多维动态模糊数据的分解模型研究

对多维数据的研究最大困难在于其数据量十分庞大,数据规模的单纯扩大,往往导致系统的复杂性成倍增加,而且在研究中发现,存在着大量具有“多维性”、“动态性”和“模糊性”的数据.对此,该文基于动态模糊理论,提出了多维动态模糊数据的分解模型,该模型的建立,为人们用计算机处理这类问题提供了理论依据.     

Mapping of the rice （Oryza sativa L.） thermo-sensitive genic male sterile gene tms5 with EST and SSR markers

VELOPING“TWO-LINE”HYBRID RICE.THE POLLEN FERTILITY OF TGMS IS REGULATED BY THE TEMPERATURE OF ENVIRONMENT.THE POLLENS OF TGMS LINES ARE STERILE WHEN THE ENVI-RONMENT TEMPERATURE IS ABOVE A CRITICAL POINT,BUT FERTILE BELOW THIS POINT.SO FAR,A NUMBER OF T…     

Thermal conductivity measurement of InGaAs/InGaAsP superlattice thin films

The thermal conductivities of InGaAs/ InGaAsP superlattices with different period lengths were measured from 100 to 320 K using 3ω method. In this temperature range, the thermal conductivities were found to decrease with an increase in temperature. For the period length-dependant thermal conductivity, the minimum value does exist at a certain period length, which demonstrates that at a short period length, superlattice thermal conductivity increases with a decrease in the period length. When the period is longer than a certain period length, the interface thermal resistance dominates in phonon transport. The experimental and theoretical results confirmed the previous predictions from the lattice dynamics analysis, i.e. with the increase in period length, the dominant mechanisms of phonon transport in superlattices will shift from wave mode to particle mode. This is crucial for the cutoff of the phonons and lays a sound foundation for the design of superlattice structures.     

经激光表面晶粒细化TiAl合金的超塑扩散连接及其数学模拟

对经TiB2(体积分数ω=0．8％)颗粒细化后Ti-45Al-2Mn-2Nb合金进行了激光表面快速熔凝处理，在其表面获得了超细晶粒组织，并在T=1073～1223K,P=40-120MPa，t=0．5-1h条件下对经激光处理后的合金进行了超塑扩散连接，研究了温度、压强和时问对连接过程的影响，建立相应的理论模型对连接过程进行了模拟,研究结果表明，在晶粒度和表面粗糙度一定的条件下，温度、压强和时间对连接过程产生重要影响，而且三者之间存在一定的关系；经过激光表面处理后的合金试样在较低温度、较低压力和较短时间下(11731K/60 MPa／Ih)可达到完全冶金结合；实验结果与模拟计算结果相一致。     

Effect of particle size distribution on the microstructure,texture,and mechanical properties of Al–Mg–Si–Cu alloy

The effect of particle size distribution on the microstructure,texture,and mechanical properties of Al–Mg–Si–Cu alloy was investigated on the basis of the mechanical properties,microstructure,and texture of the alloy.The results show that the particle size distribution influences the microstructure and the final mechanical properties but only slightly influences the recrystallization texture.After the pre-aging treatment and natural aging treatment(T4 P treatment),in contrast to the sheet with a uniform particle size distribution,the sheet with a bimodal particle size distribution of large constituent particles and small dispersoids exhibits higher strength and a somewhat lower plastic strain ratio(r) and strain hardening exponent(n).After solution treatment,the sheet with a bimodal particle size distribution of large constituent particles and small dispersoids possesses a finer and slightly elongated grain structure compared with the sheet with a uniform particle size distribution.Additionally,they possess almost identical weak recrystallization textures,and their textures are dominated by CubeND {001}310 and P {011}122 orientations.     

