Modeling of quasi-static thrust load of wind turbines based on 1 s SCADA data

Noppe, Nymfa; Weijtjens, Wout; Devriendt, Christof

doi:10.5194/wes-3-139-2018

Cited by 16 publications

(14 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NCA is a non-parametric classification model used for metric learning and linear dimensionality reduction (Goldberger et al, 2004). It is based on a modelling technique known as k-nearest neighbours (k-NNs), which is a supervised learning algorithm used for classification or regressions (Han and Kamber, 2011;Parsian, 2015). In its simplest form, the k-NN approach looks for the closest k = 1 observation to the query observation x q within the training dataset by measuring the distances to the neighbouring data points and selecting the one that satisfies the min i distance (x i , x q ).…”

Section: Dependent Variablesmentioning

confidence: 99%

See 1 more Smart Citation

Feature selection techniques for modelling tower fatigue loads of a wind turbine with neural networks

2021

View full text Add to dashboard Cite

Abstract. The rapid development of the wind industry in recent decades and the establishment of this technology as a mature and cost-competitive alternative have stressed the need for sophisticated maintenance and monitoring methods. Structural health monitoring has risen as a diagnosis strategy to detect damage or failures in wind turbine structures with the help of measuring sensors. The amount of data recorded by the structural health monitoring system can potentially be used to obtain knowledge about the condition and remaining lifetime of wind turbines. Machine learning techniques provide the opportunity to extract this information, thereby improving the reliability and cost-effectiveness of the wind industry as well. This paper demonstrates the modelling of damage-equivalent loads of the fore–aft bending moments of a wind turbine tower, highlighting the advantage of using the neighbourhood component analysis. This feature selection technique is compared to common dimension reduction/feature selection techniques such as correlation analysis, stepwise regression, or principal component analysis. For this study, recordings of data were gathered during approximately 11 months, preprocessed, and filtered by different operational modes, namely standstill, partial load, and full load. The results indicate that all feature selection techniques were able to maintain high accuracy when trained with artificial neural networks. The neighbourhood component analysis yields the lowest number of features required while maintaining the interpretability with an absolute mean squared error of around 0.07 % for full load. Finally, the applicability of the resulting model for predicting loads in the wind turbine is tested by reducing the amount of data used for training by 50 %. This analysis shows that the predictive model can be used for continuous monitoring of loads in the tower of the wind turbine.

show abstract

Section: Dependent Variablesmentioning

confidence: 99%

“…In line with the findings of Ziegler et al (2018), other authors have worked on the aforementioned data-driven approaches. Noppe et al (2018), for example, reconstructed the thrust loads history of a WT based on both simulated and measured SCADA data. The data gathered corresponded to operational 1 s and 10 min.…”

Section: Introductionmentioning

confidence: 99%

Feature selection techniques for modelling tower fatigue loads of a wind turbine with neural networks

2021

View full text Add to dashboard Cite

show abstract

“…To correctly assess the wave loads, these different loads need to be accurately separated from each other. One possible additional input is to use the turbine's operational (SCADA) data to estimate the wind-induced loads [32] in parallel to the wave load estimation.…”

Section: Future Workmentioning

confidence: 99%

Inverse estimation of breaking wave loads on monopile wind turbines

et al. 2018

View full text Add to dashboard Cite

Breaking waves occur in shallow water and cause large forces on offshore wind turbines, which can determine the design of the support structure. Due to the strong nonlinearity and statistical variability of breaking waves, the corresponding loads used in the design are generally characterized by large uncertainty. This paper presents a validation of an inverse method to estimate breaking wave loads on offshore monopile wind turbines based on wave basin tests. The accelerations measured on a scale model of a monopile wind turbine are passed to a force identification algorithm together with a dynamic model of the structure in order to identify the wave loads. In addition, the member forces in the support structure are estimated. Both the wave loads and member forces are estimated with a reasonable accuracy. The proposed methodology creates opportunities for operational monitoring of offshore wind turbines, hereby providing essential information to improve design guidelines for future wind turbines and allowing for a continuous fatigue assessment of the wind turbine during the operation.

show abstract

“…In line with the findings of Ziegler et al (Ziegler et al, 2018), other authors have worked on the aforementioned data-driven approaches. Noppe et al (Noppe et al, 2018), for example, reconstructed the thrust loads history of a WT based on both simulated and measured SCADA data. The data gathered corresponded to operational 1 s and 10 min data for 2.5 months.…”

Section: Introductionmentioning

confidence: 99%

Modelling tower fatigue loads of a wind turbine using data mining techniques on SCADA data

Movsessian¹,

Schedat²,

Faber³

2020

Preprint

View full text Add to dashboard Cite

Abstract. The rapid development of the wind industry in recent decades and the establishment of this technology as a mature and cost-competitive alternative have stressed the need for sophisticated maintenance and monitoring methods. Structural health monitoring has risen as a diagnosis strategy to detect damage or failures in wind turbine structures with the help of measuring sensors. The amount of data recorded by the structural health monitoring system can potentially be used to obtain knowledge about the condition and remaining lifetime of wind turbines. Machine learning techniques provide the opportunity to extract this information, thereby improving the reliability and cost-effectiveness of the wind industry as well. This paper demonstrates modeling damage equivalent loads of the fore-aft bending moments of a wind turbine tower with the advantage of using the neighborhood component analysis as a feature selection technique in comparison to common dimension reduction/feature selection techniques such as correlation analysis, stepwise regression or principal component analysis. For this study a one-year measuring period of data was gathered, pre-processed, and filtered by different operational modes, namely stand still, full load, and partial load. Finally, a sensitivity analysis was performed in the partial load model to determine the required length of the data collection campaign that guarantees the most precise results. The results indicate that applying neighborhood component analysis yields more conservative models regarding the number of features and equally accurate outcomes than traditional feature selection techniques.

show abstract

Modeling of quasi-static thrust load of wind turbines based on 1 s SCADA data

Cited by 16 publications

References 9 publications

Feature selection techniques for modelling tower fatigue loads of a wind turbine with neural networks

Feature selection techniques for modelling tower fatigue loads of a wind turbine with neural networks

Inverse estimation of breaking wave loads on monopile wind turbines

Modelling tower fatigue loads of a wind turbine using data mining techniques on SCADA data

Contact Info

Product

Resources

About