Ponències/Comunicacions de congressos
http://hdl.handle.net/2117/3543
2024-03-28T16:32:57Z
-
Deep learning-based partial transfer fault diagnosis methodology for electromechanical systems
http://hdl.handle.net/2117/397444
Deep learning-based partial transfer fault diagnosis methodology for electromechanical systems
Arellano Espitia, Francisco; Delgado Prieto, Miquel; Valls Pérez, Joan; Saucedo Dorantes, Juan Jose; Osornio Rios, Roque Alfredo
Recently, transfer learning technology has provided valuable solutions to problems that are present in machinery with industrial applications. Through the use of transfer learning, basic diagnostic problems have been well addressed, especially in scenarios in which the training and test data are from different distributions. However, there are scenarios that require further consideration, such as partial fault diagnosis. In this paper, a deep learning-based fault diagnosis methodology is proposed to address the partial fault diagnosis problem, in which the data from the unsupervised target domain represents a category subspace of the full machine-state-label space. Specifically, a domain adaptation with adversarial learning schemes is proposed to achieve partial domain adaptation. The experimental results on an electromechanical test bench suggest that the proposed approach offers a practical solution to this partial fault diagnosis problem.
2023-11-30T15:35:31Z
Arellano Espitia, Francisco
Delgado Prieto, Miquel
Valls Pérez, Joan
Saucedo Dorantes, Juan Jose
Osornio Rios, Roque Alfredo
Recently, transfer learning technology has provided valuable solutions to problems that are present in machinery with industrial applications. Through the use of transfer learning, basic diagnostic problems have been well addressed, especially in scenarios in which the training and test data are from different distributions. However, there are scenarios that require further consideration, such as partial fault diagnosis. In this paper, a deep learning-based fault diagnosis methodology is proposed to address the partial fault diagnosis problem, in which the data from the unsupervised target domain represents a category subspace of the full machine-state-label space. Specifically, a domain adaptation with adversarial learning schemes is proposed to achieve partial domain adaptation. The experimental results on an electromechanical test bench suggest that the proposed approach offers a practical solution to this partial fault diagnosis problem.
-
Diagnosis electromechanical system by means CNN and SAE: an interpretable-learning study
http://hdl.handle.net/2117/371442
Diagnosis electromechanical system by means CNN and SAE: an interpretable-learning study
Arellano Espitia, Francisco; Delgado Prieto, Miquel; Martínez Viol, Víctor; Saucedo Dorantes, Juan Jose; Osornio Rios, Roque A.
Cyber-physical systems are the response to the adaptability, scalability and accurate demands of the new era of manufacturing called Industry 4.0. They will become the core technology of control and monitoring in smart manufacturing processes. In this regard, the complexity of industrial systems implies a challenge for the implementation of monitoring and diagnosis schemes. Moreover, the challenges that is presented in technological aspects regarding connectivity, data management and computing are being resolved through different IT-OT (information technology and operational technology) convergence proposals. These solutions are making it possible to have large computing capacities and low response latency. However, regarding the logical part of information processing and analysis, this still requires additional studies to identify the options with a better complexity-performance trade-off. The emergence of techniques based on artificial intelligence, especially those based on deep-learning, has provided monitoring schemes with the capacity for characterization and recognition in front of complex electromechanical systems. However, most deep learning-based schemes suffer from critical lack of interpretability lying to low generalization capabilities and overfitted responses. This paper proposes a study of two of the main deep learning-based techniques applied to fault diagnosis in electromechanical systems. An analysis of the interpretability of the learning processes is carried out, and the approaches are evaluated under common performance metrics.
© 2022 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.
2022-07-28T08:29:22Z
Arellano Espitia, Francisco
Delgado Prieto, Miquel
Martínez Viol, Víctor
Saucedo Dorantes, Juan Jose
Osornio Rios, Roque A.
