Clarifying Low-Frequency Variatoions Phenomenon Using a Novel Mode Decomposition Method

2006 ◽  
Vol 22 (3) ◽  
pp. 193-198 ◽  
Author(s):  
C.-T. Wang

AbstractLow-frequency variations in wake flow are complex and many aspects of its behavior remain poorly understood. A mode decomposition method developed by Huang et al is utilized herein because it can decompose any complicated data set into a finite number of intrinsic modes without distorting their original characteristics. The results of decomposition analysis of the measured base pressure signals reveal that a finite number of various kinds of flow structure modes, with their own characteristic time scales, coexist with the residue that exhibits most of the low-frequency variations in flow at Re = 11760 and 31600, respectively. As the Reynolds number increases, the decomposition method yields more components. Results also show exactly the vortex shedding structure in an intrinsic mode and the low-frequency variations that appear in the residue during vortex shedding process.

1999 ◽  
Vol 13 (3) ◽  
pp. 339-359 ◽  
Author(s):  
J.J. MIAU ◽  
J.T. WANG ◽  
J.H. CHOU ◽  
C.Y. WEI

2018 ◽  
Vol 62 ◽  
pp. 03008 ◽  
Author(s):  
Yury Senkevich

The results of the study of the possibility of using the empirical mode decomposition method for cleaning geoacoustic emission signals from various types of noise are presented. It is shown that the application of the method allows to increase the ratio of the signal noise 3-6 dB depending on the ratio of signal dispersion and noise in the input signal. The examples demonstrate the ability to remove trends and harmonic interference, as well as the ability to highlight a useful signal when masking its powerful noise. A comparative evaluation of the method in relation to the low-frequency filtration is carried out. The limitation of the method applicability in the case of processing of pulse signals asymmetric with respect to its average value is indicated.


1997 ◽  
Author(s):  
J. Miau ◽  
J. Wang ◽  
J. Chou ◽  
J. Miau ◽  
J. Wang ◽  
...  

2006 ◽  
Vol 22 (4) ◽  
pp. 263-270 ◽  
Author(s):  
C.-T. Wang ◽  
C.-T. Chen

AbstractChaos theory has been seen as an efficient tool for studying the turbulent flow, the findings of attractor were also important and made in the study to investigate the wake flow behind the bluff body. Here, the fractal dimension value would then be found by Hurst analysis. According to the results found, the Hurst empirical formula derived by the self-similar laceration of vortex plane would be applied by self-similar property to decide the band of the frequency variations in the vortex shedding process. The three kinds of flow mode with their individual attractors and characteristics could be decomposed and shown as following: self-similar laceration, energy input and white noise band. Finally, the energy ratio for the three kinds of flow mode had been confirmed. Hence, these findings would be helpful to further study the wake flow in the vortex shedding process.


2013 ◽  
Vol 31 (4) ◽  
pp. 619 ◽  
Author(s):  
Luiz Eduardo Soares Ferreira ◽  
Milton José Porsani ◽  
Michelângelo G. Da Silva ◽  
Giovani Lopes Vasconcelos

ABSTRACT. Seismic processing aims to provide an adequate image of the subsurface geology. During seismic processing, the filtering of signals considered noise is of utmost importance. Among these signals is the surface rolling noise, better known as ground-roll. Ground-roll occurs mainly in land seismic data, masking reflections, and this roll has the following main features: high amplitude, low frequency and low speed. The attenuation of this noise is generally performed through so-called conventional methods using 1-D or 2-D frequency filters in the fk domain. This study uses the empirical mode decomposition (EMD) method for ground-roll attenuation. The EMD method was implemented in the programming language FORTRAN 90 and applied in the time and frequency domains. The application of this method to the processing of land seismic line 204-RL-247 in Tacutu Basin resulted in stacked seismic sections that were of similar or sometimes better quality compared with those obtained using the fk and high-pass filtering methods.Keywords: seismic processing, empirical mode decomposition, seismic data filtering, ground-roll. RESUMO. O processamento sísmico tem como principal objetivo fornecer uma imagem adequada da geologia da subsuperfície. Nas etapas do processamento sísmico a filtragem de sinais considerados como ruídos é de fundamental importância. Dentre esses ruídos encontramos o ruído de rolamento superficial, mais conhecido como ground-roll . O ground-roll ocorre principalmente em dados sísmicos terrestres, mascarando as reflexões e possui como principais características: alta amplitude, baixa frequência e baixa velocidade. A atenuação desse ruído é geralmente realizada através de métodos de filtragem ditos convencionais, que utilizam filtros de frequência 1D ou filtro 2D no domínio fk. Este trabalho utiliza o método de Decomposição em Modos Empíricos (DME) para a atenuação do ground-roll. O método DME foi implementado em linguagem de programação FORTRAN 90, e foi aplicado no domínio do tempo e da frequência. Sua aplicação no processamento da linha sísmica terrestre 204-RL-247 da Bacia do Tacutu gerou como resultados, seções sísmicas empilhadas de qualidade semelhante e por vezes melhor, quando comparadas as obtidas com os métodos de filtragem fk e passa-alta.Palavras-chave: processamento sísmico, decomposição em modos empíricos, filtragem dados sísmicos, atenuação do ground-roll.


Author(s):  
Hassan F Ahmed ◽  
Hamayun Farooq ◽  
Imran Akhtar ◽  
Zafar Bangash

In this article, we introduce a machine learning–based reduced-order modeling (ML-ROM) framework through the integration of proper orthogonal decomposition (POD) and deep neural networks (DNNs), in addition to long short-term memory (LSTM) networks. The DNN is utilized to upscale POD temporal coefficients and their respective spatial modes to account for the dynamics represented by the truncated modes. In the second part of the algorithm, temporal evolution of the POD coefficients is obtained by recursively predicting their future states using an LSTM network. The proposed model (ML-ROM) is tested for flow past a circular cylinder characterized by the Navier–Stokes equations. We perform pressure mode decomposition analysis on the flow data using both POD and ML-ROM to predict hydrodynamic forces and demonstrate the accuracy of the proposed strategy for modeling lift and drag coefficients.


1998 ◽  
Vol 120 (1) ◽  
pp. 89-96 ◽  
Author(s):  
R. A. Van den Braembussche ◽  
H. Malys

A lumped parameter model to predict the high frequency pressure oscillations observed in a water brake dynamometer is presented. It explains how the measured low frequency variations of the torque are a consequence of the variation in amplitude of the high frequency flow oscillations. Based on this model, geometrical modifications were defined, aiming to suppress the oscillations while maintaining mechanical integrity of the device. An experimental verification demonstrated the validity of the model and showed a very stable operation of the modified dynamometer even at very low torque.


Sign in / Sign up

Export Citation Format

Share Document