基于多模态融合技术的用户画像方法

针对当前用户画像工作中各模态信息不能被充分利用的问题,提出一种跨模态学习思想,设计一种基于多模态融合的用户画像模型。首先利用Stacking集成方法,融合多种跨模态学习联合表示网络,对相应的模型组合进行学习,然后引入注意力机制,使得模型能够学习不同模态的表示对预测结果的贡献差异性。改进后的模型具有精心设计的网络结构和目标函数,能够生成一个由特征级融合和决策级融合组成的联合特征表示,从而可以合并不同模态的相关特征。在真实数据集上的实验结果表明,所提模型优于当前最好的基线方法。     

稠油热采井补贴管管柱界面粘脱分析与防治

稠油井注汽热采过程中,热应力易使补贴管与水泥环之间出现微裂缝、微环隙,致使界面出现粘脱现象,进而导致补贴失效。为延长补贴管使用周期,首先基于井下管柱三轴热应力理论与界面破坏准则,运用ABAQUS有限元软件建立热采井补贴管热力耦合三维有限元模型,模拟实际注蒸汽热采过程,得出补贴管-水泥-旧套管-水泥-地层系统的温度场、应力场分布规律及补贴管界面粘脱情况,并研究不同注汽温度、注汽压力、胶结强度等因素对补贴管界面黏结状态的影响;随后将预应力技术应用于补贴管二次固井过程中,并在不同工况下将施加预应力技术前后补贴管界面粘脱状态进行对比,界面粘脱长度显著降低;最后确定了不同温升条件下的预应力施加值,为二次固井与套损治理提供指导意义。     

一种在驱动轮打滑情况下爬壁机器人动力学建模方法

分析了轮式爬壁机器人(WWCR)在壁面作业容易发生打滑的原因,在建立其理想纯滚动假设条件下的动力学模型基础上,基于车辆动力学理论,引入单个驱动轮的动力学方程,提出了一种在驱动轮打滑情况下WWCR的动力学建模方法.仿真算例表明,吸盘吸附力和壁面附着系数对WWCR运动轨迹影响非常大;在复杂的壁面环境下,可通过对吸盘吸附力和驱动轮力矩的协调控制,合理调节轮子的打滑程度,获得尽可能大的附着系数,并弥补打滑造成的运动轨迹偏移,从而获得良好的壁面运动性能.该建模方法为WWCR在打滑情况下的运动控制和路径规划提供了理论研究基础.     

Genetic and physical finemapping of the Sc locus conferring indica-japonica hybrid sterility in rice （Oryza sativa L.）

Hybrid sterility is a major hindrance to utilizing the heterosis in indica-japonica hybrids. To isolate a gene Sc conferring the hybrid sterility, the locus was mapped using molecular markers and an F2 population derived from a cross between near isogenic lines. A primary linkage analysis showed that Sc was linked closely with 4 markers on chromosome 3, on which the genetic distance between a marker RG227 and Sc was 0.07 cM. Chromosome walking with a rice TAC genomic library was carried out using RG227 as a starting probe, and a contig of ca. 320 kb covering the Sc locus was constructed. Two TAC clones, M45EI4 and M90J01 that might cover the Sc locus, were partially sequenced. By searching the rice sequence databases with sequences of the TACs and RG227 a japonica rice BAC sequence, OSJNBb0078P24 was identified. By comparing the TAC and BAC sequences, six new PCR-based markers were developed. With these markers the Sc locus was further mapped to a region of 46 kb. The results suggest that the BAC OSJNBb0078P24 and TAC M45EI4 contain the Sc gene. Six ORFs were predicted in the focused 46-kb region.     

Multi-Dimensional Piece-Wise Self-Affine Fractal Interpolation Model

Iterated function system (IFS) models have been used to represent discrete sequences where the attractor of the IFS is piece-wise self-affine in R 2 or R 3 (R is the set of real numbers). In this paper, the piece-wise self-affine IFS model is extended from R 3 to Rn (n is an integer greater than 3), which is called the multi-dimensional piece-wise self-affine fractal interpolation model. This model uses a "mapping partial derivative" , and a constrained inverse algorithm to identify the model parameters. The model values depend continuously on all the model parameters, and represent most data which are not multi-dimensional self-affine in Rn . Therefore, the result is very general. The class of functions obtained is much more diverse because their values depend continuously on all of the variables, with all the coefficients of the possible multi-dimensional affine maps determining the functions.