基于数据质量评价的目标估计

针对目标估计过程需要大量人工参与、自动化程度低的问题，提出了基于数据质量评价的目标估计方法。利用目标数据质量评价方法，对不同传感器得到的目标数据质量进行科学、有效的测度和评价，并根据质量得分动态调整各数据源在目标估计过程中所占的权重，从而减少人工干预，提高目标估计效能。仿真试验结果证明了该方法的有效性。     

Status and initial experiments of DRAGON-V facility for lithium-lead blanket technologies of hydrogen fusion reactors

A dual-coolant integrated experimental facility named DRAGON-V has been developed at the Institute of Nuclear Energy Safety Technology, Hefei Institute of Physical Science, Chinese Academy of Sciences, for the key technology research and performance evaluation of candidate liquid lithium-lead (PbLi) blanket of hydrogen fusion reactors. The loop is composed of a material test sub-loop and thermal-hydraulic test sub-loop, the design parameters are PbLi inventory 20 tons, PbLi temperature up to 550 °C, the maximum PbLi flow rate up to 40 kg/s. A novel cold trap system is designed to remove the suspended and crystalized impurities in PbLi fluid with three cooling zones and cross row arrangement of rod bundle filter elements. The paper describes the loop itself and its major components, initial loop testing, flow and measurement diagnostics and current experiments. The obtained test results of the loop and its components have demonstrated that the new facility is fully functioning and ready for experimental studies of material corrosion with/without a magnetic field, magnetohydrodynamic (MHD) effect, purification, heat and mass transfer phenomena in PbLi flows and can also be used in mock-up testing in conditions relevant to fusion applications.     

Co?N graphene encapsulated cobalt catalyst for H2O2 decomposition under acidic conditions

Hydrogen peroxide (H₂O₂) has been listed as one of the 100 most important chemicals in the world. However, huge amount of residual H₂O₂ is hard to timely decomposed into O₂ and H₂O under acidic condition, easily resulting in explosion hazard. Here, we reported a core–shell structure catalyst, that is graphene with Co N structure encapsulated Co nanoparticles. Co N graphene shell serves as the active site for the H₂O₂ decomposition, and Co core further enhance this decomposition. Benefiting from it, the H₂O₂ decomposition were close to 100% after 6 cycles without pH adjustment, which increased 6 orders of magnitude compared with no catalyst. At the same time, the O₂ generation reached 99.67% in 2 h with little metal leaching, and ·OH has been greatly inhibited to only 0.08%. This work can cleanly remove H₂O₂ with little deep oxidation and protect the process of H₂O₂ utilization to achieve a safer world.     

机器翻译译文质量估计综述

机器翻译译文质量估计(Quality Estimation,QE)是指在不需要人工参考译文的条件下,估计机器翻译系统产生的译文的质量,对机器翻译研究和应用具有很重要的价值。机器翻译译文质量估计经过最近几年的发展,取得了丰富的研究成果。该文首先介绍了机器翻译译文质量估计的背景与意义;然后详细介绍了句子级QE、单词级QE、文档级QE的具体任务目标、评价指标等内容,进一步概括了QE方法发展的三个阶段: 基于特征工程和机器学习的QE方法阶段,基于深度学习的QE方法阶段,融入预训练模型的QE方法阶段,并介绍了每一阶段中的代表性研究工作;最后分析了目前的研究现状及不足,并对未来QE方法的研究及发展方向进行了展望。     

Decomposition and replacement: Spatial knowledge distillation for monocular depth estimation

Knowledge distillation has become a key technique for making smart and light-weight networks through model compression and transfer learning. Unlike previous methods that applied knowledge distillation to the classification task, we propose to exploit the decomposition-and-replacement based distillation scheme for depth estimation from a single RGB color image. To do this, Laplacian pyramid-based knowledge distillation is firstly presented in this paper. The key idea of the proposed method is to transfer the rich knowledge of the scene depth, which is well encoded through the teacher network, to the student network in a structured way by decomposing it into the global context and local details. This is fairly desirable for the student network to restore the depth layout more accurately with limited resources. Moreover, we also propose a new guidance concept for knowledge distillation, so-called ReplaceBlock, which replaces blocks randomly selected in the decoded feature of the student network with those of the teacher network. Our ReplaceBlock gives a smoothing effect in learning the feature distribution of the teacher network by considering the spatial contiguity in the feature space. This process is also helpful to clearly restore the depth layout without the significant computational cost. Based on various experimental results on benchmark datasets, the effectiveness of our distillation scheme for monocular depth estimation is demonstrated in details. The code and model are publicly available at : https://github.com/tjqansthd/Lap_Rep_KD_Depth.     

