基于深度学习的激光剪切散斑干涉无损检测缺陷识别

剪切散斑干涉技术通过测量物体表面变形来推断其内部缺陷,具有高灵敏度、检测范围广、精度高等优点,是一种极具潜力的复合材料无损检测技术。目前缺陷识别主要采用人工方式,而人工识别不但检测效率低且受到专业性限制。为了提高剪切散斑干涉无损检测方法中的缺陷识别精度和效率,本文提出基于深度学习剪切散斑干涉缺陷识别方法。利用高精度四步相移技术获取剪切散斑相位条纹高质量成像;引入了应用广泛的YOLOv5和Faster R-CNN目标检测算法,通过实验采集了大量的缺陷图像,分别用YOLOv5和Faster R-CNN两种算法获得训练模型。然后将这两种模型分别应用于剪切散斑干涉无损检测中的复合材料缺陷检测。最后,实验从检测速率和检测精度方面对模型识别效果进行了对比分析。实验结果表明,激光剪切散斑干涉技术结合深度学习的方法能有效地实现剪切散斑干涉无损检测的缺陷自动识别,Faster R-CNN和YOLOv5的检测速率分别能达到11 f/s和50 f/s,并且两种深度学习算法的平均精度均能达到92%以上,验证了提出方法的可行性。     

Theoretical and experimental study of fatigue growth of interacting cracks

Life assessment of structures weakened by interacting cracks represents an important and very challenging problem. Subsequently, the main objective of this paper is to address this problem by developing a new computational technique. It is based on the classical strip-yield model and plasticity-induced crack closure concept. It also utilises the 3D fundamental solution for an edge dislocation. The crack advance scheme adopts the cycle-by-cycle calculations of the effective stress intensity factors and crack increments. The modelling results are validated against an experimental study focusing on fatigue behaviour of two closely spaced collinear cracks in wide plates with different thicknesses. It is confirmed that non-linear effects associated with crack interaction have a significant influence on fatigue life and cannot be disregarded in life and integrity assessments of structural components.     

Micro-mechanism of rolling contact fatigue in Hadfield steel crossing

The formation of the fatigue cracks was due to massive vacancy clusters in the subsurface layer of Hadfield steel crossing, which are induced by the accumulated plastic deformation under the conditions of impact and contact stresses from train wheels. The high concentration layers of vacancy clusters were formed parallel to the working surface of the crossing, which caused the initial rolling contact fatigue cracks to be parallel to the working surface with a laminar distribution in the depth direction. It can be predicted that metals containing elements with larger atomic diameter should have better rolling contact fatigue and wear performances.     

Fabrication of HAp–8YSZ composite layer on Ti/TiO2 nanoporous substrate by EPD/MAO method

Zirconia/Hydroxyapatite composites containing 20–50 wt.% 8YSZ were prepared on Ti/TiO₂ substrates by electrophoretic deposition (EPD)/micro-arc oxidation (MAO) process. Titania, as an inner layer, was grown on the Ti plates using MAO treatment in order to form a strong join between substrate and HAp. These composites were produced by EPD in ethanol containing ZrO₂/HAp particles at 50, 100 and 150 V in 1 min. As-prepared samples were sintered at 900, 1100 and 1300 °C. HAp, β-TCP, CaZrO₃ phases were identified using X-ray diffractometry analysis (XRD). Scanning electron microscopy (SEM) utilized to study the surface morphology indicated a crack free microstructure at 1300 °C.     

