Human-like autonomous car-following model with deep reinforcement learning

This study proposes a framework for human-like autonomous car-following planning based on deep reinforcement learning (deep RL). Historical driving data are fed into a simulation environment where an RL agent learns from trial and error interactions based on a reward function that signals how much the agent deviates from the empirical data. Through these interactions, an optimal policy, or car-following model that maps in a human-like way from speed, relative speed between a lead and following vehicle, and inter-vehicle spacing to acceleration of a following vehicle is finally obtained. The model can be continuously updated when more data are fed in. Two thousand car-following periods extracted from the 2015 Shanghai Naturalistic Driving Study were used to train the model and compare its performance with that of traditional and recent data-driven car-following models. As shown by this study’s results, a deep deterministic policy gradient car-following model that uses disparity between simulated and observed speed as the reward function and considers a reaction delay of 1 s, denoted as DDPGvRT, can reproduce human-like car-following behavior with higher accuracy than traditional and recent data-driven car-following models. Specifically, the DDPGvRT model has a spacing validation error of 18% and speed validation error of 5%, which are less than those of other models, including the intelligent driver model, models based on locally weighted regression, and conventional neural network-based models. Moreover, the DDPGvRT demonstrates good capability of generalization to various driving situations and can adapt to different drivers by continuously learning. This study demonstrates that reinforcement learning methodology can offer insight into driver behavior and can contribute to the development of human-like autonomous driving algorithms and traffic-flow models.     

Back-Calculation of Ground Stress Field and Rockburst Analysis for Wunvfeng Tunnel北大核心CSCD

Studying the in-situ stress distribution at a tunnel site is very important to determine surrounding rock characteristics, the engineering design and the construction scheme. By using the multiple linear regression method based on the least square algorithm, the initial geostress field is analyzed and the corresponding regression coefficients are obtained. The ground stress obtained from the proposed back analysis is reasonable and can meet the demands of the engineering applications. From the rockburst risk level distribution diagram, it is speculated that the Wunvfeng tunnel is in the high filed stress area. Field monitoring should be strengthened and emergency plans should be made to cope with the rockburst risks during the construction process. © 2018, Editorial Office of "Modern Tunnelling Technology". All right reserved.     

低排量下并联甩泵分析与对策

乌鄯原油支干线乌鲁木齐输油首站低排量下给油泵特殊工况并联运行极易出现甩泵故障,给输油生产带来困扰和设备危害。通过对几年来甩泵分析,找出原因,制定对策（提高来油温度、缩小过滤器差压报警值、流量计并行、降低外输排量等）,减少了并联运行次数,避免了并联甩泵现象的再次发生,为密闭输油生产中合理配泵提供参考。     

高温烟气膨胀节爆裂原因分析与对策

文中从失效波纹管的化学成份、力学性能、显微组织、裂纹与断口及腐蚀产物等方面进行检测,并结合波纹管工作状况对波纹管进行应力计算及检测结果分析,探讨高温烟气膨胀节爆裂失效的原因及防护措施。     

大体积承台混凝土施工

文章以新寨河特大桥主墩承台施工为例,介绍了大体积混凝土施工的质量控制要点、施工工艺及质量保障措施,并重点讲述了大体积混凝土在施工中,如何预防因温度应力而引起混凝土开裂的温度监测措施。     

屏山渡湘江大桥设计

屏山渡湘江大桥采用预应力混凝土连续箱梁,跨径组合为45 270 45m,桥墩高度仅7·4m。文章简要介绍了该桥的总体方案设计、上部结构设计、计算要点以及底板上弯钢束布置,为类似桥梁的设计提供借鉴。     

低温热水地板辐射采暖系统水力平衡分析

分析了地暖系统中重力循环作用压力对共用立管的影响,运用基尔霍夫定律分析了地暖系统中下分异程式双管系统和下分同程式双管系统的水力平衡状况,得出了在没有平衡调节装置的情况下,下分异程式双管系统更适用于地暖系统的结论,通过计算确定了地暖系统共用立管水力平衡的影响因素和平均比摩阻的范围.     

The problem of cold starts: A closer look at mobile source emissions levels

While the phenomenon of excess vehicle emissions from cold-start conditions is well known, the magnitude and duration of this phenomenon is often unclear due to the complex chemical processes involved and uncertainty in the literature on this subject. This paper synthesizes key findings regarding the influence of ambient and engine temperatures on light-duty vehicle (LDV) emissions. Existing literature, as well as analytical tools like the U.S. Environmental Protection Agency’s Motor Vehicle Emission Simulator (MOVES), indicate that while total vehicle emissions have dropped significantly in recent years, those associated with cold starts can still constitute up to 80% for some pollutant species. Starting emissions are consistently found to make up a high proportion of total transportation-related methane (CH₄), nitrous oxide (N₂O), and volatile organic compounds (VOCs). After 3–4 min of vehicle operation, both the engine coolant and the catalytic converter have generally warmed, and emissions are significantly lower. This effect lasts roughly 45 min after the engine is shut off, though the cooling rate depends greatly on the emission species and ambient temperature. Electrically (pre-)heated catalysts, using the bigger batteries available on hybrid drivetrains and plug-in vehicles, may be the most cost-effective technology to bring down a sizable share of mobile source emissions. Trip chaining (to keep engines warm) and shifting to non-motorized modes for shorter trips, where the cold start can dominate emissions, are also valuable tactics.     

隧道开挖中的应力旋转和裂隙塑性变形问题研究

通过对隧道开挖过程进行三维数值模拟,分析了隧道开挖过程中主应力方向的旋转变化,研究了主应力旋转与隧道裂隙塑性变形的关系,验证了"考虑应力主轴旋转的广义塑性原理"的正确性,并将数值模拟结果和相关试验成果进行了对比,得出了一致性的结论。研究结果表明,应力旋转基本在隧道开挖前后10 m左右的距离内完成,隧道开挖过程中的应力旋转可以引起裂隙塑性变形的突变,且对竖向和斜向裂隙的影响较大;可以根据应力旋转的变化规律合理地选择支护时机,依据应力旋转对裂隙的影响对隧道的重点部位进行超前加固。     

电动汽车动力电池低温性能研究与应用

电动汽车锂离子动力电池在低温条件下工作时,电池的内阻明显升高、功率和能量急剧下降,导致整车低温下动力性能不足、续驶里程不足、充电受限等问题。通过试验进行了动力电池的低温特性研究,提出了适合整车低温工作的解决方案,通过实车验证,达到了很好的效果。