Pasar al contenido principal

DESARROLLO DE UNA HERRAMIENTA DE AYUDA AL MANTENIMIENTO PREDICTIVO DE AEROGENERADORES

PID2021-128013OB-I00

•
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 POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 8
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Fault Detection in Induction Machines Using Learning Models and Fourier Spectrum Image Analysis

  • Barrera-Llanga, Kevin
  • Burriel-Valencia, Jordi|||0000-0002-1680-4412
  • Sapena-Bano, Angel|||0000-0002-3888-6498
  • Martinez-Roman, Javier|||0000-0001-7544-8481
[EN] Induction motors are essential components in industry due to their efficiency and cost-effectiveness. This study presents an innovative methodology for automatic fault detection by analyzing images generated from the Fourier spectra of current signals using deep learning techniques. A new preprocessing technique incorporating a distinctive background to enhance spectral feature learning is proposed, enabling the detection of four types of faults: healthy motor coupled to a generator with a broken bar (HGB), broken rotor bar (BRB), race bearing fault (RBF), and bearing ball fault (BBF). The dataset was generated from three-phase signals of an induction motor controlled by a Direct Torque Controller under various operating conditions (20¿1500 rpm with 0¿100% load), resulting in 4251 images. The model, based on a Visual Geometry Group (VGG) architecture with 19 layers, achieved an overall accuracy of 98%, with specific accuracies of 99% for RAF, 100% for BRB, 100% for RBF, and 95% for BBF. A new model interpretability was assessed using explainability techniques, which allowed for the identification of specific learning patterns. This analysis introduces a new approach by demonstrating how different convolutional blocks capture particular features: the first convolutional block captures signal shape, while the second identifies background features. Additionally, distinct convolutional layers were associated with each fault type: layer 9 for RAF, layer 13 for BRB, layer 16 for RBF, and layer 14 for BBF. This methodology offers a scalable solution for predictive maintenance in induction motors, effectively combining signal processing, computer vision, and explainability techniques., This publication is part of the R+D+i project (grant PID2021-128013OB-I00) funded by MCIN/AEI/10.13039/501100011033 and by "ERDF A way of making Europe". This work was also supported by Generalitat Valenciana (reference CIAICO/2022/042).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Transient analysis of induction machine deep bar current displacement through bar layer model

  • Terrón-Santiago, Carla|||0000-0001-8178-3331
  • Delfa-Baena, Sandra
  • Riera-Guasp, Martín|||0000-0003-1327-242X
  • Sapena-Bano, Angel|||0000-0002-3888-6498
  • Martinez-Roman, Javier|||0000-0001-7544-8481
[EN] The aim of this contribution is to advance in the description of the bar current displacement effect in induction machines with deep rotor bars. The model proposes a transient bar layer model that is based on the former steady state studies (the multiple bar layer model) to provide a state space model of the bar layer currents that can be integrated in standard transient performance models of the induction machine. The transient bar layer model results are compared with the analytical results for steady state performance showing very good agreement. The model results are compared also with a previous proposal based on uncoupled time-finite element method (FEM) calculations showing a better description of the transient bar current displacement effect that should be considered in the design of high-performance torque control systems for converter-fed induction machines. The transient bar layer model has been validated also against FEM calculation including the eddy current effect equivalent to the bar current displacement under transient conditions equivalent to those typical of converter-fed operation. The proposed model is quite time efficient compared to such coupled time-FEM models able to represent the bar current displacement and, thus, it is very well suited to the effective design and tuning of induction machine control systems., This publication is part of the R+D+i project (grant PID2021-128013OB-I00) funded by MCIN/AEI/10.13039/501100011033 and by "ERDF A way of making Europe". This work was also supported by Generalitat Valenciana, Spain (reference CIAICO/2022/042) .
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Generalized analytical modelling of rotor asymmetry faults using winding and connection tensor formalism

