力学进展  2019 , 49 (1): 201907-201907 https://doi.org/10.6052/1000-0992-17-018

力学

非定常流动模拟的时间离散方法

张伟伟, 贡伊明, 刘溢浪

西北工业大学翼型叶栅重点实验室, 西安 710072

Time discretization methods in the computation of unsteady flow

ZHANG Weiwei, GONG Yiming, LIU Yilang

School of Aeronautics, Northwestern Polytechnical University, Xi'an 710072, China

中图分类号:  V211.3

文献标识码:  A

通讯作者:  †E-mail: aeroelastic@nwpu.edu.cn

收稿日期: 2017-09-20

接受日期:  2018-05-30

网络出版日期:  2019-01-15

版权声明:  2019 中国力学学会 This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

基金资助:  国家自然科学基金项目资助(11622220,11572252).

作者简介:

作者简介:张伟伟, 西北工业大学教授, 流体系主任.主要从事复杂流固耦合力学、非定常空气动力学和流动控制研究.中国空气动力学会理事, 陕西省力学学会理事.主持国家重大专项、国家自然科学基金(3项)、863项目、教育部博士点基金、国防预研基金、航空基金等省部级以上基金13项.在国内外刊物和学术会议发表论文100余篇, 被SCI检索22篇, 其中JCRQ1区14篇, 被EI检索67篇.


展开

摘要

对于不同非定常流动问题, 采用合适的时间离散方法,可有效提高数值精度和计算效率. 本文在总结传统时间离散方法的基础上,对近些年发展的非线性频域法、谐波平衡法、经典时间谱方法、时间谱元法、时间有限差分法等进行了系统地总结.根据离散形式的不同,将上述方法分为时域推进法、频域谐波法、时域配点法和混合方法4大类.首先简要介绍了各类方法的数学思想以及研究进展,并重点比较了(准)周期性非定常流动计算中各方法的精度、效率以及适用范围.然后, 对各种时间离散格式的特点进行总结,并就不同的非定常流动问题如何选择合适的时间离散方法给予了建议.最后, 对这些新型时间离散格式在工程中的应用进行了简要介绍,并对其发展方向进行展望.

关键词: 非定常流动 ; 时间离散方法 ; 频域谐波法 ; 时域配点法 ; 时间谱方法

Abstract

For the numerical computation of unsteady flow, the computational accuracy and efficiency would have a significant difference with different time discretization methods. This paper based on the summarize of the development situation of time discretization methods at present, briefly introduces the time discretization methods developed in recent years like the nonlinear frequency domain method, harmonic balance method, time spectral method, time spectral element method, time finite difference method and so on. Based on the difference between discrete versions, the time discretization methods can be divided into four types: time domain marching method, frequency domain harmonic method, time domain collocation method and hybrid method. This paper briefly introduces the mathematical thought and study progress of each discretization method, and selective compare the accuracy, efficiency, and scope of application of each time discretization method in the computation of unsteady flow. Then we systematically summarize the characteristic of each time discretization method and advise how to choose appropriate time discretization methods in different unsteady flow problems. Finally, briefly introduce the application of current time discretization methods in the projects and discuss the development directions of the time discretization method in the future.

Keywords: unsteady flow ; time discretization method ; frequency domain harmonic method ; time domain collocation method ; time spectral method

0

PDF (9827KB) 元数据 多维度评价 相关文章 收藏文章

本文引用格式 导出 EndNote Ris Bibtex

张伟伟, 贡伊明, 刘溢浪. 非定常流动模拟的时间离散方法[J]. 力学进展, 2019, 49(1): 201907-201907 https://doi.org/10.6052/1000-0992-17-018

ZHANG Weiwei, GONG Yiming, LIU Yilang. Time discretization methods in the computation of unsteady flow[J]. Advances in Mechanics, 2019, 49(1): 201907-201907 https://doi.org/10.6052/1000-0992-17-018

1 引 言

非定常流动问题广泛存在于航空航天领域, 比如直升机旋翼 (Choi et al.2008)、叶轮机 (Guo et al. 2012)、气动弹性问题 (Farhat & Lesoinne 2000)等. 根据脉动响应的周期性特征,非定常流动可以分为周期性流动、准周期流动和非周期流动.比如直升机定直前飞时旋翼的绕流就是典型的周期性流动,机翼的气弹响应绕流为准周期流动, 湍流槽道流动属于非周期流动.对于周期性或准周期流动,根据非定常减缩频率大小分为高减缩频率流动、低减缩频率流动以及未知周期的流动.比如扑翼流动属于高减缩频率流动,刚性飞行器动导数计算中强迫振荡属于低减缩频率流动,静止非定常圆柱绕流属于未知周期的流动.

非定常流动的研究手段主要有理论模型、数值模拟和风洞试验.理论模型虽然计算量小, 但多基于线化势流理论,精度不足且无法用于跨声速或大迎角等非线性流动. 风洞试验虽然可靠性较高,但除了难度大、周期长、成本高外, 还存在洞壁及支撑干扰等复杂非定常修正问题.从国内外发展趋势来看, 通过计算流体力学方法 (computational fluid dynamic,CFD) 已经逐渐成为非定常流动研究的主要手段.

相对于定常计算, 非定常数值模拟存在计算量和存储量大的问题,很难在工程中广泛应用. 即使是在计算机配置日新月异的今天,计算机的计算能力也远远跟不上非定常流动模拟对其的需求.解决这种矛盾亟需发展高效高精度的流场求解算法. 因此,在数值计算中如何能够实现非定常流场的高效高精度求解便成了当前的研究热点.但以往多数计算流体力学研究者将关注点放在空间离散与迭代算法的改进上,忽略了时间离散方法. 近期的很多研究发现, 采用合适的时间离散方法也可使CFD计算的精度和效率有非常明显的提高. 因此,近些年尤其是进入21世纪以来时间离散方法得到了迅速发展,各种时间离散方法如雨后春笋般涌现.

当前针对非定常流场的时间离散格式主要是时域向后差分法 (backward difference format, BDF) (Butcher 2016). 其主要思想是先求解定常流场,然后通过时域推进顺次求解不同时刻的流场.最常见的就是二阶时域向后差分离散(BDF2).但这类方法在时间上只能够达到二阶精度,并且对于周期性流动需要若干周期才能达到稳定, 没有利用流动的周期性特征,计算效率不理想.

由于在实际工程应用中很多涉及到的非定常流动均具有周期性特征,于是针对周期性非定常流动, 近年来出现了线化频域法 (linear frequency domain solver, LFD) (Hall & Crawley 1989)、非线性频域方法(non-linear frequency domain method, NLFD) (Ning & He 1997,Mcmullen et al. 2001)、谐波平衡法 (harmonic balance method, HB)(Hall et al. 2002)、时间谱方法 (time spectral method, TS)(Gopinath & Jameson 2005, Van der Weide et al.2005)、时间有限差分法 (finite difference method in time, FDMT)(Leffell et al. 2016) 等时间离散格式. 对于准周期问题,发展了混合向后差分时间谱方法 (BDF/TS) (Yang & Mavriplis 2010) 以及切比雪夫伪谱法 (Dinu et al. 2006). 对于纯非周期问题,发展了时间谱元法 (spectral element method in time, SEMT) (Kurdi & Beran 2008).这些新发展的时间离散格式虽然精度和效率相比于传统的BDF有较为明显的提升,但适用范围有限. 如果时间离散格式应用不当,其计算精度和效率会反不如BDF, 甚至失效. 比如Jameson & Shankaran (2009) 将时间谱方法应用到减缩频率大的扑翼中,经过研究表明时间谱方法计算扑翼无论是精度还是收敛性均不如传统的BDF.因此, 研究并确定每一种时间离散格式的适用范围至关重要.

对于这些新发展的时间离散格式还需要较全面的横向比较. Ronch等 (2010, 2013)

对比了时域推进法、线化频域法和谐波平衡法在数值预测动导数精度、计算效率与消耗内存方面的能力.

Leffell等 (2016)

对于周期性非定常流动中的时域向后差分法、时间谱方法以及时间有限差分法做了比较详细的精度与效率对比,

并对时空并行算法进行了阐述. 但这对于所有的时间离散方法来说只是冰山一角,

有必要对近些年来蓬勃发展的多样化时间离散格式做一个系统的总结与对比.

本文总结目前非定常流场计算中主流的时间离散方法,根据每种方法时间离散形式的不同分为4大类:时域推进法、频域谐波法、时域配点法和混合方法.分别介绍各类方法的具体原理并比较各方法的计算精度、效率以及适用范围. 另外,对于新发展的时间离散格式在工程问题中的应用及效果也做了相应的归纳. 最后,对未来时间离散方法的发展方向进行了展望.

2 时域推进法

时域推进法的步骤是先算初始定常流动,之后沿物理时间向前推进顺次求出往后各时刻的流场. 主要代表就是时域向后差分法,这也是最初始且目前应用最广的时间离散方法.

时域向后差分法, 顾名思义,即对于时间导数进行向后差分离散之后进行实时间步的推进.二阶BDF关于时间导数的离散形式如下

$$\dfrac{{\rm d}\pmb U _i^n }{{\rm d}t} = \dfrac{3\pmb U _i^n - 4\pmb U _i^{n - 1} + \pmb U _i^{n - 2} }{2\Delta t}(1)$$

式中, $\pmb U $表示守恒量, $\Delta t$表示时间间隔.离散后的非定常控制方程为

$$V_i \dfrac{3\pmb U _i^n - 4\pmb U _i^{n - 1} + \pmb U _i^{n - 2} }{2\Delta t} + \pmb R \left( {\pmb U _i^n } \right)=\pmb 0(2)$$ 式中, $V_i $表示网格体积, $\pmb R $表示残差.Briley 和Mcdonald (1975)Beam 和Warming (1976)等通过线化非线性残差来隐式求解式(2).但用直接法求解如此庞大的系数矩阵的逆矩阵, 计算量难以接受,因此需要迭代求解, 即所谓的``单时间步法''. 之后, Jamson (1991)提出了"虚拟时间亚迭代法", 该方法又称"双时间步法".即在冻结的物理时间点上巧妙地引入虚拟时间迭代过程,可将定常流动计算中的加速收敛措施应用到非定常流动的计算中,思想简单、实现容易. 因此, 双时间步BDF很快得到了广泛的应用 (Briley & Mcdonald 1975). 其非定常控制方程如下所示

$$ V_i \dfrac{{\rm d}\pmb U _i^n }{{\rm d}\tau} + V_i \dfrac{3\pmb U_i^n - 4\pmb U _i^{n - 1} + \pmb U _i^{n - 2} }{2\Delta t} + \pmb R \left( {\pmb U _i^n } \right) =\pmb 0(3)$$

式中$\tau$表示伪时间. BDF的优点在于其可以适用于任意的非定常流动,方法容易理解且在定常流动计算程序上的改动工作量很小.但从计算精度上而言, 三阶或以上的BDF会导致产生非物理解甚至发散(Hull et al. 1972).因此BDF的精度很难达到高阶且需要较多的时刻点才能保证足够的时间方向准确度.

从计算效率上而言, 对于工程中常见的周期性流动,BDF需要若干周期才能达到稳定, 计算时间之长是在工程应用中难以忍受的.也正因为BDF有这些缺陷, CFD工作者需要寻找更优的时间离散方法,促使了近期各种时间离散方法的诞生.

3 频域谐波法

由于在工程中多数的非定常流动均具有周期性特征,通过利用流动的周期性特征将原非定常流场的控制方程做傅里叶变换,使之转化为频域的一组定常方程, 进而对该频域方程组进行求解获得各频域谐波分量,最后通过傅里叶反变换获得时域响应. 其计算主要在频域内进行,因而称之为频域谐波法.

3.1 线性频域谐波方法

早期频域方法在非定常气动力模型的构建 (Zeiler 2000) 中使用的,比如著名的西奥道生气动力模型 (Brunton & Rowley 2013).近期Hall等 (1989)在用欧拉方程数值模拟涡轮机的实践中发展了频域的非定常流场数值求解方法------线性谐波法,也由此拉开了时间离散方法变革的序幕.

Hall等 (1989) 假设流场原始变量由时均量以及弱扰动量叠加而成.如式(4)所示

$$\left. \begin{array}{l} \rho \left( {x,y,t} \right) = \bar {\rho }\left( {x,y} \right) + \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\rho }} \left( {x,y,t} \right) \\ u\left( {x,y,t} \right) = \bar {u}\left( {x,y} \right) + \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {u}} \left( {x,y,t} \right) \\ v\left( {x,y,t} \right) = \bar {v}\left( {x,y} \right) + \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {v}} \left( {x,y,t} \right) \\ p\left( {x,y,t} \right) = \bar {p}\left( {x,y} \right) + \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {p}} \left( {x,y,t} \right) \\ \end{array}\right\}(4)$$

将式(4)代入欧拉方程中, 把扰动部分当成小量来处理并忽略高阶项,将原欧拉方程分为与时间无关的定常欧拉方程以及与时间有关的线化扰动方程.求解定常的时均流动方程如下

$$\left. \begin{array}{l} \dfrac{\partial \pmb F }{\partial x} + \dfrac{\partial \pmb G }{\partial y} =\pmb 0 \\ \pmb F = \left[ {{\begin{array}{*{20}c} {\bar {\rho }\bar {u}} \\ {\bar {\rho }\bar {u}^2 + \bar {p}} \\ {\bar {\rho }\bar {u}\bar {v}} \\ {\dfrac{\gamma }{\gamma - 1}\bar {p}\bar {u} + \dfrac{1}{2}\bar {\rho }\bar {u}\left( {\bar {u}^2 + \bar {v}^2} \right)} \\ \end{array} }} \right] \\ \pmb G = \left[ {{\begin{array}{*{20}c} {\bar {\rho }\bar {v}} \\ {\bar {\rho }\bar {u}\bar {v}} \\ {\bar {\rho }\bar {v}^2 + \bar {p}} \\ {\dfrac{\gamma }{\gamma - 1}\bar {p}\bar {v} + \dfrac{1}{2}\bar {\rho }\bar {v}\left( {\bar {u}^2 + \bar {v}^2} \right)} \\ \end{array} }} \right] \\ \end{array}\right\}(5)$$

对应弱扰动量的非定常欧拉方程如下所示

$$\left. \begin{array}{l} \dfrac{\partial }{\partial t}\pmb B _1 \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb W}} + \dfrac{\partial }{\partial x}\pmb B _2 \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb W}} + \dfrac{\partial }{\partial y}\pmb B _3 \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb W}} =\pmb 0 \\ \pmb B _1 = \left[ {{\begin{array}{*{20}c} 1 \quad & 0 \quad & 0 \quad & 0 \\ \bar {u} \quad & \bar {\rho } \quad & 0 \quad & 0 \\ \bar {v} \quad & 0 \quad & \bar {\rho } \quad & 0 \\ {\dfrac{1}{2}\left( {\bar {u}^2 + \bar {v}^2} \right)} \quad & {\bar {\rho }\bar {u}} \quad & {\bar {\rho }\bar {v}} \quad & {\dfrac{1}{\gamma - 1}} \\ \end{array} }} \right] \\ \pmb B _2 = \left[ {{\begin{array}{*{20}c} \bar {u} \quad & \bar {\rho } \quad & 0 \quad & 0 \\ {\bar {u}^2} \quad & {2\bar {\rho }\bar {u}} \quad & 0 \quad & 1 \\ {\bar {u}\bar {v}} \quad & {\bar {\rho }\bar {v}} \quad & {\bar {\rho }\bar {u}} \quad & 0 \\ {\dfrac{1}{2}\bar {u}\left( {\bar {u}^2 + \bar {v}^2} \right)} \quad & {\dfrac{\gamma }{\gamma - 1}\bar {p} + \dfrac{3}{2}\bar {\rho }\bar {u}^2 + \dfrac{1}{2}\bar {\rho }\bar {v}^2} \quad & {\bar {\rho }\bar {u}\bar {v}} \quad & {\dfrac{\gamma \bar {u}}{\gamma - 1}} \\ \end{array} }} \right] \\ \pmb B _3 = \left[ {{\begin{array}{*{20}c} \bar {v} \quad & 0 \quad & \bar {\rho } \quad & 0 \\ {\bar {u}\bar {v}} \quad & {\bar {\rho }\bar {v}} \quad & {\bar {\rho }\bar {u}} \quad & 0 \\ {\bar {v}^2} \quad & 0 \quad & {2\bar {\rho }\bar {v}} \quad & 1 \\ {\dfrac{1}{2}\bar {v}\left( {\bar {u}^2 + \bar {v}^2} \right)} \quad & {\bar {\rho }\bar {u}V} \quad & {\dfrac{\gamma }{\gamma - 1}\bar {p} + \dfrac{1}{2}\bar {\rho }\bar {u}^2 + \dfrac{3}{2}\bar {\rho }\bar {v}^2} \quad & {\dfrac{\gamma \bar {v}}{\gamma - 1}} \\ \end{array} }} \right] \\ \end{array}\right\}(6)$$

其中$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb W}} $为原始变量的扰动分量, 即$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb W}} = \left[ {\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\rho }} ,\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {u}} ,\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {v}} ,\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {p}} } \right]^{\rm T}$. 从式(6)不难发现,由于时均量在时均流动方程中均已求出, 此扰动方程为线化方程.假设流动的弱扰动量为谐振形式, 即可表达为如下形式

$$\left.\begin{array}{l} \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\rho}} \left( {x,y,t} \right) = \tilde {\rho }\left( {x,y} \right){\rm e}^{{\rm j}wt} \\ \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {u}} \left( {x,y,t} \right) = \tilde {u}\left( {x,y} \right){\rm e}^{{\rm j}wt} \\ \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {v}} \left( {x,y,t} \right) = \tilde {v}\left( {x,y} \right){\rm e}^{{\rm j}wt} \\ \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {p}} \left( {x,y,t} \right) = \tilde {p}\left( {x,y} \right){\rm e}^{{\rm j}wt} \\ \end{array}\right\}(7)$$

式中$\omega $为角频率. 代入式(6)中整理得到

$$jwB_1 \tilde {\pmb W} + \dfrac{\partial}{\partial x}B_2 \tilde {\pmb W} + \dfrac{\partial }{\partial y}B_3 \tilde {\pmb W} =\pmb 0(8)$$

其中$\tilde {\pmb W} = \left[ {\tilde {\rho }\left( {x,y} \right),\tilde {u}\left( {x,y} \right),\tilde {v}\left( {x,y} \right),\tilde {p}\left( {x,y} \right)} \right]^{\rm T}$. 通过求解式(8)得到$\tilde {\pmb W}$进而得到弱扰动量$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb W}}$,最后进行时均量与弱扰动量的叠加得到非定常流场的时域响应.

