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。     

波瓣数对波瓣S型混合二元喷管气动热力性能影响

依据某型涡扇发动机波瓣S型混合二元喷管,保持波瓣混合器长度、内扩张角、外扩张角以及宽高比不变,依次取波瓣混合器波瓣数为12,14,16,18,20,建立了一组具有不同波瓣数的波瓣S型混合二元喷管模型.采用经过验证的CFD方法,研究了波瓣数对波瓣S型混合二元喷管气动热力性能的影响规律.结果表明:在波瓣尾缘截面至第1个S弯截面区域,波瓣数对流体混合程度产生很大影响,并且热混合效率近乎随波瓣数增加而增加.在第1个S弯截面至波瓣S型混合二元喷管出口截面区域,波瓣数为16的波瓣S型混合二元喷管模型的总压恢复系数始终最低,其余模型的总压恢复系数以及热混合效率没有明显差别.在波瓣S型混合二元喷管出口截面上,波瓣数为16的波瓣S型混合二元喷管模型的热混合效率最高,达到0.850,然而其总压恢复系数相对于该截面上最高值下降了0.289%.此外,波瓣S型混合二元喷管的渐缩型流道能够提高流向涡强迫混合效果,但同时也加速流向涡的耗散速率.      

分布式动力系统尾缘射流与边界层抽吸的数值分析

为研究带有边界层抽吸的分布式动力系统尾缘射流对机身气动性能及推进效率的影响,将机身简化为二维翼型,并加入尾缘射流及边界层抽吸的作用,利用数值模拟的手段来研究来流攻角、射流偏转角、边界层抽吸对推进效率及气动性能的影响,为分布式动力系统的设计与应用提供初步的建议.结果表明在中、小来流攻角(2°及0.6°)的情况下尾缘射流及边界层抽吸能够提高升阻比,推进效率可超过80%;而在大来流攻角(4°)情况下射流偏转角增大使翼型的阻力大幅上升,对气动性能和推进效率产生极为不利的影响.      

基于放大因子与Spalart-Allmaras湍流模型的转捩预测

为了验证放大因子输运方程与Spalart-Allmaras(S-A)湍流模型耦合对转捩现象的模拟精度,选取Schubauer and Klebanoff(S-K)平板、S809低速翼型、30p30n多段翼型以及复杂的三维HiLiftPW-1构型进行自由转捩计算,并将计算结果与实验进行比较分析,其中针对S809算例,还与Langtry-Menter(L-M)转捩模型进行了比较.算例结果表明:放大因子输运方程与S-A湍流模型的耦合能够较好的捕捉转捩位置以及转捩发展过程,对分离泡诱导的转捩的模拟相比L-M转捩模型更精确,转捩位置的捕捉精度提升了10%;对比实验,多段翼转捩位置的捕捉误差最大为6.5%;针对三维高升力增升构型,以实验作为参考,全湍流计算与考虑边界层转捩的对比显示考虑边界层转捩能够更加精确的模拟气动力系数,升力和表面摩擦阻力系数的模拟精度精度提升1%.      

二维喷管的初始流动

基于可压缩Navier-Stokes方程,采用大涡模拟方法与高精度混合WENO/TCD格式,对Ma=1.4的超声速平面射流初始流场进行了数值研究.数值结果清晰地描述了超声速平面射流初始流场的结构特征,包括主涡环与激波结构以及它们演变过程.因主涡环内存在涡导激波对,激波与涡相互作用加速射流剪切层失稳,使剪切层首次卷起形成小涡的位置出现在涡导激波对后,此与亚声速射流情况不同.小涡串卷起成后,相继与涡导激波对相互作用,使激波出现明显的变形并加速主涡环失稳.      

