Autoencoder and Mahalanobis distance for novelty detection in structural health monitoring data of an offshore wind turbine.

Weil, Maximillian; Weijtjens, Wout; Devriendt, Christof

doi:10.1088/1742-6596/2265/3/032076

Cited by 4 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…13%. The effect of anomalies on the transfer function |H (1,8)| is shown in Figure 3, we observe that this anomaly mainly affects the last mode(s). Consistent with real-world experience, the variation induced by the latent variable far exceeds the effect of the considered anomalies in the system.…”

Section: System Descriptionmentioning

confidence: 89%

“…Various features have been used for damage detection, such as cross-correlation of measured data [5], modal parameters [6], including those obtained from operational modal analysis (OMA) [7]. Auto-encoders have also been employed for novelty detection in SHM data, utilizing environmental data and modal parameters to train normality models [8].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Population-Based SHM Under Environmental Variability Using a Classifier for Unsupervised Damage Detection

BEL-HADJ,

WEIJTJENS

2023

Proceedings of the 14th International Workshop on Structural Health Monitoring

View full text Add to dashboard Cite

In this paper, we introduce a novel deep learning technique for anomaly detection in the context of Population-Based Structural Health Monitoring (PB-SHM). The proposed method eliminates manual feature engineering by utilizing Power Spectral Density (PSD) as input, allowing examination of the entire spectrum. It is based on an auxiliary classification task, wherein the model is trained to discriminate between different systems according to their dynamic response. The classifier confidence is then used during inference for damage detection. The neural network extracts discriminative features commonly impacted by damage, which are employed to create a normality model. The efficacy of our method is demonstrated on a simulated population of 20 individual 8- DOF systems influenced by a latent environmental variables, emphasizing its potential for PB-SHM under diverse conditions. Our technique achieves performance comparable to resonance frequency-based methods while potentially exhibiting higher capability in complex structures with multiple modes. Anomalies caused by a 5% decrease in stiffness are successfully detected, yielding an AUC of 0.94.

show abstract

Section: System Descriptionmentioning

confidence: 89%

Section: Introductionmentioning

confidence: 99%

Population-Based SHM Under Environmental Variability Using a Classifier for Unsupervised Damage Detection

BEL-HADJ,

WEIJTJENS

2023

Proceedings of the 14th International Workshop on Structural Health Monitoring

View full text Add to dashboard Cite

show abstract

“…The autoencoder is a widely recognized AI technique for detecting anomalies. It has been successfully applied in various domains, such as fault detection in wind turbines [19], monitoring hydroelectric power plants using shallow autoencoders coupled with hotelling control charts [20], detecting anomalies in surveillance videos [21], and identifying anomalies in MRI images [22]. The autoencoder is known for its versatility and reliability, especially when combined with a robust statistical distance.…”

Section: Related Work For Autoencodersmentioning

confidence: 99%

An Artificial Intelligence and Industrial Internet of Things-Based Framework for Sustainable Hydropower Plant Operations

Fera,

Spandonidis

2024

Smart Cities

View full text Add to dashboard Cite

Hydropower plays a crucial role in supplying electricity to developed nations and is projected to expand its capacity in various developing countries such as Sub-Saharan Africa, Argentina, Colombia, and Turkey. With the increasing demand for sustainable energy and the emphasis on reducing carbon emissions, the significance of hydropower plants is growing. Nevertheless, numerous challenges arise for these plants due to their aging infrastructure, impacting both their efficiency and structural stability. In order to tackle these issues, the present study has formulated a specialized real-time framework for identifying damage, with a particular focus on detecting corrosion in the conductors of generators within hydropower plants. It should be noted that corrosion processes can be highly complex and nonlinear, making it challenging to develop accurate physics-based models that capture all the nuances. Therefore, the proposed framework leverages autoencoder, an unsupervised, data-driven AI technology with the Mahalanobis distance, to capture the intricacies of corrosion and automate its detection. Rigorous testing shows that it can identify slight variations indicating conductor corrosion with over 80% sensitivity and a 5% false alarm rate for ‘medium’ to ‘high’ severity damage. By detecting and resolving corrosion early, the system reduces disruptions, streamlines maintenance, and mitigates unscheduled repairs’ negative effects on the environment. This enhances energy generation effectiveness, promotes hydroelectric facilities’ long-term viability, and fosters community prosperity.

show abstract

“…Additionally, the modal frequency resulting from the automated Operational Modal Analysis (OMA) strategy [8], has two main sources of variability; the Environmental and Operational Variability (EOV) and inherent uncertainty from the OMA process [9]. Different strategies exist to normalize for EOV [10,11], but it remains unknown whether the change in frequency (as e.g. caused by a scouring hole) will be hidden by this uncertainty on the tracked modal frequencies.…”

Section: Introductionmentioning

confidence: 99%

Quantifiable Scour Detection for Offshore Wind Turbines Using Resonance Frequency Monitoring and a Digital Twin

WEIL,

JURADO,

CHAAR

et al. 2023

Proceedings of the 14th International Workshop on Structural Health Monitoring

View full text Add to dashboard Cite

This paper presents a methodology for enhancing frequency-based Structural Health Monitoring (SHM) of Offshore Wind Turbines (OWTs) by incorporating digital twin models. The proposed methodology utilizes an automated operational modal analysis to extract modal frequencies from acceleration measurements. The environmental and operational variability is removed through tree-based regression models, which can adequately recognize the different operational states of the OWT. The proposed data-driven approach relies on the detection of the modal frequency changes caused by a damaging scenario in the OWT (e.g. scouring). The digital twin model accounts for the actual properties of the substructure, considering the specific foundation geometry and soil conditions at the monitored OWT location. A statistical control chart is used to identify the possible scour depth by comparing the changes in the normalized modal frequency and the digital-twin calculations. The methodology is demonstrated on 4 months of measurements from a full-scale monitored OWT. The proposed methodology advances the SHM capabilities from damage detection to damage quantification, enabling improved operations & maintenance strategies for OWTs.

show abstract

Autoencoder and Mahalanobis distance for novelty detection in structural health monitoring data of an offshore wind turbine.

Cited by 4 publications

References 9 publications

Population-Based SHM Under Environmental Variability Using a Classifier for Unsupervised Damage Detection

Population-Based SHM Under Environmental Variability Using a Classifier for Unsupervised Damage Detection

An Artificial Intelligence and Industrial Internet of Things-Based Framework for Sustainable Hydropower Plant Operations

Quantifiable Scour Detection for Offshore Wind Turbines Using Resonance Frequency Monitoring and a Digital Twin

Contact Info

Product

Resources

About