基于模糊模式识别的城市道路安全状况综合评价

针对影响城市道路安全状况的因素众多且具有模糊性的特点,采用模糊模式识别与模糊综合评价相结合的评价模型来评价城市道路安全状况.建立了城市道路安全状况综合评价指标体系,应用层次分析法确定了指标权重,应用集值统计法对指标特征值进行了量化.通过对大连市黄河路的评价实例证明,该模型实用性较强,可用于完善道路安全设计及辅助道路安全管理.     

铁道车辆运行安全评判的轮对爬轨脱轨准则

为了掌握车辆爬轨脱轨机理及主要影响因素,分析了轮对的三维空间受力,推导了轮轨横向力和垂向力比值的一般表示式,假设车轮在达到最大轮缘接触角时为脱轨的临界状态,并认为这时轮轨间出现完全摩擦滑动,导出了不考虑轮对摇头角的临界脱轨判别的二维准则与考虑轮对摇头角和轮轨蠕滑率效应的三维脱轨判别准则,给出了轮轴脱轨系数的定义,采用轮轴脱轨系数和轮重减载率进行脱轨的判别。仿真计算结果表明:二维脱轨判别准则与三维准则相比偏于保守;摇头角越小甚至变负,越有利于防止脱轨,摇头角越大,三维准则的临界脱轨曲线越接近于二维准则的;减小轮轨摩擦系数与增大轮缘角均有利于防止脱轨的发生。     

Corrosion rate measurements in steel sheet pile walls in a marine environment

Corrosion of steel structures in the marine environment is a major problem. The deterioration of this kind of structures is costly and difficult to predict both when designing new structures and when estimating the remaining service life time for existing structures. The aim of this investigation was to find indicative values for the corrosion rate of steel sheet piles on the Swedish west coast. Such corrosion rates (mm/year) can be used both when designing new structures by oversizing the steel thickness and when estimating the bearing capacity of existing sheet pile structures. Earlier investigations on the corrosion rates along the Swedish east coast – with salinity from about 0.2% to 0.8% – are still used today as guidelines for the corrosion rate of all steel structures in the Swedish maritime environment even though the salinity on the west coast can be as high as 3.0%.Steel sheet pile wharfs located in the port of Halmstad on the Swedish west coast were inspected by ultrasonic measurements. Three wharf structures with a total length of about 700 m were inspected. None of the inspected wharfs had or have had cathodic protection. The thickness measurements of the steel sheet pile structures were performed by divers.The age of the three inspected sheet pile structures ranged from 36 to 51 years. The dimensions of the original sheet pile sections are known. One of the quay structures is located along a river. The salinity at all wharfs varied from low values at the surface to approx. 2% at the bottom (also in the river outflow).The measured average corrosion rates were in the same order as the design values in the European code. However, the results indicate increased corrosion rates about 1 m below the mean water surface and at the level of the propellers from the ships berthing the most frequented of the inspected wharfs, 3–6 m below water surface.The tolerances of steel sheet thicknesses – usually in the order of ±6% – are often neglected when investigating the remaining thickness in steel sheet piles. A simple calculation model shows that the sheet pile must be almost 50 years of age before an accurate estimation on the corrosion rate can be made, considering the tolerances, if the true original sheet pile thickness is not known.     

沥青路面运营初期平整度竣工验收标准研究

采用数理统计原理,运用SPSS软件对陕西省7条高速公路交-竣工平整度数据进行分析,得出了陕南、关中和陕北地区路面平整度衰变幅度均值分别为0.39m/km、0.28m/km和0.24m/km,这也将作为计算竣工验收推荐值的依据;再基于交-竣工平整度变化规律,建立了平整度竣工验收推荐值的计算公式:IRIju=IRIji+IRIs,可为今后平整度竣工验收标准提供有效的理论依据。     

数控铣床加工低碳钢零件的进给速度优化

论述了数控铣床加工低碳钢零件时进给速度的优化问题．在基于加工效率（可等效转化为加工工时）和加工成本目标函数的基础上，对这两个目标函数运用多目标函数理论进行综合，得到了多目标综合函数．通过计算，并综合表面粗糙度，进而得到进给速度的优化范围，并对不同转速下的进给速度给出了相应的优化参数，目的更接近生产实际，使技术人员不仅了解影响生产效率和成本的因素，也可在实际生产的应用中更加快捷和方便．     

多尺度图像边缘检测技术在车辆识别中的应用

指出了图像边缘检测的定义及现有算法的不足,并对Canny边缘检测离散准则及该准则下的最优线性滤波器进行了论述。根据边缘检测离散准则,利用数值方法求出该准则下边缘检测的最优线性滤波器及对应的平滑算子,将其与Canny边缘检测技术相结合进行边缘检测,检测出不同尺度下多幅边缘图像,并根据小尺度下图像边缘细节信息丰富、边缘定位精度高,大尺度下图像边缘稳定、抗噪性好等特点,将检测到的多尺度边缘进行融合,得到抗噪性好的精确的单像素宽边缘;同时将该方法应用于汽车特征提取,检测到完整连续的汽车边缘图像,为车型识别提供条件。试验结果表明:该方法不仅能准确检测出图像边缘,而且能有效地抑制噪声,是边缘检测的实用方法。     

小样本条件下混凝土强度代表值的确定

针对小样本混凝土强度代表值确定时不能利用数理统计模型的问题,引入了信息扩散原理,提出了小样本条件下混凝土强度代表值确定的实用计算模型。2个混凝土强度检测案例的计算结果表明,利用《公路工程质量检验评定标准》(JTJ071-98)进行灌注桩芯样混凝土强度代表值确定,得到的代表值保证率在95%左右波动,具有不确定性;利用《超声回弹综合法检测混凝土强度技术规程》(CECS 02:88)进行单一构件混凝土强度推定,推定值的保证率小于95%,对结构的可靠性鉴定而言是偏于不安全的,应引起试验检测人员的充分重视。     

整车试制阶段的电器故障诊断方法

以整车装车与跑车验证案例为背景，讨论汽车开发过程中汽车电器故障诊断的方法，阐述如何应用零部件置换原则来分析故障的重要性等级（significance）。     

基于子波能量和神经网络分类器的机动车车型识别

为了使交通管理系统能进行可靠的机动车分类,研究了轿车、轻型越野车和货车3种机动车目标的声信号,提出了一种采用子波分解后不同尺度上声信号能量作为特征向量的特征提取算法,并设计了kNN(k近邻)分类器和改进BP神经网络分类器用于目标分类。目标识别和分类试验结果表明:所提出的特征提取算法能够很好地体现不同类型目标之间的差异,提取的特征向量稳健;设计的改进BP神经网络分类器的分类精度可达92.6%,且分类效果优于kNN分类器。