超声速尾迹-剪切流的混合增强

超声速条件下燃料和空气之间的高效混合是超然冲压发动机技术上的主要挑战。基于大涡模拟和流动稳定性分析,针对超声速尾迹-剪切流动开展了混合增强方法研究。尾迹的存在改变了混合层流动的速度剖面,对流动稳定性产生了重要影响,使混合层由三维最不稳定变为二维最不稳定,最不稳定扰动波频率和增长率增大。基于流动稳定性结果引入扰动的混合增强方式依然有效,根据稳定性结果设计了波纹隔板。数值结果表明：二维波纹壁引入的扰动未能增长,不具备混合强化效果,而三维波纹壁引入的扰动能够快速增长,具有混合强化效果,且波纹壁参数越接近最不稳定扰动波参数,混合强化效果越明显。     

Numerical investigation of the impact of asymmetric fuel injection on shock train characteristics

Numerical simulations are carried out to investigate the impact of asymmetric fuel injection on shock train characteristics using the commercial-code FLUENT. The asymmetry of fuel injection is examined by changing the fuel flow rates of the upper and lower wall fuel injectors. The numerical approach solves the two-dimensional Reynolds-averaged Navier–Stokes (RANS) equations, supplemented with a k-ω model of turbulence. As a result, different ways of fuel injections will always lead to shock train transitions, with the variations of shock train structure, strength and leading edge position. For symmetric fuel injection, the flowfield of the isolator is quite asymmetric with the boundary layer of the upper wall side developing much stronger than that of the lower wall, which is due to the heterogeneity of the incoming flow. Regarding to asymmetric fuel injection with more of lower wall side, though the pressures in the combustor are nearly the same, the first shock of the shock train converts between ‘Distinct symmetric X type shock’ and ‘Obscure and weaker asymmetric shock’ and the shock train leading edge moves upstream with the increase of the asymmetry level. With regard to asymmetric fuel injection with more of upper wall side, ‘incomplete asymmetric X type shock’ occurs and the shock train structures keep nearly the same with low level of fuel injection asymmetry. Unexpected results like unstart will happen when increasing the level of fuel injection asymmetry. And the isolator will come back to normal state by decreasing the differential of upper and lower wall sides fuel injections.     

Features of the occurrence of the additional stratification on the bottom-side F region over the equatorial location of Trivandrum

The general features of occurrence of an additional layer on the bottom side of F region, referred to as F0.5 layer in the pre noon period, over the magnetic equatorial location of Trivandrum (8.5° N; 77° E; dip lat of 0.5° N) in India during the period from 2004 to 2007 are presented using ionosonde observations. The F0.5 layer has a June (northern summer) solsticial maximum probability of occurrence with secondary maxima during December (northern winter) solstice. The seasonal as well as the day-to-day variability in the occurrence of F0.5 layer as mentioned in this paper seems to be a result of the variations in the amplitude and phases of the tides and gravity waves, and inventory of the metallic ions of meteoric origin. This study brings out an important manifestation of morning time F layer base region dynamics.     

Stability of the seasonal variations in diurnal and semidiurnal components of mid-latitude F2 layer parameters

We report work utilizing 15-min resolution ionospheric data obtained with DPS-4 digisonde in 2003–2011 to study the seasonal variations in amplitudes and phases of the most powerful spectral components of the F2 layer critical frequency (foF2) and peak height (hmF2) fluctuations over Irkutsk (52.5°N, 104.0°E). We show that fluctuations of both parameters contain quasi-harmonic components with periods of T_n = 24/n h (n = 1–7). The number of distinct spectral peaks varies from 3 in summer to 7 in winter. Amplitude and phase characteristics of the diurnal (n = 1) and semidiurnal (n = 2) components is studied using the data sets extracted from the original data sets with band-pass filter. It has been found that the amplitudes of diurnal/semidiurnal foF2 and diurnal hmF2 components are maximum in winter and minimum in summer. Amplitudes of the diurnal components vary gradually; those of the foF2 semidiurnal one, abruptly, thus forming a narrow winter maximum in November–January. The phase (local time of maximum) of the diurnal foF2 component increases gradually by 4–6 h from winter to summer. The phase of the semidiurnal foF2 component is nearly stable in winter/summer and sharply decreases (increases) by 2–3 h near the spring (autumn) equinox. The phase of the diurnal component of hmF2 (local time of minimum) varies slightly between 1130 and 1300 LT; that of the semidiurnal one decreases (increases) by 4–6 h from January to March (from September to November). The results obtained show that the main features of seasonal variations in the diurnal and semidiurnal components of the mid-latitude F2 layer parameters recur consistently during the solar activity growth and decline phases.