Pasar al contenido principal

DESARROLLO Y VALIDACION DE ESTRATEGIAS DE APRENDIZAJE PROFUNDO Y AUTOMATICO PARA EL MANTENIMIENTO PREDICTIVO Y DETECCION TEMPRANA DE DAÑOS ESTRUCTURALES EN AEROGENERADORES

PID2021-122132OB-C21

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprograma Subprograma Estatal de Generación de Conocimiento
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2021
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITECNICA DE CATALUNYA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 54
Encontrada(s) 1 página(s)

UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Noncontact fault diagnosis of electrical equipment using modified multiscale two-dimensional color distribution entropy and thermal imaging

  • Wang, Shun|||0000-0002-5766-6220
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
Effective health monitoring of electrical equipment is critical for industrial reliability. Although infrared thermal imaging offers a powerful noncontact diagnostic method, accurately interpreting its complex and often noisy thermal patterns remains a significant challenge. Entropy-based analysis is well suited for quantifying this complexity, but its application to images has been limited. Existing two-dimensional entropy methods are not only less developed than their one-dimensional counterparts but also typically require converting thermal images to grayscale, which discards vital diagnostic information from color channels. To overcome these limitations, this study introduces the modified multiscale two-dimensional color distribution entropy (MMCDEn_2D). This novel method directly integrates the attributes of the RGB, preserving a richer feature set for analysis. The effectiveness of the proposed method is demonstrated first through synthetic signals, showing low sensitivity to image size and high computational efficiency. The study further extends the application of entropy-based analysis to noncontact health monitoring scenarios, implementing MMCDEn_2D for thermal image-based fault diagnosis of induction motors and power transformers. The method achieves a diagnostic accuracy that exceeds 95%, significantly outperforming traditional approaches. Crucially, it demonstrates superior robustness in challenging scenarios, improving accuracy by 2%–5% under high-noise conditions and with small sample sizes. These results establish MMCDEn_2D as a highly effective and reliable tool to advance noncontact fault diagnosis in critical electrical equipment., This work was partially funded by the Spanish Agencia Estatal de Investigación (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER) through​ the research projects PID2021-122132OB-C21 and TED2021-129512B-I00, and by the Generalitat de Catalunya through the research project 2021-SGR-01044.
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Two-dimensional color distribution entropy: validation and application on non-contact fault diagnosis for induction motor

  • Wang, Shun|||0000-0002-5766-6220
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
© 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Entropy-based methods are commonly used in the f ield of vibration signal processing and fault diagnosis due to their simplicity, ease of implementation, and strong feature extraction capabilities. Despite the successful application of 1D entropy measures in vibration signal analysis, there is a gap in 2D entropy application to images, especially for color images. Therefore, this study introduces a novel concept called two-dimensional color distribution entropy (CDEn2D) and further develops its multiscale variant, modified multiscale two-dimensional color distribution entropy (MMCDEn2D). This work investigates the application of MMCDEn2D in non-contact fault diagnosis for induction motors using infrared thermal imaging. The experimental results demonstrate that MMCDEn2D outperforms traditional two-dimensional entropy methods. Moreover, MMCDEn2D shows significant reliability and superiority in scenarios with limited data samples. These findings highlight its effectiveness in the f ield of non-contact health monitoring and fault diagnosis., This work is partially funded by the Spanish Agencia Estatal de Investigaci´ on (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research project 2021-SGR-01044., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Recent advances in wind turbine condition monitoring using SCADA data: a state-of-the-art review

  • Wang, Shun|||0000-0002-5766-6220
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
With the rapid increase worldwide in the number of installed wind turbines, as an increasingly vital role in the global energy landscape, requirements and expenses of maintenance have also increased significantly. This surge in wind turbine installations worldwide has escalated maintenance demands and costs, making condition monitoring an essential focus for research and operational optimization. Built-in supervisory control and data acquisition (SCADA) systems could offer a cost-effective approach to condition monitoring in wind turbines, using existing infrastructure to reduce the need for additional sensors and simplifying system complexity. This review critically surveys state-of-the-art SCADA-based condition monitoring techniques for wind turbines. The paper first examines essential data preprocessing strategies tailored for SCADA data challenges. Subsequently, key monitoring methodologies are reviewed, including model-based and data-driven paradigms, analyzing their applications, strengths, and limitations. Finally, the challenges and future trends of SCADA-based wind turbine condition monitoring are discussed. This comprehensive overview serves as a valuable reference for researchers and industry practitioners seeking to enhance wind turbine reliability and operational efficiency through advanced monitoring techniques., This work is partially funded by the Spanish Agencia Estatal de Investigación (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER), Spain through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00, and by the Generalitat de Catalunya, Spain through the research project 2021-SGR-01044., 7 - Energia Assequible i No Contaminant, 7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el conjunt de fonts d’energia, 7.3 - Per a 2030, duplicar la taxa mundial de millora de l’eficiència energètica, 8 - Treball Decent i Creixement Econòmic, 11 - Ciutats i Comunitats Sostenibles, 13 - Acció per al Clima
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Detection of crack bar deterioration at offshore wind turbine supports using generative adversarial networks and autoencoders

  • Prieto-Galarza, Ricardo
  • Tutivén Gálvez, Christian|||0000-0001-6322-4608
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
This work focuses on the structural health monitoring of jacket-type foundations used by offshore wind turbines. A vibration-only response mechanism based on accelerometer data is specifically suggested. This work, based on an anomaly detection model, has developed a methodology to detect crack bar deterioration in the wind turbine jacket consisting of two training phases with only healthy data: training of a generative adversarial network (GAN), and encoder training of an autoencoder based on the GAN model that has already been learned. A generator and a critic may be obtained through the GAN network training process. This model is used to train an encoder that allows the mapping of healthy samples to a latent vector. Following encoder training, the encoder places the data at points in latent space that correspond to the input data’s healthy state. The mapping of the input sample space to the latent space through the encoder and the subsequent mapping of the latent space to the initial space through the generator should closely resemble the input sample in the event of a healthy input. However, when damage-state input samples are used, the model output does not resemble the input. The reconstruction error and a comparison of the residual error of the critic properties are the final two loss functions used for the identification of anomalies. The proposed strategy has been tested through laboratory experiments on a down-scaled model., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Wind turbine gearbox early fault detection using Mel-frequency cepstral coefficients of vibration data

  • Velandia Cardenas, Cristian|||0000-0002-2130-7488
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
This is the peer reviewed version of the following article: Velandia-Cardenas, C.; Vidal, Y.; Pozo, F. Wind turbine gearbox early fault detection using Mel-frequency cepstral coefficients of vibration data. "Structural control and health monitoring", 6 Juny 2024, vol. 2024, núm. article 733730, which has been published in final form at [Link to final article using the DOI]. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions. This article may not be enhanced, enriched or otherwise transformed into a derivative work, without express permission from Wiley or by statutory rights under applicable legislation. Copyright notices must not be removed, obscured or modified. The article must be linked to Wiley’s version of record on Wiley Online Library and any embedding, framing or otherwise making available the article or pages thereof by third parties from platforms, services and websites other than Wiley Online Library must be prohibited., A methodology utilizing vibration data and Mel-frequency cepstral coefficients (MFCCs) for wind turbine condition monitoring is developed to detect incipient faults in the wind turbine gearbox. This approach provides a more efficient and cost-effective solution compared to traditional condition monitoring techniques relying on physical inspections, which can be time-consuming and labor-intensive. The use of vibration data enables the identification of subtle changes in a wind turbine’s operating condition, providing early warning signs of potential issues. When the vibration data are analyzed, changes in frequency and amplitude can be detected, indicating the presence of a developing fault. Vibration-based condition monitoring systems (CMS) have already been widely used in the wind industry (mainly in new turbines). These systems utilize basic standard features, working in either the time or frequency domain, and are not optimized for nonstationary signals. In contrast, this work focuses on MFCCs, operating in both time and frequency domains, enabling the extraction of adequate information from nonstationary signals. The MFCCs are derived from vibration data signals, providing a compact representation for a more efficient analysis. These coefficients create a fingerprint of the wind turbine operating condition, compared to known healthy conditions, to identify anomalies. To underscore the practical value of this study, it is important to highlight the significant implications for the wind energy sector. The methodology developed offers an advanced, predictive tool for the early detection of gearbox faults, which is a critical aspect of optimizing the performance and longevity of wind turbines. By enabling earlier, more accurate fault detection, the proposed approach significantly reduces the likelihood of catastrophic failures and extensive downtime. This not only enhances the reliability and cost-effectiveness of wind energy systems but also contributes to sustainable energy practices by optimizing resource use and minimizing maintenance costs. The results strongly suggest that the proposed methodology is highly effective in detecting incipient faults in the wind turbine gearbox. By providing early warnings of damage, operators can address issues before significant downtime or damage occurs. The use of MFCCs offers additional benefits since data can be collected remotely, eliminating physical inspections. Analysis can be performed faster, even in real time, allowing more frequent monitoring. This provides a more complete and accurate picture of the health of the system. The approach is tested in the EISLAB dataset concerning vibration signals from six wind turbines in northern Sweden, all with three-stage gearboxes. All measurement data correspond to the axial direction of an accelerometer in the output shaft bearing housing of each turbine., This work was partially funded by the Spanish Agencia Estatal de Investigación (AEI), Ministerio de Economía, Industria y Competitividad (MINECO), the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512BI00, and the Generalitat de Catalunya through the research project 2021-SGR-01044. Open access funding was enabled and organized by CRUE-CSUC Gold., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.3 - Per a 2030, duplicar la taxa mundial de millora de l’eficiència energètica, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnolo­gies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Wind turbine modeling, Maximum Power Point Tracking (MPPT), and experimental validation

