关于流固耦合在ANSYS建模中的实现！
导师让我分析一个电厂吸收塔的抗震及动力稳定问题，该吸收塔是薄壁筒体钢结构，结构形式复杂，受力也复杂，之前只是做了动力特性分析，当时也没有考虑流固耦合问题，只是把浆液作用范围的壳体密度增大，相当于增大节点质量，把浆液作为质量加到节点上，这样只是简化分析，结果肯定是误差大！最近也看了很多这种分析，有渡槽、TLD、储液罐等，基本上用的不是直接耦合就是用的Housner模型，但在ANSYS中具体怎么实现，本人至今不是很明白，有没有谁做过这方面分析的，还望指点一下
[ 本帖最后由 rodge 于 2008-3-27 20:55 编辑 ] 我会逐渐增加我的新进展，还望大家积极讨论！
最近发现流固耦合的问题大家提的比较多，但好像都没有很好的解决，这部分确实是有点麻烦，
大家要是做过这方面分析的，还望把自己的感受写下来，这样就能让很多人得到帮助，也让我们的论坛更好的发展！
要是提出真挚见解的，有奖励哟！ 一般说来，ANSYS的流固耦合主要有4种方式：
1，sequential
这需要用户进行APDL编程进行流固耦合
2，FSI solver
流固耦合的设置过程非常简单，推荐你使用这种方式
3，multi-field solver
这是FSI solver的扩展，你可以使用它实现流体，结构，热，电磁等的耦合
4，直接采用特殊的单元进行直接耦合，耦合计算直接发生在单元刚度矩阵 以上是我在其他论坛上看到的，具体是什么意思啊？还望大家讨论讨论 流固耦合的边界应用带有SFIN标记的SF，SFA，SFE，SFL等命令来标记耦合界面，具体方法见ansys help很详细的
至今还没有找到，谁知道说一下吧 length=2
width=3
height=2
/prep7
et,1,63
et,2,30 !选用FLUID30单元，用于流固耦合问题
r,1,0.01
mp,ex,1,2e11
mp,nuxy,1,0.3
mp,dens,1,7800
mp,dens,2,1000 !定义Acoustics材料来描述流体材料-水
mp,sonc,2,1400
mp,mu,0,
!
block,,length,,width,,height
esize,0.5
mshkey,1
!
type,1
mat,1
real,1
asel,u,loc,y,width
amesh,all
alls
!
type,2
mat,2
vmesh,all
fini
/solu
antype,2
modopt,unsym,10 !非对称模态提取方法处理流固耦合问题
eqslv,front
mxpand,10,,,1
nsel,s,loc,x,
nsel,a,loc,x,length
nsel,r,loc,y
d,all,,,,,,ux,uy,uz,
nsel,s,loc,y,width,
d,all,pres,0
alls
asel,u,loc,y,width,
sfa,all,,fsi !定义流固耦合界面
alls
solv
fini
/post1
set,first
plnsol,u,sum,2,1
fini 上面是一个流固耦合的例子，大家可以讨论讨论
提出自己的理解！ 再给大家一个实例！
Example Fluid-Structural Analysis Using Physics Environments
The example in this section illustrates a steady-state fluid-structure interaction problem. This problem demonstrates the use of nonlinear large-deflection structural coupling for a fluid domain as well as the use of the "null" element type in a physics environment setting. It also demonstrates mesh morphing. 为什么没有大家的讨论？ 考虑结构在水中的自振频率：例子是一加筋板在水中的模态分析。
命令流如下：
FINISH
/CLEAR
/FILENAME,plane
/UNITS,SI
/TITLE,plane
/PREP7
!*********ELEMENT DEFINE********
ET,63,63
ET,4,beam4
et,30,fluid30
!****MATERIAL DEFINE*********
MP,EX,1,2.10E11
MP,DENS,1,7850
MP,NUXY,1,0.3
mp,dens,30,1025
mp,sonc,30,1500
mp,mu,30,0.5
!*******REAL CONSTANT***********
r,30,1e-06
r,50,0.05
r,75,0.375e-02,0.78125e-06,0.000016406
k,1
k,4,1
kfill,1,4,2,,1
kgen,4,1,4,1,,1/3,,10
a,1,2,12,11
*do,i,0,2
*do,j,0,2*10,10
a,1+i+j,2+i+j,12+i+j,11+i+j
*enddo
*enddo
!***************************fluid element****************
k,100,-14.5,-14.5
k,101,-14.5,15.5
k,102,15.5,15.5
k,103,15.5,-14.5
k,140,-14.5,-14.5,30
k,141,-14.5,15.5,30
k,142,15.5,15.5,30
k,143,15.5,-14.5,30
a,100,101,102,103,4,14,24,34,33,32,31,21,11,1
a,1,2,3,4,103,100
a,140,141,142,143
a,100,101,141,140
a,101,102,142,141
a,142,143,103,102
a,140,143,103,100
a,14,24,34,33,32,31,21,11,1,2,3,4
asel,u,,,1,
FLST,2,8,5,ORDE,2okok.org
FITEM,2,10okok.org
FITEM,2,-17okok.org
VA,P51Xokok.org
nummrg,all
alls
MSHKEY,0
MSHAPE,0
esize,1
lsel,s,loc,y,1/3
lsel,r,loc,x,0,1
lsel,r,loc,z,0
latt,1,75,4
lmesh,all
lsel,s,loc,y,2/3
lsel,r,loc,x,0,1
lsel,r,loc,z,0
latt,1,75,4
lmesh,all
lsel,s,loc,x,1/3
lsel,r,loc,y,0,1
lsel,r,loc,z,0
latt,1,75,4
lmesh,all
lsel,s,loc,x,2/3
lsel,r,loc,y,0,1
lsel,r,loc,z,0
latt,1,75,4
lmesh,all
asel,s,,,1,9
aatt,1,50,63
amesh,all
alls
MSHAPE,1,3d
esize,3
vsel,s,,,1
type,30 $mat,30 $real,30
vmesh,all
alls
FINISH
/solu
alls
!**** 求解 ***********
!*********************
ANTYPE,MODAL
MODOPT,lanb,25,0
SOLVE
FINISH
总是出现error 说矩阵不对称，不可以用lanb计算。
总结：流体单元不能用对称的解法
应该采用非对称解法。 加强流固耦合的Ansys 10.0
ANSYS Latest Release Features Integrated Flexibility and Performance
ANSYS 10.0 includes performance upgrade and superior coupled physics technology in fluid structure interaction
SOUTHPOINTE, Pa., June 2, 2005 /PRNewswire-FirstCall via COMTEX/ -- ANSYS, Inc. (Nasdaq: ANSS), a global innovator of simulation software and technologies designed to optimize product development processes, today announced Version 10.0 of its ANSYS(R) software. The new release offers major advances in performance, ease-of-use, interoperability and coupled physics technology, such as fluid structure interaction (FSI). With availability beginning in July, this latest release builds upon and is completely compatible with the current ANSYS 9.0 software.
