The electrical energy storage system faces numerous obstacles as green energy usage rises. The demand for electric vehicles (EVs) is growing in tandem with the technological advance of EV range on a single charge. To tackle the low-range EV problem, an effective electrical energy storage device is necessary. Traditionally, electric vehicles have been powered by a single source of power, which is insufficient to handle the EV’s dynamic demand. As a result, a unique storage medium is necessary to meet the EV load characteristics of high-energy density and high-power density. This EV storage system is made up of two complementing sources: chemical batteries and ultracapacitors/supercapacitors. The benefits of using ultracapacitors in a hybrid energy storage system (HESS) to meet the low-power electric car dynamic load are explored in this study. In this paper, a HESS technique for regulating the active power of low-powered EV simulations was tested in a MATLAB/Simulink environment with various dynamic loading situations. The feature of this design, as noted from the simulation results, is that it efficiently regulates the DC link voltage of an EV with a hybrid source while putting minimal load stress on the battery, resulting in longer battery life, lower costs, and increased vehicle range.
Integration of RE sources to the utility grid offers technical and operational challenges causing problems of PQ, stability, identification of operational events, etc. This article presents an algorithm to identify events including islanding, grid integration, and outage of the solar PV and WG plants in grid using a ST. Islanding event may occur in the presence of any kind of plant. Processing of negative sequence component of voltage is performed by utilizing ST based multiresolution analysis at the test node and the output matrix is evaluated. The features (F1–F4), VI and STD indexes are obtained from this matrix. These features are utilized for identifying the events and transient phenomenon. The VI and STD indexes are used to recognize the type of RE source present during the islanding and outage events. Moreover, for recognizing the type of RE source at the time of synchronization event, an SI is proposed. This is computed by the ST depended processing of voltage signals. Performance of the algorithm is found satisfactory for all incidence angles and complete voltage cycle under the noisy conditions of 10 dB SNR. As compared to the time–frequency transform based coefficients of the voltage signal, the proposed technique is found to be superior in terms of small NDZ and low computation time and least affected by noise. Further, the developed technique is also efficient to detect various events stated above and the type of RE source. Study is performed using MATLAB/Simulink software and validated in real time using RTDS.
To judge the ability of convolutional neural networks (CNNs) to effectively and efficiently transfer image representations learned on the ImageNet dataset to the task of recognizing COVID-19 in this work, we propose and analyze four approaches. For this purpose, we use VGG16, ResNetV2, InceptionResNetV2, DenseNet121, and MobileNetV2 CNN models pre-trained on ImageNet dataset to extract features from X-ray images of COVID and Non-COVID patients. Simulations study performed by us reveal that these pre-trained models have a different level of ability to transfer image representation. We find that in the approaches that we have proposed, if we use either ResNetV2 or DenseNet121 to extract features, then the performance of these approaches to detect COVID-19 is better. One of the important findings of our study is that the use of principal component analysis for feature selection improves efficiency. The approach using the fusion of features outperforms all the other approaches, and with this approach, we could achieve an accuracy of 0.94 for a three-class classification problem. This work will not only be useful for COVID-19 detection but also for any domain with small datasets.
21 February 2022 in Electric Power Systems Research
The broad application area and the accompanying challenges make machine learning-based recognition of handwritten scripts a demanding field. Individuals’ writing practices and inherent variations in the size, shape, and tilt of the characters may increase the difficulty level. The deep convolutional neural network (DCNN) models have been a great success in solving pattern recognition problems, but at the expense of a considerable number of trainable parameters and heavy computational loads. The proposed work addresses these problems by using the shifted window (SWIN) transformer method for the first time to recognize handwritten Devanagari numerals. In the presented model, the SWIN transformer is finely tuned to withstand the popular DCNN models such as VGG-16Net, ResNet-50, and DenseNet-121 in terms of recognition accuracy, space requirement, and computational complexity. The model successfully attains the recognition accuracy of 99.20% with just 0.218 million trainable parameters and 0.0912 giga floating-point operations per second (FLOPs). This suggests the validity and soundness of the proposed model for recognizing handwritten Devanagari numerals.
This paper has introduced an algorithm for the identification of islanding events in the remotely located distribution grid with renewable energy (RE) sources using the voltage signals. Voltage signal is processed using Stockwell transform (ST) to compute the median-based islanding recognition factor (MIRF). The rate of change in the root mean square (RMS) voltage is computed by differentiating the RMS voltage with respect to time to compute the voltage rate of change in islanding recognition factor (VRCIRF). The proposed voltage-based islanding recognition factor (IRFV) is computed by multiplying the MIRF and VRCIRF element to element. The islanding event is discriminated from the faulty and operational events using the simple decision rules using the peak magnitude of IRFV by comparing peak magnitude of IRFV with pre-set threshold values. The proposed islanding detection method (IDM) effectively identified the islanding events in the presence of solar energy, wind energy and simultaneous presence of both wind and solar energy at a fast rate in a time period of less than 0.05 cycles compared to the voltage change rate (ROCOV) and frequency change rate (ROCOF) IDM that detects the islanding event in a time period of 0.25 to 0.5 cycles. This IDM provides a minimum non-detection zone (NDZ). This IDM efficiently discriminated the islanding events from the faulty and switching events. The proposed study is performed on an IEEE-13 bus test system interfaced with renewable energy (RE) generators in a MATLAB/Simulink environment. The performance of the proposed IDM is better compared to methods based on the use of ROCOV, ROCOF and discrete wavelet transform (DWT).