The detection and classification of power quality (PQ) disturbances have become a pressing concern due to the increasing number of disturbing loads connected to the power line and the susceptibility of certain loads to the presence of these disturbances; moreover, they can appear simultaneously since, in any real power system, there are multiple sources of different disturbances. In this paper, a new dual neural-network-based methodology to detect and classify single and combined PQ disturbances is proposed, consisting, on the one hand, of an adaptive linear network for harmonic and interharmonic estimation that allows computing the root-mean-square voltage and total harmonic distortion indices. With these indices, it is possible to detect and classify sags, swells, outages, and harmonics-interharmonics. On the other hand, a feedforward neural network for pattern recognition using the horizontal and vertical histograms of a specific voltage waveform can classify spikes, notching, flicker, and oscillatory transients. The combination of the aforementioned neural networks allows the detection and classification of all the aforementioned disturbances even when they appear simultaneously. An experiment under real operating conditions is carried out in order to test the proposed methodology.

The development and application of techniques and methodologies for the analysis of power quality (PQ) signals that offer a more efficient and reliable analysis in terms of processing and performance are still issues for industrial and academic fields, mainly considering the quick growing of the PQ data in modern power systems. In this regard, an iterative downsampling stage fused to the empirical mode decomposition (EMD) method is proposed. It is validated and tested using synthetic and real signals. In general, the aim of the proposed method is to extract the fundamental component as the first intrinsic mode function (IMF) to simplify the remaining decomposition. The proposed method is compared to the classical EMD and the ensemble EMD (EEMD). Advantages of the proposed method include a more adequate IMF extraction than the EMD technique, and a noticeable reduction of computational burden compared to the EEMD. The results obtained from the synthetic and real signals demonstrate the reliability and efficiency of the proposed method.

Although induction motors (IMs) are robust and reliable electrical machines, they can suffer different faults due to usual operating conditions such as abrupt changes in the mechanical load, voltage, and current power quality problems, as well as due to extended operating conditions. In the literature, different faults have been investigated; however, the broken rotor bar has become one of the most studied faults since the IM can operate with apparent normality but the consequences can be catastrophic if the fault is not detected in low-severity stages. In this work, a methodology based on convolutional neural networks (CNNs) for automatic detection of broken rotor bars by considering different severity levels is proposed. To exploit the capabilities of CNNs to carry out automatic image classification, the short-time Fourier transform-based time–frequency plane and the motor current signature analysis (MCSA) approach for current signals in the transient state are first used. In the experimentation, four IM conditions were considered: half-broken rotor bar, one broken rotor bar, two broken rotor bars, and a healthy rotor. The results demonstrate the effectiveness of the proposal, achieving 100% of accuracy in the diagnosis task for all the study cases.