1.School of Physics and Electronics, Central South University, Changsha 410083, China 2.College of Advanced Interdisciplinary Studies, National University of Defense Technology, Changsha 410073, China
Fund Project:Project supported by the Equipment Pre-research Field Fund, China (Grant No. 6140415020311) and the Hunan Provincial Key Laboratory of High Energy Laser Technology Fund, China (Grant No. GNJGJS04).
Received Date:25 December 2018
Accepted Date:31 January 2019
Available Online:01 May 2019
Published Online:05 May 2019
Abstract:In aero optics, the linking equation proposed by Sutton is an important equation which can link the fluid-mechanic statistical parameters to the statistical optical degradation parameters. However, in the application of simplified linking equation (SLE) to subsonic flowfields, the weighting function is often ignored. The supersonic mixing layer flowfield is generated in the supersonic wind tunnel. The nanoparticle-based planar laser scattering technology is used to obtain the density field of flowfield. The optics errors between supersonic mixing layer wave-front variances calculated from the SLE and the generalized linking equation are analyzed. The results indicate the validity of using the SLE to estimate the wave-front variance of supersonic mixing layer flowfield. Moreover, the SLE with weighting function has better fitting accuracy than the SLE without weighting function. The weighting function for the application of SLE to the high correlated regions in the supersonic mixing layer is necessary. Keywords:aerooptics/ linking equation/ weighting function/ wave front variance
${\sigma^2 _\varphi }\left( L \right) = 2{\beta ^2}\int_0^L {\sigma^2 _{\rho′}}{\left( y \right)_{}}{l_\rho }\left( y \right)W\left( y \right){\rm{d}}y$
或高斯型
${\sigma^2 _\varphi }\left( L \right) = \sqrt {\text{π}} {\beta ^2}\int_0^L {\sigma^2 _{\rho′}}{\left( y \right)_{}}{l_\rho }\left( y \right)W\left( y \right){\rm{d}}y,$
$\begin{split}&{R_\rho }\left( {x,y,\delta \left( x \right),\delta \left( y \right)} \right) \\ = &\frac{{\overline {\left\langle {\rho′\left( {x,y,t} \right)\rho′\left( {x + \delta \left( x \right),y + \delta \left( y \right),t} \right)} \right\rangle } }}{\sqrt {\overline {\left\langle {\rho′^2\left( {x,y} \right)} \right\rangle } } \sqrt {\overline {\left\langle {\rho′^2\left( x + \delta \left( x \right),y + \delta \left( y \right) \right) } \right\rangle }}},\end{split}$
其中, $\left( {x,y} \right)$为选取的中心点坐标; $\delta \left( x \right)$, $\delta \left( y \right)$为相对于中心点的偏移量. 结合图1和图2, 选取坐标为(200, 148), (480, 140), (800, 112)和(1000, 100)中心点并计算了中心点附近区域的密度脉动相关函数, 相应的等值线分布见图4. 密度脉动特征长度可以表征流场中的涡尺度大小. 根据文献[21], 常将其定义为当相关函数值为最大值的1/e时所对应的流场尺度. 基于此, 计算出上述各中心点的密度脉动特征长度的归一化值分别为12, 44, 96和112. 图 4 密度脉动相关函数分布 (a)中心点(200, 148); (b)中心点(480, 140); (c)中心点(800, 112); (d)中心点(1000, 100) Figure4. Correlations of density fluctuations: (a) Central point (200, 148); (b) central point (480, 140); (c) central point (800, 112); (d) central point (1000, 100).
图4(b)—(d)可见, 在流场的中、下游中心点附近区域, 相关函数等值线呈倾斜近似椭圆形状, 这些倾斜的椭圆形状表示湍流大尺度结构的存在[22-24]. 同时从图4可以看到, 随着流场逐渐向下游发展, 密度脉动高度相关区域在逐渐扩大. 这是由于混合层流场开始发展的初期产生涡量堆积, 失稳之后产生大尺度结构卷起, 表征了该区域流场的高度非均匀、非各向同性性. 由(17)式可知, lρ与权重函数的计算密切相关. 结合图3选取了混合层流场中误差最大的点x/h = 772、误差最小的点x/h = 380和误差居中的点x/h = 932, 计算了(17)式所示权重函数的y方向分布, 如图5所示. 图 5 部分流向点处的权重函数分布 Figure5. Distribution of weighting functions at some stream-wise locations.
由图5可见, 在x/h = 772处, 权重函数偏离1的区域最多, 因此导致由于忽略权重函数的(12)式计算的相位方差误差也最大; 相反, x/h = 380位置处权重函数偏离1的区域最少, 因此在该位置处(12)式计算的波前方差误差值最小. 进一步, 给出了(12)和(15)式中不包含权重函数的积分核函数${\sigma^2 _{\rho′}}\left( y \right){l_\rho }\left( y \right)$和包含权重函数的积分核函数${\sigma^2 _{\rho ′}}\left( y \right){l_\rho }\left( y \right)W\left( y \right)$分布, 如图6(a)和图6(b)所示, 以及二者的差值分布如图6(c)所示. 由图6(c)可知, 权重函数影响集中在流场自由边界一侧、且位于x/h = 772处附近. 由此可见, 权重函数对于关联方程在超声速流场密度脉动高度相关区域中应用的必要性. 图 6 高斯型关联方程加入权重函数前后的积分核分布 (a)未加入权重函数; (b)加入权重函数; (c)积分核分布差 Figure6. Integral kernel distribution calculated by Gaussian linking equation before and after adding weighting function: (a) Before adding the weighting function; (b) after adding the weighting function; (c) the integral kernel distribution differences.
对权重函数(17)式中的Erf(·)函数进行进一步分析发现, 只有当$ L \gg l_\rho$, 即L/lρ较大时, W(y)值越接近1; 反之, W(y)值偏离1越远. 由于流场的厚度L是一不变量, 因此密度脉动特征长度lρ的大小决定着权重函数W(y)的值. 由于流场失稳导致大涡结构的存在, 在流场的中游附近区域L/lρ较小, 所以权重函数的值会逐渐偏离1较远, 此时忽略权重函数的影响自然会导致波前方差计算误差的增加. 最后, 基于上述分析, 加入权重函数重新计算波前方差, 绘出基于(15)式计算波前相位方差的曲线, 结果如图7所示. 可以看到, 高斯形式的关联方程(12), 在加入权重函数(17)式后, 计算得到的波前相位方差曲线的拟合效果要明显变好. (8)和(15)式计算结果对应的两条曲线的拟合优度为0.9127. 图 7 高斯型关联方程加入权重函数前后计算波前方差对比 Figure7. Wave-front variance calculated by Gaussian linking equation before and after adding weighting function.