大涡模拟进口合成湍流方法

进口瞬时湍流脉动条件直接影响大涡模拟的精度。对比研究了更适合工程应用的合成涡和合成谱方法,并通过考虑进口速度时间相关和随机数的概率密度分布,对合成谱方法进行改进。对零压力梯度平板湍流模拟结果的对比表明:不同湍流合成方法得到的合成湍流均存在发展为真实湍流的过渡过程;考虑进口湍流脉动的时间相关有助于合成湍流更快地发展为真实湍流,需要的流向距离更短;合成谱方法采用正态分布系数优于均匀分布,改进的合成谱方法优于常规合成谱方法。     

伴流速度对平行喷口射流影响的数值研究

采用大涡模拟方法计算研究平行喷口出口马赫数为0.9,伴流速度比分别为0.1、0.3和0.5时的喷射流场特性。计算时采用高精度的数值模拟方法,并结合Smagorinsky亚网格尺度模型。考察了流场统计平均特性、脉动特性以及射流流场中涡结构的发展演变过程,结果表明:伴流速度的增大使得势流核长度变长,减缓了空间剪切层的发展,转捩延迟。喷射流场速度分布具有自相似性,而湍流强度的分布则不具有相似性。通过分析剪切层中轴向速度脉动、径向速度脉动、压力脉动在空间任意两点上的时空相关性,发现随着伴流速度的增大,脉动量在空间上的相关性减弱,而脉动量向下游的传递速率增加。该研究结果为进一步揭示伴流速度对喷流声场的影响提供基础。     

Influence of leading edge with real manufacturing error on aerodynamic performance of high subsonic compressor cascades

To investigate the influence of real leading-edge manufacturing error on aerodynamic performance of high subsonic compressor blades, a family of leading-edge manufacturing error data were obtained from measured compressor cascades. Considering the limited samples, the leading-edge angle and leading-edge radius distribution forms were evaluated by Shapiro-Wilk test and quantile–quantile plot. Their statistical characteristics provided can be introduced to later related researches. The parameterization design method B-spline and Bezier are adopted to create geometry models with manufacturing error based on leading-edge angle and leading-edge radius. The influence of real manufacturing error is quantified and analyzed by self-developed non-intrusive polynomial chaos and Sobol’ indices. The mechanism of leading-edge manufacturing error on aerodynamic performance is discussed. The results show that the total pressure loss coefficient is sensitive to the leading-edge manufacturing error compared with the static pressure ratio, especially at high incidence. Specifically, manufacturing error of the leading edge will influence the local flow acceleration and subsequently cause fluctuation of the downstream flow. The aerodynamic performance is sensitive to the manufacturing error of leading-edge radius at the design and negative incidences, while it is sensitive to the manufacturing error of leading-edge angle under the operation conditions with high incidences.     

一种高光谱图像的双压缩感知模型

高光谱图像因其海量数据性,给存储、传输及后续分析处理带来了挑战。压缩感知理论提供了一种全新的信号采集框架。针对高光谱数据的三维特性,提出一种双压缩感知的采样与重构模型。该模型在采样阶段兼顾高光谱数据的空间和谱间稀疏特性,构造了能同时实现空间和谱间压缩采样的感知矩阵;重构阶段不同于传统的压缩感知重构方法直接重构高光谱数据,而是将高光谱数据分离成端元和丰度分别进行重构,然后利用重构的端元和丰度信息合成高光谱数据。实验结果表明,所提双压缩感知在低采样率下重构精度较三维压缩采样提高了10 dB以上,更为显著的是运算速度提升了3个数量级,同时该方法还便于获得端元和丰度信息。     

非均匀环境下利用单个数据集的STAP算法

对非均匀环境下利用单个数据集的机载雷达空时自适应处理(SDS-STAP)算法进行了研究。针对现有单个数据集类空时自适应算法存在的问题,提出了一种基于对待测数据集进行空时二维滑窗处理的二维幅度相位估计(2D-APES)算法。该算法利用回波的统计信息,不仅放宽了算法处理器对于雷达系统阵元数与脉冲数选取的约束,而且能够有效抑制杂波及待测单元内的干扰,同时降低了处理器的计算复杂度。最后,仿真数据及实测数据实验验证了所提算法的正确性与有效性。     

