM scheme

Simpson

Dervilis²,

Couturier³

et al. 2023

Front. Energy Res.

View full text Add to dashboard Cite

Non-linear analysis is of increasing importance in wind energy engineering as a result of their exposure in extreme conditions and the ever-increasing size and slenderness of wind turbines. Whilst modern computing capabilities facilitate execution of complex analyses, certain applications which require multiple or real-time analyses remain a challenge, motivating adoption of accelerated computing schemes, such as reduced order modelling (ROM) methods. Soil structure interaction (SSI) simulations fall in this class of problems, with the non-linear restoring force significantly affecting the dynamic behaviour of the turbine. In this work, we propose a ROM approach to the SSI problem using a recently developed ROM methodology. We exploit a data-driven non-linear ROM methodology coupling an autoencoder with long short-term memory (LSTM) neural networks. The ROM is trained to emulate a steel monopile foundation constrained by non-linear soil and subject to forces and moments at the top of the foundation, which represent the equivalent loading of an operating turbine under wind and wave forcing. The ROM well approximates the time domain and frequency domain response of the Full Order Model (FOM) over a range of different wind and wave loading regimes, whilst reducing the computational toll by a factor of 300. We further propose an error metric for capturing isolated failure instances of the ROM.

show abstract

Section: Figurementioning

confidence: 99%

Reduced order modeling of non-linear monopile dynamics via an AE-LSTM scheme

Simpson

Dervilis²,

Couturier³

et al. 2023

Front. Energy Res.

View full text Add to dashboard Cite

show abstract

“…Additionally, a gradient-descent approach permits good optimisation performance on arbitrary (differentiable) objective functions, even in the presence of large numbers of model parameters. The cycle-GAN NNMs of [14] showed improved performance over the multinomial GP approach in both decomposition and reconstruction tasks.…”

Section: Introductionmentioning

confidence: 98%

“…A recent work adopts an alternative neural architecture for 𝑓 and 𝑓 −1 [14]. In this work, a cycle-consistent generative adversarial network (cycle-GAN) is used [15].…”

Section: Introductionmentioning

confidence: 99%

“…Retaining the dimensionality of the input ensures that as many modal coordinates are defined as there are degrees-of-freedom in the nonlinear structure. The target distributions of the generator models can be chosen freely to ensure independence (in [14], independent Gaussians are used). There is also no need to learn a separate inverse mapping, because the forward and inverse mappings are learnt simultaneously.…”

Section: Introductionmentioning

confidence: 99%

“…When first proposed, the quality of the simplification was judged 'by-eye' to be the extent to which the modal displacements produced solitary peaks in their power-spectral densities (PSD). In [13], this was extended to a convolution-based approach, and in [14], to an approach based on the inner-product. The quality of the reconstruction can of course be evaluated by a normalised mean-square error (NMSE) metric.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations