The early fault detection in the rotary electrical machines,such as induction motors (IMs), has been growing in modern industry. IMs have been widely used in industrial applications due to its easy installation, reliability, and low cost. However, the increasing usage of IMs also increases the need for timely maintenance in order to ensure their operation and a longer service life. This article proposes a new diagnosis methodology based on maximal overlap discrete wavelet transform and a lightweight 1-D convolutional neural network (CNN) architecture, in order to detect mechanical and electrical faults and their combination, in adjustable speed drive (ASD)-powered IMs. Specifically, single and combined faults were studied from the next: Outer raceway bearing (mechanical), turn-to-turn short-circuit, and phase-to-ground short circuit (electrical). The presented study was developed using current signals acquired from stators of IMs of 1 hp. The current signals are measured at powered conditions introduced by a power grid with a constant frequency at 60 Hz, and an ASD at three different frequencies. The proposed diagnostic methodology reaches more than 99% of accuracy.

A methodology for automatic incipient broken rotor bar detection in induction motors (IMs) is presented. Sparse representations of signals are applied as a diagnosis technique. The novelty of this technique is that it can analyze the frequency spectra from vibration signals even when the differences among signals are small. This representation allows decomposing or reconstructing signals through a trained dictionary that has learned the features of one specific group/class. The main feature of this paper is the use of overcomplete dictionaries trained from sets of signals with faults to be detected. In this way, trained dictionaries perform the decomposition of signals using the orthogonal matching pursuit (OMP) algorithm. The decomposition is evaluated and classified by error-based criteria and a majority decision classifier, allowing the detection of early damage, ranging from 1 mm to one broken bar. The detection is performed by the decomposition of vibration signals from three axes (x, y, and z) of IMs under three load conditions (unloaded, half loaded, and three-fourths loaded) and different levels of damage (healthy or 0 mm, 1-9 mm, and one broken bar). These signals are processed by the Fourier transform and the spectrum obtained is evaluated by the OMP algorithm. Finally, the retrieved information is evaluated and the diagnosis is given. All algorithms are developed in MATLAB software and the detection accuracy is higher than 90% for damages as small as 1 mm.

Induction machines are widely used in several industries around the world; their robust design allows them to operate even under non-optimal conditions; the non-optimal operation can reduce the machine lifetime depending on the anomaly magnitude; this leads to a loss of process efficiency, which eventually generates a considerable operational costs increment. Monitoring methods, that allow an early fault detection, are getting developed currently; these methods are focused on the fault detection of the main components of the machine; one of them is the bearing fault detection that can be obtained through the phase current signal analysis. In this work, three types of Goodness-of-Fit test are studied; in these methods, the Motor Current Signature and the Motor Square Current Signature are analyzed. Furthermore, three types of bearing damage are presented and studied; the damages studied are: single point damage (bearing Outer-race damage and bearing ball damage), and distributed damage (corrosion damage). The IM signals, when working with the damages mentioned before, are measured at two powering conditions: power grid sourced (at 60Hz constant frequency), and adjustable speed drive (at six operating frequencies).

Due to its straightforward installation, operation and low cost, the Induction Motor (IM) is widely used in industry. Because of its constant service, there is a growing necessity of fault detection techniques for IM. Bearings fault (BF) is a common fault that may appear in an induction motor, and it may cause severe damage to the device. Thus, early detection is necessary to program the corresponding maintenance and extend the service of the machine and to reduce expenses for maintenance service or replacement. In literature, different techniques have been proposed to detect this fault involving a signal preprocessing step and applying time or frequency domain approaches, which require time and resource consumption. In this paper, a technique to detect BF based on Motor Current Signal Analysis (MCSA) via a statistical analysis named Kolmogorov-Smirnov Test (K-S test), is presented. K-S test determines if two samples come from the same distribution by measuring the maximum distance between them. We compared signals of two motor conditions, no damage (ND) and BF. We applied the test to the raw current signal.

Induction Motors (IMs) have been extensively used in the industry due to their low-cost, fast installation, and easy operation. However, it is necessary to have adequate maintenance programs because sudden faults, in the IM, could lead in important financial losses. Several methods have been proposed in the literature, where the thermography images analysis is an excellent candidate due to it is a non-invasive and non-contact technique, avoiding problems that exist in traditional techniques as indirect and unnoticed interference from other equipment in the signal to analyze, e.g. vibration or current. In this paper, a methodology based on the analysis of the specific region in thermography images to detect bearing damage is developed. In this manner, a study of two common damages is performed. The analysis is performed in three regions in the IM, with mechanical load, and with different supply frequencies. The difference of temperature among the regions are studied and the results are shown a difference in about 1.8°C between healthy and damaged condition, enough difference to detect the damage in the bearing taking into account that the IM has a cooling fan in the back.

The noise is the most common issue present in all signals. Depending on the process or application, the reduction of the noise is mandatory mainly in areas as signals classification, pattern recognition, and training process among others. In recent years, diverse methodologies for noise reduction have been proposed, but they have left some issues with no solving. Such are the noise reduction in signals immersed in white noise without changes in their amplitude or phase or the high performance of recovering components of the desire signals distributed among the frequency spectrum. Also, filtering techniques without significant changes in signal phase and amplitude are necessary. This paper proposes a methodology for noise reduction based on the Orthogonal Matching Pursuit algorithm with dictionaries constructed from kernel functions of the discrete cosine transform and discrete sine transform. The methodology is proven in synthetics signals and real electrical signals, reaching noise reduction of white noise (~34dB in electrical signals) and recovering without distortion.