为了验证此求解方法的精度和效率, Hall和 Crawley (1989)采用亚声速翼型俯仰算例,其压力对比如 图1所示. 从 图1可以看出, 在满足假设条件前提下,当前频域计算方法与传统时域求解方法吻合很好, 最大误差不超过0.01,而效率提高了几十倍.

图 1   翼型表面压力系数$C_p$ 对比 (Hall & Crawley 1989)

   

Hall和Clark (1991) 针对守恒量进行谐波展开, Hall和 Crawley (1994)针对跨声速流动运用激波捕获法并套用Lax-Wendroff格式推导出相应的公式如下

$$\left( {1 - {\rm e}^{{\rm j}w\Delta t}} \right)u_i - \dfrac{\Delta t}{2\Delta x}\left[ {\left. {\dfrac{\partial F}{\partial U}} \right|_{i + 1} u_{i + 1} - \left. {\dfrac{\partial F}{\partial U}} \right|_{i - 1} u_{i - 1} } \right]+ $$ \n \quad $$ \dfrac{\Delta t^2}{2\Delta x^2}\left[ {\left. {\dfrac{\partial F}{\partial U}} \right|_{i + 1 / 2} \left( {\left. {\dfrac{\partial F}{\partial U}} \right|_{i + 1} u_{i + 1} - \left. {\dfrac{\partial F}{\partial U}} \right|_i u_i } \right) - \left. {\dfrac{\partial F}{\partial U}} \right|_{i - 1 / 2} \left( {\left. {\dfrac{\partial F}{\partial U}} \right|_i u_i - \left. {\dfrac{\partial F}{\partial U}} \right|_{i - 1} u_{i - 1} } \right)} \right]+$$ \n \quad $$ \dfrac{\Delta t^2}{2\Delta x^2}\left[ {\left. {\dfrac{\partial ^2F}{\partial U^2}} \right|_{i + 1 / 2} u_{i + 1 / 2} \left( {F_{i + 1} - F_i } \right) - \left. {\dfrac{\partial ^2F}{\partial U^2}} \right|_{i - 1 / 2} u_{i - 1 / 2} \left( {F_i - F_{i - 1} } \right)} \right] = 0(9)$$

Pechloff和Laschka (2006)也发展了一种基于小扰动理论的线性谐波法来计算非定常RANS方程.在此之后Widhalm等 (2010)对于守恒量进行时均值与扰动分量的叠加并通过自动微分法求解扰动方程中的雅可比矩阵,并取名为线化频域法 (linearized frequency domain solver, LFD).

线性频域谐波方法虽然很大程度上减少了计算量,但是仅仅局限于小扰动弱非线性的周期性流动, 局限性较强从而限制了其应用,而对于一些复杂的周期性流动, 就必须考虑流动的非线性. 因此,就有了非线性频域谐波方法的发展.

3.2 非线性频域谐波方法

为了能够将频域方法应用于非线性较强的周期性流动问题, Ning 和He(1997) 基于准三维非定常欧拉方程提出了非线性谐波法,在线性谐波法时均方程中引入额外的非定常应力项,其时均方程如式(10)所示

$$\left.\begin{array}{l} \dfrac{\partial \bar{\pmb F}}{\partial x} + \dfrac{\partial \bar{\pmb G}}{\partial y} =\pmb 0 \\ \bar{\pmb F} = \left[ {{\begin{array}{*{20}c} {\overline {\rho U} - \overline {\rho U_g } } \\ {\bar {U}\left( {\overline {\rho U} - \overline {\rho U_g } } \right) + \bar {P} + \overline {\left( {\widehat{\rho U}} \right)\widehat{U}} - \overline {\left( {\widehat{\rho U_g }} \right)\widehat{U}} } \\ {\bar {V}\left( {\overline {\rho U} - \overline {\rho U_g } } \right) + \overline {\left( {\widehat{\rho U}} \right)\widehat{V}} - \overline {\left( {\widehat{\rho U_g }} \right)\widehat{V}} } \\ {\overline H \left( {\overline {\rho U} - \overline {\rho U_g } } \right) + \bar {P}\overline U _g + \overline {\widehat{H}\left( {\widehat{\rho U}} \right)} + \overline {\widehat{P}\widehat{U_g }} - \overline {\widehat{H}\left( {\widehat{\rho U_g }} \right)} } \\ \end{array} }} \right] \\ \bar{\pmb G} = \left[ {{\begin{array}{*{20}c} {\overline {\rho V} - \overline {\rho V_g } } \\ {\bar {U}\left( {\overline {\rho V} - \overline {\rho V_g } } \right) + \bar {P} + \overline {\left( {\widehat{\rho V}} \right)\widehat{U}} - \overline {\left( {\widehat{\rho V_g }} \right)\widehat{U}} } \\ {\bar {V}\left( {\overline {\rho V} - \overline {\rho V_g } } \right) + \overline {\left( {\widehat{\rho V}} \right)\widehat{V}} - \overline {\left( {\widehat{\rho V_g }} \right)\widehat{V}} } \\ {\overline H \left( {\overline {\rho V} - \overline {\rho V_g } } \right) + \bar {P}\overline V _g + \overline {\widehat{H}\left( {\widehat{\rho V}} \right)} + \overline {\widehat{P}\widehat{V_g }} - \overline {\widehat{H}\left( {\widehat{\rho V_g }} \right)} } \\ \end{array} }} \right] \\ \end{array}\right\}(10)$$

式中, $U_g $与$V_g $表示网格运动速度.采用四阶龙格库塔将时均方程与扰动方程进行耦合求解, 求解流程如 图2所示.

图 2   非线性谐波法求解流程(Ning & He 1997)

   

通过在时均方程中引入非定常应力项, 在单倍谐波情况下非线性谐波法相比于线性谐波法可以把扰动幅值对流动特征的影响更好地刻画出来, 如 图3所示. 同时非线性谐波法可以通过控制扰动源数目和傅里叶阶数来控制求解精度, 然而求解阶数越高, 计算耗费的时间也就越多. 相比于线性谐波法效率会有所折扣 (Vilmin et al. 2006), 但对于非线性较强的周期性问题非线性谐波法则有更高的精度.

图 3   跨声速槽道流中非定常压力扰动项(Ning & He 1997)

   

He 和 Ning (1998) 引入伪时间项,并将非线性谐波法扩展到二维黏性流动, 应用到叶轮机数值模拟中.之后Chen等 (2000) 将其应用到三维湍流问题,通过对德国宇航院的单级跨声速压气机研究认为二阶谐波是在计算成本和计算精度间最佳折衷.

然而, 随着谐波数增加, 非线性谐波法中通量的频域求解就会变得复杂且计算效率会严重降低. 之后基于非线性谐波法基础上, 2001年McMullen提出了非线性频域法(Non-linear frequency domain method, NLFD) (Mcmullen et al. 2001, 2002, 2006, Mcmullen 2003). 在一个周期内取$N$个采样点, 之后对守恒量和通量均进行傅里叶变换, 并引入伪时间项得到一组定常频域方程. 其频域方程如下

$$ V\dfrac{{\rm d} \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb U}} }{{\rm d}\tau } + j\omega V \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb U}} + \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb R}}_k =\pmb 0(11)$$

式中, $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb U}} $表示守恒量在频域内的分量, $\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb R}} $为残差在频域内分量. 其计算流程如{\bf 图4}所示,对于初始守恒量的谐波分量$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb U}}_k^0 $通过傅里叶逆变换得到守恒量时域值$\pmb U ( t )$,计算对应的时域残差$\pmb R ( t )$并通过傅里叶变换得到其频域分量$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb R}} _k^0 $, 进而通过式(11)求出守恒量的谐波分量$\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\frown$}}\over {\pmb U}} _k^1 $,再进行下一步迭代, 直至收敛.

图 4   非线性频域法求解流程(Mcmullen & Jameson et al. 2001)

   

此算法相比于非线性谐波法的优点在于其在时域内计算残差,可以有效地计算扰动大的周期性问题以及复杂流动问题且计算精度和效率更高.McMullen 和 Jameson (2006)对于非线性频域法的计算效率及其关于减缩频率和迎角等敏感度的研究,结果表明NLFD方法计算效率相比于时域推进法普遍提高了3$\sim $9倍.

4 时域配点法

进入21世纪后, 时域配点方法开始兴起. 最早是Hall等 (2002)提出的时域谐波平衡法,在此之后又出现了时间谱方法、时间谱元法等一系列时间离散格式.这类方法的特点在于通过将非定常流场的时间响应在给定的一组正交多项式系上投影,然后在非定常流场中选$N$个时刻,将非定常流动问题转化为$N$个时刻点相互耦合的定常问题. 这类方法叫做时域配点法.

4.1 时间谱方法

2002年, Hall等 (2002) 提出了应用于流体数值求解的谐波平衡法,将守恒量在频域内进行时间离散后代回时域方程进行求解.将周期性的非定常问题转化为$N$个时刻定常方程的耦合.利用傅里叶变换和逆变换求出谱矩阵, 最终的迭代计算仍然在时域内进行.但谐波平衡法这一概念在较早的计算力学中已经提出.其原理为假设近似解为傅里叶级数的形式代入方程通过各自谐波自相平衡来求得傅里叶系数进而求得原方程的解,是典型的频域方法. 因此, 为了区分, 后续研究将流体中谐波平衡法的概念进行推广,谐波平衡法作为一大类包括频域的谐波平衡法和时域的谐波平衡法.2002年Hall提出的谐波平衡法又被称为时域谐波平衡法或高维谐波平衡法.

2005年Gopinath等 (Gopinath & Jameson 2005, Van der Weide et al. 2005)进一步提出了时间谱方法,并针对Hall提出的谐波平衡法推出了更为简洁的时间谱矩阵的表达式.谱方法的概念是建立一个满足周期性边界条件的完备函数族所构成的谱空间,将方程的解在给定的谱空间上投影后再带回原方程进行求解,离散是在谱空间完成的. 对于足够光滑的物理问题,谱方法能达到很高的精度和效率.但对于某些完备函数族不满足周期性边界条件,则这类完备函数族所构成的谱空间称为伪谱空间,对应的求解方法称为伪谱法. 下面对时间谱方法的公式进行推导.

如果流动具有周期性, 则在每一个计算单元内的守恒量$\pmb U _i$随时间均呈周期性变化. 在一个周期$T$内设置$N$个等时间间距的采样点,若$N$为奇数, 对守恒量$\pmb U _i $进行正向傅里叶变换得

$$\widehat{\pmb U }_i^k \approx \dfrac{1}{N}\sum\limits_{n = 0}^{N - 1} {\pmb U _i^n {\rm e}^{ - {\rm j}k\omega n\Delta t}}(12)$$

对应的傅里叶反变换为

$$ \pmb U _i^n \approx \sum\limits_{k = - (N - 1) / 2}^{(N - 1) / 2} {\widehat{\pmb U }_i^k } {\rm e}^{{\rm j}k\omega n\Delta t}(13)$$

将式(13)对时间求导可得

$$\dfrac{{\rm d}}{{\rm d}t}\pmb U _i^n \approx \sum\limits_{k = - (N - 1) / 2}^{(N - 1) / 2} {jk\omega \widehat{\pmb U }_i^k } {\rm e}^{{\rm j}k\omega n\Delta t}(14)$$

将式(12)代入式(14)中经过化简可得

$$\dfrac{{\rm d}}{{\rm d}t}\pmb U _i^n \approx \sum\limits_{j_n = 0}^{N - 1} {d_n^{j_n } \pmb U _i^{j_n } }(15)$$

其中

$$d_n^{j_n } = \left\{ \begin{array}{ll} \dfrac{2\pi }{T}\cdot \dfrac{1}{2}( - 1)^{n - j_n }{\rm csc }\left(\dfrac{\pi \left( {n - j_n } \right)}{N}\right), \quad & {n \neq j_n } \\ 0, \quad & n = j_n \\ \end{array} \right. $$

若$N$为偶数, 则类似进行计算可得

$$d_n^{j_n } = \left\{ \begin{array}{ll} \dfrac{2\pi }{T}\cdot \dfrac{1}{2}( - 1)^{n - j_n }\cot \left(\dfrac{\pi \left( {n - j_n } \right)}{N}\right) , \quad & n \neq j_n \\ 0, \quad & n = j_n \\ \end{array} \right. $$

于是, 可将含有$N$个采样点在第$n$个时刻的Navier-Stokes (N-S)方程的半离散形式写为

$$\sum\limits_{j_n = 0}^{N - 1} {d_n^{j_n } V_i^{j_n } } \pmb U _i^j + \pmb R _i^n =\pmb 0,\qquad n = 0,1,{2},\cdots ,N - 1(16)$$

在式(16)中引入伪时间项, 则该方程可进一步写为

$$\dfrac{\partial }{\partial \tau _i^n }\left( {V_i^n \pmb U _i^n } \right) + \sum\limits_{j_n = 0}^{N - 1} {d_n^{j_n } V_i^j } \pmb U _i^j + \pmb R _i^n =\pmb 0,\qquad n = 0,1,{2},\cdots ,N - 1(17)$$

如果对伪时间项进行显示离散,为保证计算稳定性, 伪时间步长须限制在

$$\tau _i^n = CFL\dfrac{V_i^n }{\left\| \lambda \right\| + V_i^n k'}(18)$$

式中, $\lambda $为谱半径,$k'$为最大谐波数, 具体表达式如下

$$k' = \left\{ \begin{array}{ll} \dfrac{\pi N}{T}, \quad & {\rm mod} (N,2) = 1 \\ \dfrac{\pi \left( {N - 1} \right)}{T}, \quad & {\rm mod} (N,2) = 0 \\ \end{array} \right.$$

随着无量纲时间步长的减小, 迭代的伪时间步长会非常小,从而严重降低计算效率. 采用隐式格式可以放宽时间步长限制,因此在计算中多采用隐式格式对伪时间项进行离散.对伪时间导数进行一阶向后差分离散得到式(19)所示的迭代格式

$$\pmb A ^n\Delta \pmb U ^n = - \sum\limits_{j_n = 0}^{N - 1} {d_n^{j_n } V^{j_n }} \pmb U ^{j_n } - \pmb R ^n(19)$$

其中

$$\pmb A = \left[ \begin{array}{cccc} \dfrac{V_0 }{\Delta \tau _0 }{\pmb I} + {\pmb J}_0 \quad & {V_1 d_0^1 \pmb I } \quad & \cdots \quad & {V_{N - 1} d_0^{N - 1} \pmb I } \\ {V_0 d_1^0 \pmb I } \quad & {\dfrac{V_1 }{\Delta \tau _1 }\pmb I + {\pmb J}_1 } \quad & \cdots \quad & {V_{N - 1} d_1^{N - 1} \pmb I } \\ \vdots \quad & \vdots \quad & \vdots \quad & \vdots \\ {V_0 d_{N - 1}^0 \pmb I } \quad & {V_1 d_{N - 1}^1 \pmb I } \quad & \cdots \quad & {\dfrac{V_{N - 1} }{\Delta \tau _{N - 1} }\pmb I + {\pmb J}_{N - 1} } \\ \end{array} \right]$$

从系数矩阵$\pmb A $的表达式可以看出,时间谱方法引入了额外的非对角项, 且随着无量纲时间步长减小,非对角项的值也就越大. 这就严重削弱了系数矩阵的对角占优特性,导致了矩阵的病态.常用的伪时间推进格式比如对称高斯赛德尔迭代均要求系数矩阵对角占优,但时间谱方法引入的额外非对角项会破坏系数矩阵的对角占优特性从而导致传统的迭代方法失效.因此, 相关研究者针对这一问题展开了研究,并提出可以通过带预处理的广义极小残值(generalized minimum residual,GMRES)方法进行伪时间推进 (Su & Yin 2010),一方面通过预处理可以改善系数矩阵$\pmb A $的性质,另一方面采用GMRES算法也可以加快迭代法的收敛速度. Mundis 和Mavriplis (2013, 2014, 2015) 针对GMRES算法的预处理进行一系列改进,可以通过并行同时计算2047个采样点, 如 表1所示.