Cyber-physical systems are the response to the adaptability, scalability and accurate demands of the new era of manufacturing called Industry 4.0. They will become the core technology of control and monitoring in smart manufacturing processes. In this regard, the complexity of industrial systems implies a challenge for the implementation of monitoring and diagnosis schemes. Moreover, the challenges that is presented in technological aspects regarding connectivity, data management and computing are being resolved through different IT-OT (information technology and operational technology) convergence proposals. These solutions are making it possible to have large computing capacities and low response latency. However, regarding the logical part of information processing and analysis, this still requires additional studies to identify the options with a better complexity-performance trade-off. The emergence of techniques based on artificial intelligence, especially those based on deep-learning, has provided monitoring schemes with the capacity for characterization and recognition in front of complex electromechanical systems. However, most deep learning-based schemes suffer from critical lack of interpretability lying to low generalization capabilities and overfitted responses. This paper proposes a study of two of the main deep learning-based techniques applied to fault diagnosis in electromechanical systems. An analysis of the interpretability of the learning processes is carried out, and the approaches are evaluated under common performance metrics.
-
Indoor monitoring system based on ARQ signaling generated by a Visible Light Communication link
http://hdl.handle.net/2117/367985
Indoor monitoring system based on ARQ signaling generated by a Visible Light Communication link
Bas, Joan; Ortega Redondo, Juan Antonio; Busquets González, Martí; Dowhuszko, Alexis
Visible Light Communications (VLC) is a candidate technology that complements the benefits of radio communication networks, particularly in those situations where a low-cost solution using license-free spectrum is required to enable ultra-dense deployments of indoor small cells. However, the main drawback that VLC has when compared to wireless communications on RF bands is that, in presence of obstacles between the transmitter and the receiver, full-blockage events are likely to happen as the power received on reflections is much weaker than the power of the blocked line-of-sight link. In this paper, we take advantage of this phenomenon and study the effect that different activities performed by people in the service area to-be-monitored have on the Automatic Repeat reQuest (ARQ) signaling of the VLC link. Based on the presented experimental studies, it is possible to conclude that different relevant events are able to be detected correctly according to the statistics of the ARQ signaling that is collected from the ongoing VLC transmission.
2022-06-02T12:27:11Z
Bas, Joan
Ortega Redondo, Juan Antonio
Busquets González, Martí
Dowhuszko, Alexis
Visible Light Communications (VLC) is a candidate technology that complements the benefits of radio communication networks, particularly in those situations where a low-cost solution using license-free spectrum is required to enable ultra-dense deployments of indoor small cells. However, the main drawback that VLC has when compared to wireless communications on RF bands is that, in presence of obstacles between the transmitter and the receiver, full-blockage events are likely to happen as the power received on reflections is much weaker than the power of the blocked line-of-sight link. In this paper, we take advantage of this phenomenon and study the effect that different activities performed by people in the service area to-be-monitored have on the Automatic Repeat reQuest (ARQ) signaling of the VLC link. Based on the presented experimental studies, it is possible to conclude that different relevant events are able to be detected correctly according to the statistics of the ARQ signaling that is collected from the ongoing VLC transmission.
-
Uncertainty analysis for industries investing in energy equipment and renewable energy sources
http://hdl.handle.net/2117/365978
Uncertainty analysis for industries investing in energy equipment and renewable energy sources
Urbano González, Eva María; González Abreu, Artvin Darién; Kampouropoulos, Konstantinos; Romeral Martínez, José Luis
This paper studies the optimal design and operation of new energy equipment including renewable energy sources for prosumer industries. In order to augment the interest of industries in performing energy actions, the economic parameters of the investment are analysed and the risk related to it considering the uncertainty in energy markets is evaluated. A two-stage optimization approach is proposed considering the whole lifetime of the energy equipment and an uncertainty analysis performed through the evaluation of the deterministic model under Latin Hypercube Samples of uncertain parameters. A case study based on a real industry is presented, whose results expose the robustness of the optimization methodology and the acceptable risk of investing in renewable energy sources and energy equipment for prosumer purposes.
2022-04-14T11:41:44Z
Urbano González, Eva María
González Abreu, Artvin Darién
Kampouropoulos, Konstantinos
Romeral Martínez, José Luis
This paper studies the optimal design and operation of new energy equipment including renewable energy sources for prosumer industries. In order to augment the interest of industries in performing energy actions, the economic parameters of the investment are analysed and the risk related to it considering the uncertainty in energy markets is evaluated. A two-stage optimization approach is proposed considering the whole lifetime of the energy equipment and an uncertainty analysis performed through the evaluation of the deterministic model under Latin Hypercube Samples of uncertain parameters. A case study based on a real industry is presented, whose results expose the robustness of the optimization methodology and the acceptable risk of investing in renewable energy sources and energy equipment for prosumer purposes.
