Search

The validation of Electromagnetic Transients (EMT) models is essential for ensuring the accuracy of grid compliance and stability studies. However, there is still room for improvement in understanding the processes and a lack of standards for EMT model validation, especially after the commissioning of a wind power plant. This paper addresses this gap and presents experiences and recommendations for EMT model validation from an Original Equipment Manufacturer (OEM) perspective throughout the entire life cycle of offshore type IV wind turbines and wind power plants. The paper showcases different perspectives on model validation, including methodologies, requirements, and accuracy. It illustrates the process of wind turbine level model validation during design or pre-commissioning with practical examples and also demonstrates the wind power plant level model validation for two operational plants. In conclusion, the paper addresses various challenges and recommendations for model validation related to measurement location placement, data acquisition, aggregation methods, preprocessing, turbine connection status, and operating conditions at the time of the fault.

The Transformers architecture has recently emerged as a revolutionary paradigm in the field of deep learning, particularly excelling in Natural Language Processing (NLP) and Computer Vision (CV) applications. Despite its success, the security implications of Transformers have not been comprehensively explored, encompassing a broad spectrum of both hardware and software vulnerabilities. This paper aims to address this critical gap by conducting an extensive exploration of security challenges confronting Transformers from both software and hardware perspectives. While software-related concerns like adversarial attacks, private inference, and watermarking have been studied, the paper sheds light on previously underexplored hardware vulnerabilities such as trojans and side-channel attacks. By unraveling the intricacies of these hardware threats, the study aims to contribute to a comprehensive understanding of Transformer security. It presents an in-depth analysis of recent advancements in the security of Transformers. Additionally, it outlines existing challenges and forecasts future research trends, offering insights for researchers and practitioners aiming for the secure and resilient design and deployment of Transformers. The survey categorizes different attacks and defenses related to Transformers, helping researchers identify gaps and opportunities in this area. Furthermore, it defines a roadmap for a unified security framework, serving as a foundational starting point for developers seeking to implement robust security measures.

Grid-forming (GFM) converter control technology and synchronous condensers (SCs) have both been reported to improve the overall stability of a power system. GFM technology is being considered for offshore wind power plants (OF WPPs) due to their stabilising effect, and SCs are being integrated into OF WPPs as they increase the short circuit level of the system. This paper investigates the effect of SCs on the small-signal stability of an OF WPP with different converter control strategies, namely grid-following (GFL) and grid-forming (GFM). Primarily, the effect of SCs can be two-fold: (1) overall stability enhancement of the WPP by providing reactive power support, and (2) contribution to the effective short circuit ratio (SCR) of the WPP by fault current support. Therefore, this paper focuses on studies concerning these effects on an aggregated model of a WPP connected to the grid. This paper also tries to compare the stabilising effects of GFM converters and SCs by comparing the small-signal stability of the test system with GFL control and SC and with GFM control. To that end, a state-space model of the test system is developed for small-signal stability assessment and the SC’s effect on its stability. In addition, a mathematical explanation of SCR enhancement with an SC is provided and is verified against the fault MVA calculations done on time-domain simulation models.

Brain hemorrhage refers to a potentially fatal medical disorder that affects millions of individuals. The percentage of patients who survive can be significantly raised with the prompt identification of brain hemorrhages, due to image-guided radiography, which has emerged as the predominant treatment modality in clinical practice. A Computed Tomography Image has frequently been employed for the purpose of identifying and diagnosing neurological disorders. The manual identification of anomalies in the brain region from the Computed Tomography Image demands the radiologist to devote a greater amount of time and dedication. In the most recent studies, a variety of techniques rooted in Deep learning and traditional Machine Learning have been introduced with the purpose of promptly and reliably detecting and classifying brain hemorrhage. This overview provides a comprehensive analysis of the surveys that have been conducted by utilizing Machine Learning and Deep Learning. This research focuses on the main stages of brain hemorrhage, which involve preprocessing, feature extraction, and classification, as well as their findings and limitations. Moreover, this in-depth analysis provides a description of the existing benchmark datasets that are utilized for the analysis of the detection process. A detailed comparison of performances is analyzed. Moreover, this paper addresses some aspects of the above-mentioned technique and provides insights into prospective possibilities for future research.

