深基坑支撑体系参数优化分析

文章在总结已有研究成果的基础上,分析影响深基坑的支撑体系参数,并根据优化分析的原则,提出模糊神经网络的优化分析方法,确定了深基坑支撑体系优化分析步骤。     

面向地下停车场的轻量级目标检测算法研究

基于深度学习的目标检测算法能够取得良好的检测速度离不开高性能GPU硬件设备的支持。然而,在智能车中搭载高性能、高功耗、大尺寸的硬件设备与汽车的长续航理念不符。因此,文章以YOLOv3目标检测算法为基线模型进行改进,提出轻量化的目标检测模型Mobile-YOLO,并在采集制作的地下停车场数据集中进行训练测试。实验结果表明,提出了Mobile-YOLO相较于YOLOv3,在平均精度均值略微提升的情况下,检测速度提升了47.1%。在移动端平台TX2上每秒能够检测31张图像。     

基于3DCNN-DNN 的高空视频交通状态预测

为使用高空视频识别和预测道路交通状态，提出基于三维卷积神经网络-深度神经网络(3DCNN-DNN)的交通状态预测方法. 将道路切分为D 个路段，每个路段视频片段时长 m s，基于典型3DCNN结构C3D识别路段视频片段交通状态；建立道路? 个历史时段、D 个路段的交通状态矩阵Φ ，将道路交通状态预测问题转化为以Φ 为输入，有限交通状态为输出的分类问题，构建基于DNN的短时交通状态预测模型原型；建立交通视频数据集，对DNN预测模型原型的隐藏层数量、神经元数量及训练批大小进行测试优化，提出有4 隐藏层，各层神经元数量为64/128/128/64，训练批大小为64 的优化模型DNN*.测试结果表明：C3D视频交通态识别平均F1 值为95.71%，DNN*道路交通状态预测准确率为91.18%，比DNN线性分类、KMeans 、KNN、SVM和线性分类分别高6.86%、57.85%、62.26%、26.47%、43.14%；C3D能提供准确的交通状态矩阵，3DCNN-DNN可有效识别和预测道路视频交通状态.     

Prediction of design loads for deep water subsea lifting operations based on non-stationary time response

Two methodologies for the prediction of the design loads of deep water subsea lifting operations crossing resonance zones are presented. These methodologies are applicable to models that consider the influence of the payout speed on the dynamics of the operation, leading to a non-stationary time series for the dynamic forces on the system. The first model is based on running several random simulations of the same scenario and using these simulations as a sample in which statistical parameters are inferred. The second model uses a weighted least squares approach to predict a normalizing function that is used to evaluate the statistical parameters of the response. Both models are tested by considering the installation of a typical manifold in the Pre-Salt fields, in Brazil, and are able to predict the general form of the envelope of forces on the cable for various sea states and payout speeds. The models also provided similar results for the availability of the vessel after evaluating the weather window for this operation. Finally, the advantages of using the weighted least squares approach in comparison to the direct method are discussed, since it may considerably reduce the total number of simulations required to perform a real operation assessment, especially during preliminary design phases.     

A car-following model considering asymmetric driving behavior based on long short-term memory neural networks

Asymmetric driving behavior is a critical characteristic of human driving behaviors and has a significant impact on traffic flow. In consideration of the asymmetric driving behavior, this paper proposes a long short-term memory (LSTM) neural networks (NN) based car-following (CF) model to capture realistic traffic flow characteristics by incorporating the driving memory. The NGSIM data are used to calibrate and validate the proposed CF model. Meanwhile, three characteristics closely related to the asymmetric driving behavior are investigated: hysteresis, discrete driving, and intensity difference. The simulation results show the good performance of the proposed CF model on reproducing realistic traffic flow features. Moreover, to further demonstrate the superiority of the proposed CF model, two other CF models including recurrent neural network based CF model and asymmetric full velocity difference model, are compared with LSTM-NN model. The results reveal that LSTM-NN model can capture the asymmetric driving behavior well and outperforms other models.     

Deep subterranean railway system: Acceptability assessment of the public discourse in the Seoul Metropolitan Area of South Korea

The objective of this study is to analyze the public acceptability of deep subterranean railway systems, which will be constructed in the space 40 m below ground level and will be operated at twice the speed of the existing subway system. Although such railway systems have been feasible in terms of construction technologies and economics, public acceptability must be considered for the successful introduction of such a new public infrastructure. Therefore, to perform the analysis of public acceptability, a telephone-based survey was conducted for residents in the vicinity of the planned the deep subterranean railway systems. As a result, about 70% of the respondents answered that they took a neutral or opposing attitude to introducing the deep subterranean railway systems. Awareness of the deep subterranean railway systems has a positive impact on its acceptability. In addition, the acceptability is found to show a negative relationship with environment and inconvenience factors. Based on the analysis results, an affective approach through soft measures such as awareness campaigns and advertisements is recommended to effectively address and mitigate the concerns and issues raised by the public.     