  • Eben Zaid, Ichrak
  • Stallon, Daison
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Boussada, Moez
  • Said Nouri, Ahmed
The research presented is driven by the global increase in wind power capacity and the commitment of the scientific community to facilitate its integration into electrical grids. The focus of this study is the modeling of a wind turbine system, beginning with its mechanical components. To ensure the production of power at optimal levels, a control strategy for Maximum Power Point Tracking (MPPT) based on Optimal Torque (OT) has been adopted. The model and control method, developed in Matlab/Simulink, have demonstrated their precision and efficacy through experimental verification using SCADA data acquired from an operational wind turbine., Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Enhancing wind turbine diagnostics with SCADA-vibration fusion, contrastive learning, and linear predictive coefcients

  • Velandia Cardenas, Cristian|||0000-0002-2130-7488
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
Wind energy plays a pivotal role in the transition to sustainable power generation. However, maintaining the reliability and efficiency of wind turbine (WT) remains a significant challenge due to complex operational conditions and the high cost associated with unexpected failures. Effective condition monitoring (CM) and predictive maintenance (PM) strategies are critical to mitigate these risks. This study presents a data-driven fault detection framework that fuses supervisory control and data acquisition (SCADA) data with high-frequency vibration signals using deep learning techniques to enhance diagnostic performance. Unlike conventional normal behavior models that rely exclusively on healthy data for training, the proposed framework incorporates limited labeled fault data when available. As only a few types of faults and a few samples are typically available in real-world scenarios, the approach does not assume a complete representation of all possible fault conditions. Instead, it is designed to generalize beyond the specific faults seen during training. This is demonstrated by training the model on healthy conditions and only two known fault types (with labeled data available) and testing it on a third, previously unseen fault type. In particular, Siamese networks with contrastive and reconstruction learning are employed to improve feature representation and anomaly detection. Two distinct methodologies are compared: the first utilizes a binary cross-entropy (BCE) loss function to classify the healthy or faulty status of the WT, while the second uses a triplet loss function for multiclass representation learning. Both methodologies generate low-dimensional representations of the input features, also known as embeddings. The resulting feature embeddings are passed through a k-means clustering algorithm to improve fault separation and identification. Statistical features are extracted from SCADA data to capture key trends or event information, while the linear prediction coefficient (LPC) method, which models a signal by predicting future values based on its past samples, is applied to the vibration data for better fault characterization. The proposed approach is evaluated using the publicly available ETH Zurich dataset from an Aventa AV-7 turbine. Experimental results indicate that the fusion of SCADA and vibration-based diagnostics, in combination with contrastive and representation learning, substantially improves the predictive accuracy and generalization of fault detection models., This work was partially funded by the Spanish Agencia Estatal de Investigación (AEI), Ministerio de Economía, Industria y Competitividad (MINECO); the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research project 2021-SGR-01044. Open Access funding was enabled and organized by CRUE-CSUC Gold., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Wind turbine gearbox condition monitoring through the sequential analysis of industrial SCADA and vibration data

  • Castellani, Francesco
  • Natili, Francesco
  • Astolfi, Davide
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
The operation & maintenance expenditure for a wind farm project can reach the impressive share of 30% of the total costs. This matter of fact motivates the need for optimal operation & maintenance, which is estimated to provide up to a 10% of energy production improvement. Such potential benefit can only be achieved through an efficient condition monitoring and predictive maintenance strategy. Based on these motivations, this paper presents a real-world case study in which standard diagnostic techniques failed to detect severe faults in the planetary stage of a wind turbine gearbox in time to prevent prolonged downtime. To address this issue, a measurement data processing and fusion algorithm is developed. The approach is capable of leveraging all the information from different data sources (with low to high time resolution) using different machine and deep learning algorithms, connected between them in cascade. This enables the detection of the fault some weeks in advance, compared to the commonly used methods and with lower-level processing of industrial operational data. A qualifying feature of the proposed workflow is that it enables the identification of the faulty component, which is a well known critical point in real-world applications., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Strain virtual sensor for offshore wind turbine jacket supports: a time series transformer approach validated with Alpha Ventus wind farm data

  • Encalada-Dávila, Ángel
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Tutivén Gálvez, Christian|||0000-0001-6322-4608
© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/), Monitoring the structural health of offshore wind turbines in harsh marine environments presents significant challenges due to physical sensor limitations, such as susceptibility to high salinity, strong underwater currents, and the resulting durability and maintenance issues. These challenges increase the need for robust and cost-effective alternatives. Motivated by the urgent demand for sustainable and reliable solutions in offshore wind energy, this study introduces a virtual strain sensor based on the PatchTST architecture, a time series transformer designed to capture local and global patterns in multivariate data. By addressing the limitations of conventional physical sensors, the proposed approach uses advanced deep learning techniques to estimate strain accurately in wind turbine jacket supports. The method involves pre-training the model with self-supervised learning and fine-tuning it using real-world strain data from the Alpha Ventus offshore wind farm. Extensive evaluations using statistical metrics (Pearson correlation, root mean square error) and mechanics-based metrics (normalized fatigue damage error and equivalent strain range) demonstrate the virtual sensor’s superior predictive capabilities. Key results reveal that the model outperforms conventional approaches in reliability, adaptability, and computational efficiency, offering an innovative tool for improving structural health monitoring systems. These outcomes support the deployment of virtual sensors as a durable, cost-effective alternative to physical sensors, contributing to enhanced sustainability and efficiency in offshore wind energy systems., APC – UPC (Yolanda Vidal), Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Wind turbine modeling, maximum power point tracking (MPPT), and experimental validation

  • Eben Zaid, Ichrak
  • Samuel Raj, Daison Stallon|||0000-0002-1339-6919
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Boussada, Moez
  • Said Nouri, Ahmed
The research presented is driven by the global increase in wind power capacity and the commitment of the scientific community to facilitate its integration into electrical grids. The focus of this study is the modeling of a wind turbine system, beginning with its mechanical components. To ensure the production of power at optimal levels, a control strategy for Maximum Power Point Tracking (MPPT) based on Optimal Torque (OT) has been adopted. The model and control method, developed in Matlab/Simulink, have demonstrated their precision and efficacy through experimental verification using SCADA data acquired from an operational wind turbine., This work was supported by the Ministry of Higher Education and Scientific Research-Tunisia;by the Spanish Agencia Estatal de Investigaci\'on (AEI) -Ministerio de Econom\'ia, Industria y Competitividad (MINECO), and the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research project 2021-SGR-01044., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Condition monitoring of wind turbine drivetrains: state-of-the-art technologies, recent trends, and future outlook

  • Kestel, Kayacan
  • Chesterman, Xavier
  • Zappalá, Donatella
  • Watson, Simon
  • Li, Mingxin
  • Hart, Edward
  • Carroll, James
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
As global wind capacity expands, reducing operations and maintenance costs is critical to lowering the levelized cost of energy. This paper explores the state of the art in condition monitoring and prognostic strategies for wind turbine drivetrains, which are among the most failure-prone and maintenance-intensive subsystems. Current diagnostic methodologies are evaluated, covering supervisory control and data acquisition (SCADA) data, high-frequency vibration and acoustic analysis, machine learning and digital twin frameworks. Finally, practical challenges are identified that limit wide-scale industrial adoption, in order to guide future research and industrial efforts., This research has been supported by the Agentschap Innoveren en Ondernemen (grant no. HBC.2023.0702); the Agentschap Innoveren en Ondernemen (grant no. HBC.2024.0130); COST (action CA23155); the Research Association for Drive Technology FVA (project FVA 878 II); the German Federal Ministry for Economic Affairs and Climate Action (project CyPhEs-Sens, KK5211204); the EPSRC (fellowship EP/W037009/1); MCIN/AEI/10.13039/5011000110-33 and the European Union (grants PID2021-122132OBC21 and TED2021-129512B-I00); the Generalitat de Catalunya (grant 2021-SGR-01044); and the Dutch Research Council NWO (grant KICH1.ED02.20.004)., 7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnologies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, 7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Challenges on prognostics and health management for wind turbine components

  • Cuesta, Jokin
  • Leturiondo, Urko
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
This paper explores the applicability of prognostics and health management (PHM) for wind turbines (WTs), presenting the PHM approach along with challenges and opportunities in the context of WT components. First, the PHM framework is introduced, consisting of three blocks: observation, analysis, and action. Critical components and failure modes for WTs are identified, and data acquisition strategies using supervisory control and data adquisition (SCADA) and condition monitoring (CM) data are discussed. Prognostics, specifically remaining useful life (RUL) estimation, employs physics model-based, data-driven, and hybrid models. Finally, challenges and opportunities related to data, analysis and CM, and developing RUL prediction models have been found. Data challenges include data standardization, limited public datasets, and data quality issues. Analysis and CM challenges address new sensorless and non-intrusive techniques, as well as the fusion of data sources. Prognostics model challenges involve uncertainty management, interpretability issues, and the need for online updates. Addressing challenges requires incorporating physical knowledge, utilizing transfer learning, and improving online RUL prediction methods., Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Data-driven unsupervised stator interturn short-circuit fault detection for doubly fed induction generator wind turbines

  • Espinoza, Victor
  • Benalcazar-Parra, Carlos
  • Tutivén Gálvez, Christian|||0000-0001-6322-4608
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
Inter-turn short-circuit (ITSC) faults in doubly fed induction generators (DFIGs) pose a significant challenge for reliable wind turbine operation. Existing fault–detection approaches often rely on data that neglect key aeroelastic, electromechanical, control, and grid interactions, limiting their applicability to real operating conditions. To address this limitation, this study develops a coupled OpenFAST-Simulink simulation framework capable of generating physically realistic datasets of healthy operation and ITSC faults across multiple severity levels. Leveraging this data, an unsupervised DL fault-detection scheme based on an autoencoder is proposed, trained exclusively on healthy data and evaluated using faulty scenarios. The results show effective detection of medium and high severity faults, with longer detection times as severity decreases due to weaker system disturbances., This work is partially funded by grant PID2021-122132OB-C21 funded by MCIN/AEI/10. 13039/501100011033 and by “ERDF A way of making Europe”, by the “European Union”., 7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnologies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, 7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Digital twin technology in wind turbine components: A review

  • Leon Medina, Jersson Xavier|||0000-0002-9198-1996
  • Tibaduiza Burgos, Diego Alexander
  • Parés Mariné, Núria|||0000-0002-2914-9904
  • Pozo Montero, Francesc|||0000-0001-8958-6789
The industrial development, the advances in sensor technology and the processing of large amounts of data, have enabled the training and testing of artificial intelligence models that reproduce, with high accuracy, the behavior of some variables of interest. With the consolidation of the big data era and the proliferation of sensors that can acquire information directly from various components of a wind turbine (WT), a digital twin (DT) allows to close the gap between the physical and the digital worlds. It combines historical data, sensor readings, machine learning and physics-based modeling to replicate the behavior of the physical component accurately. This DT can simulate the performance and behavior of the physical object under different conditions and situations, allowing for predicting failures in WT components and determining their remaining useful life. This review describes the existing literature related to the use of DTs and their developments for WT applications and their components in onshore and offshore applications. This review explores various types of DTs and their approaches, aiming to cover different methods of data processing and concepts related to each DT framework. In addition, it identifies insights from various studies and reviews, particularly focusing on the components of WTs., This work has been partially funded by the Spanish Agencia Estatal de Investigación (AEI)—Ministerio de Economía, Industria y Competitividad (MINECO), and the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21, PID2021-126051OB-C44 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research projects 2021-SGR-01044 and 2021-SGR-01049. The authors also express their gratitude to Fundación Carolina and the Universidad Nacional de Colombia for supporting Diego Tibaduiza’s postdoctoral research stay through the program ‘‘Convocatoria de becas para estancias cortas postdoctorales 2023–2024 de la Fundación Carolina’’., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Robust unsupervised gearbox fault detection under variable-speed operating conditions

  • Wang, Shun|||0000-0002-5766-6220
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Zhang, Jiacong
  • Pozo Montero, Francesc|||0000-0001-8958-6789
Gear transmission systems in wind turbines and other rotating energy machinery operate under variable rotational speeds, where vibration responses are governed by speed-dependent kinematic frequency variations. This speed-induced transformation challenges the construction of condition-consistent vibration representations and undermines reliable AI-enabled condition monitoring. To address this limitation, this study proposes a mechanics-consistent speed-invariant autoencoder (SiAE) for unsupervised fault detection using only healthy-condition data. Instead of treating speed variation as a generic statistical distribution shift, the proposed framework embeds gearbox kinematics into representation learning. First, the squared envelope order spectrum is used to map vibration signals into the order domain and suppress proportional frequency migration caused by rotational speed changes. Second, cross-condition interpolation is performed after order-domain alignment to construct a continuous and physically coherent healthy manifold across discrete speeds. Third, a speed-conditioned reconstruction mechanism with latent regularization is introduced to separate structural health information from deterministic speed-induced variations. Unlike conventional autoencoder-based anomaly detectors and existing speed-normalized approaches, which treat rotational speed as a generic auxiliary variable, SiAE jointly enforces kinematic consistency at the representation, augmentation, and model levels. The physical validity of the order-domain augmentation strategy is examined through theoretical analysis and gearbox dynamic simulation. Experiments under steady-state, time-varying, and completely unseen speed conditions, evaluated under fixed-load settings so that the influence of speed-induced kinematic effects can be isolated, show that SiAE produces stable and threshold-consistent health indicators, enabling robust fault detection across diverse operating regimes. These results demonstrate that integrating mechanics-based signal representation with AI learning improves the reliability of condition monitoring for variable-speed energy equipment., 7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnologies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, 7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Modelos híbridos basados en inteligencia artificial para la estimación de vida útil remanente de componentes electromecánicos

  • Cuesta, Jokin
  • Leturiondo, Urko
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
Este estudio ha sido parcialmente financiado por la Agencia Estatal de Investigación (AEI), Ministerio de Ciencia e Innovación, Fondo Europeo de Desarrollo Regional (FEDER) a través de los proyectos de investigación PID2021-122132OB-C21, PID2021-122132OB-C22, y TED2021- 129512B-I00; y por la Generalitat de Catalunya a través del proyecto de investigación 2021-SGR- 01044., 7 - Energia Assequible i No Contaminant, 7.3 - Per a 2030, duplicar la taxa mundial de millora de l’eficiència energètica
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

A long-short-term-memory neural network approach to early fault detection of wind turbines' main bearing based only on SCADA data.

  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Puruncajas Maza, Bryan
  • Castellani, Francesco
  • Tutivén Gálvez, Christian|||0000-0001-6322-4608
Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Wind turbine fault detection through a virtual sensor using exogenous variables

  • Wang, Shun|||0000-0002-5766-6220
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
Wind turbine drivetrains face significant maintenance challenges, where early fault detection is critical to reduce downtime and cost. Most SCADAbased condition monitoring methods rely on internal variables that may co-vary with the target component signal when degradation develops, allowing the model to track faulty behavior and maintain small residuals, which delays anomaly detection. This paper proposes a virtual sensor framework that predicts component temperatures using exogenous variables that are independent of drivetrain health. A spatiotemporal forecasting model is trained offline on historical healthy data and deployed online to compute an anomaly score from the prediction residuals. Validation on a healthy turbine confirms that the exogenous virtual sensor achieves comparable prediction accuracy to internal-variable baselines with zero false alarms. A real-world gearbox bearing fault case study further demonstrates that the proposed framework detects an incipient fault approximately 93 days before the recorded failure, whereas a correlation-based baseline using the same architecture fails to trigger any alarm under the identical decision rule., 7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

A review of prognostics and health management in wind turbine components

  • Cuesta Arrillaga, Jokin
  • Leturiondo, Urko
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
Wind turbines (WTs) play an essential role in renewable energy generation, and ensuring their reliable operation is essential for sustainable energy production and reduction of levelized cost of energy. In this context, the field of prognostics and health management (PHM) is a powerful tool to predict and assess the health status of WT components, thereby enabling timely maintenance and reducing downtime. The study begins with an overview of WT components studied, including the blades, gearbox, generator, and bearings, and their common failure modes. For each component, various remaining useful life (RUL) estimation methods are explored, categorizing them into physics-based, data-driven, and hybrid methods. Despite the potential benefits, the application of PHM strategies in WTs is currently limited. Although PHM strategies have been present for years, their development in WTs remains a challenge. These key challenges are presented, including uncertainty management, integrating physical knowledge into models, variable operational conditions, data issues and system complexity., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Wind turbine gearbox condition monitoring using vibration data and Mel-Frequency Cepstral Coefficients

  • Velandia Cardenas, Cristian|||0000-0002-2130-7488
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
A new methodology for wind turbine condition monitoring that utilizes vibration data and Mel-frequency cepstral coefficients (MFCCs) is being developed to accurately assess the condition of the wind turbine gearbox. Traditional condition monitoring techniques rely on physical inspections, which can be time-consuming and labor-intensive [1]. This new approach offers a more efficient and cost-effective solution. The use of vibration data allows for the identification of subtle changes in a wind turbine's operating condition, providing early warning signs of potential issues. By analyzing the vibration data, engineers can detect changes in the frequency and amplitude of the turbine's oscillations, which can indicate the presence of mechanical problems or other issues [2]–[5]. The MFCCs are derived from the vibration data signals and provide a compact representation of the information, allowing for more efficient analysis [6]. These coefficients are used to create a fingerprint of the wind turbine's operating condition, which can then be compared to known healthy operating conditions to identify any deviations or anomalies. This new methodology has been shown to be highly effective in detecting potential issues with wind turbine components such as gearbox failure, and drive train bearing degradation. By providing early warning signs of these issues, wind farm operators can take action to address the issue before it leads to significant downtime or damage. The use of MFCCs as descriptors offers several other benefits. The data can be collected remotely, eliminating the need for physical inspections, allowing its analysis to be performed quickly, even in real time, and more frequent monitoring of wind turbine conditions, providing a more complete and accurate picture of the health of the system. In a nutshell, the use of vibration data and MFCCs offers a promising new approach for wind turbine condition monitoring. The proposed approach has been tested on the EISLAB (Lulea, Sweden) dataset concerning the vibration signals from six wind turbines in northern Sweden, being all the wind turbines of the same type and with a three-stage gearbox [7]. All measurement data corresponds to the axial direction of an accelerometer mounted on the housing of the output shaft bearing of each turbine., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Fault detection in wind turbines using health index monitoring with variational autoencoders

  • Wang, Shun|||0000-0002-5766-6220
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
As wind energy capacity expands globally, ensuring the operational reliability and economic viability of wind turbines has become a critical industrial challenge. Effective fault detection systems are essential for minimizing high maintenance costs and preventing catastrophic failures. To address this need, this paper presents a semi-supervised framework designed to identify anomalies in wind turbines using only healthy operational data. The methodology begins by extracting a comprehensive set of features from the time and frequency domains of raw vibration signals to capture a rich representation of the dynamics of wind turbines. A variational autoencoder, a deep generative model, is then trained exclusively on these features from healthy operational periods to learn a robust model of normal behavior and generate reconstruction errors as health indicators with exponentially weighted moving average smoothing to enhance robustness and reduce false alarms. The framework is evaluated using public data from the Aventa AV-7 ETH Zurich research wind turbine, which includes multiple failure events. Experimental results demonstrate effective and early detection of pitch faults, as well as accurate detection of icing events and aerodynamic imbalances. The proposed approach therefore offers a robust and practical solution to improving operational safety and enabling proactive maintenance of wind turbines., This research has been supported by the Agencia Estatal de Investigación (grant nos. PID2021-122132OB-C21 and TED2021-129512B-I00) and the Generalitat de Catalunya, Agència de Gestió d'Ajuts Universitaris i de Recerca (grant no. 2021-SGR-01044).
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

A review of prognostics and health management techniques in wind energy

  • Cuesta Arrillaga, Jokin
  • Leturiondo, Urko
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
  • Pozo Montero, Francesc|||0000-0001-8958-6789
This review aims to provide a holistic understanding of prognostics and health management (PHM) techniques in wind energy, particularly in the estimation of remaining useful life (RUL) of wind turbine (WT) components. The study begins with an introduction that discusses the principles of PHM and its critical role in the wind energy sector. This is followed by an overview of WT systems and the importance of accurate RUL predictions for specific failure modes. Then, various data sources, methods of feature extraction, and criteria for constructing health indices are explored, along with techniques for threshold determination. Degradation modeling techniques, essential for RUL prediction, are examined through three approaches: physics-based models, data-driven methods (including statistical and artificial intelligence techniques), and hybrid models. The performance of these models is evaluated using specific metrics which have been explored. Next, predictive maintenance strategies, optimized using RUL predictions, are presented to minimize downtime and maintenance costs. The paper concludes by identifying future research directions, emphasizing the need to manage uncertainty, integrate physical knowledge, address variable environmental and operational conditions, overcome data issues, and handle system complexity., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Challenges on prognostics and health management for wind turbine components

UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Two-stage fault detection and control approach for DFIG-based wind energy conversion system

  • Samuel Raj, Daison Stallon|||0000-0002-1339-6919
  • Eben Zaid, Ichrak
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
Doubly-Fed Induction Generator (DFIG)-based Wind Energy Conversion Systems (WECS) are critical in modern electricity generation due to their ability to enhance energy capture and seamlessly integrate with the electrical grid. However, maintaining reliability and minimizing maintenance costs are essential to ensure consistent energy production. This research presents an innovative method for fault detection and diagnosis in DFIG-based WECS. The approach leverages independent component analysis-based correlation coefficient for precise fault identification. Additionally, an enhanced multihead cross attention with bi-directional long short term memory classifier is employed to accurately categorize different fault types. To further improve classifier’s performance, the multi-strategy enhanced orchard algorithm is implemented, focusing on regulating active and reactive power variations, harmonics in rotor current, and voltage in the DC link. The proposed method is evaluated using MATLAB working platform and demonstrates a high accuracy rate of 98% compared to other techniques, Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnolo­gies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Structural health monitoring of jacket-type support structures in offshore wind turbines: a comprehensive dataset for bolt loosening detection through vibrational analysis

  • Valdez-Yepez, Rhandall
  • Tutivén Gálvez, Christian|||0000-0001-6322-4608
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
This dataset provides a comprehensive collection of vibrational data for the purpose of structural health monitoring, particularly focusing on the detection of bolt loosening in offshore wind turbine jacket foundations. The data set comprises 780 comma-separated values (CSV) files, each corresponding to specific experimental conditions, including various structural states of the wind turbine's support structure. These states are systematically varied considering three main aspects: the amplitude of a white noise (WN) signal, the type of bolt damage, and the level at which damage has occurred. The data were meticulously collected using eight triaxial accelerometers (PCB R Piezotronic model 356A17), strategically placed at different locations on a scaled-down replica of an offshore jacket-type wind turbine. This setup facilitated the acquisition of detailed vibrational data through a National Instruments’ data acquisition (DAQ) system, comprising six input modules (NI 9234 model) housed in a chassis (cDAQ model). The white noise signal, simulating wind disturbance at the nacelle, was produced by a modal shaker and varied in three amplitudes (0.5, 1, and 2), directly proportional to the induced vibration in the wind turbine. The dataset uniquely captures the vibrational behaviour under different scenarios of bolt loosening in the turbine's foundation. The conditions include a healthy state (bolts tightened to 12 Nm) and various degrees of loosening (bolts loosened to 9 Nm, 6 Nm, and completely absent), examined at four distinct levels of the turbine's base structure. This granular approach offers a nuanced view of how varying degrees of bolt loosening impact the vibrational characteristics of the structure. The value of this dataset lies in its potential for wide-ranging applications in the field of structural health monitoring. Researchers and engineers can leverage this data for developing and testing new methodologies for early damage detection and progressive damage assessment in offshore wind turbines. The dataset's comprehensive coverage of damage scenarios makes it a valuable resource for the validation and enhancement of existing damage detection algorithms. Furthermore, the dataset can serve as a benchmark for comparing the efficacy of different vibrational analysis techniques in the context of wind turbine maintenance and safety. Its application is not only limited to wind turbines but can extend to other structures where bolt integrity is critical for operational safety. This dataset represents a significant contribution to the field of structural health monitoring, providing a detailed and practical resource for enhancing the reliability and safety of offshore wind turbines and similar structures., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Incremental minimum distortion embedding for online structural damage classification

  • Leon Medina, Jersson Xavier|||0000-0002-9198-1996
  • Parés Mariné, Núria|||0000-0002-2914-9904
  • Pozo Montero, Francesc|||0000-0001-8958-6789
The damage detection problem in structures diminishes the maintenance cost and extends the life of the structures. Online structural damage classification is a relevant topic in structural health monitoring since monitoring the structure using sensors can help decision-makers can rely on empirical evidence for making informed choices regarding the maintenance, repair, or replacement of structural components. Sensors generate high-dimensional data. To address this challenge, employing dimensionality reduction methods becomes essential. These methods help obtain a low-dimensional representation of the data. Traditional dimensionality reduction methods are trained on a fixed data set, and incorporating new data may require retraining the entire model. This study employed the Minimum Distortion Embedding (MDE) method incrementally, ensuring the maintenance of a consistent embedding as new data becomes available. This incremental embedding took advantage of an anchor constraint to pin the existing embedding in place, then embed the new points. An online structural classification methodology was developed using the incremental MDE and a stream based hoeffding tree classifier. Stream data processing differs from batch processing, as it involves the real-time consumption of each incoming data. The developed methodology was tested with data obtained by accelerometers in a laboratory scaled jacket-type wind turbine foundation. The structural damage classification problems had 5 different classes including the undamaged class and 4 different damage classes. The damage corresponded to a 5mm crack in four different structural elements of the wind turbine foundation. The results indicate the good behavior of the damage classification methodology supported by the high values of the classification metrics obtained., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Explainable early fault detection in wind turbine generators using only SCADA data from an operational wind farm

  • Barrera Llanga, Kevin Iván|||0000-0002-1072-6142
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
This work presents a methodology for early fault detection in wind turbine generators using SCADA data from eight turbines, one of which failed from winding overheating. The methodology estimates healthy phase temperatures to establish a baseline, detecting faults via deviations from measurements. SCADA time series are transformed into Gramian angular field (GAF) images, allowing convolutional neural networks (CNNs) to capture complex temporal dependencies. To address computational cost, a two-stage approach is proposed: first, a global model is trained on data from multiple turbines. Second, a fine-tuning process adapts the model to each turbine, reducing cost and enabling efficient adaptation. Explainability techniques (integrated gradients) identify influential variables. The global model achieves a mean absolute error (MAE) of 2.016 and MSE of 8.279. Fine-tuning improves performance, reducing the MAE to 1.479 and MSE to 4.299. The method offers an efficient and scalable solution for predictive maintenance., This work was supported by Grant PID2021-122132OB-C21, funded by MCIN/AEI/10.13039/501100011033 and by “ERDF A Way of Making Europe”, by the European Union; Grant TED2021-129512B-I00, funded by MCIN/AEI/10.13039/501100011033 and by the European Union NextGen- erationEU/PRTR; and Grant 2021-SGR-01044, funded by the Generalitat de Catalunya., Peer Reviewed, 7 - Energia Assequible i No Contaminant, 9 - Indústria, Innovació i Infraestructura, 13 - Acció per al Clima
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Early detection of gearbox faults in wind turbines using a fine-tuned transformer encoder