Stroke is a dangerous medical disorder that occurs when blood flow to the brain is disrupted, resulting in neurological impairment. It is a big worldwide threat with serious health and economic implications. To solve this, researchers are developing automated stroke prediction algorithms, which would allow for early intervention and perhaps save lives. The number of people at risk for stroke is growing as the population ages, making precise and effective prediction systems increasingly critical. wo In a comparison examination with six well-known classifiers, the effectiveness of the proposed ML technique was explored in terms of metrics relating to both generalization capability and prediction accuracy. To give insight into the black-box machine learning models, we also studied two kinds of explainable techniques, namely SHAP and LIME, in this study. SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-agnostic Explanations) are well-established and reliable approaches for explaining model decision-making, particularly in the medical industry. The findings of the experiment revealed that more complicated models outperformed simpler ones, with the top model obtaining almost 91% accuracy and the other models achieving 83-91% accuracy. The proposed framework, which includes global and local explainable methodologies, can aid in standardizing complicated models and gaining insight into their decision-making, which can enhance stroke care and treatment.

Combining skeleton and RGB modalities in human action recognition (HAR) has garnered attention due to their ability to complement each other. However, previous studies did not address the challenge of recognizing fine-grained human-object interaction (HOI). To tackle this problem, this study introduces a new transformer-based architecture called Sequential Skeleton RGB Transformer (SSRT), which fuses skeleton and RGB modalities. First, SSRT leverages the strength of Long Short-Term Memory (LSTM) and a multi-head attention mechanism to extract high-level features from both modalities. Subsequently, SSRT employs a two-stage fusion method, including transformer cross-attention fusion and softmax layer late score fusion, to effectively integrate the multimodal features. Aside from evaluating the proposed method on fine-grained HOI, this study also assesses its performance on two other action recognition tasks: general HAR and cross-dataset HAR. Furthermore, this study conducts a performance comparison between a HAR model using single-modality features (RGB and skeleton) alongside multimodality features on all three action recognition tasks. To ensure a fair comparison, comparable state-of-the-art transformer architectures are employed for both the single-modality HAR model and SSRT. In terms of modality, SSRT outperforms the best-performing single-modality HAR model on all three tasks, with accuracy improved by 9.92% on fine-grained HOI recognition, 6.73% on general HAR, and 11.08% on cross-dataset HAR. Additionally, the proposed fusion model surpasses state-of-the-art multimodal fusion techniques like Transformer Early Concatenation, with an accuracy improved by 6.32% on fine-grained HOI recognition, 4.04% on general HAR, and 6.56% on cross-dataset.

The calculation of short-circuit current contributions from full-converter based type IV wind turbines and other inverter-based power plants are often simplified. They are represented as simple fixed current sources for the steady-state stage of the fault during short-circuit evaluations. As wind power penetration increases, it becomes increasingly important to factor the details of short-circuit contribution from such inverter-based power plants for long-term planning as well as equipment and protection system design. The challenge lies in the fact that this type of inverter-based power plants is fully connected through power electronic devices, and their short-circuit contributions cannot be easily represented and calculated based on traditional physical laws developed for synchronous generators. This paper addresses the main challenges and limitations of the current accepted practices in the industry. The challenges are discussed for all three stages during a fault referring to the stages of more traditional power plants, namely sub-transient, transient and steady-state. An electromagnetic transient (EMT) model validated through field tests on a full-scale prototype wind turbine was used to simulate over 20,000 scenarios and examine the three stages in detail. Results show that during the sub-transient stage, in the first milliseconds of the fault, the converter exhibits a so-called quasi-voltage source behavior with high currents limited by transient hardware limitations and high negative sequence current injection for unbalanced faults, regardless of the current priority chosen. During the transient stage, the converter starts transitioning to a grid code/voltage-dependent current source until reaching the steady-state stage of the fault. The paper also includes a theoretical analysis, a comparison with the conventional ac-dc fault representation, and thorough time-domain and statistical analyses of the relevant variables during a short-circuit. Finally, it outlines future challenges and identifies areas of improvement for enhancing the modeling of fault currents in Type IV wind turbine generators.

