车轮超声波数字化探伤系统

介绍车轮超声波探伤信号数字化处理的研究,以及系统的设计思想、构成和现场应用效果.     

Microstructures of pure nickel induced by high-current pulsed electron beam irradiation

The characteristics of surface feature and microstructure of the near-surface of nickel were investigated after high current pulsed electron beam (HCPEB) irradiation. Based on a physical model, the temperature profile was simulated. The depth of heat-affected zone, the initial melting position, heating rate and quenching rate were computed. The initial coarse-grained structure on the surface was refined, of which the size is about 70 nm. The TEM observations indicate that dislocation wall, sub-dislocation wall, twin and twin boundary steps were achieved on the surface of material after HCPEB irradiation. Furthermore, huge numbers of vacancy defect clusters including dislocation loops, stacking fault tetrahedras (SFT) and voids were also formed within the sublayer of the irradiated surface. The fact that supersaturation vacancies migrated towards the surface along grain boundaries and/or various dislocation configurations led to the formation of microporous feature on the irradiated surface.     

基于无人机的建筑立面裂缝检测

为准确、全面检测建筑立面裂缝,提出一种基于无人机近景摄影测量和计算机视觉的裂缝检测方法。首先根据建筑立面高度和结构特点,结合无人机自动航线飞行和手动贴近飞行进行影像数据采集,并基于此构建针对建筑裂缝的影像数据集。其次,借助基于深度残差神经网络的识别方法对建筑立面进行识别以提取出墙面上的裂缝区域。进一步地,构建基于坐标转换的无人机影像像素解析度计算方法,建立数字图像处理驱动的裂缝几何信息提取方法。最后,以上海市同济大学四平路校区内中德楼为例进行裂缝检测实验,实验结果表明利用本文方法可对高层建筑立面进行影像采集,重建后影像精度达到亚毫米级(0.5mm/pixel),并能在此基础上准确地进行建筑立面裂缝检测(精确率99%),同时可以合理判断裂缝的长度,检测成果可以反映裂缝在墙面上的分布和走向,为后续建筑立面裂缝病害类型的判断提供支撑,具有良好的应用价值。     

新型自复位变摩擦阻尼器力学性能研究

本文提出了一种兼具自复位、变摩擦和高耗能于一体的新型自复位变摩擦阻尼器,其自复位功能由碟簧实现,变摩擦机构提供耗能作用.基于该阻尼器独特的构造形式推导了其恢复力模型,并利用MATLAB中的Simulink工具箱建立其动态仿真模型,并引用他人试验数据对比验证.结果表明:所建立的动态仿真模型能够较为准确地计算出阻尼器的恢复...     

Experimental Study of Sub-critical Velocity in Longitudinally Ventilated Tunnels

Fire Technology - The critical velocity is the minimum velocity blowing towards the fire seat in tunnels to reduce the length of back-layering to zero, and it has been extensively adopted in the...     

Deep Convolutional Network with Pixel-Aware Attention for Smoke Recognition

Fire Technology - Deep convolutional networks have significantly improved the performance of smoke image recognition. However, the trained spatially-shared weights are applied to all pixels...     

An Improved Approach for Spatial and Temporal Individual Risk Assessment Considering Synergistic Effects of Multiple Fires Occurred Sequentially

Fire Technology - The current research on individual risks mainly focuses on the individual risk considering a single fire or the spatial distribution of multiple fires. It is more realistic to...     

飞机货舱典型可燃物燃烧火灾特征参数试验研究

摘 要：光电式火灾烟雾探测器是目前民用飞机货舱使用的主流火灾探测器,是保障飞机安全运行的关键,该探测器是基于烟雾颗粒对光的散射信号来进行报警的,但是由于空气中悬浮颗粒物（水蒸气或灰尘等）也具有散射效应,同样会触发火灾探测器发生报警,这种原理上的缺陷造成的误报情况时有发生。通过对飞机货舱典型燃烧物瓦楞纸片、榉木和棉绳的蓝光散射功率（表征为表面积浓度）、红外光散射功率（表征为体积浓度）和颗粒物索特平均粒径火灾特征参数进行分析,为复合探测器的研发优化火灾特征参数。结果表明：在常压条件下,蓝光和红外光对颗粒物浓度的变化有很好的响应,索特平均粒径可以区分火灾烟雾颗粒和非火灾烟雾颗粒,可以有效降低误报率。