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.

Condition monitoring and predictive maintenance of induction motors have great relevance in industrial applications. Nowadays, there are different techniques to analyze electronic signals from different types of sensors. At the same time, the application of FPGA-based hardware acceleration has gained traction in recent years due to the increasing demand for time and energy optimization. However, the application of hardware-accelerated algorithms in condition monitoring applications has yet to be explored despite its potential to improve online fault detection systems. This paper presents the implementation of an FPGA-based hardware-accelerated thermal image processing pipeline on the PYNQ Z2 board based on two image statistical features (mean and standard deviation) to detect five different induction motor mechanical fault conditions: misalignment, unbalanced load, bearing defect on the outer race, two broken rotor bars, and a healthy case that was used as reference. The hardware implementation of the thermal image statistical feature computation made it possible to reduce the computational load and the computation time on the development board.

Stator faults are one of the most common types of failure in induction machines. This type of motors are widely used in industry and predictive maintenance becomes crucial. Hence, this paper presents a methodology for the detection of stator asymmetries that enhances the identification of fault components and its severity. The method is based on the capture of current signals and on the subsequent application of a time-frequency transform, in this case, the Short Time Fourier Transform (STFT). The results shows clear signatures linked the fault that appear in the time-frequency map. Moreover, the intensities of these signatures increases as the fault gets worst, so different fault indicators are introduced for determining the fault severity level. The results of this work, obtained for different levels of load, allow to detect the presence of stator asymmetries and its severity

This paper shows the preliminary results of a proposed methodology for outer race fault diagnosis of contaminated rolling bearings. The proposed method is based on a low-cost thermographic sensor. This sensor allows for the acquisition of thermographic images, which are subsequently processed by means of statistical and non-statistical indicators. The obtained results shows a correct fault classification of 4 condition states as follows: healthy rolling bearing, outer race fault + 1 g of magnetite-contaminated bearing, outer race fault + 2 g of magnetite-contaminated bearing, and outer race fault + 3 g of magnetite-contaminated bearing

Induction motors play a fundamental role in most industrial processes due to their high efficiency and robust performance. For this reason the condition monitoring of these equipment is of great relevance. One of the most studied motor components is the bearing. Bearings reduce friction in the rotor allowing it to move freely, being a key factor in the correct operation of the induction motor. The most common bearing failures occur in the outer race. In this work we present the development of a system based on thermographic image processing to identify gradual failures in the outer bearing race of an induction motor. The bearing failure cases were induced by drilling holes in the outer race of metallic bearings with ascending diameters of 1 mm, 2 mm, 3 mm, 4 mm, and 5 mm. The experiments were carried out in a kinematic chain integrated by an induction motor and load elements. Thermographic images were acquired with a low-cost infrared sensor, and then segmented into three zones: motor rotor, motor body, motor backside. Subsequently, fifteen statistical parameters were calculated and processed by principal component analysis. The system obtained two characteristic features capable of effectively differentiating the bearing failures.

Within the manufacturing industry, the condition of cutting tools directly impacts the quality and costs of machining, which has driven the development of various methodologies to identify and monitor tool wear. In recent years, the stray magnetic flux has started to be used as a physical variable for wear detection, giving good results both individually and in combination with other physical quantities. The present research proposes the analysis of stray magnetic flux signals in a CNC lathe in conjunction with image analysis of the machined surfaces and the cutting tools in order to detect tool flank wear with cutting speed variation in the machining of 6061 aluminum. The results report the correct detection of the different levels of wear, demonstrating the ability of each of the methodologies to detect the level of wear regardless of the cutting speed used for machining, as well as the future improvement with the fusion of the techniques to obtain more reliable results for a robust detection system.