  • España, Sofía
  • Sánchez, Juan
  • Puruncajas Maza, Bryan
  • Castellani, Francesco
  • Tutivén Gálvez, Christian|||0000-0001-6322-4608
  • Vidal Seguí, Yolanda|||0000-0003-4964-6948
Gearbox failures remain one of the most critical and costly issues in wind turbine operation, often leading to extended downtime and increased maintenance expenses. This paper presents a novel methodology for the early detection of gearbox faults using Supervisory Control and Data Acquisition (SCADA) data and a fine-tuned Trans- former encoder model. Unlike traditional condition monitoring systems, which rely on high-frequency sensors or require extensive labeling, the proposed approach uses the existing low-frequency SCADA infrastructure and employs a semi-supervised learning strategy. A normal behavior model is first pre-trained on healthy data from a single wind turbine and subsequently fine-tuned to adapt to the operational characteristics of additional turbines in the same wind farm. Key contributions include an optimized variable selection process based on Kendall’s Tau correlation, data imputation and smoothing techniques to address missing values and noise, and a fault prognosis indicator derived from model residuals. Results obtained with real SCADA datasets demonstrate the effectiveness of the method in detecting incipient gearbox faults under realistic operating conditions., This work is partially funded by PID2021-122132OB-C21 through MCIN/AEI/10.13039/501100011033. The authors would also like to acknowledge Lucky Wind spa for providing the data used in this study., 7 - Energia Assequible i No Contaminant, 7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnologies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Incremental minimum distortion embedding for online structural damage classification

  • Leon Medina, Jersson Xavier
  • Parés Mariné, Núria
  • Pozo Montero, Francesc
The damage detection problem in structures diminishes the maintenance cost and extends the life of the structures. Online structural damage classification is a relevant topic in structural health monitoring since monitoring the structure using sensors can help decision-makers can rely on empirical evidence for making informed choices regarding the maintenance, repair, or replacement of structural components. Sensors generate high-dimensional data. To address this challenge, employing dimensionality reduction methods becomes essential. These methods help obtain a low-dimensional representation of the data. Traditional dimensionality reduction methods are trained on a fixed data set, and incorporating new data may require retraining the entire model. This study employed the Minimum Distortion Embedding (MDE) method incrementally, ensuring the maintenance of a consistent embedding as new data becomes available. This incremental embedding took advantage of an anchor constraint to pin the existing embedding in place, then embed the new points. An online structural classification methodology was developed using the incremental MDE and a stream based hoeffding tree classifier. Stream data processing differs from batch processing, as it involves the real-time consumption of each incoming data. The developed methodology was tested with data obtained by accelerometers in a laboratory scaled jacket-type wind turbine foundation. The structural damage classification problems had 5 different classes including the undamaged class and 4 different damage classes. The damage corresponded to a 5mm crack in four different structural elements of the wind turbine foundation. The results indicate the good behavior of the damage classification methodology supported by the high values of the classification metrics obtained., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Recent advances in wind turbine condition monitoring using SCADA data: a state-of-the-art review

  • Wang, Shun
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
With the rapid increase worldwide in the number of installed wind turbines, as an increasingly vital role in the global energy landscape, requirements and expenses of maintenance have also increased significantly. This surge in wind turbine installations worldwide has escalated maintenance demands and costs, making condition monitoring an essential focus for research and operational optimization. Built-in supervisory control and data acquisition (SCADA) systems could offer a cost-effective approach to condition monitoring in wind turbines, using existing infrastructure to reduce the need for additional sensors and simplifying system complexity. This review critically surveys state-of-the-art SCADA-based condition monitoring techniques for wind turbines. The paper first examines essential data preprocessing strategies tailored for SCADA data challenges. Subsequently, key monitoring methodologies are reviewed, including model-based and data-driven paradigms, analyzing their applications, strengths, and limitations. Finally, the challenges and future trends of SCADA-based wind turbine condition monitoring are discussed. This comprehensive overview serves as a valuable reference for researchers and industry practitioners seeking to enhance wind turbine reliability and operational efficiency through advanced monitoring techniques., This work is partially funded by the Spanish Agencia Estatal de Investigación (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER), Spain through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00, and by the Generalitat de Catalunya, Spain through the research project 2021-SGR-01044., 7 - Energia Assequible i No Contaminant, 7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el conjunt de fonts d’energia, 7.3 - Per a 2030, duplicar la taxa mundial de millora de l’eficiència energètica, 8 - Treball Decent i Creixement Econòmic, 11 - Ciutats i Comunitats Sostenibles, 13 - Acció per al Clima, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

A review of prognostics and health management in wind turbine components

  • Cuesta Arrillaga, Jokin
  • Leturiondo, Urko
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
Wind turbines (WTs) play an essential role in renewable energy generation, and ensuring their reliable operation is essential for sustainable energy production and reduction of levelized cost of energy. In this context, the field of prognostics and health management (PHM) is a powerful tool to predict and assess the health status of WT components, thereby enabling timely maintenance and reducing downtime. The study begins with an overview of WT components studied, including the blades, gearbox, generator, and bearings, and their common failure modes. For each component, various remaining useful life (RUL) estimation methods are explored, categorizing them into physics-based, data-driven, and hybrid methods. Despite the potential benefits, the application of PHM strategies in WTs is currently limited. Although PHM strategies have been present for years, their development in WTs remains a challenge. These key challenges are presented, including uncertainty management, integrating physical knowledge into models, variable operational conditions, data issues and system complexity., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Wind turbine modeling, maximum power point tracking (MPPT), and experimental validation

  • Eben Zaid, Ichrak
  • Samuel Raj, Daison Stallon
  • Vidal Seguí, Yolanda
  • Boussada, Moez
  • Said Nouri, Ahmed
The research presented is driven by the global increase in wind power capacity and the commitment of the scientific community to facilitate its integration into electrical grids. The focus of this study is the modeling of a wind turbine system, beginning with its mechanical components. To ensure the production of power at optimal levels, a control strategy for Maximum Power Point Tracking (MPPT) based on Optimal Torque (OT) has been adopted. The model and control method, developed in Matlab/Simulink, have demonstrated their precision and efficacy through experimental verification using SCADA data acquired from an operational wind turbine., This work was supported by the Ministry of Higher Education and Scientific Research-Tunisia;by the Spanish Agencia Estatal de Investigaci\'on (AEI) -Ministerio de Econom\'ia, Industria y Competitividad (MINECO), and the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research project 2021-SGR-01044., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Moving towards preventive maintenance in wind turbine structural control and health monitoring

  • Leon Medina, Jersson Xavier
  • Pozo Montero, Francesc
This work has been partially funded by the Spanish Agencia Estatal de Investigación (AEI)—Ministerio de Economía, Industria y Competitividad (MINECO), and the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021- 129512B-I00; and by the Generalitat de Catalunya through the research projects 2021-SGR-01044., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

A review of prognostics and health management techniques in wind energy

  • Cuesta Arrillaga, Jokin
  • Leturiondo, Urko
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
This review aims to provide a holistic understanding of prognostics and health management (PHM) techniques in wind energy, particularly in the estimation of remaining useful life (RUL) of wind turbine (WT) components. The study begins with an introduction that discusses the principles of PHM and its critical role in the wind energy sector. This is followed by an overview of WT systems and the importance of accurate RUL predictions for specific failure modes. Then, various data sources, methods of feature extraction, and criteria for constructing health indices are explored, along with techniques for threshold determination. Degradation modeling techniques, essential for RUL prediction, are examined through three approaches: physics-based models, data-driven methods (including statistical and artificial intelligence techniques), and hybrid models. The performance of these models is evaluated using specific metrics which have been explored. Next, predictive maintenance strategies, optimized using RUL predictions, are presented to minimize downtime and maintenance costs. The paper concludes by identifying future research directions, emphasizing the need to manage uncertainty, integrate physical knowledge, address variable environmental and operational conditions, overcome data issues, and handle system complexity., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Wind turbine gearbox early fault detection using Mel-frequency cepstral coefficients of vibration data

