On noise covariance estimation for Kalman filter-based damage localization

Wernitz, Stefan; Chatzi, Eleni; Hofmeister, Benedikt; Wolniak, Marlene; Shen, Wanzhou; Rolfes, Raimund

doi:10.1016/j.ymssp.2022.108808

Cited by 14 publications

(4 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In practice, collecting damage labels that describe locations of damage is expensive and often infeasible, making automatic damage localisation using a data-based approach to SHM rather challenging. In SHM, most data-driven damage localisation methods typically require a network of well-placed sensors close to the damage location in the structure of interest (Stubbs et al, 1995; Manson et al, 2003; Chesné and Deraemaeker, 2013; Wernitz et al, 2022). This approach can be expensive and time-consuming to implement.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Damage localisation using disparate damage states via domain adaptation

Wickramarachchi,

Gardner,

Poole

et al. 2024

DCE

Self Cite

View full text Add to dashboard Cite

A significant challenge of structural health monitoring (SHM) is the lack of labeled data collected from damage states. Consequently, the collected data can be incomplete, making it difficult to undertake machine learning tasks, to detect or predict the full range of damage states a structure may experience. Transfer learning is a helpful solution, where data from (source) structures containing damage labels can be used to transfer knowledge to (target) structures, for which damage labels do not exist. Machine learning models are then developed that generalize to the target structure. In practical applications, it is unlikely that the source and the target structures contain the same damage states or experience the same environmental and operational conditions, which can significantly impact the collected data. This is the first study to explore the possibility of transfer learning for damage localisation in SHM when the damage states and the environmental variations in the source and target datasets are disparate. Specifically, using several domain adaptation methods, this article localizes severe damage states at a target structure, using labeled information from minor damage states at a source structure. By minimizing the distance between the marginal and conditional distributions between the source and the target structures, this article successfully localizes damage states of disparate severities, under varying environmental and operational conditions. The effect of partial and universal domain adaptation—where the number of damage states in the source and target datasets differ—is also explored in order to mimic realistic industrial applications of these methods.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Standard SHM practices require a network of sensors placed in strategic positions in order to locate damage on a structure (Stubbs et al, 1995; Manson et al, 2003; Chesné and Deraemaeker, 2013; Wernitz et al, 2022). Consequently, large amounts of well-placed sensors and collected data are required for damage location.…”

Section: Introductionmentioning

confidence: 99%

Damage localisation using disparate damage states via domain adaptation

Wickramarachchi,

Gardner,

Poole

et al. 2024

DCE

Self Cite

View full text Add to dashboard Cite

show abstract

“…Among the most commonly used estimators in this field, one can identify the Kalman filter (KF), the Extended KF (EKF), the Particle KF (PKF), and the Dual KF (DKF) while a combination of those estimators has also been used. For example, the authors of [3] developed a hybrid framework for simultaneously tracking states and estimating system parameters. This hybrid approach combined two filters, namely the EKF and the PKF.…”

Section: Introductionmentioning

confidence: 99%

Application of fixed-lag Kalman smoother for nonlinear state estimation

Al-Hagri,

Caglio,

Stang

et al. 2024

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

During their lifetime, civil structures and infrastructures are subjected to environmental and human-induced hazards that may lead to structural damages or, in some cases, to full collapse. Such failures are associated with increased risk for life and limb, while also environmental and monetary losses may be triggered with adverse societal footprint. Consequently, it is highly relevant to screen reliably the condition of the structural systems and, if necessary, to intervene to safeguard the structural integrity and avoid undesirable failures. To this end, several structural health monitoring schemes are becoming a key element for various structures and infrastructure systems, enabling the continuous assessment of their state. The latter, in turn, can allow for detecting damages at an early age, avoiding extensive failures or even sudden collapse. In order to identify the failures due to excessive loading events, measured time-varying responses are frequently used as the basis for damage detection. However, in reality, practical restrictions exist in mounting sensors on all the locations of the structure that an engineer would prefer to know the structural response. For example, sensing data from the submerged part of an offshore platform is commonly not available, while the latter is also valid for various locations in bridges where the access is limited to apply and operate sensors. Therefore, the available measured data is mostly incomplete since only a limited number of sensors can be employed. Additionally, various noise sources usually pollute the sparsely measured responses, and the overall quality is compromised. As a remedy to this challenge, a new framework is proposed herein to aid in estimating reliably the nonlinear dynamic response of a structural system with the use of limited and noisy measurements. Especially, a Kalman smoothing technique called the fixed-lag algorithm is coupled with an augmented extended Kalman filter, and such an integrated scheme is anticipated to increase the state estimation accuracy. The current study adopts a simulation framework as the testbed for the proposed scheme. The use of the FE software OpenSees allows the numerical nonlinear modeling of the structural system subjected to time-varying loads. Compared with the ones calculated directly from nonlinear response history analysis, the estimated responses based on the proposed scheme are found to be associated with increased accuracy.

show abstract

“…Correspondingly, diferent kinds of data-driven damage-sensitive features (DSFs) have been formulated [24], including the popular modal property-based methods [25][26][27] or more indirect signal processing-based methods [28][29][30]. On the other hand, time-domain damage identifcation schemes exist, on the basis of which acceleration and displacement-based damage indices (DI) [31][32][33] can be derived. Finally, traditional approaches, such as ftting of typical constitutive models, such as the Park-Ang model [34], have also been utilized for decades.…”

Section: Introductionmentioning

confidence: 99%

Data-Driven Damage State Assessment of RC Shear Walls with Experimental Validation

Zhang

Shi

Shan

et al. 2023

Structural Control and Health Monitoring

Self Cite

View full text Add to dashboard Cite

This paper introduces a probability-based damage state evaluation methodology for shear walls deriving from a data-driven calculation. A previously proposed damage quantification index, formulated in the time domain, which is capable of tracking damage progression based on the availability of structural monitoring data, is adopted here for the quantification of structural hysteresis damages. The probability of the structure lying in a specific damage state is determined on the basis of the derived damage index and the limit state definitions, yielding valuable information for postearthquake decisions. In this study, a database of the hysteretic behavior of 1,000 shear walls, considering different structural parameters, is generated utilizing OpenSees. Four limit states of seismic performance are defined based on material properties and in relation to the simulated stress and strain data. Accordingly, the exceeding probabilities of damage can be estimated by fitting statistical models to the damage index values grouped in terms of the axial load ratio. Followingly, an informative mapping, between the monitoring-derived damage index and postearthquake damage levels, is established considering structural uncertainty. Illustrative examples, including two shear walls subjected to cyclic loading and a seven-story shear wall slice subjected to a shaking table test, are investigated to show the capability and feasibility of the proposed performance evaluation method on damage state evaluation.

show abstract

On noise covariance estimation for Kalman filter-based damage localization

Cited by 14 publications

References 29 publications

Damage localisation using disparate damage states via domain adaptation

Damage localisation using disparate damage states via domain adaptation

Application of fixed-lag Kalman smoother for nonlinear state estimation

Data-Driven Damage State Assessment of RC Shear Walls with Experimental Validation

Contact Info

Product

Resources

About