Bearing failures are among the most common faults in induction motors. In industry, their detection is usually carried out through the analysis of vibration data. However, there are applications in which this technique cannot be used. In these cases, the use of alternative quantities is primordial to reach a reliable conclusion of the bearing condition. The analysis of current and stray-flux signals has been proposed as alternative way to diagnose the bearing condition; despite it has been proven the complexity of using these signals for reaching an accurate conclusion about the condition of the bearings, it has been also demonstrated their potential for obtaining very useful information for the diagnosis of these elements. This work proposes the computation of statistical indicators obtained from stray-flux signals to detect the presence of corrosion in bearings. It is proven the usefulness of some of these indicators not only to detect this fault but also to discriminate versus other failures. In the work it is also shown the superiority of using this approach with stray-flux signals compared to other electrical quantities with regards to the detection of the considered fault.

Magnetic flux analysis is a condition monitoring technique that is drawing the interest of many researchers and motor manufacturers. The great enhancements and reduction in the costs and dimensions of the required sensors, the development of advanced signal processing techniques that are suitable for flux data analysis, along with other inherent advantages provided by this technology, are relevant aspects that have allowed the proliferation of flux-based techniques. This article reviews the most recent scientific contributions related to the development and application of flux-based methods for the monitoring of rotating electric machines. Particularly, aspects related to the main sensors used to acquire magnetic flux signals as well as the leading signal processing and classification techniques are commented on. The discussion is focused on the diagnosis of different types of faults in the most common rotating electric machines used in industry, namely: squirrel cage induction machines, wound rotor induction machines, permanent magnet machines, and wound field synchronous machines. A critical insight of the techniques developed in the area is provided and several open challenges are also discussed.

Tool condition monitoring (TCM) is one of the most relevant tasks during a machining process. The latest high-quality productivity standards make it essential to monitor the cutting tool wearing. Current TCM methodologies demand the installation of sensors near the working area, which in practical terms, it is not the most optimal solution since the final diagnosis can be disturbed by noisy signals and direct interferences with the machining process. This article proposes a novel noninvasive methodology based on the time–frequency analysis of the stray flux captured around the spindle-motor to detect and estimate the wearing level in cutting tools. Moreover, a new fault indicator based on this quantity is introduced through the application of the discrete wavelet transform. The results obtained are promising and demonstrates the effectiveness of the proposal to become a complementary source of information to classical approaches. This is validated with a Fanuc Oi mate computer numeric control turning machine for three different cutting tool wearing levels and different cutting depths.

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.

Kinematic chains have a fundamental role in the modern industry thanks to the great variety of applications where they can be found. For this reason the development of new fault detection methods has gained traction in recent years. Different physical signals have been used for the diagnosis of kinematic chains such as electric current, mechanical vibrations, or stray flux; nonetheless, the analysis of temperature signals measured by thermographic cameras has proven to be an effective way to detect certain types of failures complementing the existing work on this area. This paper presents the development of an infrared image processing system for the detection of electromechanical faults (misalignment, unbalance, broken bars, bearing defects, and gearbox wear) on a kinematic chain composed by an induction motor, an output pulley, a plastic transmission band, and an alternator. The system is based on the automatic segmentation of a region of interest associated to the kinematic chain through the implementation of image processing techniques and the calculation of statistical characteristics from the histogram of a thermal image acquired by a low-cost thermographic camera.

Critical situations such as the current pandemic produced by the Sars-Cov-2 (COVID-19) are leading to the incorporation of new technologies, such as virtual reality (VR), as a part of the teaching and training methods that facilitates the teaching task in contrast with the traditional learning methods that have been used in the academia. In this regard, in this work is proposed the design and development of a VR-based tool that is applied in the teaching and training of condition-based maintenance strategies in induction motors. The design of this tool is performed under Oculus Quest and includes the development of the 3D virtual environments, the creation of the VR learning experience, and finally the necessary adjustments for the VR application to render smoothly on standalone devices. The usability of this experience tool has been experimentally validated with different students by giving to the students practical knowledge but also a learning experience.

