This study examines the viscoelastic-plastic behavior of thermoplastic resin poly-ether-ether-ketone(PEEK)under high temperature and strain rate conditions,highlighting its potential in aerospace applications due to i...This study examines the viscoelastic-plastic behavior of thermoplastic resin poly-ether-ether-ketone(PEEK)under high temperature and strain rate conditions,highlighting its potential in aerospace applications due to its impact resistance.A dualhardening constitutive model that combines physical and phenomenological approaches is developed to simulate the mechanical behavior of PEEK.The model explicitly incorporates its marked tension-compression asymmetry in plasticity and relaxation,along with thermal softening at high strain rates,enabling accurate predictions over a wide range of temperatures and strain rates with minimal parameters.This study establishes a comprehensive workflow from experimentation to finite element(FE)simulation for thermoplastic resins.Uniaxial tensile and compression tests(23℃-180℃,0.00229s^(-1)-0,19361s^(-1))and split Hopkinson pressure bar(SHPB)tests(1094.08s^(-1)-5957.88s^(-1))are performed to capture stress-strain responses across various conditions,with small-scale specimens enhancing fracture strain measurement accuracy,and quantify the Taylor-Quinney factor of the PEEK material during the adiabatic heating process.The findings demonstrate that the proposed constitutive model effectively predicts yield points across different strain rates and temperatures,with parameters easily obtainable through simple experimental methods,enhancing its practical applications.展开更多
单晶硅生产过程中,引晶工艺产生的缺陷严重影响产品质量,传统的基于视觉的缺陷检测方法在检测引晶图像中的凸点小目标时,存在检测速度慢、参数量大、难以部署在嵌入式终端等不足。为此,提出了一种改进的YOLOv8目标检测模型,引入了Contex...单晶硅生产过程中,引晶工艺产生的缺陷严重影响产品质量,传统的基于视觉的缺陷检测方法在检测引晶图像中的凸点小目标时,存在检测速度慢、参数量大、难以部署在嵌入式终端等不足。为此,提出了一种改进的YOLOv8目标检测模型,引入了ContextGuided模块,提高了模型的推理效率;在特征融合网络中引入更为高效的DySample,优化了特征融合的效率和深度;采用轻量级网络结构,减少了模型的复杂度和计算量,使其适应计算资源有限的终端设备。在工业数据集上进行了训练和测试,实验结果表明,对凸点小目标的检测更加准确,mAP(mean average precision)达到97.7%,在精确率上相对于YOLOv8n提升了11.6%,同时参数量减少31.9%,方便部署在嵌入式终端。展开更多
非均质材料广泛应用于航空航天、电子信息、国防等领域高端装备零部件,其多相、多尺度、形貌复杂的微观结构特征决定材料宏观性能的优越性,建立准确的微观结构模型对于深入理解结构-性能关系至关重要。然而当材料表现出强烈的非均质特性...非均质材料广泛应用于航空航天、电子信息、国防等领域高端装备零部件,其多相、多尺度、形貌复杂的微观结构特征决定材料宏观性能的优越性,建立准确的微观结构模型对于深入理解结构-性能关系至关重要。然而当材料表现出强烈的非均质特性时,该过程的复杂性显著提升且难度增大。近年来,计算材料科学进步推动计算模拟方法发展,材料微观结构表征与重建(microstructure characterization and reconstruction,MCR)技术作为计算模拟过程的关键环节,为非均质材料的微观结构建模提供有力途径。目前,非均质材料的MCR主要包括两类:(1)基于统计方法的建模技术;(2)基于机器学习方法及计算机视觉的建模技术。本工作总结并梳理这两类MCR技术,阐释相关方法的特点及适用性,并分析不同方法在非均质材料微观结构表征与重建方面的研究进展,为如何选取MCR方法并将其应用于材料设计提供借鉴和指导。展开更多
文摘This study examines the viscoelastic-plastic behavior of thermoplastic resin poly-ether-ether-ketone(PEEK)under high temperature and strain rate conditions,highlighting its potential in aerospace applications due to its impact resistance.A dualhardening constitutive model that combines physical and phenomenological approaches is developed to simulate the mechanical behavior of PEEK.The model explicitly incorporates its marked tension-compression asymmetry in plasticity and relaxation,along with thermal softening at high strain rates,enabling accurate predictions over a wide range of temperatures and strain rates with minimal parameters.This study establishes a comprehensive workflow from experimentation to finite element(FE)simulation for thermoplastic resins.Uniaxial tensile and compression tests(23℃-180℃,0.00229s^(-1)-0,19361s^(-1))and split Hopkinson pressure bar(SHPB)tests(1094.08s^(-1)-5957.88s^(-1))are performed to capture stress-strain responses across various conditions,with small-scale specimens enhancing fracture strain measurement accuracy,and quantify the Taylor-Quinney factor of the PEEK material during the adiabatic heating process.The findings demonstrate that the proposed constitutive model effectively predicts yield points across different strain rates and temperatures,with parameters easily obtainable through simple experimental methods,enhancing its practical applications.
文摘单晶硅生产过程中,引晶工艺产生的缺陷严重影响产品质量,传统的基于视觉的缺陷检测方法在检测引晶图像中的凸点小目标时,存在检测速度慢、参数量大、难以部署在嵌入式终端等不足。为此,提出了一种改进的YOLOv8目标检测模型,引入了ContextGuided模块,提高了模型的推理效率;在特征融合网络中引入更为高效的DySample,优化了特征融合的效率和深度;采用轻量级网络结构,减少了模型的复杂度和计算量,使其适应计算资源有限的终端设备。在工业数据集上进行了训练和测试,实验结果表明,对凸点小目标的检测更加准确,mAP(mean average precision)达到97.7%,在精确率上相对于YOLOv8n提升了11.6%,同时参数量减少31.9%,方便部署在嵌入式终端。
文摘非均质材料广泛应用于航空航天、电子信息、国防等领域高端装备零部件,其多相、多尺度、形貌复杂的微观结构特征决定材料宏观性能的优越性,建立准确的微观结构模型对于深入理解结构-性能关系至关重要。然而当材料表现出强烈的非均质特性时,该过程的复杂性显著提升且难度增大。近年来,计算材料科学进步推动计算模拟方法发展,材料微观结构表征与重建(microstructure characterization and reconstruction,MCR)技术作为计算模拟过程的关键环节,为非均质材料的微观结构建模提供有力途径。目前,非均质材料的MCR主要包括两类:(1)基于统计方法的建模技术;(2)基于机器学习方法及计算机视觉的建模技术。本工作总结并梳理这两类MCR技术,阐释相关方法的特点及适用性,并分析不同方法在非均质材料微观结构表征与重建方面的研究进展,为如何选取MCR方法并将其应用于材料设计提供借鉴和指导。