Polycystic Ovary Syndrome (PCOS) is a critical hormonal disorder of women that significantly impacts life. In this new generation, women are more prone to PCOS. It is the cause of various problems, including infertility. Early detection of PCOS can reduce complexity. Therefore, an early and proper PCOS detection system is essential to minimize complications. Among all the detection techniques Machine Learning (ML) has an excellent performance in detection for its feature extraction capability. Therefore, considerable research has been carried out to detect PCOS using ML. Various ML approaches like Convolutional Neural Network, Support Vector Machine, K-Nearest-Neighbors, Random Forest, Logistic Regression, Decision Tree, Naive Bayes, etc., are used in detecting PCOS. This research aims to call attention to the researchers by presenting a descriptive and contextual overview of all the existing technologies on PCOS detection by ML algorithms. A comprehensive analysis is carried out of how various ML approaches have been used in PCOS detection over the last few decades, and the techniques are discussed thoroughly. A complete examination was studied on different datasets used in PCOS detection. The performance of several algorithms is compared in quantitative and qualitative approaches. Finally, the significant difficulties and future research scopes are discussed to draw a conclusion.

One of those illnesses that is most deadly to people is cancer. The only way to prevent any harm to humanity is by its early discovery and treatment. Various types of tests are conducted in the medical labs for the detection of cancer. Cancer can also be detected at the genetic level. For this distinct machine learning and deep learning methods already exist. This paper proposes a hybrid method based on Recurrent Neural Network (RNN) and Convolution Neural Network (CNN) to predict different types of cancer such as Breast, Lung, Uterine, Kidney, Prostate and colon cancer from gene expression data. The bottleneck features are extracted using the sandwich stacked method based on VGG16 and VGG19 pre-trained models. Afterward, the proposed hybrid classifier based on RNN-CNN has been used to classify the data into various classes. The proposed model performs better than the other existing methods such as VGG16, VGG19, ResNet50, Inception V3 and MobileNet classifier in terms of various performance metrics such as accuracy, Mean Square Error (MSE), precision, recall, and F1 score. RNN-CNN classifier provides the highest accuracy of 0.978 among all the other existing methods for Dataset 1 and the highest accuracy of 0.994 for Dataset 2 at 80% training data. On the other hand, RNN-CNN classifier provides the lowest MSE of 0.101 among all the other existing methods for Dataset 1 and the lowest MSE of 0.006 for Dataset 2 at 80% training data.

The world’s elderly population continues to grow at an unprecedented rate, creating a need to monitor the safety of an aging population. One of the current problems is accurately classifying elderly physical activities, especially falling down, and delivering prompt assistance to someone in need. Owing to the advancements in deep learning research, vision based solutions are employed for action recognition. One such popular approach is human pose estimation based action recognition or fall detection. Nevertheless, due to a lack of large-scale elderly fall datasets and the continuation of numerous challenges such as varying camera angles, illumination, and occlusion accurately classifying falls has been a problematic. To address these problems, this research first carried out a comprehensive study of the AI Hub dataset collected from real lives of elderly people in order to benchmark the performance of state-of-the-art human pose estimation methods. Secondly, owing to the limited number of real datasets, augmentation with synthetic data was applied and performance improvement was validated based on changes in the degree of accuracy. Third, this study shows that a Transformer network applied to elderly action recognition outperforms LSTM-based networks by a noticeable margin. Lastly, by observing the quantitative and qualitative performances of different networks, this paper proposes an efficient solution for elderly activity recognition and fall detection in the context of surveillance cameras.

