Fusing damage-sensitive features and domain adaptation towards robust damage classification in real buildings

Martakis, Panagiotis; Reuland, Yves; Stavridis, Andreas; Chatzi, Eleni

doi:10.1016/j.soildyn.2022.107739

Cited by 19 publications

(6 citation statements)

References 78 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Apart from the common system identification in terms of frequencies, modal shapes and damping ratios, transmissibility-based features, as elaborated in [25] will be investigated during the tests and implemented in the damage detection tool. The Transmissibility Assurance Criterion (TAC), which is based on the cross-spectral density between two points, whose response is recorded by accelerometers, as indicated in [13], can be used to evaluate the fitting degree between healthy and investigated (possibly damaged) state in a predefined frequency range. As long as the system remains elastic, they present a strong fitting, whereas in occurrence of damage the deviation increases.…”

Section: Damage Sensitive Indicatorsmentioning

confidence: 99%

“…The latest years, there is an increasing tendency to complete the traditional vibration-based methods with the use of machine learning (ML) applications [12]. This enables the training of different damage-sensitive indicators as classifiers to indicate the presence, location, and severity of structural damage [13] [14]. Within this approach many indicators can be weighted properly, after a training process and result in a multi-parametric identification process.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Shaking Table Tests on Scaled Two-Floor Steel Framed Structures – Design and Preliminary Numerical Investigations

Balaskas,

Vulcu,

Hoffmeister

et al. 2023

Proceedings of the 8th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering (COM

View full text Add to dashboard Cite

Structural Health Monitoring (SHM) can allow for rapid structural damage detection in the aftermath of an earthquake event. Especially, regarding industrial facilities depending on the seismic performance of critical components, automated counter measures could be activated to prevent domino effects and lead to damage mitigation. Real-time seismic damage detection through SHM is a promising, but challenging field. Although various methods, regarding structural damage detection, have been developed the previous years, a holistic approach is still missing. In the framework of an ongoing German nationally funded research project, an interdisciplinary, user-oriented earthquake early warning and rapid response system was developed. Part of the project were shaking table tests on steel space framed test structure. The objective of the investigation was to: (i) validate damage sensitive indicators, arranged based on numerical simulations and literature; (ii) investigate the effectiveness of measuring and identification methods; and (iii) correlate the measured data with the performance of the structure. Output-only and input-output methods in time and frequency domain will be used for the damage detection scheme. Three configurations of steel moment resisting frames (MRF) with different beam-to-column joints are currently being tested. This paper summarizes the preliminary numerical investigations for the design of the test structure. Beam-to-column and column base joints were analyzed using advanced finite element numerical simulations and their properties were implemented in global structural models. Response history analyses were used as a prognosis for the ongoing experimental tests. The instrumentation concept and the experimental program are also outlined in this contribution.

show abstract

Section: Damage Sensitive Indicatorsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Shaking Table Tests on Scaled Two-Floor Steel Framed Structures – Design and Preliminary Numerical Investigations

Balaskas,

Vulcu,

Hoffmeister

et al. 2023

Proceedings of the 8th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering (COM

View full text Add to dashboard Cite

show abstract

“…It is noteworthy to mention the recent research contributions of Reuland et al [21], who conducted an extensive review of data-driven damage indicators for rapid seismic structural health monitoring. Additionally, Martakis et al [22] explored the integration of traditional structural health monitoring techniques with innovative machine learning tools, offering a comprehensive perspective. Moreover, [23,24] provide an extensive review of available methods and case studies related to damage identification in bridge structures.…”

Section: Introductionmentioning

confidence: 99%

The “M and P” Technique for Damage Identification in Reinforced Concrete Bridges

Bakalis,

Makarios,

Lekidis

2024

Infrastructures

View full text Add to dashboard Cite

The seismic damage in reinforced concrete bridges is identified in this study using the “M and P” hybrid technique initially developed for planar frames, where M signifies “Monitoring” and P denotes “Pushover analysis”. The proposed methodology involves a series of pushover and instantaneous modal analyses with a progressively increasing target deck displacement along the longitudinal direction of the bridge. From the results of these analyses, the diagram of the instantaneous eigenfrequency of the bridge, ranging from the health state to near collapse, is plotted against the inelastic seismic deck displacement. By pre-determining the eigenfrequency of an existing bridge along its longitudinal direction through “monitoring and frequency identification”, the target deck displacement corresponding to the damage state can directly be found from this diagram. Subsequently, the damage can be identified by examining the results of the pushover analysis at the step where the target deck displacement is indicated. The effectiveness of this proposed technique is evaluated in the context of straight multiple span bridges with unequal pier heights, illustrated through an example of a four-span bridge. The findings demonstrate that the damage potential in bridge piers can be successfully identified by combining the results of a monitoring process and pushover analysis.

show abstract

“…In order to quantify the value of information extracted from a SHM system to implement it in a decisionmaking tool, studies were conducted by [19], in which a heuristic model was used for life-cycle optimization by the sequential updating of structural reliability based on the identification of deterioration and the estimation of its evolution using a classical Bayesian model updating method. In [20], a framework is proposed to transfer knowledge obtained through synthetic data creation from earthquake simulations for various damage classes to real data with exposure limited to a single health state via a domain adversarial neural network (DANN) architecture. Again, to overcome the limitations imposed by knowledge of the goodness-of-fit class data set alone, the authors of [21] developed a vibrationbased SHM framework for damage classification in structural systems to overcome this limitation.…”

Section: Introductionmentioning

confidence: 99%

Framework for Identification and Prediction of Corrosion Degradation in a Steel Column through Machine Learning and Bayesian Updating

Castelli

Belleri

2023

Applied Sciences

View full text Add to dashboard Cite

In recent years, structural health monitoring, starting from accelerometric data, is a method which has become widely adopted. Among the available techniques, machine learning is one of the most innovative and promising, supported by the continuously increasing computational capacity of current computers. The present work investigates the potential benefits of a framework based on supervised learning suitable for quantifying the corroded thickness of a structural system, herein uniformly applied to a reference steel column. The envisaged framework follows a hybrid approach where the training data are generated from a parametric and stochastic finite element model. The learning activity is performed by a support vector machine with Bayesian optimization of the hyperparameters, in which a penalty matrix is introduced to minimize the probability of missed alarms. Then, the estimated structural health conditions are used to update an exponential degradation model with random coefficients suitable for providing a prediction of the remaining useful life of the simulated corroded column. The results obtained show the potentiality of the proposed framework and its possible future extension for different types of damage and structural types.

show abstract

Fusing damage-sensitive features and domain adaptation towards robust damage classification in real buildings

Cited by 19 publications

References 78 publications

Shaking Table Tests on Scaled Two-Floor Steel Framed Structures – Design and Preliminary Numerical Investigations

Shaking Table Tests on Scaled Two-Floor Steel Framed Structures – Design and Preliminary Numerical Investigations

The “M and P” Technique for Damage Identification in Reinforced Concrete Bridges

Framework for Identification and Prediction of Corrosion Degradation in a Steel Column through Machine Learning and Bayesian Updating

Contact Info

Product

Resources

About