湍流模型和网格分布对水下航行体回转水动力数值计算的影响

针对水下航行体安全操纵设计的关键环节水动力预报,用FLUENT求解不可压缩RANS方程,采用定常旋转运动坐标系,运用相对运动理论和运动叠加原理,计算RNG k-ε,可实现k-ε和SST k-ε等3种湍流模型以及不同网格分布对水下航行体做单平面回转运动所受垂向力和纵倾力矩,并与试验结果比较.结果表明在Re=11.7×106条件下,垂向力和纵倾力矩预报精度较高.该数值预报方法可行、有效,具有较好的工程实用价值.     

地铁车辆轴箱吊耳动态分析和优化

针对某型地铁车辆转向架在被吊起时起保护一系悬挂系统作用的轴箱吊耳在车辆运营中多次发生断裂的问题,用HyperWorks建立吊耳的有限元模型,以试验数据作为输入分析模态频率响应,通过仿真结果与试验结果的对比调整模型,保证模型仿真的准确性.综合分析仿真结果和试验结果,找出引起吊耳模态频率响应过大的原因,利用OptiStruct进行结构拓扑优化,根据优化结果提出4种新方案,并分别对比各方案的静强度、模态频率、模态频率响应和体积等,确定最佳设计方案.     

基于车型识别的BP算法Matlab实现

车型识别具有广阔的应用前景,BP神经网络在车型识别中能够提高车型的识别率。在任何车型大致都可以抽象成一个"工"字型情况下,提取其中的顶长比、前后比和顶高比这三项相对参数作为BP神经网络的输入参数。采用三层3-8-3的BP神经网络,并用14对输入参数离线训练,再用4对新数据进行检验,均得到了预想的期望值。     

车联网中基于轨迹预测的无人机动态协同优化覆盖算法

针对城市车联网中出现的基站覆盖空洞及局部流量过载等问题,提出了一种基于车辆轨迹预测信息的动态预部署方案。首先,为了训练得到统一的seq2seq-GRU轨迹预测模型,多个携带边缘计算服务器的无人机在分布式联邦学习与区块链的架构下,去除中心聚合节点,采取改进的Raft算法,在每轮训练中根据贡献数据量的大小,选举得到节点来完成参数聚合及模型更新任务。其次,基于模型预测结果,提出了一种改进的虚拟力向导部署算法,通过各虚拟力来引导无人机进行动态地部署以提升车辆的接入率及通信质量。仿真结果表明,提出的训练架构能够加速模型的训练,部署算法在提升车辆接入率的同时提升了车辆与无人机之间的通信质量。     

基于智能垃圾桶的垃圾分类动态收运路径优化问题研究

针对城市生活垃圾分类收运过程中存在的环境二次污染和垃圾产生量不确定性等问题,提出了一种基于智能垃圾桶的动态收运车辆路径优化方法。建立以最小化碳排放成本、燃油消耗成本、固定成本和车辆延迟到达惩罚成本为目标的动态车辆路径优化模型。采用滚动时域的方式将动态问题转换为一系列静态问题,并设计两阶段算法进行求解。首先采用粒子群算法对收运车辆路径进行规划,而后在每个时域末,综合考虑待清运垃圾桶的位置和垃圾量、垃圾收运车辆的位置和装载量以动态调整现有车辆路径。研究结果表明,相较于传统的静态收运方案,动态垃圾收运方案能够在降低车辆运输成本和碳排放成本的同时,显著降低由于清运不及时造成的环境二次污染的风险。     

Seasonal variation in light penetration and subsurface chlorophyll-α in southern Lake Michigan observed by a glider

The Cooperative Institute for Great Lakes Research (CIGLR) in collaboration with the Great Lakes Observing System and National Oceanic and Atmospheric Administration, Great Lakes Environmental Research Laboratory (NOAA GLERL) deployed an autonomous underwater glider in southern Lake Michigan several times per year between 2012 and 2019 to collect offshore (>30 m depth) limnological measurements, including temperature, photosynthetically active radiation (beginning during 2015), and chlorophyll-α fluorescence. From these data, we calculated mixed layer depth, several measures of light penetration (diffuse attenuation coefficient, first optical depth, euphotic zone depth), and depth of the subsurface chlorophyll-α maxima. During summer, mean offshore mixed layer depth was typically 10–15 m, K_d for PAR was 0.1–0.17 m^?1, first optical depth was 6–9 m, euphotic zone depth was 35–40 m, and depth of subsurface chlorophyll-α maxima was 30–35 m. We also observed substantial spatial and temporal variation in these values across the basin and within and among seasons. Glider-based observations provide a wider horizontal and vertical perspective than other methods (e.g., ship- and satellite-based observations, buoys, and fixed moorings), and are therefore a valuable, complementary tool for Great Lakes limnology. The set of observations reported here provide seasonal and basin-scale information that may help to identify anomalies useful for future glider-assisted investigation into the role of biophysical processes in Great Lakes limnology and ecology.     

Conceptual digital twin modeling based on an integrated five-dimensional framework and TRIZ function model

Digital twin represents a fusion of the informational and physical domains, to bridge the material and virtual worlds. Existing methods of digital twin modeling are mainly based on modular representation, which limits guidance of the modeling process. Such methods do not consider the components or operational rules of the digital twin in detail, thereby preventing designers from applying these methods in their fields. With the increasing application of digital twin to various engineering fields, an effective method of modeling a multi-dimensional digital twin at the conceptual level is required. To such an end, this paper presents a method for the conceptual modeling of a digital twin based on a five-dimensional digital twin framework to represent the complex relationship between digital twin objects and their attributes. The proposed method was used to model the digital twin of an intelligent vehicle at the concept level.     

矿用柴油机车中新型无源液位传感器设计

针对目前矿用柴油机车液位传感器无法动态调整水箱剩余水位的报警高度、可靠性较差等问题,提出一种新型无源液位传感器的设计方案。该液位传感器采用干簧管和电阻阵列等无源器件,实现了对水箱水位高度的连续监测,并可设置报警深度,可靠性和耐高温性能较好。     

基于凌动处理器ATOM-510的车载信息终端设计

随着人们对汽车安全性、舒适性要求的不断提高,汽车电子控制功能逐渐增多,针对使用传统单片机控制方法功能单一、灵活性差等缺点,提出基于intel凌动处理器ATOM-510平台和嵌入式操作系统ubuntu的车载信息终端解决方案;实现了GPS(全球卫星定位系统)定位信息采集与解析、GPRS(通用无线分组业务)数据传输以及多媒体娱乐等功能,并基于嵌入式GUI(图形用户界面)Qt设计了良好的人机交互界面;应用结果表明,该方案运行稳定可靠、功能丰富、操作简便、易于扩展。     

模糊自适应卡尔曼滤波算法在航位推算系统中的应用

针对实际的应用中车载航位推算系统的模型参数、噪声的统计特性不确定性,影响估计效果,提出了车载航位推算的模糊自适应卡尔曼滤波模型及其滤波算法;该方法通过监视理论残差与实际残差的比值是否接近1,应用模糊推理系统不断地调整量测噪声协方差的加权,对自适应卡尔曼滤波的量测噪声协方差进行递推修正,通过该算法来抑制噪声对精度的影响,进而提高系统的导航精度;仿真结果表明,这种算法能够有效地提高系统的精度,是一种比较理想的车载DR导航滤波方法。