Recently, wireless telecommunication networks have become a promising alternative for rainfall measuring instruments that complement existing monitoring devices. Due to big dataset of rainfall and telecommunication networks data, empirical computational methods are less adequate representation of the actual data. Therefore, deep learning models are proposed for the analysis of big data and give more accurate representation of real measurements. In this study, we investigated rainfall monitoring results from experimental measurements and deep learning approaches such as artificial neural networks and long short-term memory. The experimental setups were in South Korea over terrestrial and satellite links, and in Ethiopia over terrestrial link for different frequency bands and link distances. The received signal level and rainfall data measurement covered four years in South Korea and the data were sampled at intervals of 10 seconds. In Ethiopia, the data were recorded over 10 months and sampled at intervals of 15 minutes. The received signal power data were used to derive the rainfall rate distribution and compared to actual rainfall measurements over the same time periods. Our results demonstrate that the proposed deep learning-based models generally have a good fit with the measured rainfall rates. The rainfall rate generated from terrestrial links was a better fit to the actual rainfall rate data than that generated from satellite links.

Evapotranspiration, as a combination of evaporation and transpiration of water vapour, is a primary component of global hydrological cycles. It accounts for significant loss of soil moisture from the earth to the atmosphere. Reliable methods to monitor and forecast evapotranspiration are required for decision-making. Reference evapotranspiration, denoted as ET, is a major parameter that is useful in quantifying soil moisture in a cropping system. This article aims to design a multi-stage deep learning hybrid Long Short-Term Memory (LSTM) predictive model that is coupled with Multivariate Empirical Mode Decomposition (MEMD) and Boruta-Random Forest (Boruta) algorithms to forecast ET in the drought-prone regions (i.e., Gatton, Fordsdale, Cairns) of Queensland, Australia. Daily data extracted from NASA’s Goddard Online Interactive Visualization and Analysis Infrastructure (GIOVANNI) and Scientific Information for Land Owners (SILO) repositories over 2003–2011 are used to build the proposed multi-stage deep learning hybrid model, i.e., MEMD-Boruta-LSTM, and the model’s performance is compared against competitive benchmark models such as hybrid MEMD-Boruta-DNN, MEMD-Boruta-DT, and a standalone LSTM, DNN and DT model. The test MEMD-Boruta-LSTM hybrid model attained the lowest Relative Root Mean Square Error (≤17%), Absolute Percentage Bias (≤12.5%)and the highest Kling-Gupta Efficiency (≥0.89%) relative to benchmark models for all study sites. The proposed multi-stage deep hybrid MEMD-Boruta-LSTM model also outperformed all other benchmark models in terms of predictive efficacy, demonstrating its usefulness in the forecasting of the daily ET dataset. This MEMD-Boruta-LSTM hybrid model could therefore be employed in practical environments such as irrigation management systems to estimate evapotranspiration or to forecast ET.

In this paper, an efficient and fault-proof active node selection approach for localization tasks in Internet of Things (IoT) and Mobile Crowd Sensing (MCS) systems is proposed. The proposed approach is resilient to the presence of anomalous nodes. Localization is the process of fusing data readings from multiple sensing nodes with the aim of finding the location of a specific target, such as radiation source, forest fires and noisy areas. Current active node selection systems for localization tasks perform algorithms like greedy and genetic methods over the whole Area of Interest (AoI). As such, a system that considers anomalous data is required to detect anomalies and perform localization over a large number of active nodes, which usually takes multiple iterations and is computationally costly. To overcome this, we propose a resilient localization approach which a) uses the median filter based image filtering technique to level out anomalous readings, b) uses the filtered readings to reduce the AoI to be around the target location without being influenced by anomalous nodes, c) detects and eliminates anomalies in the new AoI based on the deviation between filtered readings and original readings, and d) selects remaining nodes in new AoI for localization. As a result, there is a huge reduction in the complexity of active node selection and thus reduction in time taken by the system to perform the task of source localization. The efficacy of the proposed system is evaluated for radiation source localization tasks using simulated radiation dataset, by performing experiments for several test scenarios. The results demonstrate that the system is able to perform localization tasks in significantly reduced time and therefore generate near real-time results while also maintaining low localization error.