Search for: All records

Abstract This paper presents a systematic study that focuses on how the number of viewpoints distributed in the heliosphere affects the accuracy and uncertainty of the 3D geometric coronal mass ejection (CME) measurements. An efficient nonmanual minimization-based fitting technique that is different from the manual methods widely used in the community is developed. It uses the MPFIT minimization IDL routine and searches for the optimized model point clouds that best fit the observed CME leading edges from one, two, or three viewpoints using a set of combinations of observations provided by the Solar Terrestrial Relations Observatory and Solar and Heliospheric Observatory. The technique also provides a robust calculation of uncertainties of the CME geometric parameters that is lacking in manual methods. Three well-known geometric models, the cone, graduated cylindrical shell, and spheroid shock, are used. All three models depend on geometric parameters that govern the CME propagation direction and size. Sample cases of a halo, partial halo, and limb CMEs as seen from the Earth are used in the fitting and uncertainty calculation. It is found that, after adding a second viewpoint off the Sun–Earth line, the uncertainties drop significantly, while the addition of the third viewpoint adds limited benefits. This study shows that the minimization fitting method provides a robust, fast, and straightforward way to define the CME geometric parameters along with their uncertainties for individual events, which shall provide the necessary data constraints for ensemble predictions of CME evolution. It also underlines the importance of having a permanent observatory off the Sun–Earth line for operational space weather prediction. 

Abstract Flourished wind energy market pushes the latest wind turbines (WTs) to further and harsher inland and offshore environment. Increased operation and maintenance cost calls for more reliable and cost effective condition monitoring systems. In this article, a bi-level condition monitoring framework for interturn short-circuit faults (ITSCFs) in WT generators is proposed. A benchmark dataset, consisting of 75 ITSCF scenarios and generator current signals of a specific WT, has been created and made publicly available on Zenodo. The data are simulated at a rate of 4 kHz. Based on the time and frequency features extracted from data processing, machine learning-based severity estimation and faulty phase identification modules can provide valuable diagnostic information for wind farm operators. Specifically, the performance of long short-term memory (LSTM) networks, gated recurrent unit (GRU) networks, and convolutional neural networks (CNNs) are analyzed and compared for severity estimation and faulty phase identification. For test-bed experimental reference, various numbers of scenarios for training the models are analyzed. Numerical experiments demonstrate the computational efficiency and robust denoising capability of the CNN algorithm. The GRU network, however, achieves the highest accuracy. The overall system performance improves significantly, from 87.76% with 16 training scenarios to 99.95% with 52 training scenarios, when tested on a set containing all 76 scenarios from an unforeseen period. 

In the post-pandemic era, global working patterns have been reshaped, and the demand for online network services has increased significantly. Therefore, cross-data-center content migration has become a relevant problem to address, leading to higher attention in data backup/recovery planning. Beyond traditional pre-disaster content redundancy approaches, this work focuses on the challenge of rapid post-disaster content evacuation under the threat of cascading failures. In fact, due to the interdependence of data centers (DCs), inter-DC optical networks, and power grid networks, disasters may have a domino effect on these infrastructures, with their impact gradually expanding over time and space. In this paper, we propose two trajectory models that capture the dynamic evolution of cascading failures, and we propose a trajectory-based content evacuation (TCE) strategy that considers the spatiotemporal evolution of cascading failures to minimize content loss. Numerical results show that, when each DC needs to evacuate about 200 TB of massive content, TCE can reduce content loss by up to 25% compared to baseline strategies. 

Abstract In this work, TiO₂thin films deposited by the atomic layer deposition (ALD) method were treated with a special N₂O plasma surface treatment and used as the gate dielectric for AlGaN/GaN metal insulator semiconductor high electron mobility transistors (MISHEMTs). The N₂O plasma surface treatment effectively reduces defects in the oxide during low-temperature ALD growth. In addition, it allows oxygen atoms to diffuse into the device cap layer to increase the barrier height and thus reduce the gate leakage current. These TiO₂films exhibit a dielectric constant of 54.8 and a two-terminal current of 1.96 × 10⁻¹⁰A mm⁻¹in 2μm distance. When applied as the gate dielectric, the AlGaN/GaN MISHEMT with a 2μm-gate-length shows a high on/off ratio of 2.59 × 10⁸and a low subthreshold slope (SS) of 84 mV dec⁻¹among all GaN MISHEMTs using TiO₂as the gate dielectric. This work provides a feasible way to significantly improve the TiO₂film electrical property for gate dielectrics, and it suggests that the developed TiO₂dielectric is a promising high-κgate oxide and a potential passivation layer for GaN-based MISHEMTs, which can be further extended to other transistors. 

Abstract While preventive maintenance is crucial in wind turbine operation, conventional condition monitoring systems face limitations in terms of cost and complexity when compared to innovative signal processing techniques and artificial intelligence. In this paper, a cascading deep learning framework is proposed for the monitoring of generator winding conditions, specifically to promptly detect and identify inter-turn short circuit faults and estimate their severity in real time. This framework encompasses the processing of high-resolution current signal samples, coupled with the extraction of current signal features in both time and frequency domains, achieved through discrete wavelet transform. By leveraging long short-term memory recurrent neural networks, our aim is to establish a cost-efficient and reliable condition monitoring system for wind turbine generators. Numeral experiments show an over 97% accuracy for fault diagnosis and severity estimation. More specifically, with the intrinsic feature provided by wavelet transform, the faults can be 100% identified by the diagnosis model.