In the modern industry framework, the application of condition monitoring schemes over electromechanical systems is being subjected to demanding requirements. Currently, the massive digitalization of industrial assets allows the investigation towards multiple monitoring strategies capable of emphasize deviations over the nominal system operation. However, the most prominent techniques, such as Machine Learning, present great challenges in complex systems. In this regard, the proposed study presents the analysis of the diagnostic capabilities resulting from the classical approaches based on machine learning facing to complex electromechanical systems that implies a working environment subject to different operation condition, configurations with multiple components and the presence of faults of different nature (mechanical, electrical, electromagnetic), under isolated or combined scenarios. Discriminative feature extraction capabilities and classification accuracy will be analyzed as performance measures.

Condition monitoring applied to power quality involves several techniques and procedures for the assessment of the electrical signal. Data-driven approaches are the most common methodologies supported on data and signal processing procedures. Electrical systems in factory automation become more complex with the increase of multiple load profiles connected, and unexpected electrical events can occur causing the appearance of power quality disturbances. However, emerging technologies in the techniques related to the detection and identification of power quality disturbances are analyzed and compared according to the complexity of the current electrical system, that is, including simple and combined disturbances. These new technologies allow developing more cyber-physical systems to process the new methodologies for condition monitoring. Thus, in this study, a deep learning-based approach for the identification of power quality disturbances is implemented and their performance analyzed in front of classical disturbances defined by the International standards considered in the related literature.

During a machining process, it is very important to carry out the corresponding tasks using cutting tools in a healthy condition, since the quality of the manufactured component can be seriously deteriorated when using worn tools, leading to a possible increase of production costs. Current methodologies are capable of monitoring and diagnosing the wear on the cutting tool adequately. However, in most of them the use of sensors located near the working area is necessary and indispensable, which in practical terms is not the most optimal solution since it is desirable to rely on non-invasive techniques. In this work, the impact of the use of worn cutting tools on the stray flux captured around the induction motor (IM) driving the spindle chuck is studied using time-frequency decomposition mathematical tools (and, more specifically, the short-time Fourier transform (STFT)). The experimentally obtained results show the appearance of some specific frequency components having considerable amplitudes for worn cutting tools, and also demonstrates the potential that this technique has to provide highly relevant information to determine the presence of worn cutting tools. Therefore, it becomes an excellent informational source for the diagnosis of these types of failures, taking advantage of its non-invasive nature.

The detection of uncharacterized events during electromechanical systems operation represents one of the most critical data challenges dealing with condition-based monitoring under the Industry 4.0 framework. Thus, the detection of novelty conditions and the learning of new patterns are considered as mandatory competencies in modern industrial applications. In this regard, this article proposes a novel multifault detection and identification scheme, based on machine learning, information data-fusion, novelty-detection, and incremental learning. First, statistical time-domain features estimated from multiple physical magnitudes acquired from the electrical motor under inspection are fused under a feature-fusion level scheme. Second, a self-organizing map structure is proposed to construct a data-based model of the available conditions of operation. Third, the incremental learning of the condition-based monitoring scheme is performed adding self-organizing structures and optimizing their projections through a linear discriminant analysis. The performance of the proposed scheme is validated under a complete set of experimental scenarios from two different cases of study, and the results compared with a classical approach.

Condition monitoring in electromechanical systems represents, currently, one of the most critical challenges dealing with the advancement and modernization in intelligent manufacturing. In this regard, machine learning based algorithms widely applied in other technological fields are being considered now to face the automatic feature extraction on the electric machine monitoring. In this study, a monitoring scheme is considered for faults detection performance evaluation, where vibrations signal under different fault conditions are acquired. Thus, the common electric machine monitoring framework, that is, a set of features estimated from a limited number of measurements, is considered in front of the three main dimensionality reduction approaches: principal component analysis, linear discriminant analysis and auto-encoder based. Performance of the corresponding approaches are studied and discussed experimentally. It is revealed that, although scheme based on auto-encoder provides enhanced diagnosis results, it is still necessary to carry out a detailed study on the automatic extraction capabilities of important features for the detection of faults.

