scholarly journals Transfer Function Based Band Selection in Frequency Domain Time Series Regression

2020 ◽  
Author(s):  
Skye Griffith ◽  
Wesley S. Burr ◽  
Glen Takahara
Keyword(s):  
Time Series ◽  
Transfer Function ◽  
Frequency Domain ◽  
Band Selection ◽  
Time Series Regression

scholarly journals Frequency Domain Repercussions of Instantaneous Granger Causality

2021 ◽  
Author(s):  
Luiz Antonio Baccalá ◽  
Koichi Sameshima
Keyword(s):  
Time Series ◽  
Transfer Function ◽  
Frequency Domain ◽  
Granger Causality ◽  
Partial Directed Coherence ◽  
Directed Transfer Function ◽  
Vector Autoregressive Models ◽  
Vector Autoregressive ◽  
Standard Vector ◽  
Domain Description

Using Directed Transfer Function (DTF) and Partial Directed Coherence (PDC) in their information version, this paper extends their theoretical framework to incorporate instantaneous Granger Causality (iGC)’s frequency domain description into a single unified perspective. We show that standard vector autoregressive models allow portraying iGC’s repercussions associated with Granger Connectivity where interactions mediated without delay between time series can be easily detected.

scholarly journals Frequency Domain Repercussions of Instantaneous Granger Causality

Entropy ◽  
2021 ◽  
Vol 23 (8) ◽  
pp. 1037
Author(s):  
Luiz A. Baccalá ◽  
Koichi Sameshima
Keyword(s):  
Time Series ◽  
Transfer Function ◽  
Frequency Domain ◽  
Granger Causality ◽  
Partial Directed Coherence ◽  
Directed Transfer Function ◽  
Vector Autoregressive Models ◽  
Vector Autoregressive ◽  
Standard Vector ◽  
Domain Description

Using directed transfer function (DTF) and partial directed coherence (PDC) in the information version, this paper extends the theoretical framework to incorporate the instantaneous Granger causality (iGC) frequency domain description into a single unified perspective. We show that standard vector autoregressive models allow portraying iGC’s repercussions associated with Granger connectivity, where interactions mediated without delay between time series can be easily detected.

Error analyses of a Multi-Input-Multi-Output cantilever beam test

2021 ◽  
pp. 107754632110337
Author(s):  
Arup Maji ◽  
Fernando Moreu ◽  
James Woodall ◽  
Maimuna Hossain
Keyword(s):  
Transfer Function ◽  
Frequency Domain ◽  
Cantilever Beam ◽  
Transfer Functions ◽  
Linear Superposition ◽  
Transfer Function Matrix ◽  
Error Analyses ◽  
Pseudo Inverse ◽  
Function Matrix

Multi-Input-Multi-Output vibration testing typically requires the determination of inputs to achieve desired response at multiple locations. First, the responses due to each input are quantified in terms of complex transfer functions in the frequency domain. In this study, two Inputs and five Responses were used leading to a 5 × 2 transfer function matrix. Inputs corresponding to the desired Responses are then computed by inversion of the rectangular matrix using Pseudo-Inverse techniques that involve least-squared solutions. It is important to understand and quantify the various sources of errors in this process toward improved implementation of Multi-Input-Multi-Output testing. In this article, tests on a cantilever beam with two actuators (input controlled smart shakers) were used as Inputs while acceleration Responses were measured at five locations including the two input locations. Variation among tests was quantified including its impact on transfer functions across the relevant frequency domain. Accuracy of linear superposition of the influence of two actuators was quantified to investigate the influence of relative phase information. Finally, the accuracy of the Multi-Input-Multi-Output inversion process was investigated while varying the number of Responses from 2 (square transfer function matrix) to 5 (full-rectangular transfer function matrix). Results were examined in the context of the resonances and anti-resonances of the system as well as the ability of the actuators to provide actuation energy across the domain. Improved understanding of the sources of uncertainty from this study can be used for more complex Multi-Input-Multi-Output experiments.

A Frequency-Domain Filtering Technique for Triple Decomposition of Unsteady Turbulent Flow

1992 ◽  
Vol 114 (1) ◽  
pp. 45-51 ◽  
Author(s):  
G. J. Brereton ◽  
A. Kodal
Keyword(s):  
Time Series ◽  
Turbulent Flow ◽  
Frequency Domain ◽  
Measurement Data ◽  
Averaging Technique ◽  
Phase Averaging ◽  
Spectral Estimates ◽  
Improved Performance ◽  
A New Technique ◽  
Flow Variables

A new technique is presented for decomposing unsteady turbulent flow variables into their organized unsteady and turbulent components, which appears to offer some significant advantages over existing ones. The technique uses power-spectral estimates of data to deduce the optimal frequency-domain filter for determining the organized and turbulent components of a time series of data. When contrasted with the phase-averaging technique, this method can be thought of as replacing the assumption that the organized motion is identically reproduced in successive cycles of known periodicity by a more general condition: the cross-correlation of the organized and turbulent components is minimized for a time series of measurement data, given the expected shape of the turbulence power spectrum. The method is significantly more general than the phase average in its applicability and makes more efficient use of available data. Performance evaluations for time series of unsteady turbulent velocity measurements attest to the accuracy of the technique and illustrate the improved performance of this method over the phase-averaging technique when cycle-to-cycle variations in organized motion are present.

scholarly journals Time series regression studies in environmental epidemiology

2013 ◽  
Vol 42 (4) ◽  
pp. 1187-1195 ◽  
Author(s):  
Krishnan Bhaskaran ◽  
Antonio Gasparrini ◽  
Shakoor Hajat ◽  
Liam Smeeth ◽  
Ben Armstrong
Keyword(s):  
Time Series ◽  
Environmental Epidemiology ◽  
Time Series Regression

Prediction of Surge Motion Response of Floating Structure Using Kalman Smoother Adaptive Filters Based Time-Domain Volterra Model

2014 ◽  
Vol 660 ◽  
pp. 799-803
Author(s):  
Edwar Yazid ◽  
M.S. Liew ◽  
Setyamartana Parman ◽  
V.J. Kurian ◽  
C.Y. Ng
Keyword(s):  
Time Series ◽  
Transfer Function ◽  
Time Domain ◽  
Adaptive Filters ◽  
Low Frequency ◽  
Volterra Model ◽  
Floating Structure ◽  
Frequency Responses ◽  
Kalman Smoother ◽  
System Output

This work presents an approachto predict the low frequency and wave frequency responses (LFR and WFR) of afloating structure using Kalman smoother adaptive filters based time domain Volterramodel. This method utilized time series of a measured wave height as systeminput and surge motion as system output and used to generate the linear andnonlinear transfer function (TFs). Based on those TFs, predictions of surgemotion in terms of LFR and WFR were carried out in certain frequency ranges ofwave heights. The applicability of the proposed method is then applied in ascaled 1:100 model of a semisubmersible prototype.

Sign in / Sign up

Export Citation Format

Share Document