空间科学任务协同设计论证平台

空间科学任务协同设计论证存在设计方案耦合强,数据一致性差,数据变更难,数据与流程脱节等问题.为了解决上述问题,建立了典型空间科学任务的多层数字化模型;采用图形化方式对论证流程进行建模,并建立流程与数据的映射关系;通过共享数据池的方式为多岗位用户提供多方案数据协同机制;采用消息总线对数据变更进行及时提醒;利用方案依赖关系矩阵来判别方案耦合关系,最终自动合并任务总体设计方案.作为一个分布式平台,空间科学任务协同设计平台采用Eclipse RCP和Spring技术架构,整合了Hibernate、工作流Activiti5等中间件,提供统一门户,支持多岗位、多任务、多方案、多版本的管理能力,提供论证流程监控、数据协同交互等功能.结合某空间科学任务论证验证了该平台的有效性.      

地基伪卫星系统多址测距误差的校正算法

相对于卫星导航,地基伪卫星与接收机距离较近,测距中不涉及电离层传播等误差,易实现高精度测距和定位。但由于地基伪卫星位置固定,在信号连续发射模式下伪卫星信号间的多址干扰(MAI)会引起相对恒定的测距误差,针对地基伪卫星系统这一特点,提出一种基于反馈串行干扰抵消(SIC)的测距算法。首先对全部伪卫星信号进行捕获跟踪和功率估计,然后通过串行干扰抵消算法降低强功率伪卫星信号对期望测距信号的干扰,通过反馈模式进一步提高干扰抵消算法的性能。仿真结果表明,该方法可有效改善地基伪卫星系统强远近效应区域内的测距性能,不考虑其他误差下在强远近效应区域内基于码环可将伪卫星测距性能从10 m以上降低到约0.3 m。     

基于眼动数据的网络搜索行为预测方法

预测用户的网络搜索行为对改进搜索引擎和提升用户体验十分重要.现有大多数方法是基于用户的交互数据,如查询、点击和鼠标移动等.提出一种基于眼动数据的用户网络搜索行为预测方法.通过眼动实验,采集用户在网络搜索任务时的眼睛运动数据,将这些数据转化成两种数据格式:直方图和序列.直方图数据描述用户注意力的分布情况,序列数据呈现用户的扫视路径.使用4种学习算法对用户决策或用户意图进行预测,同时研究两种数据格式的性能.结果显示,两种数据格式均适合于预测用户决策,而序列数据更适合于预测用户意图.该结果表明,利用眼动数据能够有效预测网络搜索行为.      

一种改进的多点融合非均匀校正技术

针对空间宽幅相机在进行非均匀性校正过程中相机镜头与积分球输出口的相对位置发生变换后非均匀性校正精度下降的原因进行了分析,提出了一种改进的融合多点位置的非均匀校正新方法,即为了消除积分球的角均匀性和面均匀性的不足,采用获取多个位置的大量定标图像,融合求取平均值的方法得到校正系数来削弱这种影响,并进行试验验证分析。结果表明,多点融合的校正方法在积分球输出口光强不均匀、又无法固定相机与积分球相对位置的前提下,减少了积分球光强的不均匀性对系数可靠性的影响,相对辐射定标的非均匀校正精度得到大幅提高。     

Influence of lunar topography on simulated surface temperature

The surface temperature of the Moon is one of the essential parameters for the lunar exploration, especially to evaluate the Moon thermophysical features. The distribution of the temperature is heavily influenced by the Moon topography, which, however, is rarely studied in the state-of-art surface temperature models. Therefore, this paper takes the Moon topography into account to improve the surface temperature model, Racca model. The main parameters, such as slopes along the longitude and latitude directions, are estimated with the topography data from Chang’E-1 satellite and the Horn algorithm. Then the effective solar illumination model is then constructed with the slopes and the relative position to the subsolar point. Finally, the temperature distribution over the Moon surface is obtained with the effective illumination model and the improved Racca model. The results indicate that the distribution of the temperature is very sensitive to the fluctuation of the Moon surface. The change of the surface temperature is up to 150 K in some places compared to the result without considering the topography. In addition, the variation of the surface temperature increases with the distance from the subsolar point and the elevation, along both latitude and longitude directions. Furthermore, the simulated surface temperature coincides well with the brightness temperature in 37 GHz observed by the microwave sounder onboard Chang’E-2 satellite. The corresponded emissivity map not only eliminates the influence of the topography, but also hints the inherent properties of the lunar regolith just below the surface. Last but not the least, the distribution of the permanently shadowed regions (PSRs) in the lunar pole area is also evaluated with the simulated surface temperature result.