The demand of electric power supply has increase in the industry due to most of its processes are involved with the use of electrical equipment and machines. Electric power generation due to integration of new energy sources has become an important area of continuous development in which Power Quality (PQ) problems must be faced. In this paper is proposed a condition monitoring strategy based on the estimation of statistical time domain-based features and Linear Discriminant Analysis to identify different PQ disturbances in a cogeneration system. The proposed method is first evaluated with a set of synthetics signals that include different PQ disturbances and then evaluated a real data acquired from a cogeneration system. The final diagnosis outcome is performed by means of a Neural Network. The obtained results shown that the proposed method is suitable for being applied in cogeneration system to identify different PQ disturbances.

The analysis of the external magnetic field has been proven to be an effective way to diagnose different types of faults in induction motors. Classical methods based on permanent regime are being complemented by recent approaches that rely on the analysis of the magnetic field during transient operation of the machine. Most of these approaches have been applied with success to detect failures in cage induction motors, but few works have dealt with their wound rotor counterparts. This paper proposes the analysis of the magnetic field in the vicinity of the machine to detect rotor failures in wound rotor induction motors. Classical methods based on the Fast Fourier transform of steady-state signals are compared with recent methods relying on the analysis of signals captured under starting. Different positions for the considered flux sensors are considered and novel indicators are presented for the determination of the level of fault severity.

Induction motors are fundamental elements in the industry field since they perform several tasks under a broad variety of conditions, a situation that affects its performance making them susceptible to fail despite its robustness. Over several decades, many techniques and methodologies to assess the healthiness state of electrical motors have been proposed and implemented successfully. In this regard new advances in the signal processing field has taken a great importance since these emerging tools are used to accomplish fault diagnosis tasks in order to increase the reliability of such processes. So, it is very important to explore the use of new signal processing techniques to detect and diagnose in a timely manner faults in electrical motors. In this work, it is proposed to use the wavelet entropy of the stray flux signal captured by a coil sensor to estimate the winding insulation status of induction motors, a very common failure presented on this type of drives, that if not attended on time, it can end in a catastrophic and irreversible fault in a matter of minutes. The proposal uses suitable time-frequency decomposition (TFD) tools to isolate the studied fault. The results obtained show that the methodology proposed here can be considered as an excellent alternative for estimating the winding insulation healthiness state of an induction motor online and efficiently, as well as being possible to implement it in a programmable logic device.

Infrared thermography (IRT) is a noninvasive technique that is drawing an increasing attention in industry. The spectacular advancement in the features of the infrared cameras that has come together with their progressive cost reduction has expanded the use of this technique to many industrial applications that were unfeasible just a few years ago. This paper compiles and comments the most recent scientific contributions related to the application of this technique in the industrial context. The paper classifies the analyzed references into three main groups: electrical, mechanical, and other applications. Especial emphasis is made on induction-motor-related applications of the IRT due to the extensive participation of these machines in the industrial context. The paper provides a critical review of most of the analyzed references, emphasizes the way in which the infrared technique is applied to the specific application and presents the limitations and pending issues as well as future challenges regarding the application of the technique.

This study is focused on the current challenges dealing with electromechanical system monitoring applied in industrial frameworks, that is, the presence of unknown events and the limitation to the nominal healthy condition as starting knowledge. Thus, an industrial machinery condition monitoring methodology based on novelty detection and classification is proposed in this study. The methodology is divided in three main stages. First, a dedicated feature calculation and reduction over each available physical magnitude. Second, an ensemble structure of novelty detection models based on one-class support vector machines to identify not previously considered events. Third, a diagnosis model supported by a feature fusion scheme in order to reach high fault classification capabilities. The effectiveness of the fault detection and identification methodology has been compared with classical single model approach, and verified by experimental results obtained from an electromechanical machine.