  • Terrón-Santiago, Carla|||0000-0001-8178-3331
  • Delfa-Baena, Sandra
  • Burriel-Valencia, Jordi|||0000-0002-1680-4412
  • Sapena-Bano, Angel|||0000-0002-3888-6498
[EN] Purpose The main objective of this purpose is to provide a general and systematic method for obtaining the inductance matrices of induction machines under rotor asymmetry faults without requiring a full redefinition of the winding model for each case.Design/methodology/approach This paper combines the winding tensor formalism with the use of connection tensors to automatically generate the inductances corresponding to any configuration of broken bars.Findings Firstly, it introduces a unified mathematical framework capable of modelling arbitrary rotor asymmetries using only the inductances of the healthy machine. Secondly, it demonstrates how connection tensors can encode the structural effects of faults in a compact and algebraically tractable manner. Thirdly, it provides a scalable and computationally efficient procedure that avoids the cumbersome derivation of fault-specific inductance expressions, thereby facilitating both numerical simulation and the development of diagnostic tools.Originality/value The methodology presented in this paper offers a significant simplification of the modelling process while preserving the accuracy required for the analysis of fault-induced harmonic phenomena., This publication is part of the R+D+I project (grant PID2021-128013OBI00) funded by MCIN/AEI/10.13039/501100011033 and by "ERDF A way of making Europe". This work was also supported by Generalitat Valenciana (reference CIAICO/2022/042).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Contribución de congreso (ConferenceOutput)

Development an Artificial Intelligence Diagnostic System for Low-Profile Wind Turbine Equipment

  • Terrón-Santiago, Carla|||0000-0001-8178-3331
  • Puche-Panadero, Rubén|||0000-0003-2090-1941
  • Burriel-Valencia, Jordi|||0000-0002-1680-4412
  • Manuel Pineda-Sanchez|||0000-0001-7844-8831
[EN] This research focus on improving diagnostic methods for low-profile wind turbines, which, due to their characteristics or price, do not justify the addition of a sensor array for monitoring. Therefore, the possibility of training an artificial intelligence AI system for wind turbine diagnosis is investigated. Unlike other research, this study focuses on developing a low-profile, cost-effective AI expert diagnostic system. The process involves training various AIs to verify their ability to detect faults in wind turbines, using a virtual wind turbine model for the database samples., This work was supported by the Spanish Ministerio de Ciencia, Innovación y Universidades (MCIU) , Agencia Estatal de Investigación (AEI) and the Fondo Europeo de Desarrollo Regional (FEDER) in the framework of the Proyectos I+D+I Retos de Investigación 2018 , project reference PID2021-128013OB-I00 (MCIU/AEI/FEDER_UE) and also by Generalitat Valenciana (reference CIAICO/2022/042).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Identification of abnormal conditions in induction motors from current spectrum images using a two-stage approach with progressive learning

  • Barrera-Llanga, Kevin Iván
  • Burriel-Valencia, Jordi|||0000-0002-1680-4412
  • Sapena-Bano, Angel|||0000-0002-3888-6498
  • Martinez-Roman, Javier|||0000-0001-7544-8481
[EN] Background and objectives: This study presents a fault diagnosis system for induction machines based on a two-stage architecture using Convolutional Neural Networks (CNN). The aim is to improve fault identification by simplifying the classification process through sequential modeling. Methods: A dataset of 5404 images was generated from the Fourier spectra of current signals acquired from a test bench under five conditions: healthy machine, rotor asymmetry fault, broken rotor bar, race bearing fault, and ball bearing fault. A CNN based on the Visual Geometry Group (VGG) architecture was trained from scratch and then adapted using transfer learning. The classification strategy follows two steps: first, distinguishing healthy from faulty machines; then, identifying the specific fault type. Results: The system reached 96.5% accuracy in the first stage and 98.5% in the second. All main performance metrics (sensitivity, specificity, precision, F1-score) remained above 95%. The behavior of the models was examined using Uniform Manifold Approximation and Projection (UMAP), which showed clearer separation between conditions in the latent space when using the sequential approach. In addition, Gradient-Weighted Class Activation Mapping (Grad-CAM) visualizations provided insights into the image regions influencing each decision, showing consistent focus on spectral areas related to each condition. Conclusions: The combination of image-based preprocessing, sequential classification, and model interpretation techniques leads to accurate predictions and helps to understand how the models behave. These features support its use in predictive maintenance tasks for industrial applications., The article is funded by the R+D+i project (PID2021-128013OB-I00) through the MCIN/AEI/10.13039/-501100011033 and ERDF A way of making Europe, as well as by Generalitat Valenciana, Spain (CIAICO/2022/042). Additionally, I personally acknowledge funding from the Secretary of Higher Education, Science, Technology and Innovation of Ecuador as an individual grant holder.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Condition Monitoring of In-Service DFIGs Working Under Non-Stationary Conditions via NsHOTA: A Motor Current Signature Approach