  • Velandia Cardenas, Cristian
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
This is the peer reviewed version of the following article: Velandia-Cardenas, C.; Vidal, Y.; Pozo, F. Wind turbine gearbox early fault detection using Mel-frequency cepstral coefficients of vibration data. "Structural control and health monitoring", 6 Juny 2024, vol. 2024, núm. article 733730, which has been published in final form at [Link to final article using the DOI]. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions. This article may not be enhanced, enriched or otherwise transformed into a derivative work, without express permission from Wiley or by statutory rights under applicable legislation. Copyright notices must not be removed, obscured or modified. The article must be linked to Wiley’s version of record on Wiley Online Library and any embedding, framing or otherwise making available the article or pages thereof by third parties from platforms, services and websites other than Wiley Online Library must be prohibited., A methodology utilizing vibration data and Mel-frequency cepstral coefficients (MFCCs) for wind turbine condition monitoring is developed to detect incipient faults in the wind turbine gearbox. This approach provides a more efficient and cost-effective solution compared to traditional condition monitoring techniques relying on physical inspections, which can be time-consuming and labor-intensive. The use of vibration data enables the identification of subtle changes in a wind turbine’s operating condition, providing early warning signs of potential issues. When the vibration data are analyzed, changes in frequency and amplitude can be detected, indicating the presence of a developing fault. Vibration-based condition monitoring systems (CMS) have already been widely used in the wind industry (mainly in new turbines). These systems utilize basic standard features, working in either the time or frequency domain, and are not optimized for nonstationary signals. In contrast, this work focuses on MFCCs, operating in both time and frequency domains, enabling the extraction of adequate information from nonstationary signals. The MFCCs are derived from vibration data signals, providing a compact representation for a more efficient analysis. These coefficients create a fingerprint of the wind turbine operating condition, compared to known healthy conditions, to identify anomalies. To underscore the practical value of this study, it is important to highlight the significant implications for the wind energy sector. The methodology developed offers an advanced, predictive tool for the early detection of gearbox faults, which is a critical aspect of optimizing the performance and longevity of wind turbines. By enabling earlier, more accurate fault detection, the proposed approach significantly reduces the likelihood of catastrophic failures and extensive downtime. This not only enhances the reliability and cost-effectiveness of wind energy systems but also contributes to sustainable energy practices by optimizing resource use and minimizing maintenance costs. The results strongly suggest that the proposed methodology is highly effective in detecting incipient faults in the wind turbine gearbox. By providing early warnings of damage, operators can address issues before significant downtime or damage occurs. The use of MFCCs offers additional benefits since data can be collected remotely, eliminating physical inspections. Analysis can be performed faster, even in real time, allowing more frequent monitoring. This provides a more complete and accurate picture of the health of the system. The approach is tested in the EISLAB dataset concerning vibration signals from six wind turbines in northern Sweden, all with three-stage gearboxes. All measurement data correspond to the axial direction of an accelerometer in the output shaft bearing housing of each turbine., This work was partially funded by the Spanish Agencia Estatal de Investigación (AEI), Ministerio de Economía, Industria y Competitividad (MINECO), the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512BI00, and the Generalitat de Catalunya through the research project 2021-SGR-01044. Open access funding was enabled and organized by CRUE-CSUC Gold., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.3 - Per a 2030, duplicar la taxa mundial de millora de l’eficiència energètica, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnolo­gies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Structural health monitoring of jacket-type support structures in offshore wind turbines: a comprehensive dataset for bolt loosening detection through vibrational analysis

  • Valdez-Yepez, Rhandall
  • Tutivén Gálvez, Christian
  • Vidal Seguí, Yolanda
This dataset provides a comprehensive collection of vibrational data for the purpose of structural health monitoring, particularly focusing on the detection of bolt loosening in offshore wind turbine jacket foundations. The data set comprises 780 comma-separated values (CSV) files, each corresponding to specific experimental conditions, including various structural states of the wind turbine's support structure. These states are systematically varied considering three main aspects: the amplitude of a white noise (WN) signal, the type of bolt damage, and the level at which damage has occurred. The data were meticulously collected using eight triaxial accelerometers (PCB R Piezotronic model 356A17), strategically placed at different locations on a scaled-down replica of an offshore jacket-type wind turbine. This setup facilitated the acquisition of detailed vibrational data through a National Instruments’ data acquisition (DAQ) system, comprising six input modules (NI 9234 model) housed in a chassis (cDAQ model). The white noise signal, simulating wind disturbance at the nacelle, was produced by a modal shaker and varied in three amplitudes (0.5, 1, and 2), directly proportional to the induced vibration in the wind turbine. The dataset uniquely captures the vibrational behaviour under different scenarios of bolt loosening in the turbine's foundation. The conditions include a healthy state (bolts tightened to 12 Nm) and various degrees of loosening (bolts loosened to 9 Nm, 6 Nm, and completely absent), examined at four distinct levels of the turbine's base structure. This granular approach offers a nuanced view of how varying degrees of bolt loosening impact the vibrational characteristics of the structure. The value of this dataset lies in its potential for wide-ranging applications in the field of structural health monitoring. Researchers and engineers can leverage this data for developing and testing new methodologies for early damage detection and progressive damage assessment in offshore wind turbines. The dataset's comprehensive coverage of damage scenarios makes it a valuable resource for the validation and enhancement of existing damage detection algorithms. Furthermore, the dataset can serve as a benchmark for comparing the efficacy of different vibrational analysis techniques in the context of wind turbine maintenance and safety. Its application is not only limited to wind turbines but can extend to other structures where bolt integrity is critical for operational safety. This dataset represents a significant contribution to the field of structural health monitoring, providing a detailed and practical resource for enhancing the reliability and safety of offshore wind turbines and similar structures., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Wind turbine modeling, Maximum Power Point Tracking (MPPT), and experimental validation

  • Eben Zaid, Ichrak
  • Stallon, Daison
  • Vidal Seguí, Yolanda
  • Boussada, Moez
  • Said Nouri, Ahmed
The research presented is driven by the global increase in wind power capacity and the commitment of the scientific community to facilitate its integration into electrical grids. The focus of this study is the modeling of a wind turbine system, beginning with its mechanical components. To ensure the production of power at optimal levels, a control strategy for Maximum Power Point Tracking (MPPT) based on Optimal Torque (OT) has been adopted. The model and control method, developed in Matlab/Simulink, have demonstrated their precision and efficacy through experimental verification using SCADA data acquired from an operational wind turbine., Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Noncontact fault diagnosis of electrical equipment using modified multiscale two-dimensional color distribution entropy and thermal imaging

  • Wang, Shun
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
Effective health monitoring of electrical equipment is critical for industrial reliability. Although infrared thermal imaging offers a powerful noncontact diagnostic method, accurately interpreting its complex and often noisy thermal patterns remains a significant challenge. Entropy-based analysis is well suited for quantifying this complexity, but its application to images has been limited. Existing two-dimensional entropy methods are not only less developed than their one-dimensional counterparts but also typically require converting thermal images to grayscale, which discards vital diagnostic information from color channels. To overcome these limitations, this study introduces the modified multiscale two-dimensional color distribution entropy (MMCDEn_2D). This novel method directly integrates the attributes of the RGB, preserving a richer feature set for analysis. The effectiveness of the proposed method is demonstrated first through synthetic signals, showing low sensitivity to image size and high computational efficiency. The study further extends the application of entropy-based analysis to noncontact health monitoring scenarios, implementing MMCDEn_2D for thermal image-based fault diagnosis of induction motors and power transformers. The method achieves a diagnostic accuracy that exceeds 95%, significantly outperforming traditional approaches. Crucially, it demonstrates superior robustness in challenging scenarios, improving accuracy by 2%–5% under high-noise conditions and with small sample sizes. These results establish MMCDEn_2D as a highly effective and reliable tool to advance noncontact fault diagnosis in critical electrical equipment., This work was partially funded by the Spanish Agencia Estatal de Investigación (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER) through​ the research projects PID2021-122132OB-C21 and TED2021-129512B-I00, and by the Generalitat de Catalunya through the research project 2021-SGR-01044., Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Data-driven unsupervised stator interturn short-circuit fault detection for doubly fed induction generator wind turbines

  • Espinoza, Victor
  • Benalcazar-Parra, Carlos
  • Tutivén Gálvez, Christian
  • Vidal Seguí, Yolanda
Inter-turn short-circuit (ITSC) faults in doubly fed induction generators (DFIGs) pose a significant challenge for reliable wind turbine operation. Existing fault–detection approaches often rely on data that neglect key aeroelastic, electromechanical, control, and grid interactions, limiting their applicability to real operating conditions. To address this limitation, this study develops a coupled OpenFAST-Simulink simulation framework capable of generating physically realistic datasets of healthy operation and ITSC faults across multiple severity levels. Leveraging this data, an unsupervised DL fault-detection scheme based on an autoencoder is proposed, trained exclusively on healthy data and evaluated using faulty scenarios. The results show effective detection of medium and high severity faults, with longer detection times as severity decreases due to weaker system disturbances., This work is partially funded by grant PID2021-122132OB-C21 funded by MCIN/AEI/10. 13039/501100011033 and by “ERDF A way of making Europe”, by the “European Union”., 7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnologies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, 7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Two-dimensional color distribution entropy: validation and application on non-contact fault diagnosis for induction motor

  • Wang, Shun
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
© 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Entropy-based methods are commonly used in the f ield of vibration signal processing and fault diagnosis due to their simplicity, ease of implementation, and strong feature extraction capabilities. Despite the successful application of 1D entropy measures in vibration signal analysis, there is a gap in 2D entropy application to images, especially for color images. Therefore, this study introduces a novel concept called two-dimensional color distribution entropy (CDEn2D) and further develops its multiscale variant, modified multiscale two-dimensional color distribution entropy (MMCDEn2D). This work investigates the application of MMCDEn2D in non-contact fault diagnosis for induction motors using infrared thermal imaging. The experimental results demonstrate that MMCDEn2D outperforms traditional two-dimensional entropy methods. Moreover, MMCDEn2D shows significant reliability and superiority in scenarios with limited data samples. These findings highlight its effectiveness in the f ield of non-contact health monitoring and fault diagnosis., This work is partially funded by the Spanish Agencia Estatal de Investigaci´ on (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research project 2021-SGR-01044., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Wind turbine fault detection through a virtual sensor using exogenous variables

