Deep convolutional tree-inspired network: a decision-tree-structured neural network for hierarchical fault diagnosis of bearings

The fault diagnosis of bearings is crucial in ensuring the reliability of rotating machinery. Deep neural networks have provided unprecedented opportunities to condition monitoring from a new perspective due to the powerful ability in learning fault-related knowledge. However, the inexplicability and low generalization ability of fault diagnosis models still bar them from the application. To address this issue, this paper explores a decision-tree-structured neural network, that is, the deep convolutional tree-inspired network (DCTN), for the hierarchical fault diagnosis of bearings. The proposed model effectively integrates the advantages of convolutional neural network (CNN) and decision tree methods by rebuilding the output decision layer of CNN according to the hierarchical structural characteristics of the decision tree, which is by no means a simple combination of the two models. The proposed DCTN model has unique advantages in 1) the hierarchical structure that can support more accuracy and comprehensive fault diagnosis, 2) the better interpretability of the model output with hierarchical decision making, and 3) more powerful generalization capabilities for the samples across fault severities. The multiclass fault diagnosis case and cross-severity fault diagnosis case are executed on a multicondition aeronautical bearing test rig. Experimental results can fully demonstrate the feasibility and superiority of the proposed method.     

On-line tracking of pulverized coal and biomass fuels through flame spectrum analysis

This paper presents a new approach to the on-line tracking of pulverized coal and biomass fuels through flame spectrum analysis.A flame detector containing four photodiodes is used to derive multiple signals covering a wide spectrum of the flame from visible,near-infrared and mid-infrared spectral bands as well as a part of far-infrared band.Different features are extracted in time and frequency domains to identify the dynamic ＂fingerprints＂ of the flame.Fuzzy logic inference techniques are employed to combine typical features together and infer the type of fuel being burnt.Four types of pulverized coal and five types of biomass are burnt on a laboratory-scale combustion test rig.Results obtained demonstrate that this approach is capable of tracking the type of fuel under steady combustion conditions.     

Cyclic stress–strain response and crystal plasticity finite element analysis of AISI 9310 steel in biaxial fatigue loading

There are still many gaps in the research on the multiaxial fatigue failure mechanism of the gear shaft. In this paper, cyclic stress–strain response and biaxial fatigue damage characteristics of gear steel AISI 9310 were investigated. The specimens showed obvious cyclic softening characteristics at all phase angles, and the softening rate was directly associated with the initiation and propagation of cracks. The fractographies at different phase angles revealed that the specimens under out-of-phase loading suffered fatigue failure caused by a single crack source on the surface, while the fatigue crack under in-phase loading was gathered together by the propagation of different crack sources. Finally, the established crystal plastic finite element model showed a good prediction of the plastic strain energy density at different phase angles, and the maximum error was 13.03%. Furthermore, a biaxial fatigue life prediction method was proposed, with a maximum error of 39.5%.     

多风格融合的复杂森林场景自适应可视化

为了保证森林场景可视化时的真实感,同时保持动态森林场景生成时间的恒定性,本文提出了一种多风格融合的复杂森林场景自适应可视化方法。该方法利用基于视距的模型分布函数来控制树木模型的分布比例,从而建立多风格融合的森林可视化模型;在此基础上,根据复杂森林场景中树木生长模型的计算时间、三维树木绘制时间的估算结果,以及树木的视觉重要性,确定生成森林场景的最佳方案。该方法能够使复杂动态森林场景的生成时间保持较好的稳定性,并且在可视化过程中根据仿真效果动态调整绘制策略。为了验证该方法的有效性和实用性,在动态生长的森林仿真场景中对本文的方法进行了实验和应用。应用结果表明,多风格融合的森林场景自适应可视化方法能在保证森林场景可视化真实感的基础上,有效地提升复杂森林场景的绘制速度,使森林场景的快速漫游具有更好的稳定性和流畅性。     

基于数字孪生的城市交通流智能预测与导引策略

物联网和人工智能等信息技术的快速发展极大地推动了交通系统的变革,同样也带来了机遇与挑战。针对现有导航系统忽略交通流时空特征而产生的策略重复性拥堵问题,对宏观交通流和微观车辆驾驶分别建模,并挖掘其耦合关系,进而提出一种基于数字孪生的城市智能交通流分层预测与导引策略,为减缓交通拥堵提供新思路。在该策略中,虚拟空间中的上层道路孪生通过扩散卷积递归神经网络预测时空交通流量,并显式作用于车辆路径规划决策。在此基础上,提出一种时空协同深度强化学习方法,用于实现车辆面向未来的协作式路径规划,指导虚拟空间中的下层车辆孪生选出最优策略反馈于真实世界。基于SUMO仿真平台进行了仿真验证。实验结果表明,本文所提方法在提高出行达成率、缓解拥堵等方面显著优于现有算法,能够有效提升城市交通出行效率。     

