Non-probabilistic reliability research on multi-variable hydraulic turbine blade model

Based on the interval mathematics and possibility theory, the variables existing in hydraulic turbine blade are described. Considering the multi-failure mode in turbine blade, multi-variable model is established to meet the actual situation. Thus, non-probabilistic reliability index is presented by comparing with the output range and the given range.     

基于减法聚类模糊神经网络高速公路混沌控制

研究了基于减法聚类的高速公路混沌系统模糊神经网络控制方法.提出通过数据挖掘技术建立交通流混沌控制器知识库的思想,设计了以密度、上游流量和最大李亚普诺夫指数作为输入,红灯时间作为输出的T-S模糊神经网络混沌控制器.采用减法聚类确定控制器结构提取模糊规则、控制器初始参数;应用模糊神经网络方法对控制器参数进行优化;结合遗传算法对聚类半径进行优化.仿真实验分析了该控制方法的控制效果,证明了该混沌控制方法的有效性.     

SPSS软件在隧道围岩监控量测数据分析中的应用

为了保证隧道工程的安全施工,并为围岩稳定性评价提供较为可靠的数据预测,应用数据挖掘软件SPSS对研究区隧道监测断面的数据资料进行了分析得出:隧道围岩接触压力、位移及时间等监测值之间存在着显著的相关关系;通过对隧道围岩各项监测值进行的回归分析,得到了各监测值即围岩压力、位移与监测时间之间的非线性回归方程,在此基础上,对隧道围岩稳定性进行了评价。证实SPSS软件在分析隧道围岩监测数据方面具有较高的预测精度,可用来指导隧道的设计与施工。     

结合交通部公路设计新理念谈山区公路路线设计

路线设计是公路工程设计的基础和灵魂,路线设计的好坏直接影响着整个工程优劣和造价,山区复杂的地形、地质等条件更决定了其路线设计的复杂性。该文从交通部公路设计新理念理论体系着手,分析了山区公路路线设计的特点,结合具体设计实践,总结了新理念下山区公路路线设计的基本原则,并提出在山区公路路线设计中贯彻和落实新理念所需注重的问题。     

微型客车气动阻力性能研究及优化设计

综合利用计算流体力学数值仿真分析与风洞试验验证相结合的方法对某量产微型客车进行整车全细节空气动力学仿真分析和研究。对比分析前保险杠、后视镜、车窗等零部件对气动阻力的影响。研究表明,采用优化前保险杠并增加导流板、修改后视镜造型、车窗一体化设计等方案,可使该车气动阻力系数减少约6%。     

汽车制动尖叫的稳健性设计综述

制动尖叫是全球范围内的汽车制动器的技术难题,如何通过稳健性设计有效控制制动尖叫成为制动器的研发重点。在系统分析盘式制动器的制动盘、制动块、制动钳和保持架的关键结构参数因素对制动尖叫影响的基础上,对制动尖叫稳健性设计的研究现状与未来趋势进行了综述。同时,为了方便读者理解和开展相关研究,对稳健性设计方法的研究进展也进行了综述。     

有机涂层中性盐雾试验及其结果评价的探讨

简述了涂层中性盐雾试验的影响因素及结果评价方法,着重讲述了对变色、生锈、起泡的评价方法。     

Ant colony optimization for feature selection in software product lines

Software product lines （SPLs） are important software engineering techniques for creating a collection of similar software systems. Software products can be derived from SPLs quickly. The process of software product derivation can be modeled as feature selection optimization with resource constraints, which is a non- deterministic polynomial-time hard （NP-hard） problem. In this paper, we present an approach that using ant colony optimization to get an approximation solution of the problem in polynomial time. We evaluate our approach by comparing it to two important approximation techniques. One is filtered Cartesian flattening and modified heuristic （FCF＋M-HEU） algorithm, the other is genetic algorithm for optimized feature selection （GAFES）. The experimental results show that our approach performs 6% worse than FCF＋M-HEU with reducing much running time. Meanwhile, it performs 10% better than GAFES with taking more time.     

黄土地区公路路基压实技术研究

系统介绍了振动碾压、冲击碾压、强夯技术的压实机理,并对试验点各处治方案下湿陷性黄土路基的压实度、承载力、湿陷性进行了比较分析,提出了湿陷性黄土路基经济有效的施工处治方案。     

Market based framework for multiple AUVs cooperation

A“Market” based framework for multiple AUVs team is introduced in this paper. It is a distributed meta-level task allocation framwork. The formulation and the basic concepts of the “Market” such as “goods” and “price” are discussed first, then the basic algorithm of the “auction”. The loosely coupled v-MDTSP tasks are considered as an example of the task allocation mission. A multiple AUV team controller and a detailed algorithm are developed for such applications. The simulation results show that the controller has the advantages such as robustness and low complexity and it can achieve better optimization results than the classical central controller （ such as GA） in some tasks. And the comparison of two different local solvers also implies that we should get the reasonable task allocation even not using the high quality algorithm, which can considerably decrease the cooperation computation.