-
Anomaly detection in electromechanical systems by means of deep-autoencoder
http://hdl.handle.net/2117/360838
Anomaly detection in electromechanical systems by means of deep-autoencoder
Arellano Espitia, Francisco; Delgado Prieto, Miquel; Martínez Viol, Víctor; Fernández Sobrino, Ángel; Osornio Rios, Roque A.
Anomaly detection in manufacturing processes is one of the main concerns in the new era of the Industry 4.0 framework. The detection of uncharacterized events represents a major challenge within the operation monitoring of electrical rotatory machinery. In this regard, although several machine learning techniques have been classically considered, the recent appearance of deep-learning approaches represents an opportunity in the field to increase the anomaly detection capabilities in front of complex electromechanical systems. However, each anomaly detection technique considers its own data interpretability and modelling strategy, which should be analyzed in front of the specificities of the data generated in an industrial environment and, specifically, by an electromechanical actuator. Thus, in this study, a comparison framework is considered including multiple fault scenarios in order to analyze the performance of four representative anomaly detection techniques, that is, one-class support vector machine, k-nearest neighbor, Gaussian mixture model and, finally, deep-autoencoder. The experimental results suggest that the use of the deep-autoencoder in the task of detecting anomalies of operation in electromechanical systems has a higher performance compared to the state of the art methods.
© 2021 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
2022-01-27T11:23:39Z
Arellano Espitia, Francisco
Delgado Prieto, Miquel
Martínez Viol, Víctor
Fernández Sobrino, Ángel
Osornio Rios, Roque A.
Anomaly detection in manufacturing processes is one of the main concerns in the new era of the Industry 4.0 framework. The detection of uncharacterized events represents a major challenge within the operation monitoring of electrical rotatory machinery. In this regard, although several machine learning techniques have been classically considered, the recent appearance of deep-learning approaches represents an opportunity in the field to increase the anomaly detection capabilities in front of complex electromechanical systems. However, each anomaly detection technique considers its own data interpretability and modelling strategy, which should be analyzed in front of the specificities of the data generated in an industrial environment and, specifically, by an electromechanical actuator. Thus, in this study, a comparison framework is considered including multiple fault scenarios in order to analyze the performance of four representative anomaly detection techniques, that is, one-class support vector machine, k-nearest neighbor, Gaussian mixture model and, finally, deep-autoencoder. The experimental results suggest that the use of the deep-autoencoder in the task of detecting anomalies of operation in electromechanical systems has a higher performance compared to the state of the art methods.
-
Performance assessment of a wide-bandgap-semiconductor dual-active-rridge converter for electrical vehicles
http://hdl.handle.net/2117/358984
Performance assessment of a wide-bandgap-semiconductor dual-active-rridge converter for electrical vehicles
Berbel Artal, Néstor; Capellá Frau, Gabriel José; Zaragoza Bertomeu, Jordi; Romeral Martínez, José Luis
Dc-dc converters can be found in different kinds of electric vehicles (EVs). Their main function is to accommodate voltages and currents to the motor or other EV systems requirements. The use of wide-bandgap (WBG) devices can improve the efficiency of silicon-based power converters, qualifying also for higher switching frequencies. In this article the features of a dual active bridge (DAB) converter are studied. The high voltage side of the DAB is implemented with Silicon Carbide (SiC) MOSFETs. For the low voltage side two types of devices are used: either Gallium Nitride (GaN) enhancement high-electronmobility transistors (e-HEMTs) or SiC MOSFETs. The influence of switching frequency and output power on the efficiency are evaluated. The parallel connection of GaN devices is proposed to overcome the device current limits and thus increase the overall DAB converter output power. A feedback controller has been designed to reduce the effects on the output voltage of load changes. The DAB converter evaluation has been realized by using MATLAB/Simulink and PLECS software.