表1   跨声速AGARD5算例时间谱方法GMRES-AF推进的收敛性
(Mundis & Mavriplis 2016)

   

采样点数非线性迭代步BCGS迭代步Krylov向量并行计算时间BCGS CFL
1524913838484.43.0
312490593801073.0
632495154001813.0
1272489153742913.0
2552495634015753.0
51124965940511513.0
76724970740717523.0
102324965940530243.0
153524963540450583.0
204724946739777863.0

新窗口打开

时间谱方法的优势在于将非定常周期性流动数值求解转化为$N$个时刻的全耦合定常流场的求解.对于时间响应光滑及低减缩频率的周期性流动, 其效率可以提高一个量级左右,并且时间谱方法可以实现时间与空间的同时并行计算.以亚声速翼型简谐强迫俯仰振荡为例, 从 图5可以看出只需要5个点就能与BDF吻合,图6给出了不同采样点数情况下时间谱方法的效率. 可以看出, 在这个算例中,相比于传统时域推进法, 时间谱方法在保证精度同时可以将效率提高大约一个量级.

图 5   亚声速算例力矩系数$C_m$ 时间谱方法TS与BDF对比

   

图 6   总计算时间对比

   

时间谱方法的缺点在于随着无量纲时间步长的减小, 会带来严重的矩阵病态问题,计算效率急剧下滑.这也就决定了时间谱方法一般只能适用于减缩频率低于0.4流动的数值求解,且对于由大迎角动态失速等引起的气动力随时间响应的高频分量,基于傅里叶变换的时间谱在进行数值模拟时会有明显的吉布斯效应, 如 图7所示.

图 7   时间谱方法中吉布斯效应$(C_l$ 为升力系数)

   

4.2 时间伪谱方法

继时间谱方法提出之后, 各种时间伪谱方法也发展起来并应用到流体数值求解当中.这些方法与时间谱方法的本质区别在于投影的谱空间不同.谱方法投影的谱空间须具有完备的周期性边界条件, 若不具备该类方法称为伪谱法.比较典型的有时间切比雪夫伪谱法 (Dong et al. 2015) 和时间有限差分法(Leffell 2016).

切比雪夫多项式前六项图形见 图8.

图 8   切比雪夫多项式前六项图形

   

切比雪夫伪谱法的谱矩阵如下所示

$${\pmb D}_{\rm ch} = \tilde {\pmb D}_T \pmb {GD}_T(20)$$

其中

$$\begin{array}{l} G _{ij} = \left\{ \begin{array}{ll} \dfrac{2j}{c_i } , \quad & {\rm if} \quad {i + j} \quad {\rm even} \\ 0, \quad & {\rm otherwise} \\ \end{array} \right. \\ c_i = \left\{ \begin{array}{ll} 2, \quad & {\rm if} \quad {i = 0,N} \\ 1,\quad & {\rm otherwise} \\ \end{array} \right.\\ \end{array}$$

$$\tilde {\pmb D}_T = \left[ {{\begin{array}{*{20}c} {T_0 \left( {t_0 } \right)} \quad & {T_1 \left( {t_0 } \right)} \quad & \cdots \quad & {T_N \left( {t_0 } \right)} \\ {T_0 \left( {t_1 } \right)} \quad & {T_1 \left( {t_1 } \right)} \quad & \cdots \quad & {T_N \left( {t_1 } \right)} \\ \vdots \quad & \vdots \quad & \ddots \quad & \vdots \\ {T_0 \left( {t_N } \right)} \quad & {T_1 \left( {t_N } \right)} \quad & \cdots \quad & {T_N \left( {t_N } \right)} \\ \end{array} }} \right]$$

$$\pmb D _T = \dfrac{2}{N} \left[ {{\begin{array}{*{20}c} {\dfrac{T_0 \left( {t_0 } \right)}{4}} \quad & {\dfrac{T_0 \left( {t_1 }\right)}{2}} \quad & \cdots \quad & {\dfrac{T_0 \left( {t_N }\right)}{4}} \\ {\dfrac{T_1 \left( {t_0 } \right)}{2}} \quad & {T_1 \left( {t_1 } \right)}\quad & \cdots \quad & {\dfrac{T_1 \left( {t_N } \right)}{2}} \\ \vdots \quad & \vdots \quad & \ddots \quad & \vdots \\ {\dfrac{T_N \left( {t_0 } \right)}{4}} \quad & {\dfrac{T_N \left( {t_1 } \right)}{2}} \quad & \cdots \quad & {\dfrac{T_N \left( {t_N } \right)}{4}} \\ \end{array} }} \right]$$

切比雪夫伪谱法的基函数如下:

时间切比雪夫伪谱法的好处在于其可以应用于非周期的非定常流动,但对于周期性非定常流动切比雪夫伪谱法的效果并没有时间谱方法好.

时间有限差分法中比较经典的是时间八阶中心差分法(FDMT8), 其谱矩阵如下所示

$$\dfrac{{\rm d}}{{\rm d}t}\pmb U ^n = \sum\limits_{j_n = 1}^M {d_{j_n }^p } \left( {\pmb U ^{n + j_n } - \pmb U ^{n - j_n }} \right)(21)$$

$$d^{\rm VIII} = \dfrac{\omega N}{2\pi }\left\{ {\dfrac{4}{5},-\dfrac{1}{5},\dfrac{4}{105},-\dfrac{1}{280}} \right\}$$

时间八阶中心差分法的好处在于每个时刻的时间导数仅与相邻的8个点有关,计算时间随采样点数增加近似呈线性增长.对于时间响应有高频分量的流动仍能较好地刻画出来.但缺点在于与时间谱方法相比精度较差,时间八阶中心差分法需要更多的采样点才能达到与时间谱方法相当的精度. 图9为时间有限差分法与时间谱方法的精度对比. 图10为不同采样点数下时间有限差分法与时间谱方法每一步迭代所耗用的时间.从 图10可以看出, 随着采样点数增加,时间有限差分法每一步迭代所耗用的时间基本不变, 而时间谱方法由于是全耦合,因此随着采样点数增加, 每一步迭代所用的时间会显著延长.

图 9   时间有限差分法FDMT与时间谱方法TS精度对比 (Leffell 2016), $\|\pmb q_{\rm ex}-\pmb q\|_2$ 为数值解与精确解的误差二范数

   

图 10   时间有限差分法FDMT与时间谱方法TS每步计算时间对比 (Leffell 2016)

   

4.3 时间谱元法

时间谱元法是一种比较特殊的时间离散手段, Mundis和Mavriplis (2015)首次将该方法应用到欧拉方程的时间离散当中, 并取得了很好的效果.与其他时域配点方法不同的是,时间谱元法需要先对欧拉方程沿时间积分之后进行离散. 具体步骤如下.

对守恒量进行拉格朗日插值可得

$$ \left( {V\pmb U } \right)^{\left( m \right)}\left(\varsigma \right) = \sum\limits_{k = 0}^N {\left( {V\pmb U } \right)} ^{\left( m \right)}\left( {\varsigma _k } \right)\psi _k^{\left( m \right)} \left( \varsigma \right)(22)$$

式中,$\psi _k \left( \varsigma \right)$为拉格朗日基函数,对欧拉方程沿时间进行积分可得

$$\int_{ - 1}^1 {v\left( \zeta \right)} \left[{\dfrac{{\rm d}\left( {V\pmb U } \right)^{\left( m \right)}}{{\rm d}\zeta } + \dfrac{T^{\left( m \right)}}{2}\pmb R \left( {\pmb U ^{\left( m \right)},\dot {x}^{\left( m \right)},\tilde {n}^{\left( m \right)}} \right)} \right]{\rm d}\zeta =\pmb 0(23)$$

其中权函数取为拉格朗日基函数, 通过分步积分可得

$$\left. {\left( {V\pmb U } \right)^{\left( m \right)}\psi _p^{\left( m \right)} } \right|_{ - 1}^1 - \int_{ - 1}^1 {\left( {\left( {V\pmb U } \right)^{\left( m \right)}\dfrac{{\rm d}\psi _p^{\left( m \right)} }{{\rm d}\zeta } - \dfrac{T^{\left( m \right)}}{2}\pmb R \left( {U^{\left( m \right)},\dot {x}^{\left( m \right)},\tilde {n}^{\left( m \right)}} \right)\psi _p^{\left( m \right)} } \right)}{\rm d}\zeta =\pmb 0(24)$$

对于式(24)中积分通过高斯点进行离散, 有

$$\int_{ - 1}^1 \pmb Q {\rm d}\zeta = \sum\limits_{p = 0}^N \pmb Q \left( {\zeta _p } \right)\omega _p(25)$$

整理可得

$$\pmb \varPsi ^{\left( m \right)}\left[ {{\begin{array}{*{20}c} {V\pmb U \left( {\zeta _0 } \right)} \\ \vdots \\ {V\pmb U \left( {\zeta _N } \right)} \\ \end{array} }} \right]^{\left( m \right)} = \pmb I _\omega ^{\left( m \right)} \left[ {{\begin{array}{*{20}c} {\pmb R \left( {\pmb U \left( {\zeta _0 } \right),\dot {\pmb x}\left( {\zeta _0 } \right), \tilde {\pmb n}\left( {\zeta _0 } \right)} \right)} \\ \vdots \\ {\pmb R \left( {\pmb U \left( {\zeta _N } \right), \dot {\pmb x}\left( {\zeta _N } \right), \tilde {\pmb n}\left( {\zeta _N } \right)} \right)} \\ \end{array} }} \right]^{\left( m \right)}(26)$$

$$ \pmb \varPsi ^{\left( m \right)} = \left[ {{\begin{array}{*{20}c} {\left( {\left. {\dfrac{{\rm d}\psi _0 }{{\rm d}\zeta }} \right|_{\zeta _0 } \omega _0 + 1} \right)} & {\left. {\dfrac{{\rm d}\psi _0 }{{\rm d}\zeta }} \right|_{\zeta _1 } \omega _1 } & \cdots & {\left. {\dfrac{{\rm d}\psi _0 }{{\rm d}\zeta }} \right|_{\zeta _{N - 1} } \omega _{N - 1} } & {\left. {\dfrac{{\rm d}\psi _0 }{{\rm d}\zeta }} \right|_{\zeta _N } \omega _N } \\ {\left. {\dfrac{{\rm d}\psi _1 }{{\rm d}\zeta }} \right|_{\zeta _0 } \omega _0 } \! \! \! &\! \! \! {\left. {\dfrac{{\rm d}\psi _1 }{{\rm d}\zeta }} \right|_{\zeta _1 } \omega _1 } \! \! \! &\! \! \! \cdots \! \! \! &\! \! \! {\left. {\dfrac{{\rm d}\psi _1 }{{\rm d}\zeta }} \right|_{\zeta _{N - 1} } \omega _{N - 1} } \! \! \! &\! \! \! {\left. {\dfrac{{\rm d}\psi _1 }{{\rm d}\zeta }} \right|_{\zeta _N } \omega _N } \\ \vdots \! \! \! &\! \! \! \vdots \! \! \! &\! \! \! \ddots \! \! \! &\! \! \! \vdots \! \! \! &\! \! \! \vdots \\ {\left. {\dfrac{{\rm d}\psi _{N - 1} }{{\rm d}\zeta }} \right|_{\zeta _0 } \omega _0 } \! \! \! &\! \! \! {\left. {\dfrac{{\rm d}\psi _{N - 1} }{{\rm d}\zeta }} \right|_{\zeta _1 } \omega _1 } \! \! \! &\! \! \! \cdots \! \! \! &\! \! \! {\left. {\dfrac{{\rm d}\psi _{N - 1} }{{\rm d}\zeta }} \right|_{\zeta _{N - 1} } \omega _{N - 1} } \! \! \! &\! \! \! {\left. {\dfrac{{\rm d}\psi _N }{{\rm d}\zeta }} \right|_{\zeta _N } \omega _N } \\ {\left. {\dfrac{{\rm d}\psi _N }{{\rm d}\zeta }} \right|_{\zeta _0 } \omega _0 } \! \! &\! \! {\left. {\dfrac{{\rm d}\psi _N }{{\rm d}\zeta }} \right|_{\zeta _1 } \omega _1 } \! \! &\! \! \cdots \! \! &\! \! {\left. {\dfrac{{\rm d}\psi _N }{{\rm d}\zeta }} \right|_{\zeta _{N - 1} } \omega _{N - 1} } \! \! &\! \! {\left( {\left. {\dfrac{{\rm d}\psi _N }{{\rm d}\zeta }} \right|_{\zeta _N } \omega _N - 1} \right)} \\ \end{array} }} \right] $$

$$\pmb I _\omega ^{\left( m \right)} = \dfrac{T^{\left( m \right)}}{2}\left[ {{\begin{array}{*{20}c} {\omega _0 } \quad & 0 \quad & \cdots \quad & 0 \quad & 0 \\ 0 \quad & {\omega _1 } \quad & \cdots \quad & 0 \quad & 0 \\ \vdots \quad & \vdots \quad & \ddots \quad & \vdots \quad & \vdots \\ 0 \quad & 0 \quad & \cdots \quad & {\omega _{N - 1} } \quad & 0 \\ 0 \quad & 0 \quad & \cdots \quad & 0 \quad & {\omega _N } \\ \end{array} }} \right] $$

时间谱元法巧妙地结合了谱方法与有限元方法, 对时间进行分块求解,可以应用于任意非定常流动的时间响应.当周期非定常流动时间响应中有高频分量时,时间谱元法的效果比时间谱方法的效果好. 图11 (Mundis & Mavriplis 2015) 给出了时间谱元法的时间精度验证. 从 图11}可以看出, 当$N = 1$时精度为1.85阶, 当$N = 2$时精度为2.42阶,而当$N > 2$时精度并没有明显提升. 因此,如何提升时间谱元法的精度也需要进一步研究.

图 11   时间谱元法精度图

   

5 时域推进与时域配点混合方法

时间谱方法只能应用于周期性非定常流动, 而很多流场并不是周期性的,而是具有准周期的特征, 如机动旋翼、气动弹性响应等.非定常流动的时间响应中含有周期性分量和时变分量但周期性分量占主导,这类流动就叫做准周期流动. 针对这类准周期流动, Yang 等 (2010)提出了混合BDF/TS方法.将准周期流场表示为周期性分量与时变分量的叠加, 如式(27)所示

$$\pmb U ( t ) = \sum\limits_{k = - \frac{N}{2}}^{\frac{N}{2} - 1} {\widehat{\pmb U }_k {\rm e}^{{\rm j}k\frac{2\pi }{T}n\Delta t}} + \overline{ \pmb U} ( t )(27)$$

式中, 等式右端首项为周期性分量, $\overline{ \pmb U} ( t )$表示时变分量. 对于时变分量的处理以一阶精度为例, 表达式如下

$$\overline{ \pmb U} ( t ) = \varphi _{12} ( t )\pmb U ^{m + 1} + \varphi _{11} ( t )\pmb U ^m(28)$$

式中, $\pmb U ^{m + 1}$和$\pmb U ^m$表示该计算周期结束时刻点与初始时刻点.$\varphi _{12} $与$\varphi _{11} $表达式如下

$$\varphi _{12} ( t ) = \dfrac{t - t^m}{T}, \quad \varphi _{11} ( t ) = \dfrac{t^{m + 1} - t}{T}$$

对式(27)中等式右端周期性分量利用3.1节中时间谱方法进行离散,等式右端时变分量用BDF方法进行离散. 因此可得

$$\dfrac{{\rm d}}{{\rm d}t}\left( {\pmb U ^n} \right) = \sum\limits_{j_n = 1}^{N - 1} {d_n^{j_n } } \widetilde{\pmb U }^{j_n } + \varphi' _{12} \left( {t_n } \right)\pmb U ^{m + 1} + \varphi' _{11} \left( {t_n } \right)\pmb U ^m(29)$$

式中,$\widetilde{\pmb U }$为周期性分量. 且有

$$t_j = t_m + \dfrac{j}{N}\left( {t_{m + 1} - t_m } \right), \quad j = 0, \cdots,N - 1(30)$$

整理可得到最终方程形式

$$ \sum\limits_{j_n = 1}^{N - 1} {d_n^{j_n } V^{j_n }\pmb U ^{j_n }} - \left( {\sum\limits_{j_n = 1}^{N - 1} {d_n^{j_n } \varphi _{12} \left( {t_{j_n } } \right)} - \varphi' _{12} \left( {t_n } \right)} \right)V^{m + 1}\pmb U ^{m + 1} $$

$$- \left( {\sum\limits_{j = 1}^{N - 1} {d_n^j \varphi _{11} \left( {t_j } \right)} - \varphi' _{11} \left( {t_n } \right)} \right)V^m\pmb U ^m + \pmb R \left( {\pmb U ^n,\overline {\pmb x} ^n,\overrightarrow {\pmb n} ^n} \right) + {\rm {\bf S}}\left( {\pmb U ^n,\overrightarrow {\pmb n} ^n} \right) =\pmb 0 $$

$$n = 0,1,2 , \cdots,N - 1$$

混合BDF/TS方法与BDF方法的区别在于BDF是以每一个时刻点进行推进,而混合BDF/TS方法是以每一个周期进行推进,在每一个周期内用时间谱方法进行求解.由于对时变分量进行BDF离散时的时间步长为一个周期,这也就决定了混合BDF/TS方法不能保证时变分量的准确度,因此只有时变分量较小且变化平缓时该方法才适用. 图12(Yang & Mavriplis 2015) 和 图13(Mundis & Mavriplis 2013)分别为混合BDF/TS方法计算直升机爬升和气动弹性问题并与BDF进行对比,从图中可看出吻合得很好.虽然混合BDF/TS方法的计算效率相比于传统BDF方法并没有太大的提升,但混合BDF/TS方法具有良好的时空并行特性,可以实现时间与空间的同时并行计算.