Study on vapour dispersion and explosion from compressed hydrogen spill: Risk assessment on a hydrogen plant

Hydrogen produced from renewable resources is one of the cleanest fuels and could be used to store intermittent solar, wind and other energies. The main concern about using hydrogen is its hazards, such as high storage pressure, wide-range flammability, low mass density, and high diffusion. This study investigated the hazards of compressed hydrogen storage by developing a CFD model to understand the gas dispersion behaviour. The model was validated using the past experimental data and showed a good agreement, which could demonstrate the diffusion characteristics and gas stratification of a buoyant gas. A case study of an accidental release of compressed hydrogen from a storage tank was investigated to evaluate the risk of a hydrogen plant. A mathematical model of the jet spill was used to account for the choking effect from a high-pressure release to ensure the input velocity in CFD simulation is suitable for modelling gas dispersion using verified spatial and temporal scales, then the simulation results were used as inputs of vapour cloud explosions (VCEs) to investigate the potential overpressure effect. It was found the CFD model could predict a more reasonable flammable gas amount in cloud than using the bulk hydrogen release rate. The safety distance based on the overpressure prediction was reduced by 35%. The method proposed in this study can provide more validity for the consequence analysis as part of risk assessment.     

A comparison of shear stress estimation methods for a single geobag on a rough bed

There are several methods for estimating bed shear stress in the literature, but comprehensive comparisons among them are limited and under specific conditions. This study compared these methods first on a bare smooth bed, and then for a single geobag on a rough bed in the interest of determining the stability of geobags used in riverbank protection structures. The geobag was filled with cement or sand and tested under different open channel flow conditions. The turbulent kinetic energy method appeared to best represent the local bed shear stress on the geobag when using the newly calibrated proportionality constants. The Reynolds stress method via extrapolation was relatively unaffected by changes to the geobags shape and measurement locations, suggesting this method inadequately represents the local bed shear stress. The Patel method and the universal law of the wall method failed to represent local bed shear stress in the rough bed cases due to instrument limitations and the breakdown of the law of the wall. This study highlights the impact of different methods on the bed shear stress estimation.     

Convolutional Neural Network Auto Encoder Channel Estimation Algorithm in MIMO-OFDM System

Higher transmission rate is one of the technological features of prominently used wireless communication namely Multiple Input Multiple Output-Orthogonal Frequency Division Multiplexing (MIMO–OFDM). One among an effective solution for channel estimation in wireless communication system, specifically in different environments is Deep Learning (DL) method. This research greatly utilizes channel estimator on the basis of Convolutional Neural Network Auto Encoder (CNNAE) classifier for MIMO-OFDM systems. A CNNAE classifier is one among Deep Learning (DL) algorithm, in which video signal is fed as input by allotting significant learnable weights and biases in various aspects/objects for video signal and capable of differentiating from one another. Improved performances are achieved by using CNNAE based channel estimation, in which extension is done for channel selection as well as achieve enhanced performances numerically, when compared with conventional estimators in quite a lot of scenarios. Considering reduction in number of parameters involved and re-usability of weights, CNNAE based channel estimation is quite suitable and properly fits to the video signal. CNNAE classifier weights updation are done with minimized Signal to Noise Ratio (SNR), Bit Error Rate (BER) and Mean Square Error (MSE).     

Fire safe,sustainable loose furnishing

The aim of this exploratory study has been to investigate the fire properties and environmental aspects of different upholstery material combinations, mainly for domestic applications. An analysis of the sustainability and circularity of selected textiles, along with lifecycle assessment, is used to qualitatively evaluate materials from an environmental perspective. The cone calorimeter was the primary tool used to screen 20 different material combinations from a fire performance perspective. It was found that textile covers of conventional fibres such as wool, cotton and polyester, can be improved by blending them with fire resistant speciality fibres. A new three‐dimensional web structure has been examined as an alternative padding material, showing preliminary promising fire properties with regard to ignition time, heat release rates and smoke production.