2021-12-21T14:47:18Z
Berbel Artal, Néstor
Capellá Frau, Gabriel José
Zaragoza Bertomeu, Jordi
Romeral Martínez, José Luis
Dc-dc converters can be found in different kinds of electric vehicles (EVs). Their main function is to accommodate voltages and currents to the motor or other EV systems requirements. The use of wide-bandgap (WBG) devices can improve the efficiency of silicon-based power converters, qualifying also for higher switching frequencies. In this article the features of a dual active bridge (DAB) converter are studied. The high voltage side of the DAB is implemented with Silicon Carbide (SiC) MOSFETs. For the low voltage side two types of devices are used: either Gallium Nitride (GaN) enhancement high-electronmobility transistors (e-HEMTs) or SiC MOSFETs. The influence of switching frequency and output power on the efficiency are evaluated. The parallel connection of GaN devices is proposed to overcome the device current limits and thus increase the overall DAB converter output power. A feedback controller has been designed to reduce the effects on the output voltage of load changes. The DAB converter evaluation has been realized by using MATLAB/Simulink and PLECS software.
-
Support vector machine based novelty detection and FDD framework applied to building AHU systems
http://hdl.handle.net/2117/339885
Support vector machine based novelty detection and FDD framework applied to building AHU systems
Martínez Viol, Víctor; Martínez Viol, Víctor; Urbano González, Eva María; Urbano González, Eva María; Kampouropoulos, Konstantinos; Kampouropoulos, Konstantinos; Delgado Prieto, Miquel; Delgado Prieto, Miquel; Romeral Martínez, José Luis; Romeral Martínez, José Luis
The increasing energy consumption of heating, ventilation and air conditioning (HVAC) systems is one of the main concerns in the building sector. Fault detection technologies are now indispensable for energy efficiency and performance improvement. In this paper, a methodology for the robust and reliable fault detection and diagnosis is presented as a two-stage framework composed by an offline stage where the models are built and an online stage that is constantly receiving new samples. The system includes a novelty detection scheme developed using one-class support vector machines (OC-SVM) and a classifier built using SVM. The proposed strategy is applied to a dataset for a single-zone constant air volume air handling unit. The experimental results show that the novelty detection stage adds robustness layer to the typical classification scheme.
© 2020 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
2021-02-17T10:41:11Z
Martínez Viol, Víctor
Martínez Viol, Víctor
Urbano González, Eva María
Urbano González, Eva María
Kampouropoulos, Konstantinos
Kampouropoulos, Konstantinos
Delgado Prieto, Miquel
Delgado Prieto, Miquel
Romeral Martínez, José Luis
Romeral Martínez, José Luis
The increasing energy consumption of heating, ventilation and air conditioning (HVAC) systems is one of the main concerns in the building sector. Fault detection technologies are now indispensable for energy efficiency and performance improvement. In this paper, a methodology for the robust and reliable fault detection and diagnosis is presented as a two-stage framework composed by an offline stage where the models are built and an online stage that is constantly receiving new samples. The system includes a novelty detection scheme developed using one-class support vector machines (OC-SVM) and a classifier built using SVM. The proposed strategy is applied to a dataset for a single-zone constant air volume air handling unit. The experimental results show that the novelty detection stage adds robustness layer to the typical classification scheme.
-
Future European energy markets and Industry 4.0 potential in energy transition towards decarbonization
http://hdl.handle.net/2117/339874
Future European energy markets and Industry 4.0 potential in energy transition towards decarbonization
Urbano González, Eva María; Martínez Viol, Víctor; Kampouropoulos, Konstantinos; Romeral Martínez, José Luis
Climate change, economic growth and fossil fuel price volatility are forcing governments and thus society to adopt economical and technical measures in the energy sector to reach sustainability. These actions can be seen as opportunities for the stakeholders that form the energy market and also for new actors that may enter as a consequence of the energy transition that is taking place. In this paper, a description of the energy targets and potential market scenarios in Europe is carried out, together with a review of the policies implemented to achieve these objectives. Within this framework, the possibility of the industry to adopt a crucial role in the development of the new energy market is also analysed. The potential tools for its achievement are also presented, together with some of the techniques and mechanisms that make it feasible. From this study, it can be concluded that the industrial sector will become a major distributed prosumer, providing services to the energy market and facilitating the energy transition towards the decarbonization of the society.
