The large-scale integration of renewable energy sources into the power grid has significantly changed traditional grid characteristics, posing great challenges to the stability of inverters. To ensure high stability and superior dynamic performance, grid-connected inverters shall dynamically adapt their control strategies from grid-following (GFL) to grid-forming (GFM) control. While existing studies primarily focus on mode transitions between GFL control and Virtual Synchronous Generator (VSG) control with voltage and current dual-loop structures, seamless switching between GFL and alternative GFM control, such as direct voltage control, remains unexplored. To address this gap, this paper proposes a smooth switching method for transitioning between GFL control and direct voltage control in the grid-connected operation. The proposed approach integrates a novel modulation voltage based pre-synchronization strategy with an adaptive adjustment method to simultaneously minimize the impact current and switching time during mode transition. Time-domain simulations are carried out in MATLAB/Simulink to validate the effectiveness and robustness of the proposed switching method under various operational scenarios. Results demonstrate that the method ensures an impact current ≤ 20% of the rated current, and a switching time ≤ 200 ms across varying operational conditions.

Ground-to-air real aperture radar scans the airspace to acquire range-azimuth-pitch 3D echoes for imaging. In real situations, the echo data may be corrupted by interference, making it difficult to reconstruct the original scene using echoes with missing data. To overcome this problem, a fast superresolution imaging method under the non-uniform sampling model is proposed in this paper. Firstly, a non-uniform sampling model is proposed to well model the echoes with missing data, which makes it possible to reconstruct the real target distribution from the echoes with missing data. Secondly, we use the sparse regularization (SR) super-resolution imaging method to reconstruct the real target distribution. Since the high dimension of the dictionary matrix leads to the expensive computational cost of the SR method, we propose a fast superresolution imaging algorithm based on low-rank approximation to reconstruct the targets quickly. Simulation results show that the original scene can be effectively reconstructed from the echoes with missing data based on our proposed model, and the proposed algorithm greatly improves the computational efficiency compared with the traditional methods while not leading to a loss of imaging performance.

Land Surface Temperature (LST) is one of the important parameters in thermal environment monitoring. Satellite thermal remote sensing is the major way to obtain spatial-temporal information of LST. However, limited by the cloud contamination and the trade-off between spatial and temporal resolution, current temperature products are difficult to provide all-weather LST. In this paper, a method is proposed to obtain all-weather hourly LST. It consists of two main steps: 1) reconstruction of LST under cloudy-sky by using enhanced annual temperature cycle (ATCE) model and 2) establishment of relationship between LST and air temperature which is used for the acquisition of hourly LST. In the end, the performance of the method is analyzed through the artificial data which is created by masking the origin images. And the results show that the proposed method is valuable for generating all-weather hourly LST.

Web attacks such as Cross-Site Scripting and SQL Injection are serious Web threats that lead to catastrophic data leaking and loss. Because attack payloads are often short segments hidden in URL requests/posts that can be very long, classical machine learning approaches have difficulties in learning useful patterns from them. In this study, we propose a novel Locate-Then-Detect (LTD) system that can precisely detect Web threats in real-time by using attention-based deep neural networks. Firstly, an efficient Payload Locating Network (PLN) is employed to propose most suspicious regions from large URL requests/posts. Then a Payload Classification Network (PCN) is adopted to accurately classify malicious regions from suspicious candidates. In this way, PCN can focus more on learning malicious segments and highly increase detection accuracy. The noise induced by irrelevant background strings can be largely eliminated. Besides, LTD can greatly reduce computational costs (82.6% less) by ignoring large irrelevant URL content. Experiments are carried out on both benchmarks and real Web traffic. The LTD outperforms an HMM-based approach, the Libinjection system, and a leading commercial rule-based Web Application Firewall. Our method can be efficiently implemented on GPUs with an average detection time of about 5ms and well qualified for real-time applications.

Social networks are effective tools for analyzing many social topics in sociology. In the past few decades, a great deal of efforts have been made to study the balance property of social networks. This paper presents a novel bi-objective model for social network structural balance, and a multiobjective discrete particle swarm optimizer is used to optimize the bi-objective model. Each single run of the algorithm can yield a set of Pareto solutions, each of which represents a certain network partition that divides a signed network into many clusters. Consequently, by simultaneously optimizing the objectives in the proposed model, one may have many choices to analyze the balance problem. Extensive experiments compared against several other models and algorithms have been done. All the experiments indicate that the proposed model is helpful for social network structural balance analytics, and that the algorithm is effective.