A Novel Approach for Moving Window Size Selection utilizing recursive PCA

Tao, Songbing; Li, Benqi; Li, Nan; Chai, Yi; Liu, Bowen

doi:10.23919/chicc.2018.8483813

Cited by 2 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The process of extraction of residual components/nonlinear sensitive components is given by

X_{res} goodbreak= ψ * sans-serif-italicS ((), N_{lin} goodbreak+ sans-serif1 :) * V^{T}

ψ goodbreak= sans-serif-italicU ((), : N_{lin} goodbreak+ sans-serif1); N_{lin} goodbreak= sans-serif-italicp sans-seriffor which {CCR}_{p} \geq sans-serif99.95 %

where

ψ

(i.e., of size N × 1) is the first residual PC/lower order PC sensitive to nonlinear damage. The most popularly used guideline of cumulative contribution ratio (CCR) > 0.995 44–50 is employed in the present work to isolate active PCs/linear components of the total response.…”

Section: Nonlinear Sensitive Feature Component Using Dpcamentioning

confidence: 99%

“…DPCA [44][45][46][47][48][49][50] is similar to PCA and involves applying a moving time window of fixed size. The DPCA performs PCA only on a subset (windowed) of time series.…”

Section: Dynamic Principal Component Analysismentioning

confidence: 99%

“…DPCA algorithm processes the vibration data in real-time and computes the PCs for each time window or at each instant. As recommended by earlier researchers, [44][45][46][47][48][49][50] a nonoverlapping fixed moving window is employed in the proposed DPCA due to the nonlinear behavior of the vibration response of the structure with a breathing crack. The nonoverlapping fixed moving window technique helps reduce the correlations of windows in conventional DPCA.…”

Section: Dynamic Principal Component Analysismentioning

confidence: 99%

“…where ψ (i.e., of size N Â 1) is the first residual PC/lower order PC sensitive to nonlinear damage. The most popularly used guideline of cumulative contribution ratio (CCR) > 0.995 [44][45][46][47][48][49][50] is employed in the present work to isolate active PCs/linear components of the total response.…”

Section: Nonlinear Sensitive Feature Component Using Dpcamentioning

confidence: 99%

See 3 more Smart Citations

Real‐time reference‐free breathing crack identification using ambient vibration data

Prawin

2021

Structural Contr & Hlth

View full text Add to dashboard Cite

Summary A novel dynamic principal component analysis (DPCA)‐based baseline‐free damage diagnostic technique addressing breathing crack detection, localization, and characterization is proposed using ambient vibration data in the time domain. The nonlinear components sensitive to breathing crack, buried in the noisy responses, are first reconstructed by the elimination of the active principal components contributing to the total response using DPCA. A temporal damage sensitive feature based on the evolution of the variation of the residual principal component over time is proposed for confirming the presence and identifying the exact time instant of breathing crack in the structure. Besides detection, three new damage localization indices based on the fractal dimensional analysis of the residual response, first residual principal component vector, and directional angle are proposed for breathing crack localization. The effectiveness of the proposed DPCA approach is verified using the synthetic datasets of the benchmark simply supported beam with a breathing crack, provided by Helsinki Metropolia University of Applied Sciences and a numerically simulated cantilever beam with varied spatial locations and different depths of breathing crack. Finally, experimental investigations have been carried out to demonstrate the practical viability of the proposed DPCA approach.

show abstract

“…The process of extraction of residual components/nonlinear sensitive components is given by

X_{res} goodbreak= ψ * sans-serif-italicS ((), N_{lin} goodbreak+ sans-serif1 :) * V^{T}

ψ goodbreak= sans-serif-italicU ((), : N_{lin} goodbreak+ sans-serif1); N_{lin} goodbreak= sans-serif-italicp sans-seriffor which {CCR}_{p} \geq sans-serif99.95 %

where

ψ

Section: Nonlinear Sensitive Feature Component Using Dpcamentioning

confidence: 99%

“…DPCA [44][45][46][47][48][49][50] is similar to PCA and involves applying a moving time window of fixed size. The DPCA performs PCA only on a subset (windowed) of time series.…”

Section: Dynamic Principal Component Analysismentioning

confidence: 99%