2021-02-17T10:24:22Z
Urbano González, Eva María
Martínez Viol, Víctor
Kampouropoulos, Konstantinos
Romeral Martínez, José Luis
Climate change, economic growth and fossil fuel price volatility are forcing governments and thus society to adopt economical and technical measures in the energy sector to reach sustainability. These actions can be seen as opportunities for the stakeholders that form the energy market and also for new actors that may enter as a consequence of the energy transition that is taking place. In this paper, a description of the energy targets and potential market scenarios in Europe is carried out, together with a review of the policies implemented to achieve these objectives. Within this framework, the possibility of the industry to adopt a crucial role in the development of the new energy market is also analysed. The potential tools for its achievement are also presented, together with some of the techniques and mechanisms that make it feasible. From this study, it can be concluded that the industrial sector will become a major distributed prosumer, providing services to the energy market and facilitating the energy transition towards the decarbonization of the society.
-
HVAC early fault detection using a fuzzy logic based approach
http://hdl.handle.net/2117/339728
HVAC early fault detection using a fuzzy logic based approach
Martínez Viol, Víctor; Urbano González, Eva María; Delgado Prieto, Miquel; Romeral Martínez, José Luis
The need for improving the energy efficiency of existing buildings has driven to the implementation of building energy management systems (BEMS) that can help facilities manager to discover and identify problems that may cause energy wastage or affect to occupants’ comfort. Modern data-driven fault detection and diagnosis (FDD) make use of the data collected by the building BEMS to provide high accuracy in the revelation of heating, ventilation and air-conditioning (HVAC) system faults. However, these methods need a large amount of faulty data samples during the training, which is an uncommon situation in the real world. The main focus of this paper is to present a methodology to detect faults when the number of faulty samples is low. For this purpose, a regression-based methodology based on an adaptative neuro-fuzzy inference system (ANFIS) chiller model is developed using the data collected from a real use case. The model presents good results, that can be used for benchmarking the machine operation and detect the abnormal operation states.
2021-02-16T10:32:29Z
Martínez Viol, Víctor
Urbano González, Eva María
Delgado Prieto, Miquel
Romeral Martínez, José Luis
The need for improving the energy efficiency of existing buildings has driven to the implementation of building energy management systems (BEMS) that can help facilities manager to discover and identify problems that may cause energy wastage or affect to occupants’ comfort. Modern data-driven fault detection and diagnosis (FDD) make use of the data collected by the building BEMS to provide high accuracy in the revelation of heating, ventilation and air-conditioning (HVAC) system faults. However, these methods need a large amount of faulty data samples during the training, which is an uncommon situation in the real world. The main focus of this paper is to present a methodology to detect faults when the number of faulty samples is low. For this purpose, a regression-based methodology based on an adaptative neuro-fuzzy inference system (ANFIS) chiller model is developed using the data collected from a real use case. The model presents good results, that can be used for benchmarking the machine operation and detect the abnormal operation states.
-
Vector control of crosswise saturating five-phase PMaSynRM in wide speed range
http://hdl.handle.net/2117/337112
Vector control of crosswise saturating five-phase PMaSynRM in wide speed range
Michalski, Tomasz Dobromir; Acosta Cambranis, Fernando Geovany; Romeral Martínez, José Luis; Zaragoza Bertomeu, Jordi
This paper deals with the realization of a five-phase cross saturating permanent magnet assisted synchronous reluctance motor (PMaSynRM) drive with vector control based on maximum torque per ampere (MTPA) and flux weakening (FW) control strategies derived from the identified flux maps originating from the finite element analysis (FEM). Tracking of the reference currents in dq1 and dq3 axes is guaranteed with the proposed approach and the voltage and current constraints are not exceeded at any working condition.
2021-02-09T08:49:05Z
Michalski, Tomasz Dobromir
Acosta Cambranis, Fernando Geovany
Romeral Martínez, José Luis
Zaragoza Bertomeu, Jordi
This paper deals with the realization of a five-phase cross saturating permanent magnet assisted synchronous reluctance motor (PMaSynRM) drive with vector control based on maximum torque per ampere (MTPA) and flux weakening (FW) control strategies derived from the identified flux maps originating from the finite element analysis (FEM). Tracking of the reference currents in dq1 and dq3 axes is guaranteed with the proposed approach and the voltage and current constraints are not exceeded at any working condition.