超音速喷射实验分析论文

nacwu

Case Study: Flow field generated by supersonic jet impingement

nacwu

Understanding the velocity field generated by twin supersonic twin jet impingement plays a pivotal
in the design of next generation fighter aircraft, such as the Joint Strike Fighter. This feature rich
flow encompasses phenomena such as the acoustic interaction between the jets, the jets and the
structure and the unsteady loading due to the fountain flow. To achieve a competent design that
survives the loads generated by the complex aero-acoustic interactions it is essential to fully
characterize and understand the velocity field.
An experimental study, aimed at the measurement of the flow field generated by twin supersonic
jet impingement was conducted in the STOVL facility at the Fluid Mechanics Research
Laboratory at the Florida State University. The configuration in this experiment used two circular
jet nozzles with a throat diameter of 18mm, designed for an exit Mach number of 1.5. The exit
diameter of the C-D nozzle is 19.5 mm and the nozzle separation is about three nozzle
diameters. The air for the jets was supplied from high-pressure (2000 psi) storage tanks coupled
to a high displacement air compressor. The experiments were conducted at NPR=3.7 (stagnation
pressure/ambient pressure), which is the ideally expanded jet condition of the nozzle. The
nominal exit Reynolds number at the exit of the nozzle was about 5.5x105. The jet flow was
seeded with sub-micron oil droplets generated by a Wright nebulizer, whereas the ambient air
was seeded with smoke particles produced by a Rosco fog generator. For the PIV recordings a
Nd:Yag Laser (New-Wave Gemini) producing 140 mJ/pulse was used. A Kodak ES1.0 camera
fitted with a 110 mm F/1.8 Nikon lens recorded the images. While the maximum area covered in
this experiment was 100x100 mm, only part of the image contained the flow spanning 100x24
mm. The camera was used in the double-triggered exposure mode for the recording of
successive illumination pulses with a pulse separation of 1.4 microseconds.

nacwu

[这个贴子最后由nacwu在 2005/05/20 11:47am 第 1 次编辑]

Several image sequences with 40 image pairs were acquired at a rate of 15 Hz. Figure 1 is one
representative example. In the figure the two corresponding images are displayed together to
facilitate the visualization. The images clearly show these salient features: The two jets with their
associated shear layers and stand-off shock structure, the re-circulating flow downstream of the
shock and the highly three-dimensional and unsteady fountain flow. The images were analyzed
using a rectangular grid with 100x30 cells. Figure 2 shows one of the computed vector fields
superposed to its corresponding images. The vectors were color coded to represent the original
data and the data that was obtained using an interpolation scheme. According to the scheme a
dropout is obtained when there is not sufficient image data to construct the correlation, or when a
no-valid correlation is computed. In this case the velocity at the grid points are interpolated using
the least squares algorithm with the information provided by the neighbor grid points. However, it
is noteworthy that even in the difficult conditions of large velocity and density gradients the
number of interpolations was kept below a very minimal level of 150 out of 3000 grid points. For
clarity the velocity field in figure 2 is displayed, superposed with iso-velocity contours, in figure 3.
In this figure the unsteady nature of the fountain flow is further documented as well as the
structure of the flow downstream of the shock. As it was evidenced by the particle images, the
velocity field confirms the existence of small, albeit strong, re-circulating regions. Analysis of
several frames of a sequence reveals that these structures are quite unsteady in nature,
shedding with a frequency that is coupled with that of the oscillation of the fountain flow (as
evidenced by the attached animation TwinJet.avi )

nacwu

The level of detail depicted in figures 3 and 4 was made possible by the highly accurate and
spatially resolving characteristics of the processing scheme. For example the re-circulation zone
downstream of the shock wave, extending about 2.5x2.5 mm (figure 3), were clearly captured.
This stringent test case evidences the ability of the algorithm to perform accurate computations in
spite of strong velocity and density gradients. Figure 4 shows the average flow features obtained
using 40 velocity field samples. It is noteworthy that in spite of the highly unsteady and threedimensional
nature of the flow, the average flow field is captured with great fidelity. This is a
consequence of keeping the appropriate image acquisition parameters (particle images
occupying three or more pixels) and the properties of the processing scheme.

nacwu

Further confirmation that the processing scheme produces velocity distributions without biases or
with peak-lock effects is given in figure 5. This figure is the histogram of the cross-stream velocity
component of a single flow realization. The data in the figure shows a continuous distribution of
the velocity in highly turbulent flow.

nacwu

费点力气帖上来，方便大家观看：）

liuwen509

这都贴出来了^_^  厉害^_^

liuwen509

在清华大学做的PIV展示怎么样^_^  拍了什么好东西没有啊^_^

robin

超音速流体的PIV测量的确是非常具有挑战性的工作。楼主为什么不把完整的文章贴出来呢？莫非是机密报告？
我这里贴一篇外国人的文章，希望能借楼主这个帖子，大家交流一下超音速流体PIV测量方面的信息。

yxjbuaa

[这个贴子最后由yxjbuaa在 2005/11/06 02:09pm 第 1 次编辑]

这里给出的粒子图像位移pdf图不够细致啊。
你图像中的粒子图像直径大约是多少？
粒子图像位移pdf分布应当给的细致一些，也就是亚象素的间距给的小一些，此外，给出亚象素的分布pdf或许会更直观一些。
粒子分布好像很浓。对于这样高速的情况，影响因素太复杂了，精度分析显然欠充分。
此外，按照你的粒子图像位移分布，得到的结果误差会较大，当然这样的实验存在很大的难度，就是速度大小分布变化太剧烈。实验中的粒子图像位移主要集中在+-2pixels之间，这样小的位移，如果亚象素分辨率不高，可想而知误差是多大。还有，对于有激波的情况，你实验结果中的激波很不清晰，这是物理上的原因还是图像处理上的原因？可能后者居多吧。对于激波的情况简单的精度分析是远远不够的。