  • Wang, Shun
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
Wind turbine drivetrains face significant maintenance challenges, where early fault detection is critical to reduce downtime and cost. Most SCADAbased condition monitoring methods rely on internal variables that may co-vary with the target component signal when degradation develops, allowing the model to track faulty behavior and maintain small residuals, which delays anomaly detection. This paper proposes a virtual sensor framework that predicts component temperatures using exogenous variables that are independent of drivetrain health. A spatiotemporal forecasting model is trained offline on historical healthy data and deployed online to compute an anomaly score from the prediction residuals. Validation on a healthy turbine confirms that the exogenous virtual sensor achieves comparable prediction accuracy to internal-variable baselines with zero false alarms. A real-world gearbox bearing fault case study further demonstrates that the proposed framework detects an incipient fault approximately 93 days before the recorded failure, whereas a correlation-based baseline using the same architecture fails to trigger any alarm under the identical decision rule., 7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Challenges on prognostics and health management for wind turbine components

  • Cuesta, Jokin
  • Leturiondo, Urko
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
This paper explores the applicability of prognostics and health management (PHM) for wind turbines (WTs), presenting the PHM approach along with challenges and opportunities in the context of WT components. First, the PHM framework is introduced, consisting of three blocks: observation, analysis, and action. Critical components and failure modes for WTs are identified, and data acquisition strategies using supervisory control and data adquisition (SCADA) and condition monitoring (CM) data are discussed. Prognostics, specifically remaining useful life (RUL) estimation, employs physics model-based, data-driven, and hybrid models. Finally, challenges and opportunities related to data, analysis and CM, and developing RUL prediction models have been found. Data challenges include data standardization, limited public datasets, and data quality issues. Analysis and CM challenges address new sensorless and non-intrusive techniques, as well as the fusion of data sources. Prognostics model challenges involve uncertainty management, interpretability issues, and the need for online updates. Addressing challenges requires incorporating physical knowledge, utilizing transfer learning, and improving online RUL prediction methods., Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Two-stage fault detection and control approach for DFIG-based wind energy conversion system

  • Samuel Raj, Daison Stallon
  • Eben Zaid, Ichrak
  • Vidal Seguí, Yolanda
Doubly-Fed Induction Generator (DFIG)-based Wind Energy Conversion Systems (WECS) are critical in modern electricity generation due to their ability to enhance energy capture and seamlessly integrate with the electrical grid. However, maintaining reliability and minimizing maintenance costs are essential to ensure consistent energy production. This research presents an innovative method for fault detection and diagnosis in DFIG-based WECS. The approach leverages independent component analysis-based correlation coefficient for precise fault identification. Additionally, an enhanced multihead cross attention with bi-directional long short term memory classifier is employed to accurately categorize different fault types. To further improve classifier’s performance, the multi-strategy enhanced orchard algorithm is implemented, focusing on regulating active and reactive power variations, harmonics in rotor current, and voltage in the DC link. The proposed method is evaluated using MATLAB working platform and demonstrates a high accuracy rate of 98% compared to other techniques, Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnolo­gies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Enhancing wind turbine diagnostics with SCADA-vibration fusion, contrastive learning, and linear predictive coefcients

  • Velandia Cardenas, Cristian
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
Wind energy plays a pivotal role in the transition to sustainable power generation. However, maintaining the reliability and efficiency of wind turbine (WT) remains a significant challenge due to complex operational conditions and the high cost associated with unexpected failures. Effective condition monitoring (CM) and predictive maintenance (PM) strategies are critical to mitigate these risks. This study presents a data-driven fault detection framework that fuses supervisory control and data acquisition (SCADA) data with high-frequency vibration signals using deep learning techniques to enhance diagnostic performance. Unlike conventional normal behavior models that rely exclusively on healthy data for training, the proposed framework incorporates limited labeled fault data when available. As only a few types of faults and a few samples are typically available in real-world scenarios, the approach does not assume a complete representation of all possible fault conditions. Instead, it is designed to generalize beyond the specific faults seen during training. This is demonstrated by training the model on healthy conditions and only two known fault types (with labeled data available) and testing it on a third, previously unseen fault type. In particular, Siamese networks with contrastive and reconstruction learning are employed to improve feature representation and anomaly detection. Two distinct methodologies are compared: the first utilizes a binary cross-entropy (BCE) loss function to classify the healthy or faulty status of the WT, while the second uses a triplet loss function for multiclass representation learning. Both methodologies generate low-dimensional representations of the input features, also known as embeddings. The resulting feature embeddings are passed through a k-means clustering algorithm to improve fault separation and identification. Statistical features are extracted from SCADA data to capture key trends or event information, while the linear prediction coefficient (LPC) method, which models a signal by predicting future values based on its past samples, is applied to the vibration data for better fault characterization. The proposed approach is evaluated using the publicly available ETH Zurich dataset from an Aventa AV-7 turbine. Experimental results indicate that the fusion of SCADA and vibration-based diagnostics, in combination with contrastive and representation learning, substantially improves the predictive accuracy and generalization of fault detection models., This work was partially funded by the Spanish Agencia Estatal de Investigación (AEI), Ministerio de Economía, Industria y Competitividad (MINECO); the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research project 2021-SGR-01044. Open Access funding was enabled and organized by CRUE-CSUC Gold., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Robust unsupervised gearbox fault detection under variable-speed operating conditions

  • Wang, Shun
  • Vidal Seguí, Yolanda
  • Zhang, Jiacong
  • Pozo Montero, Francesc
Gear transmission systems in wind turbines and other rotating energy machinery operate under variable rotational speeds, where vibration responses are governed by speed-dependent kinematic frequency variations. This speed-induced transformation challenges the construction of condition-consistent vibration representations and undermines reliable AI-enabled condition monitoring. To address this limitation, this study proposes a mechanics-consistent speed-invariant autoencoder (SiAE) for unsupervised fault detection using only healthy-condition data. Instead of treating speed variation as a generic statistical distribution shift, the proposed framework embeds gearbox kinematics into representation learning. First, the squared envelope order spectrum is used to map vibration signals into the order domain and suppress proportional frequency migration caused by rotational speed changes. Second, cross-condition interpolation is performed after order-domain alignment to construct a continuous and physically coherent healthy manifold across discrete speeds. Third, a speed-conditioned reconstruction mechanism with latent regularization is introduced to separate structural health information from deterministic speed-induced variations. Unlike conventional autoencoder-based anomaly detectors and existing speed-normalized approaches, which treat rotational speed as a generic auxiliary variable, SiAE jointly enforces kinematic consistency at the representation, augmentation, and model levels. The physical validity of the order-domain augmentation strategy is examined through theoretical analysis and gearbox dynamic simulation. Experiments under steady-state, time-varying, and completely unseen speed conditions, evaluated under fixed-load settings so that the influence of speed-induced kinematic effects can be isolated, show that SiAE produces stable and threshold-consistent health indicators, enabling robust fault detection across diverse operating regimes. These results demonstrate that integrating mechanics-based signal representation with AI learning improves the reliability of condition monitoring for variable-speed energy equipment., 7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnologies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, 7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Digital twin technology in wind turbine components: A review

  • Leon Medina, Jersson Xavier
  • Tibaduiza Burgos, Diego Alexander
  • Parés Mariné, Núria
  • Pozo Montero, Francesc
The industrial development, the advances in sensor technology and the processing of large amounts of data, have enabled the training and testing of artificial intelligence models that reproduce, with high accuracy, the behavior of some variables of interest. With the consolidation of the big data era and the proliferation of sensors that can acquire information directly from various components of a wind turbine (WT), a digital twin (DT) allows to close the gap between the physical and the digital worlds. It combines historical data, sensor readings, machine learning and physics-based modeling to replicate the behavior of the physical component accurately. This DT can simulate the performance and behavior of the physical object under different conditions and situations, allowing for predicting failures in WT components and determining their remaining useful life. This review describes the existing literature related to the use of DTs and their developments for WT applications and their components in onshore and offshore applications. This review explores various types of DTs and their approaches, aiming to cover different methods of data processing and concepts related to each DT framework. In addition, it identifies insights from various studies and reviews, particularly focusing on the components of WTs., This work has been partially funded by the Spanish Agencia Estatal de Investigación (AEI)—Ministerio de Economía, Industria y Competitividad (MINECO), and the Fondo Europeo de Desarrollo Regional (FEDER) through the research projects PID2021-122132OB-C21, PID2021-126051OB-C44 and TED2021-129512B-I00; and by the Generalitat de Catalunya through the research projects 2021-SGR-01044 and 2021-SGR-01049. The authors also express their gratitude to Fundación Carolina and the Universidad Nacional de Colombia for supporting Diego Tibaduiza’s postdoctoral research stay through the program ‘‘Convocatoria de becas para estancias cortas postdoctorales 2023–2024 de la Fundación Carolina’’., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

A long-short-term-memory neural network approach to early fault detection of wind turbines' main bearing based only on SCADA data.

Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Fault detection in wind turbines using health index monitoring with variational autoencoders

  • Wang, Shun
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
As wind energy capacity expands globally, ensuring the operational reliability and economic viability of wind turbines has become a critical industrial challenge. Effective fault detection systems are essential for minimizing high maintenance costs and preventing catastrophic failures. To address this need, this paper presents a semi-supervised framework designed to identify anomalies in wind turbines using only healthy operational data. The methodology begins by extracting a comprehensive set of features from the time and frequency domains of raw vibration signals to capture a rich representation of the dynamics of wind turbines. A variational autoencoder, a deep generative model, is then trained exclusively on these features from healthy operational periods to learn a robust model of normal behavior and generate reconstruction errors as health indicators with exponentially weighted moving average smoothing to enhance robustness and reduce false alarms. The framework is evaluated using public data from the Aventa AV-7 ETH Zurich research wind turbine, which includes multiple failure events. Experimental results demonstrate effective and early detection of pitch faults, as well as accurate detection of icing events and aerodynamic imbalances. The proposed approach therefore offers a robust and practical solution to improving operational safety and enabling proactive maintenance of wind turbines., This research has been supported by the Agencia Estatal de Investigación (grant nos. PID2021-122132OB-C21 and TED2021-129512B-I00) and the Generalitat de Catalunya, Agència de Gestió d'Ajuts Universitaris i de Recerca (grant no. 2021-SGR-01044)., Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Wind turbine gearbox condition monitoring through the sequential analysis of industrial SCADA and vibration data

  • Castellani, Francesco
  • Natili, Francesco
  • Astolfi, Davide
  • Vidal Seguí, Yolanda
The operation & maintenance expenditure for a wind farm project can reach the impressive share of 30% of the total costs. This matter of fact motivates the need for optimal operation & maintenance, which is estimated to provide up to a 10% of energy production improvement. Such potential benefit can only be achieved through an efficient condition monitoring and predictive maintenance strategy. Based on these motivations, this paper presents a real-world case study in which standard diagnostic techniques failed to detect severe faults in the planetary stage of a wind turbine gearbox in time to prevent prolonged downtime. To address this issue, a measurement data processing and fusion algorithm is developed. The approach is capable of leveraging all the information from different data sources (with low to high time resolution) using different machine and deep learning algorithms, connected between them in cascade. This enables the detection of the fault some weeks in advance, compared to the commonly used methods and with lower-level processing of industrial operational data. A qualifying feature of the proposed workflow is that it enables the identification of the faulty component, which is a well known critical point in real-world applications., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Early detection of gearbox faults in wind turbines using a fine-tuned transformer encoder

  • España, Sofía
  • Sánchez, Juan
  • Puruncajas Maza, Bryan
  • Castellani, Francesco
  • Tutivén Gálvez, Christian
  • Vidal Seguí, Yolanda
Gearbox failures remain one of the most critical and costly issues in wind turbine operation, often leading to extended downtime and increased maintenance expenses. This paper presents a novel methodology for the early detection of gearbox faults using Supervisory Control and Data Acquisition (SCADA) data and a fine-tuned Trans- former encoder model. Unlike traditional condition monitoring systems, which rely on high-frequency sensors or require extensive labeling, the proposed approach uses the existing low-frequency SCADA infrastructure and employs a semi-supervised learning strategy. A normal behavior model is first pre-trained on healthy data from a single wind turbine and subsequently fine-tuned to adapt to the operational characteristics of additional turbines in the same wind farm. Key contributions include an optimized variable selection process based on Kendall’s Tau correlation, data imputation and smoothing techniques to address missing values and noise, and a fault prognosis indicator derived from model residuals. Results obtained with real SCADA datasets demonstrate the effectiveness of the method in detecting incipient gearbox faults under realistic operating conditions., This work is partially funded by PID2021-122132OB-C21 through MCIN/AEI/10.13039/501100011033. The authors would also like to acknowledge Lucky Wind spa for providing the data used in this study., 7 - Energia Assequible i No Contaminant, 7.a - Per a 2030, augmentar la cooperació internacional per tal de facilitar l’accés a la investigació i a les tecnologies energètiques no contaminants, incloses les fonts d’energia renovables, l’eficiència energètica i les tecnologies de combustibles fòssils avançades i menys contaminants, i promoure la inversió en infraestructures energètiques i tecnologies d’energia no contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Strain virtual sensor for offshore wind turbine jacket supports: a time series transformer approach validated with Alpha Ventus wind farm data

  • Encalada-Dávila, Ángel
  • Vidal Seguí, Yolanda
  • Tutivén Gálvez, Christian
© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/), Monitoring the structural health of offshore wind turbines in harsh marine environments presents significant challenges due to physical sensor limitations, such as susceptibility to high salinity, strong underwater currents, and the resulting durability and maintenance issues. These challenges increase the need for robust and cost-effective alternatives. Motivated by the urgent demand for sustainable and reliable solutions in offshore wind energy, this study introduces a virtual strain sensor based on the PatchTST architecture, a time series transformer designed to capture local and global patterns in multivariate data. By addressing the limitations of conventional physical sensors, the proposed approach uses advanced deep learning techniques to estimate strain accurately in wind turbine jacket supports. The method involves pre-training the model with self-supervised learning and fine-tuning it using real-world strain data from the Alpha Ventus offshore wind farm. Extensive evaluations using statistical metrics (Pearson correlation, root mean square error) and mechanics-based metrics (normalized fatigue damage error and equivalent strain range) demonstrate the virtual sensor’s superior predictive capabilities. Key results reveal that the model outperforms conventional approaches in reliability, adaptability, and computational efficiency, offering an innovative tool for improving structural health monitoring systems. These outcomes support the deployment of virtual sensors as a durable, cost-effective alternative to physical sensors, contributing to enhanced sustainability and efficiency in offshore wind energy systems., APC – UPC (Yolanda Vidal), Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Wind turbine gearbox condition monitoring using vibration data and Mel-Frequency Cepstral Coefficients

  • Velandia Cardenas, Cristian
  • Vidal Seguí, Yolanda
  • Pozo Montero, Francesc
A new methodology for wind turbine condition monitoring that utilizes vibration data and Mel-frequency cepstral coefficients (MFCCs) is being developed to accurately assess the condition of the wind turbine gearbox. Traditional condition monitoring techniques rely on physical inspections, which can be time-consuming and labor-intensive [1]. This new approach offers a more efficient and cost-effective solution. The use of vibration data allows for the identification of subtle changes in a wind turbine's operating condition, providing early warning signs of potential issues. By analyzing the vibration data, engineers can detect changes in the frequency and amplitude of the turbine's oscillations, which can indicate the presence of mechanical problems or other issues [2]–[5]. The MFCCs are derived from the vibration data signals and provide a compact representation of the information, allowing for more efficient analysis [6]. These coefficients are used to create a fingerprint of the wind turbine's operating condition, which can then be compared to known healthy operating conditions to identify any deviations or anomalies. This new methodology has been shown to be highly effective in detecting potential issues with wind turbine components such as gearbox failure, and drive train bearing degradation. By providing early warning signs of these issues, wind farm operators can take action to address the issue before it leads to significant downtime or damage. The use of MFCCs as descriptors offers several other benefits. The data can be collected remotely, eliminating the need for physical inspections, allowing its analysis to be performed quickly, even in real time, and more frequent monitoring of wind turbine conditions, providing a more complete and accurate picture of the health of the system. In a nutshell, the use of vibration data and MFCCs offers a promising new approach for wind turbine condition monitoring. The proposed approach has been tested on the EISLAB (Lulea, Sweden) dataset concerning the vibration signals from six wind turbines in northern Sweden, being all the wind turbines of the same type and with a three-stage gearbox [7]. All measurement data corresponds to the axial direction of an accelerometer mounted on the housing of the output shaft bearing of each turbine., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Detection of crack bar deterioration at offshore wind turbine supports using generative adversarial networks and autoencoders

  • Prieto-Galarza, Ricardo
  • Tutivén Gálvez, Christian
  • Vidal Seguí, Yolanda
This work focuses on the structural health monitoring of jacket-type foundations used by offshore wind turbines. A vibration-only response mechanism based on accelerometer data is specifically suggested. This work, based on an anomaly detection model, has developed a methodology to detect crack bar deterioration in the wind turbine jacket consisting of two training phases with only healthy data: training of a generative adversarial network (GAN), and encoder training of an autoencoder based on the GAN model that has already been learned. A generator and a critic may be obtained through the GAN network training process. This model is used to train an encoder that allows the mapping of healthy samples to a latent vector. Following encoder training, the encoder places the data at points in latent space that correspond to the input data’s healthy state. The mapping of the input sample space to the latent space through the encoder and the subsequent mapping of the latent space to the initial space through the generator should closely resemble the input sample in the event of a healthy input. However, when damage-state input samples are used, the model output does not resemble the input. The reconstruction error and a comparison of the residual error of the critic properties are the final two loss functions used for the identification of anomalies. The proposed strategy has been tested through laboratory experiments on a down-scaled model., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant::7.2 - Per a 2030, augmentar substancialment el percentatge d’energia renovable en el con­junt de fonts d’energia, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Postprint (published version)