Induction motors (IM) play a fundamental role in the industrial sector because they are robust, efficient, and low-cost machines. Changes in the environment, installation errors, or modifications to working conditions can generate faults in induction motors. The trend on IM fault detection is focused on the design techniques and sensors capable of evaluating multiple faults with various signals using non-invasive analysis. The methodology is based on processing electric current signals by applying the short-time Fourier transform (STFT). Additionally, the computation of the mean and standard deviation of infrared thermograms is proposed as main indicators. The proposed system combines both parameters by means of Support Vector Machine and k-nearest-neighbor classifiers. The development of the diagnostic system was done with digital hardware implementations using a Xilinx PYNQ Z2 card that integrates an FPGA with a microprocessor, thus taking advantage of the acquisition and processing of digital signals and images in hardware. The proposed method has proved to be effective for the classification of healthy (HLT), misalignment (MAMT), unbalance (UNB), damaged bearing (BDF), and broken rotor bar (BRB) faults with an accuracy close to 99%.

Bearings are one of the most essential elements in an induction motor, and they are built with different materials and constructions according to their application. These components are usually one of the most failure-prone parts of an electric motor, so correct and accurate measurements, instrumentation, and processing methods are required to prevent and detect the presence of different failures. This work develops a methodology based on the fusion of current and magnetic stray flux signals, calculation of statistical and non-statistical indicators, genetic algorithms (GAs), linear discriminant analysis (LDA), and neural networks. The proposed approach achieves a diagnostic effectiveness of 99.8% for detecting various damages in the outer race at 50 Hz frequency and 96.6% at 60 Hz. It also demonstrates 99.8% effectiveness for detecting damages in the presence of contaminants in lubrication at 50 Hz and 97% at 60 Hz. These results apply across metallic, ceramic, and hybrid bearings.

Early fault diagnosis in induction motors is important to maintain correct operation in terms of energy and efficiency, as well as to achieve a reduction in costs associated with maintenance or unexpected stoppages in production processes. These motors are widely used in industry due to their reliability, low cost, and great robustness; however, over time, they may be exposed to wear that can affect their performance, endanger the integrity of operators, or cause unexpected shutdowns that generate economic losses. Corrosion in the bearings is one of the most common failures, which is mainly triggered by high humidity in combination with high temperatures. However, despite its relevance, it has not been widely explored as a cause of failure in induction motors. Unlike failures that occur in specific or localized areas, corrosion in bearings does not manifest through specific frequencies associated with the phenomenon, since the corrosion occurs extensively on the surface of the raceway, making early diagnosis difficult with conventional techniques based on spectral analysis. Therefore, this work proposes an approach for the analysis of magnetic stray flux and vibration signals under different levels of corrosion using statistical and non-statistical parameters to capture variations in the dynamic behavior of the motors while employing genetic algorithms to select the most relevant parameters for each signal and optimize the configuration of an ensemble learning algorithm. The classification of the bearing condition is achieved using support vector machines in combination with the bagging method, which increases the robustness and accuracy of the model in the presence of signal variability. A classification accuracy between the healthy state and two gradualities greater than 99% was obtained, indicating that the proposed approach is reliable and efficient for corrosion diagnosis.

Dry-type electrical transformers are essential components in commercial, industrial, and residential power distribution systems, as they adapt voltage levels required by a broad range of load types. Although they are robustly constructed, they are exposed to adverse operational and environmental conditions such as dust, humidity, and electrical disturbances that may cause premature winding damage, such as inter-turn short circuits. This study focuses on the detection of inter-turn short-circuit faults in a 15 kVA commercial dry-type transformer, where a fault equivalent to 11.54% of short-circuited turns was induced in the tap changers. Axial, radial, and rotational leakage magnetic flux signals were captured using a low-cost, non-invasive triaxial Hall-effect magnetic flux sensor. During data processing, Fisher Score feature selection was applied to identify the most relevant indicators. Subsequently, feature extraction techniques, including Linear Discriminant Analysis, Principal Component Analysis (PCA), Uniform Manifold Approximation and Projection, and Isometric Mapping, were evaluated. The technique that best preserved global and local data structures was selected using Trustworthiness, Spearman’s correlation, and Kruskal’s stress metrics. PCA was selected as the optimal technique based on these quality metrics, achieving the highest classification performance. The resulting subspace data were classified using support vector machines and applying K-fold cross-validation. The proposed system achieved classification accuracies above 95%, with high recall and F1-score values, for inter-turn fault detection in each winding, confirming its effectiveness for reliable inter-turn fault detection in each transformer winding.