  • Delfa-Baena, Sandra
  • Terrón-Santiago, Carla|||0000-0001-8178-3331
  • Burriel-Valencia, Jordi|||0000-0002-1680-4412
  • Artigao, Estefania
  • Honrubia-Escribano, Andres
  • Gomez-Lazaro, Emilio
[EN] The reliability of wind turbines largely depends on the ability to detect electrical and mechanical faults under variable operating conditions. This paper applies the Non-steady-state Harmonic Order Tracking Analysis (NsHOTA) method to the diagnosis of doubly-fed induction generators (DFIGs) in real wind turbines. Unlike other steady-state and transient techniques, NsHota stabilizes and enhances fault components in any operating regime, allowing for more in-depth analysis. Therefore, this method enables highly accurate fault diagnosis, allowing the measurement and analysis of small degradations over time. The method is validated using eight months of field data from an 850 kW DFIG previously diagnosed with mixed eccentricity. The results demonstrate that NsHOTA improves the consistency and quality of fault feature extraction, reduces background noise, and avoids false negatives under steady and non-steady regimes. In the real data test, NsHOTA is also compared with the steady-state HOTA (SsHOTA) method. These findings confirm the robustness of NsHOTA for real-world wind turbine condition monitoring and highlight its potential integration into predictive maintenance systems., This publication is part of the R+D+i project (grant PID2021-128013OB-I00) funded by MCIN/AEI/10.13039/501100011033 and by "ERDF A way of making Europe". This work was also supported by Generalitat Valenciana (reference CIAICO/2022/042).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Two-Level FEM-Based Inductance Modeling for Real-Time Simulation of Bar Breakages in Squirrel Cage Induction Machines

  • Delfa-Baena, Sandra
  • Terrón-Santiago, Carla|||0000-0001-8178-3331
  • Manuel Pineda-Sanchez|||0000-0001-7844-8831
  • Sapena-Bano, Angel|||0000-0002-3888-6498
[EN] This paper presents an innovative two-level modeling approach for accurately simulating the inductance matrix of induction machines (IMs) under fault conditions. The first level uses Finite Element Method (FEM) simulations to extract the inductance matrix L(¿) of the healthy machine, achieving high spatial resolution with minimal computational resources. The second level enables the introduction of rotor asymmetry faults using tensor algebra, thus eliminating the need for repeated FEM simulations. This approach enhances the efficiency of fault simulations, making it suitable for real-time diagnostic applications. The model is validated through both steady-state and transient analysis, demonstrating its effectiveness in reproducing fault-related harmonics and offering significant improvements in computational performance compared to traditional timestepping FEM models., This publication is part of the R+D+i project (grant PID2021-128013OBI00) funded by MCIN/AEI/10.13039/501100011033 and by ERDF A way of making Europe . This work was also supported by Generalitat Valenciana (reference CIAICO/2022/042).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A Comparative Analysis of Deep Learning Convolutional Neural Network Architectures for Fault Diagnosis of Broken Rotor Bars in Induction Motors

  • Barrera-Llanga, Kevin
  • Burriel-Valencia, Jordi|||0000-0002-1680-4412
  • Sapena-Bano, Angel|||0000-0002-3888-6498
  • Martinez-Roman, Javier|||0000-0001-7544-8481
[EN] Induction machines (IMs) play a critical role in various industrial processes but are susceptible to degenerative failures, such as broken rotor bars. Effective diagnostic techniques are essential in addressing these issues. In this study, we propose the utilization of convolutional neural networks (CNNs) for detection of broken rotor bars. To accomplish this, we generated a dataset comprising current samples versus angular position using finite element method magnetics (FEMM) software for a squirrel-cage rotor with 28 bars, including scenarios with 0 to 6 broken bars at every possible relative position. The dataset consists of a total of 16,050 samples per motor. We evaluated the performance of six different CNN architectures, namely Inception V4, NasNETMobile, ResNET152, SeNET154, VGG16, and VGG19. Our automatic classification system demonstrated an impressive 99% accuracy in detecting broken rotor bars, with VGG19 performing exceptionally well. Specifically, VGG19 exhibited high accuracy, precision, recall, and F1-Score, with values approaching 0.994 and 0.998. Notably, VGG19 exhibited crucial activations in its feature maps, particularly after domain-specific training, highlighting its effectiveness in fault detection. Comparing CNN architectures assists in selecting the most suitable one for this application based on processing time, effectiveness, and training losses. This research suggests that deep learning can detect broken bars in induction machines with accuracy comparable to that of traditional methods by analyzing current signals using CNNs., K Barrera-Llanga appreciates the financial support of the Secretary of Higher Education, Science, Technology and Innovation of Ecuador as a personal sponsor entity.