图 12   UH60A爬升时混合BDF/TS方法与BDF计算结果对比. (a)单桨叶$x$方向力响应,(b)单桨叶$y$方向力响应, (c) 单桨叶$z$方向力响应, (d) 总升力 (1 bs=4.448 2 N)

   

图 13   气弹响应混合BDF/TS方法与BDF计算结果对比

   

6 时间离散格式特点及选择

前几节主要介绍了各种时间离散格式及其特点, 如 表2所示.对于非定常流动的CFD计算,根据其非定常特征要合理地选择时间离散格式才能达到更好的效果. 图14 给出了不同非定常流动特征所对应的合适的时间离散格式,为在实际工程应用中非定常流动数值计算时间离散格式的选择提供了参考.

表2   不同时间离散格式特点

   

时间离散格式求解域离散基函数优点缺点
时域BDF时域泰勒级数 (后插)(1) 适用于任意 非定常流动(1) 对于周期性流动 计算效率低
推进法(2) 实现容易(2) 时间精度一般最
高二阶
线性频域傅里叶基.(1)计算效率高(1)谐波法适用于小扰动
谐波法
频域非线性频域傅里叶基(2)方程求解稳 定性较高(2) 相同精度下,比时域 配点法计算量大
谐波法谐波法(3) 对程序改动量大
非线性时域/ 频域傅里叶基
频域法
时间谱时域傅里叶基(1)计算效率高(1)随着时间步长减小稳
方法
时域时间伪时域切比雪夫基 泰勒基(中心)(2) 计算精度高定性和效率明显下降 (2)对于响应历程不光滑的 问题很难精确仿真
配点法谱方法(3) 相比于频域法对 程序改动量较小
时间谱时域高斯积分
元法

新窗口打开

图 14   不同非定常流动特征时间离散格式选取参考

   

对于周期性非定常流动来说, 要充分利用其周期性特征.因此频域谐波法和时域配点法更为适合. 对于低减缩频率流场的求解,时域配点法可以通过在时域内各时刻点的耦合求解达到高效高精度的要求,与频域谐波法相比不需要时域与频域来回转换且对程序的改动相对较小,因此应选用时域配点法. 然而对于高减缩频率流场的求解, 时域配点法的矩阵病态会导致计算效率严重下滑,此时时域配点法就失去了其原有的优势,而频域谐波法计算效率随着减缩频率变化没有太大影响,仍然能保证流场求解的高效和高精度要求, 因此应选用频域谐波法.在频域谐波法中, 对于扰动大的问题,采用非线性频域法(NLFD)能够保证更高的精度和效率. 对于小扰动问题,根据脉动载荷的动力学特征的非线性大小决定采用线性或非线性谐波法.在时域配点法中, 则需要预先判断其脉动响应是否光滑,如果脉动响应是光滑的, 则可以采用时间谱方法求解.若脉动响应不光滑或含有高阶谐波分量,传统时间谱方法需要很多时刻点且容易出现吉布斯现象,此时需要采用时间谱元法或时间有限差分法.

对于准周期流动, 其流动既有周期分量也有时变分量, 但周期分量占主导,此时频域谐波法和时间谱方法由于都是基于周期性的傅里叶变换因而失效.但是混合BDF/TS方法以及切比雪夫伪谱法等仍然适用. 在减缩频率较小时,可以选用上述方法, 虽然计算量并没有明显减少,但是其并行性相比于时域推进法有较大的改善. 当减缩频率较大时,同样是由于系数矩阵病态的原因, 选用时域推进法的计算效率会更高.

对于非周期流动, 由于其时变分量占主导, 周期性特征已经很弱.频域谐波法已然失效,时域配点法中混合BDF/TS方法以及伪谱方法等虽然有效但其计算精度和效率会明显下滑甚至不如时域推进法.因此, 对于非周期无规则脉动响应, 传统的时域推进法具有优势.

以上是针对强迫运动中减缩频率已知的流动. 而对于非定常流动中的稳定性问题,其流动频率不是事先给定的, 而是待求解的参数.无论是频域谐波法还是时域配点法都需要显式表达流动频率,因此需要对流动频率进行搜索, 而搜索的效率与给定的初值和算法有关.这方面的选择还需进一步研究.

7 新型时间离散格式在工程中的应用

近年来, 以频域谐波法、时间谱方法等为代表的新型时间离散格式发展非常迅速,并以其高精度和高效率的特点在工程中得到广泛的应用.下面以叶轮机内流、动导数计算、旋翼非定常优化设计和气动弹性动力学仿真为例,阐述上述新型时间离散格式在工程中的应用和效果.

7.1 叶轮机内部流动的数值模拟

叶轮机械内部的流动在本质上是非定常的, 由于相邻叶排之间的相对周向运动,叶轮机械内流可以看成是周期性的. 对于传统的时域推进法来说,没有利用流动的周期性特征, 其计算量庞大. 计算规模越大, 计算的过渡时间越长,从而计算耗时越多, 因此很难在工程设计中应用.

目前在叶轮机的非定常计算中以频域谐波法和时域配点法为主. Chen等(2000)应用非线性谐波法求解叶轮机内流并指出在与传统时域推进法的计算结果差别小于10%的情况下,非线性谐波法可以将效率提高20倍左右. 之后Ekici 和Hall (2007, 2008)将谐波平衡法应用到叶轮机颤振和强迫响应的计算中,效率提高了一个量级以上. He (2008, 2010)改进线性谐波法并应用于叶轮机大尺度分离流动,有很高的计算效率,并总结了傅里叶快速预测在叶轮机中的应用. Frey等(2015)采用谐波平衡法求解叶轮机内流并将线化方法和时域推进法的计算结果进行比较.Gopinath等 (2007)将时间谱方法应用到了二维和三维叶轮机的数值模拟中, {\bf图15}为用时间谱计算叶轮机叶片扭转力矩的数值计算结果, 可以看出当$N> 7$时与BDF计算结果基本吻合. Dufour等 (2012)将时域谐波平衡法应用到叶轮机叶片转子、定子的数值模拟.还有Guédeney等 (2013)、Salles等 (2014) 以及Custer等(2012)均对时域配点法在叶轮机中的应用做了相关研究.

图 15   用时间谱TS与时域推进法BDF计算叶轮机叶片扭转力矩的结果及对比(Gopinath & Van et al. 2007 )

   

国内有Du和 Ning (2012), Su 和 Yuan (2010), Yuan 和 Su (2009) 和王丁喜等 (Rahmati & He et al. 2014, Huang & Wang 2016) 也开展了相关的研究工作. 杜鹏程(Du & Ning 2012)对比时间推进和高维谐波平衡法对涡轮机械中流场的模拟能力,苏欣荣(Su & Yuan 2010)用时间谱隐式求解叶轮机流场并对计算方法的稳定性和无反射边界条件处理方法进行研究.王丁喜对于多排叶栅频域方法的边界条件进行改进同时也对谐波平衡法在叶轮机内流计算中的快速算法进行了研究.也有许多相关研究者对于非线性谐波法在叶轮机内流中的应用进行研究(赵军和刘宝杰 2008), 在此不做赘述.

7.2 飞行器的动导数计算

动导数常规数值计算方法是采用CFD模拟飞行器的长周期微振幅强迫简谐振动.由于预测动导数需要计算周期性非定常气动响应,计算效率是CFD在动导数方面应用的瓶颈.

时域配点法刚好适用于长周期的非定常简谐振荡问题,通过求解一个周期内几个时刻的流场重建整个周期的非定常流动,具有很高的计算精度和效率,从而在动导数计算领域得到广泛的关注和应用. Murman等 (Murman et al.2004, Murman 2012) 将谐波平衡法应用于预测Finner及SDM模型的动导数;Hassan和Sicot (011) 将谐波平衡法应用于动导数的快速预测.Ronch等(Ronch et al. 2010, 2013)对比了时域推进法、线性频域方法和谐波平衡法在数值预测动导数精度、计算效率与消耗内存方面的能力.图16 为NACA0012强迫简谐振动不同时间离散方法的计算效率对比.其中加速比为1对应的是时域推进法. LFD为线性频域法,PMB-HB对应的是Liverpool大学PMB程序中的谐波平衡方法 (Badcock et al.2000); COSA-HB对应的是Glasgow大学COSA程序中的谐波平衡方法(Bonfiglioli et al. 2009).

图 16   NACA0012强迫简谐振动不同时间离散格式的计算效率(Ronch et al. 2010)

   

国内的研究相对较少, 谢立军等 (2015)采用时间谱方法对高超声速HBS标模和超声速Finner标模进行动导数计算,并与时域推进法的精度和效率进行对比.陈琦等采用谐波平衡法对翼型和纯锥的非定常绕流进行数值模拟 (陈琦等 2014a),并预测带翼导弹的俯仰动导数 (陈琦等 2014b).

7.3 旋翼非定常绕流数值模拟与优化设计

对于直升机旋翼非定常绕流数值模拟与设计优化, 传统的时域推进法CFD十分耗时,很难在工程设计中应用. 而直升机在前飞过程中旋翼流场具有明显的周期性特征,这就意味着频域谐波法和时域配点法有明显的优势.

Butsuntorn等 (Butsuntorn & Jameson 2008a, 2008b; Butsuntorn 2008) 将时间谱方法应用到三维直升机旋翼前飞流场的数值模拟,通过引入涡修正将多旋翼桨叶简化为一个旋翼桨叶的流场进行求解,采用时间谱方法效率相比于传统BDF方法效率可提高一个量级. Choi (Choi & Datta 2008, Yi et al. 2015)用时间谱方法求解直升机旋翼流固耦合并分析了时间谱方法的精度和效率.Yang (Yang et al. 2011, 2012)通过混合BDF/TS方法求解直升机做准周期机动飞行时的旋翼流场模拟.

旋翼非定常设计优化也非常耗时,在定常优化设计中广泛应用的伴随方法很难应用到非定常设计优化当中.随着频域谐波法和时域配点法的发展,为伴随方法在非定常设计优化中的应用搭建了桥梁. Thomas等 (2005)提出将伴随方法与二维黏性谐波平衡方程结合计算翼型的网格灵敏度.Nadaraja和Jameson (2007) 提出结合ADjoint方法与非线性频域法对跨声速振荡翼型进行优化. Choi等(2008)将时间谱方法与ADjoint方法结合进行直升机旋翼桨叶优化的梯度求解.Woodgate分别通过时间线化法和时域谐波平衡法与伴随方法结合进行旋翼设计并进行对比(Woodgate & Barakos 2012). 图17给出了谐波平衡法与时域推进法在UH60前飞时旋翼流场对比.关于伴随方法与频域谐波法或时域配点法结合进行非定常优化设计或灵敏度分析还有很多文献(Cherif et al. 2012, Ekici & Beran 2014), 这里不作赘述.

图 17   谐波平衡法与时域推进法计算流场对比. (a) 二阶谐波平衡法,(b) 四阶谐波平衡法,(c) 时域推进法 (Woodgate & Barakos 2012)

   

7.4 气动弹性动力学仿真

对于翼型或机翼的气动弹性问题, 无论是线性颤振还是非线性的极限环,流场的流动均有明显的准周期特征. 在气弹的数值模拟中,如果要获取某一状态下的气弹响应, 则需要采取准周期的时间离散格式;但如果要获取颤振边界, 由于在颤振边界处幅值不变,流场的非定常响应是周期的,因此可以利用周期性的时间离散格式对颤振边界进行搜索.这也是近年来新型时间离散格式在气动弹性动力学仿真中的主要应用.

针对简化的气动力模型, Liu (Liu & Dowell 2004, Liu et al. 2007) 和Lee等 (2005)使用高阶谐波平衡法研究二元机翼的Hopf分岔行为以及预测极限环的振幅与频率,Lau和Cheung (1981)还提出了增量谐波平衡法通过引入增量项来简化方程的处理难度.

对于基于CFD和CSD技术的气动弹性仿真, 如果要获取准周期的气弹响应,需要气动和结构方程均要进行相同的时间离散. Mundis等 (Mundis {\rm \&} Mavriplis 2013, Mundis et al. 2013)利用混合BDF/TS方法求解气动弹性响应如图18 (Mundis & Mavriplis 2013) 所示, 可以看出混合BDF/TS方法与BDF2吻合的很好,同时具有更好的并行性.

图 18   BDF/TS方法与BDF2气弹响应计算结果对比. (a)BDF/TS方法与BDF2气弹响应计算结果对比$(M_\infty = 0.875$, $V_{\rm f} = 0.4)$, (b) BDF/TS方法与BDF2气弹响应计算结果对比$(M_\infty = 0.875$, $V_{\rm f} = 0.537)$, (c)BDF/TS方法与BDF2气弹响应计算结果对比$(M_\infty = 0.875$, $V_{\rm f} = 0.65)$

   

如果要获取颤振边界或极限环响应, 其减缩频率并没有显式给出,采用频域谐波法或时域配点法需要对减缩频率进行搜寻. Thomas 等(2002, 2004)利用谐波平衡法求解流场, 在频域内通过Newton-Raphson迭代获取极限环频率和响应.这种求解方法高效并且有很好的鲁棒性. Gopinath等 (2006)将时间谱方法应用到极限环求解中并与时域推进法和谱元法对比计算精度. 此外,Besem等 (2016) 将谐波平衡法应用于圆柱涡致振动中对锁频现象进行探究.

在国内, 刘南等 (刘南等 2016, 刘南 2016)将高阶谐波平衡法与$v$-$g$法结合进行高效的颤振边界预测,通过高阶谐波平衡法求解器对各个结构模态的小扰动简谐强迫运动进行数值模拟进而获得广义力影响系数矩阵$\pmb F ( {jw} )$, 之后通过$v$-$g$法求出对应的颤振速度和颤振频率.

可见, 时域配点法在航空工程中有较为广泛的应用且效果很好, 因此,时域配点法作为一种新型的时间离散方法有很好的发展前景.

8 结论

本文总结了近年来计算流体力学中时间离散格式的研究进展,并根据离散形式和数学思想的不同本文将非定常时间离散格式分为4大类:时域推进法、频域谐波法、时域配点法和混合方法.主要包括传统时域向后差分法、线性谐波法、非线性谐波法、非线性频域法,以及近年来发展的时间谱方法、混合BDF/TS方法、时间伪谱法以及时间谱元法等.在综述过程中比较了各方法的计算精度、效率和适用范围以及对新发展的时间离散格式在航空航天工程中应用做了概述.

近年来虽然时间离散格式得到了迅速的发展, 其中时域配点法也逐渐趋于多元化.但是时间离散格式目前仍有一些难题有待解决, 也是时间离散格式下一步发展的方向:

(1)新发展的时间离散格式需要确定并拓宽其适用范围.一些时间离散格式的适用范围和效果仍不明确, 比如时间谱元法. 同时,后期发展的方法都有局限性,对于不同的非定常流动要选择对应的时间离散格式.新型时间离散格式的计算稳定性问题也一定程度限制了其应用.因此需要寻找鲁棒性更好的迭代算法并通过预处理解决时域配点法带来的矩阵病态问题.虽然Nathan在这方面做了不少工作, 但距离理想的效果还有一定距离.

(2)对于频率未知的流动稳定性问题,无论是频域谐波法还是时域配点法都需要一个快速搜索流动频率的算法,这方面还需要进一步探索和改进. 对于准周期流动问题,频域谐波法如何应用以及时域配点法如何进一步提升计算的精度和效率也需要研究.

(3)时域配点法在气动优化设计当中的应用.对于非定常问题的优化一直以来是航空领域的一大难题.而时域配点法的出现,使得采用定常优化处理手段解决非定常优化问题成为可能.比如时间谱方法与Adjoint方法结合进行非定常优化设计.因此时域配点法在这一方面的应用有很好的前景.

(4)频域谐波法与时域配点法在多学科耦合中的应用研究. 例如,在气动弹性力学、虚拟飞行等领域中,新型时间离散格式可以精确、高效地获取所需的非定常气动力,并且十分便于非定常气动力的建模.

因此, 尽管近期时间离散格式得到快速发展, 但距离实际工程应用仍有很多工作要做.需要CFD界更多的关注和共同努力,我们相信在不久的将来新型的时间离散格式在复杂非定常流动的数值仿真和优化设计上会发挥巨大的作用.

致 谢

The authors have declared that no competing interests exist.


