Liverpool College

Online acquisition of accurate battery electrical, thermal, and aging states is crucial for monitoring and management of batteries, especially in cloud-edge collaboration. In this paper, an online joint estimation method for state of charge (SOC), 3-dimensional state of temperature distribution (SOTD), and state of health (SOH) of batteries based on an electro-thermal-aging coupled model is proposed. Firstly, three well-chosen charging durations are leveraged as the aging features to achieve SOH estimation with the assistance of the extreme learning machine (ELM). Secondly, SOT is estimated through a lumped parameter thermal model (LPTM) and adaptive extended Kalman filter (AEKF), and SOTD is reconstructed using a well-designed fractal theory method (FTM) derived from the iterated function system (IFS) method and a Lindenmayer system (L-system), which ensures precise thermal state monitoring. Then, based on the parameter updating mechanism, the parameters of the equivalent circuit model (ECM) are adjusted in real time according to the accurate SOT and SOH, and the SOC is estimated by the AEKF. Subsequently, the proposed method is validated under different dynamic conditions and different temperatures, the validation results show that the proposed method can achieve accurate SOC-SOTD-SOH co-estimation with an error of no more than 2.14%, 1.37°C, and 1.26%, respectively. Lastly, the investigation of the SOT and SOH influence on SOC estimation reveals the important significance of the co-estimation framework.

The neutral-point voltage imbalance of a three-phase VIENNA rectifier can cause problems such as voltage asymmetry on the DC side, harmonic distortion and device overvoltage, seriously restricting the reliability of the system. This paper finds that the mismatch of positive/negative small vector duty cycles in traditional Space Vector Modulation (SVPWM) is the deep-seated mechanism causing this problem, but SVPWM has high computational complexity and poor real-time performance. To this end, an SPWM modulation method equivalent to a space vector is proposed. The active balance of the neutral-point potential is achieved by dynamically adjusting the action time of positive and negative small vectors by injecting zero-order components: Firstly, the quantization relationship between the space vector and the neutral-point current is established; Convert the vector action time calculation of SVPWM into the zero-order component analytical solution of SPWM to reduce the operation dimension. Finally, the simulation shows that the proposed method significantly improves the balance of the neutral-point potential compared with the traditional SVPWM, and the modulation effect is relatively close to that of the traditional SVPWM. This study provides a cost-effective engineering solution for the Vienna topology by replacing vector synthesis with modulated wave reconstruction

Comprehensive Summary Hydride superconductors are promising candidates for high‐temperature superconductivity across a wide pressure range. This review presents a comprehensive review of their structural, electronic, and superconducting properties, with a focus on how pressure influences phase stability and enhances critical temperature ( T c ). We categorize hydrides into three pressure regimes: ambient pressure, low pressure (<100 GPa), and high pressure (>100 GPa). Ambient pressure compounds, such as perovskite‐like hydrides and SM 2 TMH 6 structures, exhibit moderate T c values. Low‐pressure hydrides benefit from unique strategies like molecular doping and electron precompression to improve their T c . The high‐pressure hydrides exhibit higher T c values, including room‐temperature superconductivity, but require extreme conditions for synthesis and characterization. We also highlight recent theoretical and experimental advances, outlining current challenges and prospects. This review not only highlights the potential of hydride superconductors but also provides a roadmap for future research in this exciting and rapidly developing field. Key Scientists

In the pursuit of advancing autonomous maritime navigation, this study aimed to develop a novel architecture designed to enhance the detection accuracy of distant and small targets under the constraints of real-time performance and robustness. Through the innovative integration of the Convolutional Block Attention Module (CBAM) into the detection model's backbone, the study achieved superior feature extraction capabilities tailored for the complexities of maritime environments. Further optimization of the Spatial Pyramid Pooling (SPP) module ensured model compactness and computational efficiency, vital for deployment on edge devices. A key methodological novelty lay in the incorporation of the S-IoU loss function, which offers superior bounding box regression accuracy over the traditional Generalized Intersection over Union, directly contributing to more precise navigation and effective obstacle avoidance. The proposed enhancements collectively yielded a 5.1 % increase in mAP@50 %, accompanied by an 11.2 % reduction in model parameters and a 12.6 % decrease in computational complexity (GFLOPs). These findings underscore the potential of the presented architecture to significantly contribute to maritime safety, presenting an optimized solution for collision avoidance and navigation assistance in congested sea routes and adverse weather conditions. • This research introduces a novel target detection architecture from the ship bridge perspective, achieving a notable improvement in the detection accuracy of distant and small targets while ensuring real-time performance and robustness. The proposed framework effectively addresses the inherent challenges of maritime navigation, enhancing the precision of visual perception in complex navigational environments. • The integration of the Convolutional BlockAttention Module (CBAM) into the model's backbone significantly enriches feature extraction, tailored for maritime scenarios. Coupled with the refinement of the Spatial Pyramid Pooling (SPP) module, the architecture attains model compactness and computational efficiency. These enhancements are crucial for deployment on edge devices. • Incorporating a Scaled Generalized Intersection over Union (S-IoU) as an improvement over the conventional Generalized Intersection over Union (GIoU) for bounding box regression loss function, the study demonstrates a substantial boost in the model's detection accuracy of small and distant targets. This methodological refinement led to a notable 5.1 % increase in mean Average Precision (mAP) at 50 %, along with an 11.2 % reduction in model parameters.

High-stakes domains are exploring the transformative potential of Large Language Models (LLMs) for complex decision-making tasks. However, LLM deployment in real-world settings presents challenges in data security, evaluation of its capabilities outside controlled environments, and accountability attribution in the event of adversarial decisions. This paper proposes a framework for responsible deployment of LLM-based decision-support systems through active human involvement. It integrates interactive collaboration between human experts and developers through multiple iterations at the pre-deployment stage to assess the uncertain samples and judge the stability of the explanation provided by post-hoc XAI techniques. Local LLM deployment within organizations and decentralized technologies, such as Blockchain and IPFS, are proposed to create immutable records of LLM activities for automated auditing to enhance security and trace back accountability. It was tested on Bert-largeuncased, Mistral, and LLaMA 2 & 3 models to assess the capability to support responsible financial decisions on business lending.