An Efficient Intercalation Supramolecular Structure for Photocatalytic CO2 Reduction to Ethylene Under Visible Light

Photocatalytic CO₂ reduction (CO₂PR) into multi-carbon products (especially C₂H₄) is a highly attractive route for global carbon cycle, however, which is seriously limited by sluggish C-C coupling kinetics and competitive hydrogen evolution reaction (HER) and so on. Herein, the fabrication of a novel supramolecular assembly of NiAl-Fe-TCPP is reported by intercalating iron porphyrin (Fe-TCPP) into NiAl-layered double hydroxide (NiAl-LDH), and the resultant NiAl-Fe-TCPP exhibit superior catalytic performance on CO₂PR to C₂H₄ under visible light irradiation in presence of photosensitizer. A high C₂H₄ selectivity up to 93.4% in the carbon-containing products with the production rate as high as 24.7 µmol h⁻¹ can be achieved over NiAl-Fe-TCPP. The ex/in situ X–ray absorption spectoscopy (XAS) indicates that the electron transfer between NiAl-LDH and Fe-TCPP can promote the generation of low-valence of Fe sites, resulting in the efficient production of C₂H₄. The spin-polarized density functional theory (DFT) calculations find that the synergistic mechanism that CO₂ molecules are activated to CO on NiAl-LDH and then spilled to Fe-TCPP and coupled to COCHO^#, which is further reduced to C₂H₄, are feasible in the perspective of Gibbs free energy. Moreover, the strong host-guest interactions between NiAl-LDH and Fe-TCPP lead to the promoted photocatalytic activity and superior cycle stability.     

基于ATT&CK的APT攻击语义规则构建

从自然语言描述文本中提取网络攻击知识存在语义鸿沟,导致TTPs威胁情报自动化利用低。为提高威胁情报自动分析效率,设计并实现了基于ATT&CK的APT攻击语义规则。首先,构建带标签的有向图语义规则模型,对自然语言文本描述的攻击技术进行知识化描述;其次,定义语义规则,阐释网络实体属性及其逻辑运算关系的形式化描述方法;最后,利用关键词组识别、知识抽取等自然语言处理技术,从攻击技术文本中抽取形成123个APT攻击语义规则,涵盖ATT&CK的115项技术和12种战术。利用模拟场景采集的APT攻击日志数据,对语义规则进行验证,实验结果表明,语义规则检出率达到93.1%,并具备一定的攻击上下文信息还原能力,可有效支撑威胁检测分析。     

High strength and toughness of double physically cross-linked hydrogels composed of polyvinyl alcohol and calcium alginate

The weak mechanical properties of hydrogels, especially physically cross-linked hydrogels are usually a major factor to hinder their application. To solve this problem, in this work, we prepared a high strength and toughness of double physically cross-linked (PDN) hydrogels composed of crystalline domain cross-linked polyvinyl alcohol (PVA) and Ca²⁺-cross-linked alginate (Alg). With a further annealing treatment, the noncovalent cross-linked network via the formed crystalline promote the as-prepared PDN PVA/Alg hydrogel to exhibit well mechanical properties with the tensile strength of ~1.94 MPa, elongation at break of ~607% and Young's modulus of ~0.45 MPa (above 70 wt% of water content). By analyzing the mechanism of improving the hydrogel mechanical properties, it is found that annealing can effectively improve the crystallinity of PVA in the hydrogel, and then greatly improve the mechanical properties of the hydrogel. This provides a general method for improving the mechanical properties of PVA PDN hydrogels. In addition, the PDN PVA/Alg hydrogel was also proved to have good ionic conductivity of 1.70 S m⁻¹. These desirable properties make the prepared physically cross-linked hydrogels promising materials for medical and biosensing fields.     

水厂清水池的水力特性模拟与结构优化

以某水厂清水池为案例,构建7种不同结构的清水池几何模型进行CFD模拟,对比水力效率和粒子扩散特征。结果表明,清水池前段流态接近混流,后段逐渐均匀,近似平稳推流模式。进口第一个廊道区,导流板背水面和弯道拐角区域易出现回流、大面积循环流等水力状态。增加导流板数目,延长导流板长度,将进出水口设置在进出水廊道宽边中间位置可使清水池更接近理想推流状态。此外,当进水流速为1.2 m/s和2 m/s时,导流板垂直布置的清水池模型因其廊道数更多,比平行布置时的水力效率评价指标值T10/T分别提升23%和13%。CFD模拟显示,实际工程中用于池体稳定的结构柱会显著影响清水池流场流线和消毒剂粒子停留时间,当进水流速为1.2 m/s和2 m/s时,结构柱的存在会使T10/T分别提升31%和38%,但同时会引起流线的纵向扰动,是影响清水池水力条件的重要因素。研究成果可以为水厂清水池的设计优化提供参考借鉴。