参考文献

[1] 陈琦, 陈坚强, 谢昱飞, 袁先旭. 2014

a. 谐波平衡法在非定常流场中的应用

. 航空学报, 35: 736-743

DOI      URL      [本文引用: 1]      摘要

=2 case being four times that of the dual time-stepping method. Yet it can reduce the computing time greatly with only about 1/5 in =2 case of that needed by the dual time-stepping method to obtain similar results. Therefore the harmonic balance method is an engineeringly applicable method with bright prospects. For problems with a long period, it has a more obvious advantage.

(Chen Q, Chen J Q, Xie Y F, et al.2014

a. Application of harmonic balance method to unsteady flow field

. Acta Aeronautica et Astronautica Sinica, 35: 736-743).

DOI      URL      [本文引用: 1]      摘要

=2 case being four times that of the dual time-stepping method. Yet it can reduce the computing time greatly with only about 1/5 in =2 case of that needed by the dual time-stepping method to obtain similar results. Therefore the harmonic balance method is an engineeringly applicable method with bright prospects. For problems with a long period, it has a more obvious advantage.
[2] 陈琦, 陈坚强, 袁先旭, 谢昱飞. 2014

b. 谐波平衡法在动导数快速预测中的应用研究

. 力学学报, 46: 183-190

DOI      URL      [本文引用: 1]      摘要

Harmonic balance method is constructed on the base of the Fourier series expansion. In the method, the unsteady solution process of a periodically unsteady flow is transformed into coupled solution processes of several steady flows, and then the whole time history of the unsteady aerodynamic forces and moments are rebuilt from the steady results. A rapid prediction method for dynamic stability derivatives is established, utilizing the harmonic balance method. In this way, the dynamic supersonic flow around a pitching winged missile is numerically simulated, with the dynamic derivative in pitch obtained via integration method. The results agree well with those from experiments. The computational efficiency could achieve about thirteen times of that of the dual-time-stepping method, while the order of computational accuracy being the same. The harmonic balance method is then utilized to investigate the effects of reduced frequency on the computation of dynamic derivative in pitch within a broad range. The results show that, with the decrease of the reduced frequency, the variation of the dynamic derivative in pitch is great, and in some cases even sign change could happen. At last, the applicability of the present method at high angles of attack flow is also studied, aiming at the strong nonlinearity of dynamic flow fields in this case. Results show that good agreement could be achieved with the wind tunnel data.

(Chen Q, Chen J Q, Yuan X X, et al.2014

b. Application of a harmonic balance method in rapid predictions of dynamic stability derivatives

. Chinese Journal of Theoretical and Applied Mechanics, 46: 183-190).

DOI      URL      [本文引用: 1]      摘要

Harmonic balance method is constructed on the base of the Fourier series expansion. In the method, the unsteady solution process of a periodically unsteady flow is transformed into coupled solution processes of several steady flows, and then the whole time history of the unsteady aerodynamic forces and moments are rebuilt from the steady results. A rapid prediction method for dynamic stability derivatives is established, utilizing the harmonic balance method. In this way, the dynamic supersonic flow around a pitching winged missile is numerically simulated, with the dynamic derivative in pitch obtained via integration method. The results agree well with those from experiments. The computational efficiency could achieve about thirteen times of that of the dual-time-stepping method, while the order of computational accuracy being the same. The harmonic balance method is then utilized to investigate the effects of reduced frequency on the computation of dynamic derivative in pitch within a broad range. The results show that, with the decrease of the reduced frequency, the variation of the dynamic derivative in pitch is great, and in some cases even sign change could happen. At last, the applicability of the present method at high angles of attack flow is also studied, aiming at the strong nonlinearity of dynamic flow fields in this case. Results show that good agreement could be achieved with the wind tunnel data.
[3] 刘南, 白俊强, 刘艳, 华俊.2016.

基于谐波平衡法和V-g法的高效颤振预测分析(英文)

.空气动力学学报, 34: 631-637

[本文引用: 1]      摘要

提出了一种高效的颤振预测分析方法,将之应用于跨声速颤振边界分析及结构参数影响研究中。本方法基于频域颤振分析 V-g 方法,为气弹系统提供一定的人工阻尼使之保持简谐运动状态,从而将结构动力学方程转换到频域内。然后通过高效的谐波平衡法得到一系列简谐运动频率下的气动力描述函数矩阵,结合频域结构方程,将气弹系统的稳定性问题转化为广义特征值求解。结果表明:本方法计算得到的颤振边界与高精度的时间推进方法非常吻合,分析效率有了明显的提升,而且当结构参数发生变化后,只需进行若干次广义特征值求解即可得到新的颤振边界,无需像时间推进方法一样开展大量的气动/结构耦合数值模拟。

(Liu N, Bai J Q, Liu Y, et al.2016.

Efficient flutter prediction based on harmonic balance and V-g methods

. Acta Aerodynamica Sinica, 34: 631-637).

[本文引用: 1]      摘要

提出了一种高效的颤振预测分析方法,将之应用于跨声速颤振边界分析及结构参数影响研究中。本方法基于频域颤振分析 V-g 方法,为气弹系统提供一定的人工阻尼使之保持简谐运动状态,从而将结构动力学方程转换到频域内。然后通过高效的谐波平衡法得到一系列简谐运动频率下的气动力描述函数矩阵,结合频域结构方程,将气弹系统的稳定性问题转化为广义特征值求解。结果表明:本方法计算得到的颤振边界与高精度的时间推进方法非常吻合,分析效率有了明显的提升,而且当结构参数发生变化后,只需进行若干次广义特征值求解即可得到新的颤振边界,无需像时间推进方法一样开展大量的气动/结构耦合数值模拟。
[4] 刘南. 2016.

机翼跨声速非线性颤振及高校分析方法研究. [博士论文]

. 西安: 西北工业大学

[本文引用: 1]     

(Liu N.2016.

Investigation of transonic nonlinear flutter and efficient analysis approach. [PhD Thesis]

. Xi'an: Northwestern Polytechnical University).

[本文引用: 1]     

[5] 谢立军, 杨云军, 刘周, 周伟江. 2015.

基于时间谱方法的飞行器动导数高效计算技术

. 航空学报, 36: 2016-2026

DOI      URL      [本文引用: 1]      摘要

The engineering application requires the solver of periodic unsteady flows to be of high efficiency and precision. Based on solving the Reynolds-averaged Navier-Stokes (RANS) equations, a fully implicit version of the time spectral method (TSM) is established and the stability problem that occurs as the sampling number grows is improved. The Menter shear-stress transport (SST) turbulence model is discretized with the TSM, which makes the TSM more applicable to practical engineering problems. The TSM is applied to simulate the forced oscillation of NACA0015. The calculated results are in good agreement with the experimental data and the results of the dual time stepping (DTS) method, validating the ability of the TSM to simulate the periodic motion. The hypersonic HBS and the supersonic Finner are taken as typical examples of computing dynamic derivatives using the TSM. Influences of the angle of attack and Mach number on dynamic derivatives are analyzed. Results show that the TSM can acquire the same order of accuracy compared with the DTS method. But the relative efficiency of the TSM improves as the Mach number increases. The efficiency advantage in hypersonic range could be up to an order of magnitude.

(Xie L J, Yang Y J, Liu Z, Zhou W J.2015.

A high efficient method for computing dynamic derivatives of aircraft based on time spectral method

. Acta Aeronautica et Astronautica Sinica, 36: 2016-2026).

DOI      URL      [本文引用: 1]      摘要

The engineering application requires the solver of periodic unsteady flows to be of high efficiency and precision. Based on solving the Reynolds-averaged Navier-Stokes (RANS) equations, a fully implicit version of the time spectral method (TSM) is established and the stability problem that occurs as the sampling number grows is improved. The Menter shear-stress transport (SST) turbulence model is discretized with the TSM, which makes the TSM more applicable to practical engineering problems. The TSM is applied to simulate the forced oscillation of NACA0015. The calculated results are in good agreement with the experimental data and the results of the dual time stepping (DTS) method, validating the ability of the TSM to simulate the periodic motion. The hypersonic HBS and the supersonic Finner are taken as typical examples of computing dynamic derivatives using the TSM. Influences of the angle of attack and Mach number on dynamic derivatives are analyzed. Results show that the TSM can acquire the same order of accuracy compared with the DTS method. But the relative efficiency of the TSM improves as the Mach number increases. The efficiency advantage in hypersonic range could be up to an order of magnitude.
[6] 袁新, 苏欣荣. 2009.

谱方法用于非定常流动计算的隐式求解

. 工程热物理学报, 30: 2010-2012

DOI      URL      [本文引用: 1]      摘要

本文对谱方法用于周期性非定常流动的隐式求解方法进行了探讨,分析了影响计算稳定性和收敛速度的因素.提出了结合多重网格的隐式求解方法并对算法进行了验证,初步计算表明本文算法具有良好的稳定性和收敛速度.对于周期性非定常流动,结合本文提出的隐式求解的时域谱方法可以达到很高的精度且具有良好的计算效率.

(Yuan X, Su X Y.2009.

Implicit solution of time spectral method for periodic unsteady flows

. Journal of Engineering Thermophysics, 30: 2010-2012).

DOI      URL      [本文引用: 1]      摘要

本文对谱方法用于周期性非定常流动的隐式求解方法进行了探讨,分析了影响计算稳定性和收敛速度的因素.提出了结合多重网格的隐式求解方法并对算法进行了验证,初步计算表明本文算法具有良好的稳定性和收敛速度.对于周期性非定常流动,结合本文提出的隐式求解的时域谱方法可以达到很高的精度且具有良好的计算效率.
[7] 赵军, 刘宝杰. 2008.

非线性谐波法的进一步校验分析

. 航空动力学报, 23: 680-686

URL      [本文引用: 1]      摘要

基于定常Denton程序发展 了二维定常、非定常、非线性谐波法计算程序,以二维跨声压气机叶栅为例,通过详细对比其定常、非定常和非线性谐波法的计算结果,包括确定应力场和叶片表面 速度分布等,验证了非线性谐波法具有建模精度较高的优点的同时,也分析了该方法的不足,包括一些难以克服的主要误差来源,以及计算时间较长等弱点;并指出 了掺混面处理对于非线性谐波法的重要性.

(Zhao J, Liu B J.2008.

Further verification of nonlinear harmonic method

. Journal of Aerospace Power, 23: 680-686).

URL      [本文引用: 1]      摘要

基于定常Denton程序发展 了二维定常、非定常、非线性谐波法计算程序,以二维跨声压气机叶栅为例,通过详细对比其定常、非定常和非线性谐波法的计算结果,包括确定应力场和叶片表面 速度分布等,验证了非线性谐波法具有建模精度较高的优点的同时,也分析了该方法的不足,包括一些难以克服的主要误差来源,以及计算时间较长等弱点;并指出 了掺混面处理对于非线性谐波法的重要性.
[8] Badcock K J, Richards B E, Woodgate M A.2000.

Elements of computational fluid dynamics on block structured grids using implicit solvers

. Progress in Aerospace Sciences, 36: 351-392.

DOI      URL      [本文引用: 1]      摘要

This paper reviews computational fluid dynamics (CFD) for aerodynamic applications. The key elements of a rigorous CFD analysis are discussed. Modelling issues are summarised and the state of modern discretisation schemes considered. Implicit solution schemes are discussed in some detail, as is multiblock grid generation. The cost and availability of computing power is described in the context of cluster computing and its importance for CFD. Several complex applications are then considered in light of these simulation components. Verification and validation is presented for each application and the important flow mechanisms are shown through the use of the simulation results. The applications considered are: cavity flow, spiked body supersonic flow, underexpanded jet shock wave hysteresis, slender body aerodynamics and wing flutter. As a whole the paper aims to show the current strengths and limitations of CFD and the conclusions suggest a way of enhancing the usefulness of flow simulation for industrial class problems.
[9] Beam R M, Warming R F.1976.

An implicit finite-difference algorithm for hyperbolic systems in conservation-law form

. Journal of Computational Physics, 22: 87-110.

DOI      URL      [本文引用: 1]      摘要

An implicit finite-difference scheme is developed for the efficient numerical solution of nonlinear hyperbolic systems in conservation law form. The algorithm is second-order time-accurate, noniterative, and in a spatially factored form. Second- or fourth-order central and second-order one-sided spatial differencing are accommodated within the solution of a block tridiagonal system of equations. Significant conceptual and computational simplifications are made for systems whose flux vectors are homogeneous functions (of degree one), e.g., the Eulerian gasdynamic equations. Conservative hybrid schemes, which switch from central to one-sided spatial differencing whenever the local characteristic speeds are of the same sign, are constructed to improve the resolution of weak solutions. Numerical solutions are presented for a nonlinear scalar model equation and the two-dimensional Eulerian gasdynamic equations.
[10] Besem F M, Thomas J P, Kielb R E, Dowell E H.2016.

An aeroelastic model for vortex-induced vibrating cylinders subject to frequency lock-in

. Journal of Fluids & Structures, 61: 42-59.

DOI      URL      [本文引用: 1]      摘要

61We characterized the shedding patterns for transverse and in-line lock-in regions.61We used a fluid-structure interaction model with a frequency domain CFD code.61The transverse lock-in region is more accurately captured and matches experiments.61The in-line amplitude is insignificant except at very low mass and damping.61For very low mass ratios, it is necessary to model both degrees of freedom.
[11] Bonfiglioli A, Carpentieri B, Campobasso S.2009.

Parallel unstructured three-dimensional turbulent flow analyses using efficiently preconditioned Newton-Krylov Solvers//19th AIAA Computational Fluid Dynamics, San Antonio

.

DOI      URL      [本文引用: 1]      摘要

No abstract available.
[12] Briley W R, Mcdonald H.1975.

Solution of the three-dimensional compressible Navier Stokes equations by an implicit technique

. Lecture Notes in Physics, 35: 105-110.

DOI      URL      [本文引用: 2]     

[13] Brunton S L, Rowley C W.2013.

Empirical state-space representations for Theodorsen's lift model

. Journal of Fluids & Structures, 38: 174-186.

DOI      URL      [本文引用: 1]      摘要

In this work, we cast Theodorsen's unsteady aerodynamic model into a general form that allows for the introduction of empirically determined quasi-steady and added-mass coefficients as well as an empirical Theodorsen function. An empirically determined Theodorsen model is constructed using data from direct numerical simulations of a flat plate pitching at low Reynolds number, Re=100. Next, we develop low-dimensional, state-space realizations that are useful for either the classical Theodorsen lift model or the empirical model. The resulting model is parameterized by pitch-axis location and has physically meaningful states that isolate the effect of added-mass and quasi-steady forces, as well as the effect of the wake. A low-order approximation of Theodorsen's function is developed based on balanced truncation of a model fit to the analytical frequency response, and it is shown that this approximation outperforms other models from the literature. We demonstrate the utility of these state-space lift models by constructing a robust controller that tracks a reference lift coefficient by varying pitch angle while rejecting gust disturbances.
[14] Butsuntorn N, Jameson A.2008

a. Time spectral method for rotorcraft flow. [PhD Thesis]

. Stanford CA: Stanford University.

[本文引用: 2]     

[15] Butsuntorn N, Jameson A.2008

b. Time spectral method for rotorcraft flow with vorticity confinement

//26th AIAA Applied Aerodynamics Conference.

[本文引用: 1]     

[16] Butsuntorn N.2008.

Time spectral method for rotorcraft flow with vorticity confinement. [PhD Thesis]

. California: Stanford University.

[17] Butcher J C.2016.

Numerical Methods for Ordinary Differential Equations

. (First Edition). John Wiley & Sons.

URL      [本文引用: 1]      摘要

Features a simplified presentation of numerical methods by introducing and implementing SAGE programsAn Introduction to SAGE Programming: With Applications to SAGE Interacts for Numerical Methods emphasizes how to implement numerical methods using SAGE Math and SAGE Interacts and also addresses the fundamentals of computer programming, including if statements, loops, functions, and interacts. The book also provides a unique introduction to SAGE and its computer algebra system capabilities, discusses second and higher order equations and estimate limits, and determines derivatives, integrals, and summations. Providing critical resources for developing successful interactive SAGE numerical computations, the book is accessible without delving into the mathematical rigor of numerical methods.The author illustrates the benefits of utilizing the SAGE language for calculus and the numerical analysis of various methods such as bisection methods, numerical integration, Taylor's expansions, and Newton's iterations. Providing an introduction to the terminology and concepts involved, An Introduction to SAGE Programming: With Applications to SAGE Interacts for Numerical Methods features:* An introduction to computer programming using SAGE* Many practical examples throughout to illustrate the application of SAGE Interacts for various numerical methods* Discussions on how to use SAGE Interacts and SAGE Cloud in order to create mathematical demonstrations* Numerous homework problems and exercises that allow readers to practice their programming skillset* A companion website that includes related SAGE programming code and select solutions to the homework problems and exercisesAn Introduction to SAGE Programming: With Applications to SAGE Interacts for Numerical Methods is an ideal reference for applied mathematicians who need to employ SAGE for the study of numerical methods and analysis. The book is also an appropriate supplemental textbook for upper-undergraduate and graduate-level courses in numerical methods.
[18] Chen T, Vasanthakumar P, He L.2000.

