Three-phase induction motors are the main elements for converting electrical energy into mechanical energy and are extensively used in industry. Reducing maintenance costs becomes an incentive for developing systems capable of identifying defects. This research proposes a framework for recommending machine learning algorithms that diagnose and detect broken bar defects in three-phase induction motors under transient operation based on artificial intelligence. Employing experimental data, features were extracted and selected based on current, voltage, and vibration. A protocol of insertion of white noise showed that the proposed framework admitted 80% of noise without losing the predictive capacity based on a multicriteria performance measure. Copyright (C) 2024 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

The inclusion of photovoltaic systems in distribution networks has raised the importance of the prediction of photovoltaic power for safe planning and operation. Artificial neural networks (ANNs) have been used in this task due to its capacity of representing nonlinearities. However, the profile of the data used may affect the forecast accuracy. This manuscript reports on a comparative analysis of the performance of four neural network models for photovoltaic power forecast regarding their input dataset. Four sets composed of photovoltaic power data (local measurements) and external weather data (remote measurements) were used, and the networks were validated through actual measurements from a photovoltaic micro plant. The ANN that dealt with only weather data showed a good level of accuracy, being a useful tool for the feasibility analysis of new photovoltaic projects. In addition, the approach that used only photovoltaic power data has excelled and can be used in electric sector companies.

Studies that evaluate the monitoring of the condition of power insulators and the malfunction of these devices are especially focused on the main variables involved with their aging process. The early degradation of power insulators, which is more common in highly polluted locations, results in risks to the operation of the electrical system and can financially impact power utilities due to unplanned service interruptions and premature maintenance. Many techniques have been proposed in the literature to evaluate the condition of power insulators. Among these techniques, intelligent systems or machine learning techniques stand out, being pointed out as one of the most promising tools for the early detection of malfunctions in such equipment. However, there is a lack of studies that address this problem more broadly, using the full capacity of intelligent techniques to compile a complete and expert monitoring system that can make automatic decisions or provide subsidies to the operator for more assertive maintenance actions. Based on the studies found in the literature and on the shortcomings identified on the subject, this work presents an investigation into the use of intelligent techniques for monitoring the condition of power insulators in transmission lines, mainly focusing on the early detection of malfunctions of these devices.

Partial discharge (PD) investigations can identify and localise incipient failures in power transformers early, thus avoiding considerable financial losses. The feature extraction of PD signals is a fundamental step for the development of such location techniques since it directly influences the performance of a location method. This study presents a detailed comparative analysis of four traditional approaches for the obtaining of attributes towards a better set of signal features for the location of PDs. The approaches were critically compared regarding their ability to locate experimentally generated discharges between adjacent layers of a prototype winding. In order to perform such analysis, a localisation structure based on logistic regression models was elaborated, capable of determining both layers and sections of the winding affected by PDs and easily applicable in practice. The results show energy features of wavelet coefficients, obtained through the decomposition of high-frequency current signals acquired at the winding endings, achieve better performance in the PD localisation, accurately indicating discharge occurrence points among layers and sections of the winding.