Vulnerable objects detection for autonomous driving: A review

Object detection performed by Autonomous Vehicles (AV)s is a crucial operation that comes ahead of various autonomous driving tasks, such as object tracking, trajectories estimation, and collision avoidance. Dynamic road elements (pedestrians, cyclists, vehicles) impose a greater challenge due to their continuously changing location and behaviour. This paper presents a comprehensive review of the state-of-the-art object detection technologies focusing on both the sensory systems and algorithms used. It begins with a brief introduction on the autonomous driving operations and challenges. Then, different sensory systems employed on existing AVs are elaborated while illustrating their advantages, limitations and applications. Also, sensory systems employed by different research are reviewed. Moreover, due to the significant role Deep Neural Networks (DNN)s are playing in object detection tasks, different DNN-based networks are also highlighted. Afterwards, previous research on dynamic objects detection performed by AVs are reviewed in tabular forms. Finally, a conclusion summarizes the outcomes of the review and suggests future work towards the development of vehicles with higher automation levels.     

高速列车转向架构架损伤、等效应力及寿命分布特性研究

为分析高速列车转向架构架损伤、等效应力及寿命分布特性，对构架疲劳关键测点进行动应力线路实测并对测点实测时域数据波形进行解析；基于实测应力时间历程及雨流计数法编制二维应力谱，利用Goodman等寿命方程将二维应力谱等效转换为一维应力谱；阐述线性累积损伤及非线性累积损伤模型的建立方法并对实测数据的线性累积损伤及非线性累积损伤进行了计算及对比分析；分别基于线性累积损伤理论及非线性累积损伤理论推导出各理论下的等效应力，基于实测数据对两种等效应力进行了计算及对比分析；通过结合非线性累积损伤及线性累积损伤理论计算的等效应力及不同可靠度下的材料S-N曲线计算并对比分析构架结构的疲劳寿命。研究结果表明，与非线性疲劳分析理论相比，线性疲劳分析理论对高速列车转向架构架的疲劳特性评估偏于保守。     

Identification and idle speed control of internal combustion engines

This paper presents an integrated approach for the identification and control of internal combustion engines in idle-speed conditions. The inputs of the nonlinear identification model are the position of the idle speed air actuation system and the spark advance, while its outputs are the pressure inside the air intake manifold and the crank shaft speed. The estimated model is then used to synthesize an idle speed controller with the linear quadratic technique. The design procedure outlined here is currently being used by Magneti Marelli for the synthesis of commercial idle-speed regulators. Some identification and control results obtained by applying this method to a 1400 cm³ commercial engine are reported to witness the effectiveness of the proposed approach.     

碳化硅纳米纤维改性C/C复合材料的力学性能

以电镀Ni颗粒为催化剂,采用催化化学气相沉积(CCVD)法,在单向炭纤维(CF)表面原位生长碳化硅纳米纤维(SiCNF),制备出SiCNF/CF共增强毡体。以此共增强毡体为前驱体,化学气相沉积碳后得到密度为1.7g/cm3的SiCNF改性C/C复合材料。复合材料力学性能测试表明,SiCNF改性可使C/C复合材料的抗弯强度、抗压强度和显微硬度明显提高。扫描电子显微镜和偏光显微镜观察表明,SiCNF改性处理改变了C/C复合材料中基体炭的结构,使其成为类似粗糙层(RL)或高织构的结构。     

Moldflow在注塑成型翘曲优化中的应用

以汽车副仪表板装饰盖为例,借助Moldflow软件进行模拟,预测其翘曲变形量,通过分析材料、熔体温度和模具温度对翘曲的影响,从而得到最佳的工艺参数组合,对改善产品质量、降低成本具有重要的指导意义.     

半固态AZ91D镁合金流动性的研究

根据金属凝固理论,建立了半固态镁合金流动性的计算模型,采用该模型预测了半固态AZ91D镁合金的充型能力.结果证实,预测结果与试验结果相吻合.通过流动性试验,对半固态AZ91D镁合金的充型能力进行了研究.结果表明,剪切后的半固态镁合金浆料的充型能力明显要比未经剪切的镁合金浆料的充型能力好.随着充型温度降低,剪切前后镁合金浆料充型长度增加值逐渐增大,而随着充型温度的继续降低,剪切前后的充型长度的增加值又逐渐降低.半固态镁合金在560～570℃时充型长度增加值最大.