A great deal of investigations are being carried out towards the effective implementation of the 4.0 Industry new paradigm. Indeed, most of the machinery involved in industrial processes are intended to be digitalized aiming to obtain enhanced information to be used for an optimized operation of the whole manufacturing process. In this regard, condition monitoring strategies are being also reconsidered to include improved performances and functionalities. Thus, the contribution of this research work lies in the proposal of an incremental learning framework approach applied to the condition monitoring of electromechanical systems. The proposed strategy is divided in three main steps, first, different available physical magnitudes are characterized through the calculation of a set of statistical-time based features. Second, a modelling of the considered conditions is performed by means of self-organizing maps in order to preserve the topology of the data; and finally, a novelty detection is carried out by a comparison among the quantization error value achieved in the data modelling for each of the considered conditions. The effectiveness of the proposed novelty fault identification condition monitoring methodology is proved by means of the evaluation of a complete experimental database acquired during the continuous working conditions of an electromechanical system.

In recent decades, the use of infrared thermography (IRT) has proliferated in a wide diversity of industrial applications. The spectacular enhancement of the infrared data acquisition equipment together with the significant decrement in their cost has enabled the utilization of this technology in applications and processes where it was not even considered years ago. This Special Section was conceived to attract recent investigations proposing the use of IRT in different cases concerning the monitoring and diagnosis in industrial machines, materials, and facilities. The 15 selected papers are illustrative of the dynamic activity in this research area, as well as the diversity of new applications in which the technology can be employed as the main detection tool or in some applications as a complementary tool of diagnosis.

The procedures for condition monitoring of electromechanical systems are undergoing a reformulation, mainly, due to the current thermographic affordability of infrared cameras to be incorporated in industrial applications. However, high-performing multifault data-driven methodologies must be investigated in order to infer reliable condition information from the thermal distribution of not only electrical motors but also of shafts and couplings. To address this issue, a novel thermography-based methodology is proposed. First, the infrared capture is processed to obtain a thermographic residual image of the kinematic chain. Second, the thermal distribution of the image's regions of interest is characterized by means of statistical features. Finally, a distributed self-organizing map structure is used to model the nominal thermal distribution to subsequently perform a fault detection and identification. The method provides a reliability quantification of the resulting condition assessment in order to avoid misclassifications and identify the actual fault root-causes. The performance and the effectiveness of the proposed methodology is validated experimentally and compared with the classical maximum temperature gradient procedure.

Power quality (PQ) and PQ disturbances (PQD) are relevant for the industry due to the implied costs in most industrial processes. Besides, it is necessary to maintain the quality standards of the electrical grid to avoid damages in the equipment that is connected to the grid. Due to the nature and characteristics of the PQD present in the voltage and current signals, several studies have focused on detecting and classifying particular disturbances, or simple combinations between two or three of them, without presenting a methodology that describes all of them automatically. Hence, this paper proposes a hybrid approach integrating genetic algorithms (GA) and particle swarm optimization (PSO) with other techniques that make use of their individual capabilities to automatically find a wide range of PQD present in a voltage or current signal, regardless of their nature. To achieve this hybrid approach parameterization, a full PQD model is adopted to automate the search of every one of their parameters. The proposed approach is validated through synthetic signals, real data from the IEEE data base, and through data readings from a real process. A comparison using other recent heuristic techniques is made to show the robustness of the proposed hybrid approach.

Induction motors and electric machines in general are important components in most industries. The condition monitoring of these machines is relevant at the industrial facilities for reducing costs and minimizing maintenance downtime. Infrared thermography is a non-invasive technique that allows on-line monitoring of electric machines; characteristic that makes it a suitable solution for industrial environments. However, some external parameters like the camera positioning and location can affect the estimations carried out by the camera. This work presents a methodology for thermal analysis on induction motors, showing how the location of the camera affects the thermography imaging process. Images are taken from different locations, relative to the monitored machine, and a comparative of the results is carried out for an experimental case of study consisting in a healthy induction motor and a motor with a damaged bearing. Results show that despite the fact that a change on the camera location produces a variation on the estimation of the temperature; the difference is not significant for applications on electric machines.

