微信息大数据粗糙集的近似约简

为了应对微信息舆情数据的格式复杂、价值稀疏和收集困难等大数据处理技术难题,基于隐含语义分析和粗糙集近似约简理论,设计微信息的数据区间值集和近似匹配分类算法.在不影响数据主要关联关系的原则下,提炼核心属性、消减次要属性,实现一种微信息异常主题倾向的发现方法.结果表明,该近似约简算法能在完成微信息兴趣倾向主题分类的前提下,将数据集属性大幅度缩减,提高微信息的信息挖掘效率,为微信息大数据舆情处理工作提供了新的思路和案例.     

环形刀高度变化刀路轨迹规划方法

为提高汽车覆盖件模具加工效率,将环形刀应用于汽车模具小曲率面的半精及精加工过程.通过对三坐标铣削小曲率面的刀具切触几何分析,确定环形刀最佳进给方向应沿工件高度变化方向,此时可使有效切削带宽达到最大.基于此结论,分别对两类特殊型面进行了刀路轨迹规划.对于单一方向曲率变化面采用垂直截面法,截面方向沿最大主曲率方向,实验表明:垂直截面法与等高线法相比可提高加工效率25%,并且有较好的表面质量;对于圆鼓面采用放射精加工方法,仿真结果表明:此方法与三维偏置方法相比可提高加工效率22.64%,并且避免了周边欠切.     

Combined damaged elasticity and creep modeling of ceramics with wedge splitting tests

During the crack propagation in common refractory ceramics at high temperatures, creep may occur in the wake of a process zone and in front of a crack tip. To account for this phenomenon, an integrated material constitutive model was developed by combining the mechanical behavior following isotropic damaged elasticity concept and Norton-Bailey creep. The post peak fracture behavior followed the bilinear softening law and a simple criterion was defined to consider the creep asymmetricity in uniaxial tension and compression. The material constitutive model was applied to inversely identify mode I fracture parameters with wedge splitting tests of an alumina spinel material at 1200 °C. It showed that the mean ratio of the nominal notch tensile strength to the actual tensile strength was 1.93 and the mean pure fracture energy was 297.6 N/m. In addition, the creep contributed 12.9% on average into the total fracture energy.     

On the calibration of cohesive parameters for refractories from notch opening displacements in wedge splitting tests

Cohesive elements are commonly used to describe crack propagation in heterogeneous materials with toughening mechanisms. This work aims to provide a guideline on how these fracture parameters can be calibrated using notch opening displacements (NODs) measured via digital image correlation and force data from wedge splitting tests (WSTs). Weighted finite element model updating was applied to calibrate material and boundary condition parameters in the same framework. The influence of each parameter on force and NOD data are given together with uncertainties for the calibrated parameters. Numerical results were in very good agreement in terms of splitting force, NOD, displacement and gray level residual fields. It is shown that images obtained during WSTs focusing on the crack path (i.e., hiding the loading region) can be used to drive numerical simulations and obtain cohesive parameters.     

两线切分轧制钢筋头部不齐问题的解决

针对安钢260机组小规格棒材实施两线切分轧制后，两线轧件在冷床上头部不易对齐问题进行了分析，认为其产生原因是切分后的2根轧件存在尺寸差，为此提出了具体的改进措施和调整办法。改进后，定尺率、成材率均提高0．5％以上。     

A finite volume–finite difference method with a stiff ordinary differential equation solver for advection–diffusion–reaction equation

A model for atmospheric pollutant transport is proposed considering an advection–diffusion–reaction equation. A splitting method is used to decouple the advection, diffusion and reaction parts. A scheme based on finite volume, finite difference and backward differentiation formula is used for solving an atmospheric transport-chemistry problem.     

基于区间多目标量子文化算法的无线传感器网络覆盖能效优化

针对不确定环境下无线传感器网络覆盖能效优化问题,提出一种传感器节点区间感知模型;进而考虑网络覆盖率和节点冗余率,将其转化为区间多目标优化问题。基于一种新型区间个体依可能度占优关系,提出区间多目标量子文化算法,根据区间占优个体信息提取隐含知识,用于指导量子个体更新及进化个体变异与选择。不同环境下的仿真结果表明：基于所提出算法获得的Pareto解具有更好的收敛性、分布性和延展性;相应的无线传感器网络布局更合理。     

基于概率分布距离的多响应模型确认度量

针对模型确认中的确认度量问题,构造实验观测数据经验概率分布的置信包络。通过计算其与模型响应概率分布之间距离的上/下确界,给出基于概率分布距离确认度量的置信区间。通过构造与实验观测数据有关的协方差矩阵,给出基于概率分布距离的多响应模型确认度量及其置信区间的求解方式。该度量利用了模型输出与实验观测的完整概率分布信息,并且考虑了各模型响应间的相关性。算例仿真结果表明其确认错误率低于现有的其他两种确认度量。     

累积量测序列的区间云变换及识别

目标多属性序列类数据不能直接与数据库中的区间类数据融合识别, 对此, 提出一种基于云变换的序列-区间异类数据识别算法. 对目标多属性序列数据进行频数计算形成频率分布函数, 并进行虚警检测, 实施云变换形成云簇, 提取云簇特征, 再根据3 En 准则形成云滴区间, 实现了序列型数据的区间化表示. 进一步, 利用一种区间多属性识别判定准则进行识别判定, 得到识别结果, 解决了序列-区间异类数据的识别问题. 仿真实验结果验证了该算法对序列-区间异类数据识别的有效性.

     

基于簇半径差异化和节点能量区间的分簇算法

提出一种基于簇半径差异化和节点能量区间的分簇算法——CRDNE算法.该算法用于解决电子导游系统能量均衡问题,从而尽可能的延长网络寿命.算法首先利用路由跳数将整个网络划分热点和非热点区域,并根据不同区域的通信开销不同的特点,定义了不同的区域簇半径.然后,根据区域簇半径和节点剩余能量所处的能量区间,分布式地选取簇首.最后,将CRDNE算法与LDCR、Leach-V算法在Matlab平台进行多个方面仿真比较,进一步证明了本文算法的优越性.