Search

The emergence of long-range (LoRa) technology, together with the expansion of uncrewed aerial vehicles (UAVs) use in civil applications have brought significant advances to the Internet of Things (IoT) field. In this way, these technologies are used together in different scenarios, especially when it is necessary to have connectivity in remote and difficult-to-access locations, providing coverage and monitoring of greater areas. In this sense, this article seeks to determine the best positioning for the LoRa gateway coupled to the drone and the optimal spreading factor (SF) for signal transmission in a LoRa network, aiming to improve the connected devices signal-to-noise ratio (SNR), considering a suburban and densely wooded environment. Then, multi-layer perceptron (MLP) networks and generalized regression neural networks (GRNN) were trained to predict the signal behavior and determine the best network to represent this behavior. The MLP network presented the lowest RMSE, 2.41 dB, and was selected for use jointly with the bioinspired Grey-Wolf optimizer (GWO). The optimizer proved its effectiviness being able to adjust the number of UAVs used to obtain 100% coverage and determine the best SF used by the endnodes, guaranteeing a higher transmission rate and lower energy consumption. [ABSTRACT FROM AUTHOR]

In this paper, we present an unprecedented method based on Kohonen networks that is able to automatically recognize partial discharge (PD) classes from phase-resolved partial discharge (PRPD) diagrams with features of various simultaneous PD patterns. The PRPD diagrams were obtained from the stator windings of a real-world hydro-generator rotating machine. The proposed approach integrates classification probabilities into the Kohonen method, producing self-organizing probability maps (SOPMs). For building SOPMs, a group of PRPD diagrams, each containing a single PD pattern for training the Kohonen networks and single- and multiple-class-featured samples for obtaining final SOPMs, is used to calculate the probabilities of each Kohonen neuron to be associated with the various PD classes considered. At the end of this process, a self-organizing probability map is produced. Probabilities are calculated using distances, obtained in the space of features, between neurons and samples. The so-produced SOPM enables the effective classification of PRPD samples and provides the probability that a given PD sample is associated with a PD class. In this work, amplitude histograms are the features extracted from PRPDs maps. Our results demonstrate an average classification accuracy rate of approximately 90% for test samples. [ABSTRACT FROM AUTHOR]

Partial discharge (PD) monitoring is widely used in rotating machines to evaluate the condition of stator winding insulation, but its practice on a large scale requires the development of intelligent systems that automatically process these measurement data. In this paper, it is proposed a methodology of automatic PD classification in hydro-generator stator windings using neural networks. The database is formed from online PD measurements in hydro-generators in a real setting. Noise filtering techniques are applied to these data. Then, based on the concept of image projection, novel features are extracted from the filtered samples. These features are used as inputs for training several neural networks. The best performance network, obtained using statistical procedures, presents a recognition rate of 98%. [ABSTRACT FROM AUTHOR]