Analysis of unsteady blade row interaction using nonlinear harmonic approach

.Journal of Propulsion & Power, 17: 651-658.

DOI      URL      [本文引用: 2]      摘要

Abstract An efficient non-linear harmonic methodology has been developed for predicting unsteady blade row interaction effects in multistage axial flow compressors. Flow variables are decomposed into time averaged variables and unsteady perturbations, resulting in time averaged equations with deterministic stress terms depending on the unsteady perturbation. The non-linear interaction between the time averaged flow field and the unsteady perturbations are included by a simultaneous pseudotime integration approach, leading to a strongly coupled solution. The stator/rotor interface treatment follows a flux averaged characteristic based mixing plane approach and includes the deterministic stress terms due to upstream running potential disturbances and downstream running wakes, resulting in the continuous nature of all parameters across the interface. The basic computational methodology is applied to the three-dimensional Navier-Stokes equations and validated against several cases. Results show that this method is much more efficient than the non-linear time-marching methods while still modeling the nonlinear unsteady blade row interaction effects.
[19] Cherif M A, Emamirad H, Mnif M.2012.

Derivatives for time-spectral computational fluid dynamics using an automatic differentiation adjoint

. AIAA Journal, 50: 2809-2819.

DOI      URL      [本文引用: 1]      摘要

Abstract In computational fluid dynamics, for problems with periodic flow solutions, the computational cost of spectral methods is significantly lower than that of full, unsteady computations. As is the case for regular steady-flow problems, there are various interesting periodic problems, such as those involving helicopter rotor blades, wind turbines, or oscillating wings, that can be analyzed with spectral methods. When conducting gradient-based numerical optimization for these types of problems, efficient sensitivity analysis is essential. The authors developed an accurate and efficient sensitivity analysis for time-spectral computational fluid dynamics. By combining the cost advantage of the spectral-solution methodology with an efficient gradient computation, the total cost of optimizing periodic unsteady problems can be significantly reduced. The efficient gradient computation takes the form of an automatic differentiation discrete adjoint method, which combines the efficiency of an adjoint method with the accuracy and rapid implementation of automatic differentiation. To demonstrate the method, the authors computed sensitivities for an oscillating ONERA M6 wing. The sensitivities are shown to be accurate to 8-12 digits, and the computational cost of the adjoint computations is shown to scale well up to problems of more than 41 million state variables.
[20] Choi S, Lee K H, Alonso J J, Datta A.2008.

Preliminary study on time-spectral and adjoint-based design optimization of helicopter rotors//AHS specialist meeting, San Francisco, CA

.

[本文引用: 3]     

[21] Choi S, Datta A.2008.

CFD prediction of rotor loads using time-spectral method and exact fluid-structure interface//26th AIAA Applied Aerodynamics Conference

.

[22] Choi S, Lee K, Potsdam M M, Alonso J J.2008.

Helicopter rotor design using a time-spectral and adjoint-based method//Aiaa/issmo Multidisciplinary Analysis and Optimization Conference

, 412-423.

[23] Custer C H, Weiss J M, Subramanian V, Clark W S, Hall K C.2012.

Unsteady simulation of a 1.5 stage turbine using an implicitly coupled nonlinear harmonic balance method

//ASME Turbo Expo: Turbine Technical Conference and Exposition, 2303-2317.

[本文引用: 1]     

[24] Dinu A D, Botez R, Cotoi I.2006.

Chebyshev polynomials for unsteady aerodynamic calculations in aeroservoelasticity

.Journal of Aircraft, 43: 165-171.

DOI      URL      [本文引用: 1]      摘要

The aerodynamic forces in the reduced frequency domain have to be approximated in the Laplace domain, in order to study the effects of the control laws on the flexible aircraft structure. This type of approximation of unsteady generalized aerodynamic forces from the frequency domain into the Laplace domain, acting on a Fly-By-Wire aircraft, presents an important challenge in the aeroservoelasticity area. In this paper we present a new method for approximation of the generalized aerodynamic forces, using Chebyshev polynomials and their orthogonality properties. A comparison of results expressed in terms of flutter speeds and frequencies obtained by this new method and by the Pad e method is presented here. This new approximation method gives excellent results with respect to Pad e method and is validated on the aircraft test model from NASA Dryden Flight Research Center.
[25] Dong K I, Choi S, Mcclure J E, Skiles F.2015.

Mapped Chebyshev pseudospectral method for unsteady flow analysis

. AIAA Journal, 53: 1-16.

DOI      URL      [本文引用: 1]     

[26] Du P, Ning F.2012.

The development and application of a time-domain harmonic balance flow solver

// ASME 2012 Fluids Engineering Division Summer Meeting Collocated with the ASME 2012 Heat Transfer Summer Conference and the ASME 2012, International Conference on Nanochannels, Microchannels, and Minichannels.

[本文引用: 2]     

[27] Dufour G, Gourdain N, Sicot F.2012.

A time-domain harmonic balance method for rotor/stator interactions

. Journal of Turbomachinery, 134: 11001-11001.

DOI      URL      [本文引用: 1]     

[28] Ekici K, Hall K C.2007.

Nonlinear analysis of unsteady flows in multistage turbomachines using harmonic balance

. AIAA Journal, 45: 1047-1057.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT A harmonic balance technique for the analysis of two-dimensional linear (small-disturbance) and nonlinear unsteady flows in multistage turbomachines is presented. The present method uses a mixed time-domain/frequency-domain approach that allows one to compute the unsteady aerodynamic response of multistage machines to both blade vibration (the flutter problem) and wake interaction (the forced response problem). In general, the flowfield may have multiple excitation frequencies that are not integer multiples of each other, so that the unsteady flow is (sometimes) aperiodic in time. Using our approach, we model each blade row using a computational grid spanning a single blade passage. In each blade row, we store several subtime level solutions. For flows that are periodic in time, these sublime levels span a single time period. For aperiodic flows, the temporal period spanned by these sublime level solutions is sufficiently long to sample the relevant discrete frequencies contained in the aperiodic How. In both cases, these subtime level solutions are related to each other through the time-derivative terms in the Enter or Navier-Stokes equations and boundary conditions; complex periodicity conditions connect the subtime levels within a blade passage, and interrow boundary conditions connect the solutions among blade rows. The resulting discretized equations, which are mathematically steady because. time derivatives have been replaced by a pseudospectral operator in which the excitation frequencies appear as parameters, can be solved very efficiently using multigrid acceleration techniques. In this paper, we apply the technique to both flutter and wake-interaction problems and illustrate the influence of neighboring blade rows on the unsteady aerodynamic response of a blade row.
[29] Ekici K, Hall K C.2008.

Nonlinear frequency-domain analysis of unsteady flows in turbomachinery with multiple excitation frequencies

. AIAA Journal, 46: 1912.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT A harmonic balance technique for the analysis of nonlinear unsteady flows in turbomachiney arising from several excitation sources, possibly with frequencies with irrational ratios, is presented in this paper. This method uses a mixed time-domain/frequency-domain approach that allows one to model the blade row on a computational grid spanning just a single blade passage, no matter the interblade phase angles of the original disturbances. Using this approach, we compute several solutions, each one corresponding to one of several times over one period for periodic flows, or over a representative time interval for aperiodic flows. These time levels are coupled to each other through a spectral time derivative operator in the interior of the computational domain, and through the far-field and periodic boundary conditions around the boundary of the domain. In this paper, we apply the method to the two-dimensional Euler equations (although the method can be applied to three-dimensional and viscous flows), and examine the nonlinear interaction of wake passing with blade vibration.
[30] Ekici K, Beran P.2014.

Adjoint sensitivity analysis of low-speed flows using an efficient harmonic balance technique

. AIAA Journal, 52: 1330-1336.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT Computational fluid dynamics (CFD) has become a mainstream tool to help develop a deeper understanding of complex flow phenomena in engineering. Especially, in the field of aerospace engineering, the use of CFD has proven to be instrumental in advancing the state of the art, and it will continue to play an important role in the design of future technologies that are more reliable and efficient. Despite the advancement of computational engineering, computational aeroelasticity and aerodynamics of micro air vehicles (MAVs) continue to be one of the most challenging problems because of the complex fluid structure interaction phenomena involved. Sensitivity information of design variables to a given objective function can be used to step toward a point in the design space for optimization. In the case of aerodynamic optimization, the objective function may be the optimum power for flapping wings, and the design variables may be the surface points that define the geometry.
[31] Farhat C, Lesoinne M.2000.

Two efficient staggered algorithms for the serial and parallel solution of three-dimensional nonlinear transient aeroelastic problems

. Computer Methods in Applied Mechanics and Engineering, 182: 499-515.

DOI      URL      [本文引用: 1]      摘要

Partitioned procedures and staggered algorithms are often adopted for the solution of coupled fluid/structure interaction problems in the time domain. In this paper, we overview two sequential and parallel partitioned procedures that are popular in computational nonlinear aeroelasticity, and address their limitation in terms of accuracy and numerical stability. We propose two alternative serial and parallel staggered algorithms for the solution of coupled transient aeroelastic problems, and demonstrate their superior accuracy and computational efficiency with the flutter analysis of the AGARD Wing 445.6. We contrast our results with those computed by other investigators and validate them with experimental data.
[32] Frey C, Ashcroft G, Kersken H P.2015.

Simulations of unsteady blade row interactions using linear and non-linear frequency domain methods//ASME Turbo Expo 2015: Turbine Technical Conference and Exposition, V02BT39A037

.

[本文引用: 1]     

[33] Gopinath A K, Beran P S, Jameson A.2006.

Comparative analysis of computational methods for limit-cycle oscillations

. AIAA Paper, 2076: 1-4.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT Various methods are explored in the computation of time-periodic solutions for au- tonomous systems. The purpose of the work is to illuminate the capabilities and limitations of methods, including a new method developed as part of the work, not based on time inte- gration for the fast computation of limit-cycle oscillations (LCO). Discussion will focus on methodology, robustness, accuracy, and frequency prediction. Results for a model problem are shown in which temporal discetization errors during LCO are taken to machine zero. Treatment of sharp transients during LCO is also discussed.
[34] Gopinath A K, Jameson A.2005.

Time spectral method for periodic unsteady computations over two-and three-dimensional bodies

. AIAA Paper, 1220: 10-13.

DOI      URL      [本文引用: 2]      摘要

ABSTRACT The Time Spectral Algorithm is proposed for the fast and ecient computation of time- periodic turbulent Navier-Stokes calculations past two- and three-dimensional bodies. The eciency of the approach derives from these attributes: 1)time discretization based on a Fourier representation in time to take advantage of its periodic nature, 2)multigrid to drive a fully implicit time stepping scheme. The accuracy and eciency of this technique is verified by Euler and Navier-Stokes calculations for a pitching airfoil and a pitching wing. Results verify the small number of time intervals per pitching cycle required to capture the flow physics. Spectral convergence is demonstrated for pitching motion characterised by many frequencies.
[35] Gopinath A K, Van Der Weide E, Alonso J J, Jameson A.2007.

Three-dimensional unsteady multi-stage turbomachinery simulations using the harmonic balance technique//45th AIAA Aerospace Sciences Meeting and Exhibit

, 892.

DOI      URL      [本文引用: 2]      摘要

In this paper, we propose an extension of the Harmonic Balance method for threedimensional, unsteady, multi-stage turbomachinery problems modeled by the Unsteady Reynolds-Averaged Navier-Stokes (URANS) equations. This time-domain algorithm simulates the true geometry of the turbomachine (with the exact blade counts) using only one blade passage per blade row, thus leading to drastic savings in both CPU and memory requirements. Modified periodic boundary conditions are applied on the upper and lower boundaries of the single passage in order to account for the lack of a common periodic interval for each blade row. The solution algorithm allows each blade row to resolve a specified set of frequencies in order to obtain the desired computation accuracy; typically, a blade row resolves only the blade passing frequencies of its neighbors. Since every blade row is setup to resolve different frequencies the actual Harmonic Balance solution in each of these blade rows is obtained at different instances in time or time levels. The interaction between blade rows occurs through sliding mesh interfaces in physical time. Space and time interpolation are carried out at these interfaces and can, if not properly treated, introduce aliasing errors that can lead to instabilities. With appropriate resolution of the time interpolation, all instabilities are eliminated. This new procedure is demonstrated using both two-and three dimensional test cases and can be shown to significantly reduce the cost of multi-stage simulations while capturing the dominant unsteadiness in the problem. I.
[36] Guédeney T, Gomar A, Gallard F, Sicot F.2013.

Non-uniform time sampling for multiple-frequency harmonic balance computations

. Journal of Computational Physics, 236: 317-345.

DOI      URL      [本文引用: 1]      摘要

A time-domain harmonic balance method for the analysis of almost-periodic (multi-harmonics) flows is presented. This method relies on Fourier analysis to derive an efficient alternative to classical time marching schemes for such flows. It has recently received significant attention, especially in the turbomachinery field where the flow spectrum is essentially a combination of the blade passing frequencies. Up to now, harmonic balance methods have used a uniform time sampling of the period of interest, but in the case of several frequencies, non-necessarily multiple of each other, harmonic balance methods can face stability issues due to a bad condition number of the Fourier operator. Two algorithms are derived to find a non-uniform time sampling in order to minimize this condition number. Their behavior is studied on a wide range of frequencies, and a model problem of a 1D flow with pulsating outlet pressure, which enables to prove their efficiency. Finally, the flow in a multi-stage axial compressor is analyzed with different frequency sets. It demonstrates the stability and robustness of the present non-uniform harmonic balance method regardless of the frequency set. (C) 2012 Elsevier Inc. All rights reserved.
[37] Guo Y, Keller J, Parker R G.2012.

Dynamic analysis of wind turbine planetary gears using an extended harmonic balance approach: Preprint

. Office of Scientific & Technical Information Technical Reports, 8: 4329-4343.

DOI      URL      [本文引用: 1]      摘要

The dynamics of wind turbine planetary gears with gravity effects are investigated using an extended harmonic balance method that extends established harmonic balance formulations to include simultaneous internal and external excitations. The extended harmonic balance method with arc-length continuation and Floquet theory is applied to a lumped-parameter planetary gear model including gravity, fluctuating mesh stiffness, bearing clearance, and nonlinear tooth contact to obtain the planetary gear dynamic response. The calculated responses compare well with time domain integrated mathematical models and experimental results. Gravity is a fundamental vibration source in wind turbine planetary gears and plays an important role in system dynamics, causing hardening effects induced by tooth wedging and bearing-raceway contacts. Bearing clearance significantly reduces the lowest resonant frequencies of translational modes. Gravity and bearing clearance together lowers the speed at which tooth wedging occurs lower than the resonant frequency.
[38] Hall K C, Clark W S.1991.

Prediction of unsteady aerodynamic loads in cascades using the linearized Euler equations on deforming grids//27th Joint Propulsion Conference

.

[本文引用: 1]     

[39] Hall K C, Crawley E F.1989.

Calculation of unsteady flows in turbomachinery using the linearized Euler equations

. AIAA Journal, 27: 777-787.

DOI      URL      [本文引用: 5]      摘要

A method for calculating unsteady flows in cascades is presented. The model, which is based on the linearized unsteady Euler equations, accounts for blade loading shock motion, wake motion, and blade geometry. The mean flow through the cascade is determined by solving the full nonlinear Euler equations. Assuming the unsteadiness in the flow is small, then the Euler equations are linearized about the mean flow to obtain a set of linear variable coefficient equations which describe the small amplitude, harmonic motion of the flow. These equations are discretized on a computational grid via a finite volume operator and solved directly subject to an appropriate set of linearized boundary conditions. The steady flow, which is calculated prior to the unsteady flow, is found via a Newton iteration procedure. An important feature of the analysis is the use of shock fitting to model steady and unsteady shocks. Use of the Euler equations with the unsteady Rankine-Hugoniot shock jump conditions correctly models the generation of steady and unsteady entropy and vorticity at shocks. In particular, the low frequency shock displacement is correctly predicted. Results of this method are presented for a variety of test cases. Predicted unsteady transonic flows in channels are compared to full nonlinear Euler solutions obtained using time-accurate, time-marching methods. The agreement between the two methods is excellent for small to moderate levels of flow unsteadiness. The method is also used to predict unsteady flows in cascades due to blade motion (flutter problem) and incoming disturbances (gust response problem).
[40] Hall K C, Crawley E F.1994.

A linearized Euler analysis of unsteady flows in turbomachinery

. Massachusetts Inst.of Tech.Report, 116: V001T03A035.

DOI      URL      [本文引用: 1]      摘要

A computational method for efficiently predicting unsteady transonic flows in two- and three-dimensional cascades is presented. The unsteady flow is modeled using a linearized Euler analysis whereby the unsteady flow field is decomposed into a nonlinear mean flow plus a linear harmonically varying unsteady flow. The equations that govern the perturbation flow, the linearized Euler equations, are linear variable coefficient equations. For transonic flows containing shocks, shock capturing is used to model the shock impulse (the unsteady load due to the harmonic motion of the shock). A conservative Lax-Wendroff scheme is used to obtain a set of linearized finite volume equations that describe the harmonic small disturbance behavior of the flow. Conditions under which such a discretization will correctly predict the shock impulse are investigated. Computational results are presented that demonstrate the accuracy and efficiency of the present method as well as the essential role of unsteady shock impulse loads on the flutter stability of fans.
[41] Hall K C, Thomas J P, Clark W S.2002.

