In this work, the authors present a novel method to identify and differentiate between levels of bearing corrosion in soft-started induction motors. This method relies on the computation of the permutation entropy of the stray-flux and current transient signals. In the case of soft-started induction motors, the detection of the fault related patterns is more difficult, due to the harmonics introduced in the signals by those devices. Nevertheless, the experimental results show that the fault studied can be not only identified but also the level of corrosion can be differentiated using the proposed method. The main advantages of this method are that it can be used online and its low computational requirements.

Induction motors are robust machines that are present in various sectors, including industry, transportation, mining, and others, employing a wide range of applications such as pumps, presses, and mills, among others. Despite their robustness, induction motors can exhibit both electrical and mechanical faults, with mixed eccentricity being a mechanical rotor fault that can occur due to errors in the manufacturing process or incorrect bearing mounting, which in turn can cause adverse consequences for the machine. In recent years, various methodologies have been studied and presented for fault diagnosis in induction motors, although many of them present certain limitations, as in some cases, some phenomena may mask the fault signals, and even in some cases, the load may affect the fault-related frequency amplitudes, thus leading to false diagnoses. This paper analyzes the effect of the load on the amplitude of the rotor rotational frequency components of the current signals of an induction motor operating under mixed eccentricity fault by applying the Fast Fourier Transform and examining the frequency components at four different load levels.

The monitoring of vibration signals acquired in non-rotating parts represents a widespread methodology for induction machine health assessment. These signals provide direct insights about mechanical and electromagnetic signatures. However, detecting mechanical defects such as shaft misalignment remains challenging due to the complex interactions between mechanical and electromagnetic domains. The accurate measurement and acquisition of vibration signals plays a key role on elucidating possible fault indicators, which stem from different multi-physical mechanisms. The present paper presents an experimental study to analyze the effect of sensor location and placement within an induction machine housing for accurate mechanical fault identification. The comparison is performed for classical time and frequency-domain fault indicators. In addition, different time-frequency evolutions of start-up transient signals are introduced and compared.

DC electrical motors and generators continue to play an important role in many industrial applications despite their declining use. A poor maintenance may yield significant economic losses for the industries where they operate. One of the common symptoms of failure in these machines is the presence of sparks in the interface between brushes and commutator. This phenomenon is usually linked with a diversity of anomalies and defects which may even lead to the forced motor outage. Therefore, its detection in its incipient stage and the quantification of its severity becomes crucial to adopt proper maintenance actions in due time to prevent the corresponding fault. This paper explores an innovative method to detect and assess the sparking phenomenon in DC machines. The method is based on the analysis of steady-state and transient stray flux signals and the further identification of low and high frequency harmonics linked with the fault. The results prove the high potential of the approach for becoming a reliable tool for the assessment of the commutation quality in DC machines.

Stray flux analysis has proven an effective method for electrical machine, and especially induction motor, diagnosis due to its fully non-invasive, affordable approach. State-of-the-art has successfully combatted the most important shortcomings via signal processing and sensor spatial arrangement techniques. The resulting time-frequency domain spectrogram is an information rich, ideal candidate for feature extraction. This work proposes a complete, adaptable stray flux analysis setup for collection of the most common faults' pertinent information with respect to machine learning requirements, compliant with perceived industrial limitations. A representative experimental setup is utilized as proof of concept.

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

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.

The widespread use of induction motors in industry has led to an increase in the importance of predictive maintenance to avoid untimely shutdowns. On the other hand, their extensive utilization has yielded the proliferation of soft starters that are employed to avoid high currents at start-up. Fault diagnosis techniques such as stray-flux analysis are of interest to diagnose and prevent motor failures. In this regard, this paper studies the validity of the stray-flux-based techniques in motors started via soft-starters. More specifically, it is proposed the analysis of the electromotive forces induced under the starting by the stray-flux in external sensors attached to the motor frame. Various levels of failure at different sensor positions are studied. Finally, a fault severity indicator is introduced based on components energies that appear, and different fault levels are compared. The results enable us to prove the potential of the stray-flux based technique for the rotor condition monitoring of soft-started motors.

Stray flux analysis is an interesting source of information for the diagnosis of Induction Motors (IMs). The widespread use of these motors in industry leads to a necessity of additional tools and methods for their predictive maintenance. On the other hand, soft-starters are increasingly used to reduce the high consumption of IMs at start-up. In this work, AI techniques based on convolutional neural networks are applied to detect rotor faults in soft-started motors. The objective is the automatic early detection of broken bars, avoiding the necessity of user intervention to interpret the obtained results. This work proves the potential of the methodology, including a successful set of experimental results.