Vibration-based damage detection for a population of nominally identical structures via Random Coefficient Gaussian Mixture AR model based methodology

2021

A novel, unsupervised, hypersphere-based healthy subspace method for robust damage detection under non-quantifiable uncertainty via a limited number of random vibration response sensors is postulated. The method is based on the approximate construction, within a proper feature space, of a healthy subspace representing the healthy structural dynamics under uncertainty as the union of properly selected hyperspheres. This is achieved via a fully automated algorithm eliminating user intervention, and thus subjective selections, or complex optimization procedures. The main asset of the proposed method lies in combining simplicity and full automation with high performance. Its performance is systematically assessed via two experimental case studies featuring various uncertainty sources and distinct healthy subspace geometries, while interesting comparisons with three well-known robust damage detection methods are also performed. The results indicate excellent detection performance, which also compares favorably to that of alternative methods.

Section: Introductionmentioning

confidence: 99%

An automated hypersphere-based healthy subspace method for robust and unsupervised damage detection via random vibration response signals

Kyriakos

Fassois

2021

“…6,8 On the other hand, ''explicit'' methods attempt to directly model the effects of varying EOCs on healthy dynamics via deterministic or stochastic techniques. Representative methods from this category are the global statistical model-based method, 21,22 random coefficient (RC) model-based methods within a Gaussian 23 or non-Gaussian 24,25 framework, and multiple model (MM)-based methods. 6,[26][27][28][29][30] These are often effective, however, they have their own limitations.…”

Section: Introductionmentioning

confidence: 99%

“…For instance, some require the availability of a high number of data records for proper model construction in their training phase 25 -including the construction of a highdimensional probability density function. 6,26,29 Some others require the precise selection of critical parameters by the user, or the solution of a highdimensional and non-convex optimization problem, 25 while other methods require the continuous measurement of the varying EOCs (such as temperature) in the inspection, real-time, phase, which is often not practical or feasible. 31,32 For overcoming such difficulties, the present authors have, in a series of recent studies [33][34][35][36] and within an exclusively data-based (no physics-based models required) statistical time series framework, 37,38 postulated a novel functional model (FM)-based method for robust damage detection under varying and non-measurable EOCs.…”

Section: Introductionmentioning

confidence: 99%

“…6,26–30 These are often effective, however, they have their own limitations. For instance, some require the availability of a high number of data records for proper model construction in their training phase 25 – including the construction of a high-dimensional probability density function. 6,26,29 Some others require the precise selection of critical parameters by the user, or the solution of a high-dimensional and non-convex optimization problem, 25 while other methods require the continuous measurement of the varying EOCs (such as temperature) in the inspection, real-time, phase, which is often not practical or feasible.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the functional model–based method for vibration-based robust damage detection: versions and experimental assessment

Aravanis

Fassois

2020

The problem of random vibration–based robust damage detection for structures operating under varying and non-measurable environmental and operating conditions is considered via a novel unsupervised functional model–based method. Two versions of the method are employed based on either the residual variance or uncorrelatedness (whiteness) of a proper functional model that incorporates the varying environmental and operating conditions in a scheduling vector. This article constitutes a proof-of-concept study in which a comprehensive laboratory assessment of the functional model–based method is undertaken using hundreds of experiments with a composite tail structure of an unmanned aerial vehicle and two early-stage damages under a considerable number of different environmental and operating conditions. Comparisons with two alternative state-of-the-art statistical time series type methods, that is, a multiple model–based method and a principal component analysis–based method, are also performed. The results indicate ideal detection performance for the functional model–based and multiple model–based methods, with the true positive rate reaching 100% at 0% false positive rate, but degraded performance for the pricipal component analysis–based method.

“…The difficulties encountered in such a damage detection problem using conventional response-only statistical time series methods including transmittance function as well, have been demonstrated in the recent studies. 37,38 In addition, a relative damage detection problem has been investigated for five aluminium beams under various damage scenarios and limited variability among the healthy beams in Vanlanduit et al, 39 while the capability of using the natural frequencies, the mode shapes and the FRF magnitude for potential damage detection among two pairs of metallic aircraft parts is explored in Papatheou et al 40,41 Damage detection for a set of significantly non-uniform seemingly identical composite beams has been pursued by the second author and/or co-workers via a series of studies, 38,[42][43][44][45][46][47] where the parameters of multiple and random coefficient AR models are used as the sensitive to damage characteristic quantity, presenting very encouraging results when the proper number of healthy structures are used in the methods' training.…”

Section: Introductionmentioning

confidence: 99%

A transmittance-based methodology for damage detection under uncertainty: An application to a set of composite beams with manufacturing variability subject to impact damage and varying operating conditions

Poulimenos

2018

Oftentimes, the complexity in manufacturing composite materials leads to corresponding structures which although they may have the same design specifications they are not identical. Thus, composite parts manufactured in the same production line present differences in their dynamics which combined with additional uncertainties due to different operating conditions may lead to the complete concealment of effects caused by small, incipient, damages making their detection highly challenging. This damage detection problem in nominally identical composite structures is pursued in this study through a novel data-based response-only methodology that is founded on the autoregressive with exogenous (ARX) excitation parametric representation of the transmittance function between vibration measurements at two different locations on the structure. This is a statistical time series methodology within which two schemes are formulated. In the first, a single-reference transmittance model representing the healthy structure is employed, while multiple transmittance models from a sample of available healthy structures are used in the second. The model residual signal constitutes for both schemes the damage detection characteristic quantity that is used in appropriate hypothesis testing procedures with the likelihood ratio test. The methodology is experimentally assessed via damage detection for a population of composite beams which are manufactured in the same production line representing the half of the tail of a twin-boom unmanned aerial vehicle. The damage detection results demonstrate the superiority of the multiple transmittance models based scheme that may effectively detect damages under significant manufacturing variability and varying boundary conditions.