Computation of unsteady nonlinear flows in cascades using a harmonic balance technique

. AIAA Journal, 40: 879-886.

DOI      URL      [本文引用: 3]      摘要

ABSTRACT A harmonic balance technique for modeling unsteady nonlinear flows in turbomachinery is presented. The analysis exploits the fact that many unsteady flows of interest in turbomachinery are periodic in time. Thus, the unsteady flow conservation variables may be represented by a Fourier series in time with spatially varying coefficients. This assumption leads to a harmonic balance form of the Eider or Navier-Stokes equations, which, in turn, can be solved efficiently as a steady problem using conventional computational fluid dynamic (CFD) methods; including pseudotime time marching with local time stepping and multigrid acceleration. Thus, the method is computationally efficient, at least one to two orders of magnitude faster than conventional nonlinear time-domain CFD simulations. Computational results for unsteady, transonic, viscous flow in the front stage rotor of a high-pressure compressor demonstrate that even strongly nonlinear flows can be modeled to engineering accuracy with a small number of terms retained in the Fourier series representation of the flow. Furthermore, in some cases, fluid nonlinearities are found to be important for surprisingly small blade vibrations.
[42] Hassan D, Sicot F.2011.

A time-domain harmonic balance method for dynamic derivatives predictions

.//49th AIAA Aerospace Sciences Meeting including the New Horizons and Aerospace Exposition, 4-7.

DOI      URL      摘要

McMullen et al.5 solve a subset of these equations up to mode N, -N 00 k 00 N, yielding the Non-Linear Frequency Domain method. The present HB method1 is implemented in the elsA solver4, which is a time-domain solver. It uses a Discrete Inverse Fourier Transform (DIFT) to cast back in the time domain this subset of 2N + 1 equations from Eq. (6). The DIFT induces linear relations between Fourier's coefficients Wk and a uniform sampling of W within the period. This leads to a time discretization with a new time operator Dt as follows
[43] He L.2008.

Harmonic solution of unsteady flow around blades with separation

. AIAA Journal, 46: 1299-1307.

DOI      URL      [本文引用: 1]      摘要

http://arc.aiaa.org/doi/abs/10.2514/1.28656
[44] He L.2010.

Fourier methods for turbomachinery applications

. Progress in Aerospace Sciences, 46: 329-341.

DOI      URL      [本文引用: 1]      摘要

Rapid increase in computing power has made a huge difference in scales and complexities of the problems in turbomachinery that we can tackle by use of computational fluid dynamics (CFD). It is recognised, however, that there is always a need for developing efficient methods for applications to blade designs. In a design cycle, a large number of flow solutions are sought to interact iteratively or concurrently with various options, opportunities and constraints from other disciplines. This basic requirement for fast prediction methods in a multi-disciplinary design environment remains unchanged, regardless of computer speed. And it must be recognised that the multi-disciplinary nature of blading design increasingly influences outcomes of advanced gas turbine and aeroengine developments. Recently there has been considerable progress in the Fourier harmonic modelling method development for turbomachinery applications. The main driver is to develop efficient and accurate computational methodologies and working methods for prediction and analysis of unsteady effects on aerothermal performance (loading and efficiency) and aeroelasticity (blade vibration due to flutter and forced response) in turbomachinery. In this article, the developments and applications of this type of methods in the past 20 years or so are reviewed. The basic modelling assumptions and various forms of implementations for the temporal Fourier modelling approach are presented and discussed. Computational examples for realistic turbomachinery configurations/flow conditions are given to illustrate the validity and effectiveness of the approach. Although the major development has been in the temporal Fourier harmonic modelling, some recent progress in use of the spatial Fourier modelling is also described with demonstration examples.
[45] He L, Ning W.1998.

Efficient approach for analysis of unsteady viscous flows in turbomachines

. AIAA Journal, 36: 2005-2012.

DOI      URL      [本文引用: 1]      摘要

Abstract A nonlinear harmonic methodology has been developed to calculate unsteady viscous flows through turbomachinery blades. Flow variables are decomposed into time-averaged variables and unsteady perturbations, resulting in the time-averaged equations with extra nonlinear stress terms depending on the unsteady perturbations. An efficient evaluation of an unsteady flowfield is obtained by solving the first-order harmonic equations. The nonlinear interactions between the time-averaged flowfield and the unsteady perturbations were included by a strong coupling approach. The basic computational methodology was applied to the two-dimensional Navier-Stokes equations, and the method was validated against several test cases. Computational results show that this method is much more efficient than the nonlinear time-marching methods while still modeling dominant nonlinear effects.
[46] Huang X, Wang D X.2016.

Stabilizing and accelerating solution of harmonic balance equation system using the LU-SGS and block Jacobi methods

//ASME Turbo Expo 2016: Turbomachinery Technical Conference and Exposition, V02DT44A026.

[本文引用: 1]     

[47] Hull T E, Enright W H, Fellen B M, Sedgwick A E.1972.

Comparing numerical methods for ordinary differential equations

. SIAM Journal on Numerical Analysis, 9: 603-637.

DOI      URL      [本文引用: 1]      摘要

Numerical methods for systems of first order ordinary differential equations are tested on a variety of initial value problems. The methods are compared primarily as to how well they can handle relatively routine integration steps under a variety of accuracy requirements, rather than how well they handle difficulties caused by discontinuities, stiffness, roundoff or getting started. According to criteria involving the number of function evaluations, overhead cost, and reliability, the best general-purpose method, if function evaluations are not very costly, is one due to Bulirsch and Stoer. However, when function evaluations are relatively expensive, variable-order methods based on Adams formulas are best. The overhead costs are lower for the method of Bulirsch and Stoer, but the Adams methods require considerably fewer function evaluations. Krogh's implementation of a variable-order Adams method is the best of those tested, but one due to Gear is also very good. In general, Runge-Kutta methods are not competitive, but fourth or fifth order methods of this type are best for restricted classes of problems in which function evaluations are not very expensive and accuracy requirements are not very stringent. The problems, methods and comparison criteria are specified very carefully. One objective in doing so is to provide a rigorous conceptual basis for comparing methods. Another is to provide a useful standard for such comparisons. A program called DETEST is used to obtain the relevant statistics for any particular method.
[48] Jameson A.1991.

Time dependent calculations using multigrid, with applications to unsteady flows past airfoils and wings

//Computational Fluid Dynamics Conference.

DOI      URL      [本文引用: 1]      摘要

A major factor leading to the widespread acceptance of computational fluid dynamics in the design environment has been the steady and continuing reduction of computational costs, due both to improvements in computer hardware and to improvements in algorithms. The
[49] Jameson A, Shankaran S.2009.

An assessment of dual-time stepping, time spectral and artificial compressibility based numerical algorithms for unsteady flow with applications to flapping wings

//19th AIAA Computational Fluid Dynamics, 4273.

DOI      URL      [本文引用: 1]      摘要

Abstract The objective of this study is to compare and contrast three numerical algorithms that can be used to estimate the forces and pressure distribution on wings in flapping motion. All algorithms are used to solve the unsteady Navier-Stokes equations in two dimensions at low Reynolds Numbers. The four algorithms are a) an A-stable, implicit discretization b) the time-spectral algorithm that implicitly assumes that the flow-field in temporally periodic, c) incompressible formulations of a) and d) incompressible formulations of b) using the artificial compressibility method. The methods in a) and b) have been reported earlier in literature but their application to flapping wing flows at low Reynolds number is new. The algorithms introduced in c), and d) are new and previously not reported in literature. In this abstract, the four algorithms are used for roughly similar test cases to obtain preliminary estimates for their merits and demerits. The final version of the paper will use the same test case for all the algorithms to enable even-handed comparison of the different numerical methods.
[50] Kurdi M H, Beran P S.2008.

Spectral element method in time for rapidly actuated systems

. Journal of Computational Physics, 227: 1809-1835.

DOI      URL      [本文引用: 1]      摘要

In this paper, the spectral element (SE) method is applied in time to find the entire time-periodic or transient solution of time-dependent differential equations. The time-periodic solution is computed by enforcing periodicity of the element set. Of particular interest are periodic forcing functions possessing high frequency content. To maintain the spectral accuracy for such forcing functions, an h-refinement scheme is employed near the semi-discontinuity without increasing the number of degrees of freedom. Time discretization by spectral elements is applied initially to a standard form of a set of linear, first-order differential equations subject to harmonic excitation and an excitation admitting rapid variation. Other case studies include the application of the SE approach to parabolic and hyperbolic partial differential equations. The first-order form of these equations is obtained through semi-discretization using conventional finite-element, spectral element and finite-difference schemes. Element clustering (h-refinement) is applied to maintain the high accuracy and efficiency in the region of the forcing function admitting rapid variation. The convergence in time of the method is demonstrated. In some cases, machine precision is obtained with 25 degrees of freedom per cycle. Finally the method is applied to a weakly nonlinear problem with time-periodic solution to demonstrate its future applicability to the analysis of limit-cycle oscillations in aeroelastic systems.
[51] Lau S L, Cheung Y K.1981.

Amplitude incremental variational principle for nonlinear vibration of elastic systems

. Journal of Applied Mechanics, 48: 959.

DOI      URL      [本文引用: 1]     

[52] Lee B H K, Liu L, Chung K W.2005.

Airfoil motion in subsonic flow with strong cubic nonlinear restoring forces

. Journal of Sound & Vibration, 281: 699-717.

DOI      URL      [本文引用: 1]      摘要

Limit cycle oscillations of a two-degree-of-freedom airfoil motion with cubic nonlinearity in the restoring forces are investigated. The harmonic balance method is used to derive a frequency relation that depends only on airfoil parameters. The amplitudes of the pitch and plunge motions can be computed from analytical expressions once the frequency is known. The method is extended to higher harmonics. Illustrative examples are given and the results are compared with numerical computations.
[53] Leffell J, Sitaraman J, Lakshminarayan V, Wissink A.2016.

Towards efficient parallel-in-time simulation of periodic flows//54th AIAA Aerospace Sciences Meeting (AIAA Paper 2016-0066), San Diego, CA, January 4--8

.

[本文引用: 5]     

[54] Liu L, Dowell E H.2004.

The secondary bifurcation of an aeroelastic airfoil motion: Effect of high harmonics

. Nonlinear Dynamics, 37: 31-49.

DOI      URL      [本文引用: 1]      摘要

The nonlinear dynamical response of a two-degree-of-freedom aeroelastic airfoil motion with cubic restoring forces is investigated. A secondary bifurcation after the primary Hopf (flutter) bifurcation is detected for a cubic hard spring in the pitch degree-of-freedom. Furthermore, there is a hysteresis in the secondary bifurcation: starting from different initial conditions the motion may jump from one limit cycle to another at different fluid flow velocities. A high-order harmonic balance method is employed to investigate the possible bifurcation branches. Furthermore, a numerical time simulation procedure is used to confirm the stable and unstable bifurcation branches.
[55] Liu L, Dowell E H, Thomas J P.2007.

A high dimensional harmonic balance approach for an aeroelastic airfoil with cubic restoring forces

. Journal of Fluids & Structures, 23: 351-363.

DOI      URL      [本文引用: 1]      摘要

This is a study of a two-dimensional airfoil including a cubic spring stiffness placed in an incompressible flow. A new formulation of the harmonic balance method is employed for the aeroelastic airfoil to investigate the amplitude and frequency of the limit cycle oscillations. The results are compared with the results from the classical harmonic balance approach and from the conventional time marching integration method.
[56] Mavriplis D J, Yang Z, Mundis N.2012.

Extensions of time spectral methods for practical rotorcraft problems

//50th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition paper AIAA, 423: 24.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT For flows with strong periodic content, time-spectral methods can be used to obtain time-accurate solutions at substantially reduced cost compared to traditional time-implicit methods which operate directly in the time domain. In their original form, time spec-tral methods are applicable only to purely periodic problems and formulated for single grid systems. A wide class of problems involve quasi-periodic flows, such as maneuvering rotorcraft problems, where a slow transient is superimposed over a more rapid periodic motion. Additionally, the most common approach for simulating combined rotor-fuselage interactions is through the use of a dynamically overlapping mesh system. Thus, in order to represent a practical approach for rotorcraft simulations, time spectral methods that are applicable to quasi-periodic problems and capable of operating on overlapping mesh systems need to be formulated. In this paper, we propose separately an extension of time spectral methods to quasi-periodic problems, and an extension for overlapping mesh con-figurations. In both cases, the basic implementation allows for two levels of parallelism, one in the spatial dimension, and another in the time-spectral dimension, and is implemented in a modular fashion that minimizes the modifications required to an existing steady-state solver. Results are given for three-dimensional quasi-periodic problems on a single mesh, and for two-dimensional periodic overlapping mesh systems.
[57] McMullen M S.2003.

The application of non-linear frequency domain methods to the Euler and Navier-Stokes equations.[PhD Thesis]

. California: Stanford University.

DOI      URL      [本文引用: 1]      摘要

This research demonstrates the accuracy and efficiency of the Non-Linear Frequency Domain (NLFD) method in applications to unsteady flow calculations. The basis of the method is a pseudo-spectral approach to recast a non-linear unsteady system of equations in the temporal domain into a stationary system in the frequency domain. The NLFD method, in principle, provides the rapid convergence of a spectral method with increasing numbers of modes, and, in this sense, it is an optimal scheme for time-periodic problems. In practice it can also be effectively used as a reduced order method in which users deliberately choose not to resolve temporal modes in the solution. The method is easily applied to problems where the time period of the unsteadiness is known a priori. A method is proposed that iteratively calculates the time period when it is not known a priori . Convergence acceleration techniques like local time-stepping, implicit residual averaging and multigrid are used in the solution of the frequency-domain equations. A new method, spectral viscosity is also introduced. In conjunction with modifications to the established techniques this produces convergence rates equivalent to state-of-the-art steady-flow solvers. Two main test cases have been used to evaluate the NLFD method. The first is vortex shedding in low Reynolds number flows past cylinders. Numerical results demonstrate the efficiency of the NLFD method in representing complex flow field physics with a limited number of temporal modes. The shedding frequency is unknown a priori, which serves to test the application of the proposed variable-time-period method. The second problem is an airfoil undergoing a forced pitching motion in transonic flow. Comparisons with experimental results demonstrate that a limited number of temporal modes can accurately represent a non-linear unsteady solution. Comparisons with time-accurate codes also demonstrate the efficiency gains realized by the NLFD method.
[58] Mcmullen M S, Jameson A.2006.

The computational efficiency of non-linear frequency domain methods

. Journal of Computational Physics, 212: 637-661.

DOI      URL      [本文引用: 1]      摘要

This paper demonstrates efficient solver technologies applied to the non-linear frequency domain (NLFD) method. The basis of the NLFD method is to assume the time period of the solution oscillation and to transform both the solution and residual using a discrete Fourier transform. An unsteady residual is formed in the frequency domain and iteratively driven to a negligible value. This method is amenable to many of the convergence acceleration techniques used for steady state flows including pseudo-local time stepping, implicit residual averaging, coarse grid viscosity and multigrid. This paper will address the implementation of these techniques such that convergence rates of the modified unsteady solver are equivalent to those of the original steady-state techniques.
[59] Mcmullen M S, Jameson A, Alonso J J.2001.

Acceleration of convergence to a periodic steady state in turbomachinery flows

. AIAA Journal, 28: 100-152.

DOI      URL      [本文引用: 3]      摘要

ABSTRACT This paper present s at echnique usedt o accelerat et he convergence ofunst eady calculat ions oft ime-periodic flowst o a periodic st eady st at e. The basis oft he procedure is t e use of t e discret e Fouriert sform int ime, and is similart ot he harmonic balance proceduret hat has been pursued by Hallet . al. Thet echnique is amenablet o parallel processing, and convergence accelerat t echniques such as mult grid and implicit residual averaging. The comput at ional e#ciency oft his met hod is compared wit h dual t imest epping algorit hms. Sample calculat ions are provided, and a comparison bet ween solutI s wit varying t mporal resolutI is present d. The result show tat t e comput al e#ciency of t e harmonic balance t chnique is largely a funct of t et emporal resolut ion. Init ial experiment s confirmt he promise oft he harmonic balance met hodt o achieve significant reduct ions in comput at ional cost .
[60] Mcmullen M, Jameson A, Alonso J J.2002.

Application of a non-linear frerquency domain solver to the Euler and Navier-stokes equations//40th AIAA Aerospace Sciences Meeting & Exhibit

.

[本文引用: 1]     

[61] Mcmullen M, Jameson A, Alonso J J.2006.

Demonstration of nonlinear frequency domain methods

. AIAA Journal, 44: 1428-1435.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT method in which users deliberately choose not to resolve temporal modes in the solution. A variable-time-period method has been proposed such that the nonlinear frequency domain method can be applied to problems where the time period of the unsteadiness is either known or unknown a priori. To validate the latter case, results from this method have been compared with experimental results of vortex shedding in low Reynolds number flows past cylinders. Validation of the first case utilizes experimental data of a pitching airfoil in transonic flow. These comparisons demonstrate the efficiency of the nonlinear frequency domain method in representing complex nonlinear flow field physics with a limited number of temporal modes.
[62] Mundis N L, Mavriplis D J.2013

a. GMRES applied to the time-spectral and quasi-periodic time-spectral methods

. AIAA Paper, 3084: 24-27.

