Revealing the effects of damping on the flow-induced vibration of flexible cylinders

Contrary to most signal decomposition methods that usually decompose an original signal into a series of components simultaneously, a novel approach based on repeated extraction of Maximum Energy Component (MEC) is proposed. The approach starts from determination of the MEC referring to the estimated Power Spectral Density (PSD) function, and then represents the MEC by employing an exponential function to fit the original signal. By defining a stopping criterion based on two adjacent estimated PSDs, each MEC can be accurately extracted with an improved performance throughout the entire signal decomposition. To verify the proposed method, a single degree-of-freedom system subject to harmonic loads has been examined. Numerical results show that the analytical response can not only be decomposed into four MECs corresponding to the excitation and the system, respectively, but also provide an accurate estimation of natural frequency and damping ratio of the system. Meanwhile, by observing results from the Ensemble Empirical Mode Decomposition (EEMD), Variational Mode Decomposition (VMD) and Prony based on state-space model (Prony-SS), an improved decomposition accuracy has been achieved from the proposed approach. Furthermore, experimental data from the Norwegian Deepwater Programme and two sets of field-test data from one fixed offshore platform and an offshore wind turbine have been used to demonstrate the correctness of the developed signal decomposition method. It is noted that divergence in results by Prony-SS can be observed when a very large model order is used, while the proposed method provides the better decomposition and reconstruction of signals.

show abstract

“…At the same time, the frequency of the inline response is twice the frequency of the crossflow motion [30]. tained, as shown in Fig.…”

Section: Comparison With Traditional Signal Decomposition Methodsmentioning

confidence: 86%

A signal decomposition method based on repeated extraction of maximum energy component for offshore structures

et al. 2020

View full text Add to dashboard Cite

show abstract

“…More details can be found in references [41,42]. As mentioned in [43,44], the motions between the CF and IL were coupled, and the frequency of the IL response was twice that of CF. This difference allowed us to verify the correctness of the time-frequency analysis results.…”

Section: Scenario 2: Noisy Signalmentioning

confidence: 99%

A new time-frequency analysis method based on single mode function decomposition for offshore wind turbines

et al. 2020

View full text Add to dashboard Cite

The Hilbert-Huang transform (HHT) has been widely applied and recognised as a powerful time-frequency analysis method for nonlinear and non-stationary signals in numerous engineering fields. One of its major challenges is that the HHT is frequently subject to mode mixing in the processing of practical signals such as those of offshore wind turbines, as the frequencies of offshore wind turbines are typically close and contaminated by noise. To address this issue, this paper proposes a new time-frequency analysis method based on single mode function (SMF) decomposition to overcome the mode mixing problem in the structural health monitoring (SHM) of offshore wind turbines. In this approach, the structural vibration signal is first decomposed into a set of window components using complex exponential decomposition. A state-space model is introduced in the signal decomposition to improve the numerical stability of the decomposition, and then a novel window-alignment strategy, named energy gridding, is proposed and the signals are constructed in the corresponding gridding. Furthermore, energy recollection is implemented in each gridding, and the reassembling of these components yields an SMF that is comparable to the intrinsic mode function (IMF) of the HHT, but with a significant improvement in terms of mode mixing. Four case studies are conducted to evaluate the performance of the proposed method. The first case attempts to detect three different frequencies in a simulated time-invariant signal. The second case attempts to test a synthesised signal with segmental time-varying frequencies (each segment contains three different frequencies components). The analysis results in these two cases indicate that mode mixing can be reduced by the proposed method. Furthermore, a synthesised signal with slowly varying frequencies is used. These analysis results demonstrate the effective suppression of non-relevant frequency components using SMF decomposition. In the third case, the experimental data from vortex-induced vibration (VIV) experiments sponsored by the Norwegian Deepwater Programme (NDP) are used to evaluate the proposed SMF decomposition for vibration mode identification. In the final case, field data acquired from an offshore wind turbine foundation 1 and offshore wind turbine are analysed. The mode identification results obtained using SMF decomposition are compared with those produced by the HHT. The comparison demonstrates superior performance of the proposed method in identifying the vibration modes of the VIV experimental and field data.

show abstract

“…For cylinder flow-induced vibration, extensive analysis from numerical simulations and controlled laboratory experiments revealed the effect of several important dimensionless parameters. For example, increasing the amount of damping present results in a decrease of the VIV response amplitude (Scruton, 1955;Griffin, 1975; Govardhan & Williamson, 2006;Vandiver, 2012;Vandiver et al, 2018). Conversely, the peak response amplitude shows an increasing trend with an increase in Reynolds number in the subcritical regime (Govardhan & Williamson, 2006;Swithenbank, et al 2008;Resvanis, et al 2012).…”

Section: Introductionmentioning

confidence: 98%

Enhancing Machine Learning Models With Prior Physical Knowledge to Aid in VIV Response Prediction

Resvanis

Vandiver

2021

Volume 8: CFD and FSI

Self Cite

View full text Add to dashboard Cite

Practical engineering prediction models for flow-induced vibration are needed in the design of structures in the ocean. Research has shown that structural vibration response may be influenced by a large number of physical input parameters, such as damping and Reynolds number. Practical response prediction tools used in design are inevitably a compromise between complexity and simplicity of use. Modern machine learning tools may be used to identify which input parameters are most important. Standard machine learning techniques enable the researcher to compile a list of the most important input parameters, ranked or ordered by the effect of each on the prediction error of the model. When all inputs are treated as equals, blind application of machine learning may lead to predictions that are inconsistent with prior physical knowledge. To address this problem, we conducted a parameter selection process using a prior knowledge-based, trend-informed neural network architecture. This approach was used to identify features important to the prediction of the cross-flow vibration response amplitude of long flexible cylinders, given the known prior effect of Reynolds number and damping. The model balances the usual goal of minimizing the model prediction error, but doing so in a manner that closely follows the extensive knowledge we have of the influence of Reynolds number and damping parameter on response. The resulting neural network model was able to reveal additional insights, including the role of mode number shifting, mode dominance and travelling waves in the regulation of VIV response amplitude.

show abstract

Revealing the effects of damping on the flow-induced vibration of flexible cylinders

Cited by 23 publications

References 21 publications

A signal decomposition method based on repeated extraction of maximum energy component for offshore structures

A signal decomposition method based on repeated extraction of maximum energy component for offshore structures

A new time-frequency analysis method based on single mode function decomposition for offshore wind turbines

Enhancing Machine Learning Models With Prior Physical Knowledge to Aid in VIV Response Prediction

Contact Info

Product

Resources

About