Industrial machinery makes extensive use of induction motors as primary motion supplies for the associated kinematic chain. These motors and the kinematic chain are susceptible to failures in one or several of the components making the detection of the faults a major issue for industries. Thermography is a technique that has been used for monitoring and diagnosis in industrial facilities and it is suitable for the monitoring of induction motors and the associated kinematic chain. This technique is an aid for the detection of faults and the diagnosis of the operating condition of industrial machinery. Several research works have used thermography for this purpose, but the problem is the manual adjustment that needs to be done to the thermal camera in order to obtain thermal images, named thermograms, that give the true temperature readings of the objects in focus. This paper presents a novel methodology that makes the adjustment of the thermal camera in an automated way, using additional external temperature sensors to calibrate the thermal images provided by the low-cost thermal camera to give readings of the true temperature of the objects. Experimentation is performed on an induction motor with an associated kinematic chain to test the efficiency of the proposed methodology.

Gearboxes are important components in industrial applications and its condition monitoring is relevant in industries for reducing costs and minimizing maintenance downtime. Diagnosing wear in gearboxes saves time to prepare appropriate corrective actions, and to ensure that the system does not deteriorate critically. Nowadays, for condition monitoring in gearboxes, vibration analysis is commonly used due to its high reliability. In this work, a reliable methodology for diagnosing different levels of wear in a gearbox through vibration signals, and supported by a theoretical model, is proposed. The theoretical model is based on calculating the characteristic frequencies of the gearbox, with the aim for locating the spectral components of the faults in the vibration signal. Experimentation is done to a healthy gearbox and three wear levels. Results show the reliability of this method that makes it suitable to be used in diagnosing industrial machinery such as in automotive manufacturing applications.

Asymmetrical loads in power systems leads to unbalance voltage/current signals and consequently, adverse effects such as overheating, lifetime, torque, and speed reduction, among others. In order to compensate the level of unbalance and avoid the above mentioned effects, the instantaneous unbalance and symmetrical components estimations have to be carried out. In this work, a methodology for estimating the instantaneous unbalance and symmetrical components through the Hilbert transform (HT) in three-phase voltage and current systems is proposed. First, a bandpass filter is used to remove the harmonics/interharmonics components; afterwards, the HT is applied to obtain the instantaneous amplitude and phase of each voltage/current signal. Finally, the symmetrical components and unbalance percentage are computed. The proposed approach is validated and tested through synthetic signals and under real operating conditions. Besides, a field programmable gate array (FPGA)-based implementation is developed as a low-cost and portable System-on-Chip (SoC) solution.

Nowadays industry pays much attention to prevent failures that may interrupt production with severe consequences in cost, product quality, and safety. The most-analyzed parameters for monitoring dynamic characteristics and ensuring correct functioning of systems are electric current, voltage, and vibrations. System-on-chip (SoC) design is an approach to increase performance and overcome costs during equipment monitoring. This work presents the design and implementation of a low-cost SoC design that utilizes reconfigurable hardware and a customized embedded processor for time-frequency analysis on industrial equipment through short-time Fourier transform and discrete wavelet transform. Three study cases (electric current supply to an induction motor during startup transient, voltage supply to an induction motor through a variable speed drive, and vibration signals from industrial-robot links) show the suitability of the proposed monitoring system for time-frequency analysis of different signals in distinct industrial applications, and early diagnosis and prognosis of abnormalities in monitored systems.

Speed observation of rotating machines under different operational conditions is required in applications from fault detection to traction. Sensorless rotating speed observation has become popular in recent years. The MRAS rotor-flux-oriented and synchronous models are common schemes used for sensorless speed observation. These techniques are affected under low rotating frequencies and variable load conditions. This work presents an experimental analysis on the performance of both schemes during rotating speed estimation under different rotational frequencies and variable-load conditions. Obtained results show performance deficiencies on their estimations according to the rotational speed and load. A double scheme methodology that mixes the MRAS-based and the synchronous speed observers through a compensation function is proposed. Its performance analysis indicates an improvement on the estimated speed at different rotational frequencies with variable load. An FPGA-based implementation is developed as a low-cost, portable system-on-chip solution for real-time online estimation of rotating speed on induction motors.