DOI      URL      [本文引用: 2]      摘要

Abstract The time-spectral method applied to the Euler equations theoretically offers significant computational savings for purely periodic problems when compared to standard time-implicit methods. A recently developed quasi-periodic time-spectral (BDFTS) method extends the time-spectral method to problems with fast periodic content and slow mean flow transients, which should lead to faster solution of these types of problems as well. However, attaining superior efficiency with TS or BDFTS methods over traditional time-implicit methods hinges on the ability to rapidly solve the large non-linear system resulting from TS discretizations which become larger and stiffer as more time instances are employed. In order to increase the efficiency of these solvers, and to improve robustness, particularly for large numbers of time instances, the TS and BDFTS methods are reworked such that the Generalized Minimal Residual Method (GMRES) is used to solve the implicit linear system over all coupled time instances. The use of GMRES as the linear solver makes these methods more robust, allows them to be applied to a far greater subset of time-accurate problems, including those with a broad range of harmonic content, and vastly improves the efficiency of time-spectral methods.
[63] Mundis N L, Mavriplis D J.2013

b. Quasi-periodic time spectral method for aeroelastic flutter analysis

//AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition, 2013-0638.

DOI      URL      [本文引用: 1]      摘要

A recently developed quasi-periodic time spectral method is applied to the demanding problem of aeroelastic flutter. Both a standard time-implicit method and a quasi-periodic time spectral method are developed that take into account the coupling among the three fundamental aspects of computational aeroelastic calculations: unsteady flow equations, time dependent structural response to aerodynamics loads, and dynamically moving meshes. These two methods are then compared in order to demonstrate the capability of the quasi-periodic time spectral method to solve aeroelastic flutter problems. Finally, it is demonstrated that the quasi-periodic time spectral method can be used to solve aeroelastic flutter problems.
[64] Mundis N L, Mavriplis D J.2014.

An efficient flexible GMRES solver for the fully-coupled time-spectral aeroelastic system

. AIAA Paper, 1427: 13-17.

[本文引用: 1]     

[65] Mundis N L, Mavriplis D J.2015

a. Wave number independent preconditioning for GMRES time-spectral solvers

//53rd AIAA Aerospace Sciences Meeting, American Institute of Aeronautics and Astronautics, Kissimmee, Florida, 1-21.

DOI      URL      [本文引用: 3]      摘要

react-text: 112 The time-spectral method applied to the Euler and coupled aeroelastic equations theoretically offers significant computational savings for purely periodic problems when compared to standard time-implicit methods. However, attaining superior efficiency with time-spectral methods over traditional time-implicit methods hinges on the ability rapidly to solve the large non-linear system resulting... /react-text react-text: 113 /react-text [Show full abstract]
[66] Mundis N L, Mavriplis D J.2015

b. Finite-element time discretizations for the unsteady Euler equations

//53rd AIAA Aerospace Sciences Meeting, 0569.

[本文引用: 1]     

[67] Mundis N L, Mavriplis D J.2016.

Toward an optimal solver for time-spectral solutions on unstructured meshes//54th AIAA Aerospace Sciences Meeting

, 0069.

[本文引用: 1]     

[68] Mundis N L, Mavriplis D J, Sitaraman J.2013.

Quasi-periodic time-spectral methods for flutter and gust response//69th Forum of the American Helicopter Society, AHS International, Alexandria

.

[本文引用: 2]     

[69] Murman S M.2012.

Reduced-frequency approach for calculating dynamic derivatives

. AIAA Journal, 45: 2005-2840.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT A novel method of calculating dynamic stability derivatives using Computational Fluid Dynamics is presented. This method uses a non-linear, reduced-frequency ap-proach to simulate the response to a forced oscillation using a single frequency com-ponent at the forcing frequency. This provides an order of magnitude improvement in computational efficiency over similar time-dependent schemes without loss of gener-ality. The reduced-frequency approach is implemented with an automated Cartesian mesh scheme. This combination of Cartesian meshing and reduced-frequency solver enables damping derivatives for arbitrary flight condition and geometric complexity to be efficiently and accurately calculated. The method is validated for 3-D reference missile and aircraft dynamic test configurations through the transonic and high-alpha flight regimes. Comparisons with the results of time-dependent simulations are also included.
[70] Murman S M, Aftosmis M J, Berger M J.2004.

Numerical simulation of rolling airframes using a multilevel Cartesian method

. Journal of Spacecraft and Rockets, 41: 426-435.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT A supersonic rolling missile with two synchronous canard control surfaces is analyzed using an automated, inviscid, Cartesian method. Sequential-static and time-dependent dynamic simulations of the complete motion are computed for canard dither schedules for level flight, pitch, and yaw maneuver. The dynamic simulations are compared directly against both high-resolution viscous simulations and relevant experimental data, and are also utilized to compute dynamic stability derivatives. The results show that both the body roll rate and canard dither motion influence the roll-averaged forces and moments on the body. At the relatively, low roll rates analyzed in the current work these dynamic effects are modest, however the dynamic computations are effective in predicting the dynamic stability derivatives which can be significant for highly-maneuverable missiles.
[71] Nadarajah S, Jameson A.2007.

Optimum shape design for unsteady three-dimensional viscous flows using a nonlinear frequency-domain method

. Journal of Aircraft, 44: 1513-1527.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT This paper presents an adjoint method for the optimum shape design of unsteady three-dimensional viscous flows. The goal is to develop a set of discrete unsteady adjoint equations and the corre sponding boundary condition for the nonlinear frequency-domain method. First, this paper presents the complete formulation of the time-dependent optimal design problem. Second, we present the nonlinear frequency-domain adjoint equations for three-dimensional viscous transonic flows. Third, we present results that demonstrate the application of the theory to a three-dimensional wing.
[72] Ning W, He L.1997.

Computation of unsteady flows around oscillating blades using linear and nonlinear harmonic Euler methods//ASME 1997 International Gas Turbine and Aeroengine Congress and Exhibition

. American Society of Mechanical Engineers, V004T14A039-V004T14A039.

[本文引用: 4]     

[73] Pechloff A N, Laschka B.2006.

Small disturbance Navier-Stokes method: Efficient tool for predicting unsteady air loads

. Journal of Aircraft, 43: 17-29.

DOI      URL      [本文引用: 1]      摘要

Abstract Regarding industrial aeroelastic applications, the otherwise common yet computationally expensive procedure of time-accurately solving the unsteady Reynolds-averaged Navier-Stokes (RANS) equations is presently not feasible. However, a numerical method based on the small disturbance formulation of these governing equations can provide an efficient means for predicting aerodynamic loading. With this alternative approach the initial unsteady problem is reduced to a formally steady one for the perturbation part under specification of a time law, thus allowing a solution in the frequency domain. A detailed derivation of the small disturbance Navier-Stokes equations on the theoretical basis of the triple decomposition is presented for high-Reynolds-number flow restricted to harmonic behavior. Validation of the resultant numerical method FLM-SD.NS is performed with a transonic test case for each the NACA 64A010 and NLR 7301 airfoil, respectively, undergoing sinusoidal pitch oscillations. Computational results are in good agreement with experimental data, as well as time-accurate RANS solutions provided by the comparative solvers FLM-NS and FLOWer. Reductions in computational time up to an order of magnitude in relation to FLM-NS are observed.
[74] Ronch A D, Ghoreyshi M, Badcock K J, Gortz S, Widhalm M, Dwight R P, Campobasso M S.2010.

Linear frequency domain and harmonic balance predictions of dynamic derivatives//AIAA Applied Aerodynamics Conference

.

[本文引用: 3]     

[75] Ronch A D, Mccracken A J, Badcock K J, Widhalm M, Campobasso M S.2013.

Linear frequency domain and harmonic balance predictions of dynamic derivatives

. Journal of Aircraft, 50: 694-707.

DOI      URL      [本文引用: 2]      摘要

Dynamic derivatives are used to represent the influence of the aircraft motion rates on the aerodynamic forces and moments needed for studies of flight dynamics. The use of computational fluid dynamics has potential to supplement costly wind-tunnel testing. The paper considers the problem of the fast computation of forced periodic motions using the Euler equations. Three methods are evaluated. The first is computation in the time domain, which provides the benchmark solution in the sense that the time-accurate solution is obtained. Two acceleration techniques in the frequency domain are compared. The first uses a harmonic solution of the linearized problem, referred to as the linear frequency-domain approach. The second uses the harmonic balance method, which approximates the nonlinear problem using a number of Fourier modes. These approaches are compared for the ability to predict dynamic derivatives and for computational cost. The NACA 0012 aerofoil and the DLR-F12 passenger jet wind-tunnel model are the test cases. Compared to time-domain simulations, an order of magnitude reduction in computational costs is achieved and satisfactory predictions are obtained for cases with a narrow frequency spectrum and moderate amplitudes using the frequency-domain methods.
[76] Rahmati M T, He L, Wang D X, Li Y S, Wells R G, Krishnababu S K.2014.

Nonlinear time and frequency domain methods for multirow aeromechanical analysis

. Journal of Turbomachinery, 136: 041010.

[本文引用: 1]     

[77] Su X, Yuan X.Implicit solution of time spectral method for periodic unsteady flows. Journal of Engineering Thermophysics, 63: 860-876.

[本文引用: 2]     

[78] Salles L, Blanc L, Gouskov A, Jean P.2014.

Dual time stepping algorithms with the high order harmonic balance method for contact interfaces with fretting-wear

. Journal of Engineering for Gas Turbines & Power, 134: 913-921.

DOI      URL      摘要

Contact interfaces with dry friction are frequently used in turbomachinery. Dry friction damping produced by the sliding surfaces of these interfaces reduces the amplitude of bladed-disk vibration. The relative displacements at these interfaces lead to fretting-wear which reduces the average life expectancy of the structure. Frequency response functions are calculated numerically by using the multi-harmonic balance method (mHBM). The dynamic Lagrangian frequency-time method is used to calculate contact forces in the frequency domain. A new strategy for solving nonlinear systems based on dual time stepping is applied. This method is faster than using Newton solvers. It was used successfully for solving Nonlinear CFD equations in the frequency domain. This new approach allows identifying the steady state of worn systems by integrating wear rate equations a on dual time scale. The dual time equations are integrated by an implicit scheme. Of the different orders tested, the first order scheme provided the best results.
[79] Thomas J P, Dowell E H, Hall K C.2002.

Nonlinear inviscid aerodynamic effects on transonic divergence, flutter, and limit-cycle oscillations

. AIAA Journal, 40: 638-646.

DOI      URL      [本文引用: 1]      摘要

Abstract By the use of a state-of-the-art computational fluid dynamic (CFD) method to model nonlinear steady and unsteady transonic flows in conjunction with a linear structural model, an investigation is made into how nonlinear aerodynamics can effect the divergence, flutter, and limit-cycle oscillation (LCO) characteristics of a transonic airfoil configuration. A single-degree-of-freedom (DOF) model is studied for divergence, and one- and two-DOF models are studied for flutter and LCO. A harmonic balance method in conjunction with the CFD solver is used to determine the aerodynamics for finite amplitude unsteady excitations of a prescribed frequency. A procedure for determining the LCO solution is also presented. For the configuration investigated, nonlinear aerodynamic effects are found to produce a favorable transonic divergence trend and unstable and stable LCO solutions, respectively, for the one- and two-DOF flutter models.
[80] Thomas J P, Dowell E H, Hall K C.2004.

Modeling viscous transonic limit cycle oscillation behavior using a harmonic balance approach

. Journal of Aircraft, 41: 1266-1274.

DOI      URL      [本文引用: 1]      摘要

ABSTRACT Presented is a harmonic-balance computational fluid dynamic approach for modeling limit-cycle oscillation behavior of aeroelastic airfoil configurations in a viscous transonic flow. For the NLR 7301 airfoil configuration studied, accounting for viscous effects is shown to significantly influence computed limit-cycle oscillation trends when compared to an inviscid analysis. A methodology for accounting for changes in mean angle of attack during limit-cycle oscillation is also developed.
[81] Thomas J P, Hall K C, Dowell E H.2005.

Discrete adjoint approach for modeling unsteady aerodynamic design sensitivities

. AIAA Journal, 43: 1931.

DOI      URL      [本文引用: 1]      摘要

Abstract A discrete adjoint approach is presented for computing steady and unsteady aerodynamic design sensitivities for compressible viscous flows about airfoil configurations. The nominal flow solver method is based on a harmonic balance solution technique, which is capable of modeling both steady and nonlinear periodic unsteady flows. The computer code for the discrete adjoint solver, which is derived from the nominal harmonic balance solver, has been generated with the aid of the advanced automatic differentiation software tool known as TAF (Transformation of Algorithms in FORTRAN).
[82] Van Der Weide E, Gopinath A, Jameson A.2005.

Turbomachinery applications with the time spectral method

. AIAA Paper, 4905: 2005.

DOI      URL      [本文引用: 2]      摘要

This paper presents the Time Spectral Method, which is capable of significantly reducing the CPU requirements for time periodic flows. Here the main focus is turbomachinery computations for which the standard formulation must be adapted to take sector periodicity into account if the problem is solved in the Cartesian frame. Odd-even decoupling is avoided by choosing an odd number of time intervals in a period. Results are presented for the NASA Stage 35 compressor test case and these show that engineering accuracy is obtained with 11 time intervals per blade passing. Especially for high RPM applications, like the Stage 35, the Time Spectral Method leads to a reduction of an order of magnitude in CPUtime compared to a standard second order backward difference time integration scheme. I.
[83] Vilmin S, Lorrain E, Hirsch C, Swoboda M.2006.

Unsteady flow modeling across the rotor/stator interface using the nonlinear harmonic method//ASME Turbo Expo 2006: Power for Land, Sea, and Air

.

[本文引用: 1]     

[84] Widhalm M, Dwight R, Thormann R, Hubner A.2010.

Efficient computation of dynamic stability data with a linearized frequency domain solver//Eccomas Cfd

. DLR.

[本文引用: 1]     

[85] Woodgate M A, Barakos G N.2012.

Implicit computational fluid dynamics methods for fast analysis of rotor flows

. AIAA Journal, 50: 1217-1244.

DOI      URL      [本文引用: 2]      摘要

Computational fluid dynamics (CFD) based on the Navier鈥揝tokes equations is by far the most useful predictivemethod available today for helicopter analysis and design. The main drawback of CFD, and perhaps the reason forits slow acceptance by design offices of helicopter manufacturers, is apparently due to the substantial requirements ofCPU time and the relatively slow turnaround times in comparison to lower-order methods. However, progress withCFD algorithms and parallel computing has allowed CFD analyses to be used more routinely. Typical applicationsinclude computations of aerofoil data that feed rotor performance codes and analyses of rotors in hovering flight. Thecomputation of unsteady flow cases is, however, still challenging. This paper presents alternative ways of tacklingunsteady flow problems pertinent to rotorcraft using methods that aim to reduce the time-marching unsteadycomputations to more manageable steady-state solutions. The techniques investigated so far by the CFD laboratoryof Liverpool include time-linearized and harmonic-balance methods. The details of the methods are presented alongwith their implementation in the framework of the helicopter multiblock CFD solver. Results were obtained forseveral flow cases, ranging from pitching/translating aerofoils to complete rotors. The results highlight some of thelimitations of the time-linearized method and the potential of the harmonic-balance technique. It was found that thetime-linearized method can provide adequate results for cases where the unsteady flow is a rather small perturbationof a known mean. The harmonic-balance method proved to have a larger range of applicability and providedadequate results for the analysis of unsteady flows. The required CPU time was reduced, and the required corecomputer memory was increased. Overall, the harmonic-balance method appears to be a possible alternative to timemarchingCFD for a wide range of problems.
[86] Yang Z, Mavriplis D.2010.

Time spectral method for periodic and quasi-periodic unsteady computations on unstructured meshes//40th AIAA Fluid Dynamics Conference

, 28-1.

[本文引用: 2]     

[87] Yang Z, Mavriplis D, Sitaraman J. Yang Z, Mavriplis D, Sitaraman J.2011.

Prediction of helicopter maneuver loads using BDF/time spectral method on unstructured meshes//49th AIAA Aerospace Sciences Meeting

, 4-7.

URL      [本文引用: 1]     

[88] Yi S, Im D, Choi S, Lee D.2015.

An efficient fluid-structure interaction analysis based on time-spectral approaches//AIAA 53rd Aerospace Science Meeting

, 5-9.

[本文引用: 1]     

[89] Zeiler T A.2000.

Results of Theodorsen and Garrick revisited

. Journal of Aircraft, 37: 918-920.

DOI      URL      [本文引用: 1]      摘要

Evidence that some of the flutter boundaries found in the NACA reports by Theodorsen and Garrick are in error, and that many others might also be in error. Further, some of the erroneous plots have found their way into some classic texts on aeroelasticity. It is suspected that the state of computational capabilities that existed at the time the original NACA reports were prepared contributed in large measure to the errors. The exact errors made have not been identified, however, and may never be.

/

Baidu
map