Deep water mass characteristics and interannual variability in the North and Central Aegean Sea

The characteristics and interannual variability of the deep water masses in the North and Central Aegean Sea are being investigated through the data sets of the Hellenic Navy Hydrographic Service (HNHS) and the MEDATLAS 1997 project. In the period between 1987 and 1993, the densest deep water in the Mediterranean has been produced in the Aegean Sea (with σ_θ densities reaching up to 29.6 kg/m³), contributing to what has been called the Eastern Mediterranean Transient. The examination of time series of mean integrated values of θ, S and σ_θ below the depth of 500 dbar reveals the significant deep water density increase after 1987 in all of the deep basins in the area. Data suggest that the density increase of 1987–1988 is mainly attributed to a temperature drop, while in 1993, an even more intense density increase is observed, characterized this time by an abrupt salinity increase. We assume that the increased salinity necessary to produce deep water masses with the observed characteristics was not locally produced but rather advected from the Levantine through the South Aegean. After 1993, no new deep water formation episodes have been observed. A series of Θ–S diagrams derived from HNHS CTD casts covering the period between 1993 and 2000, depict the different characteristics of the deep water masses in the area. As 1993 marks the end of the formation period, observed differences between basins in that year must be attributed to different deep water formation sites. Thereafter, the stagnating deep water in the North and Central Aegean basins has been slowly gaining buoyancy by losing salt and gaining heat. The rate at which this phenomenon takes place varies between different deep basins. It is suggested that these variations are linked to the different volumes of each basin as well as to the general circulation features of the Aegean Sea.     

基于PointNet++的船体分段合拢面智能识别方法

船体分段合拢面的精度检测是分段总组合拢过程中的重要环节。在船体分段合拢面的精度检测方面,三维扫描仪相对全站仪有着巨大优势。然而,三维扫描仪在扫描过程中会记录很多与合拢面无关的点,因此本文对三维扫描仪扫描出的点云数据,进行合拢面的智能识别。通过采用深度学习理论对PointNet++点云网络进行适合本文的改进,使用CAD模型导出的点云数据构建有标注的船体分段点云数据集,进而使用Adam优化算法对网络进行优化训练。最终网络模型对分段合拢面的识别在验证集上获得精确率73%,召回率90%的效果。     

Hydrological changes in the Mediterranean Sea in relation to changes in the freshwater budget: A numerical modelling study

The response of the Mediterranean Sea and the various sub-basins to changes in the freshwater budget are investigated in a process-oriented study, using the POM model. The model is first integrated using values of the Nile and Ebro rivers runoff, as well as of the Dardanelles freshwater input, typical of the fifties. The model reaches a steady state representative of that existing in the Mediterranean prior to the major damming period after 90 years of integration. Then the model is integrated using the reduced river runoff values typical of the after-damming period. The additional impact of decadal scale trends in the precipitation rate as well as of intense surface cooling periods/events on the thermohaline circulation during the last 40 years were also examined. The model results show that the dramatic reduction of the Nile freshwater input and to a lesser extent the reduction of the freshwater input from the Dardanelles Straits induced a large increase in the sea surface salinity in the Aegean and Levantine basins in the late sixties/early seventies, in agreement with observations. Furthermore, the Ebro runoff reduction during the same period further enhanced the salinity increase in the Levantine basin as higher salinity surface waters of the western basin reached the eastern basin via the Atlantic Water circulation. This saltier surface layer in the vicinity of the Rhodes Gyre favoured the preconditioning for the formation of the Levantine Intermediate Water, resulting in about 40% increase of its formation rate. This in turn resulted in the production of saltier and larger amounts of deep waters in the various deep-water formation sites. According to the model, the river damming and decreased precipitation since the eighties explain about 95% of the observed salinity increase in the Western Mediterranean Deep Water over the last 40 years. The major contributor to this increase was proved to be the Nile damming. The salt increase in the surface layer is proved to be insufficient to produce alone the two climatic transient events in the deep waters of the Eastern Mediterranean in the late sixties and early nineties, respectively. Surface cooling was found to be important, resulting in large deep water formation and thus allowing the propagation of the increased surface salinity signal to the deep layers. However, model results demonstrate that the river damming played an important role in the long-term salt preconditioning of the surface/intermediate layers, thus contributing in triggering the two events.     

地铁换乘站深基坑内支撑体系的优化设计与分析

采用土体卸载条件下的Hardening-Soil有限元模型,研究苏州某地铁换乘站深基坑3种支撑方案对支护结构内力和基坑变形的影响。结果表明:与初始方案比较,方案A(去掉第4道支撑)的墙身最大侧移增大了35.28%,最大正负剪力分别增加41.87%和71.11%,最大正负弯矩分别增加145.32%和45.39%,最后一道支撑轴力达到1018kN,比初始方案增加了近1倍,对基坑土体变形影响增长近10%;方案B(第3、4道支撑合并)的墙身最大侧移增大了10.49%,最大正负剪力则增加6.89%和33.97%,最大正负弯矩增加34.46%和10.97%,合并后的第3、4道支撑轴力降低了15%,对基坑土体变形影响较小。方案B既能保证基坑安全和环境影响的要求,又能降低造价。