Continuing to improve upon its Multiphysics technology, ANSYS 10.0 offers superior capabilities for complex FSI problems. Representing a complete FSI solution from a single vendor, it utilizes each of the best-in-class technologies in stress and fluid flow analysis. A single geometry is used for both, with meshing appropriate to the specific physics required. A high-speed protocol provides the communication necessary to perform dynamic FSI analysis. No other FSI offering in the market today offers this degree of robustness and fidelity. Furthermore, it parallelizes on multi-machine clusters for very large problem sizes.
"ANSYS 10.0 represents the latest in integrated CAE capability, building upon the significant advances we delivered in ANSYS 9.0," said Jim Cashman, president and CEO at ANSYS, Inc. "We have continued to increase the breadth and depth of the ANSYS simulation technology, while at the same time establishing an unprecedented level of interoperability between all forms of simulation and analysis. Taking full advantage of the ANSYS(R) Workbench(TM) infrastructure for integrated CAE, we have created a more seamless approach for meshing and modeling, simulation and analysis, and post-processing. Our world-class capabilities in each of the forms of analysis such as structural, thermal and fluids are all together as part of ANSYS 10.0."
Further strengthening the industry-specific capabilities in the Workbench environment, ANSYS introduces design tools for rotating machinery and blade design. ANSYS 10.0 features ANSYS(R) BladeModeler(TM), an efficient 3-D rotating machinery design tool for bladed components and ANSYS(R) TurboGrid(TM), a high quality hexahedral meshing tool for blade design.
"Combined with ANSYS(R) CFX(R) and specialized turbo pre- and post- processing CFD capabilities, this creates a comprehensive solution set for turbomachinery design and analysis," said Chris Reid, vice president and general manager at ANSYS, Inc. "Models for stress analysis, computational fluid dynamics or fluid structure interaction can be created, expanded to include upstream and downstream components via CAD system connectivity, and then fully analyzed. ANSYS(R) Workbench(TM) is a unique environment to provide for this, such that the aerodynamics engineer can perform CFD-intensive design, while also confirming the structural characteristics. This offers the potential for significant reduction in design cycle times."
In the mechanical application area, ANSYS 10.0 includes thermal transients, fully implemented within ANSYS Workbench. This enables users to perform very sophisticated time-based simulations while the ANSYS Workbench integration automates many of the model setup and solver tasks. The result is a fast and easy-to-use solution for studying a product's thermal performance over the expected operating time period.
To address the broader need for efficient and timely execution of analysis problems of ever-increasing size and complexity, the ANSYS 10.0 parallel solver now supports new choices in both communications technology as well as processors. In addition to Ethernet and Gigabit Ethernet (GigE), ANSYS 10.0 supports Myrinet and InfiniBand. This allows for clusters of very high performance at significantly less cost than previous configurations.
Furthering our solutions on high-performance and low-cost hardware configurations, ANSYS Workbench now is supported on Windows XP 64-bit for AMD and EMT64 chipsets. This support eliminates the 2GB memory limit that many users face when running large models on Windows. In addition, it provides the opportunity to add extra physical memory which ANSYS uses to complete the entire solution without writing to the hard disk -- usually the source of extended computation times.
For customers, this enables large model sizes, such as those found in low- frequency static and full transient electromagnetic analyses, to be addressed more efficiently and economically. Plus, the ANSYS 10.0 parallel solver now supports these physics, making ANSYS the leading CAE vendor able to address large electromagnetic problem sizes that are greater than 100 million degrees of freedom.
In the area of high-frequency electromagnetics a new modal port definition is available. This port enhancement greatly simplifies the model setup for transmission line ports needed for many types of Integrated Circuit (IC), Radio Frequency IDentification (RFID) and Radio Frequency (RF) MicroElectroMechanical Systems (MEMS) device analysis. Benchmarks show typical reduction in model size resulting in a 30 to 50 percent reduction in solution times and memory requirements for a given model while still providing more accurate frequency-dependent results.
Other developments in this version include the addition of gyroscopic effects that enhance the performance of ANSYS for rotor dynamics application of turbomachinery and other rotating structures. In the area of direct-coupled field physics, structural-thermal-electric coupling has been enhanced to include thermoelastic damping (TED), an important internal loss mechanism in metals, ceramics and MEMS (resonator beams).
ANSYS continues to deliver powerful capabilities that enable our customers their choice of full automation as well as individual control against our ANSYS Workbench theme. We have made significant enhancements to core meshing architecture to share meshes across ANSYS Workbench applications and improvements in CAD robustness. ANSYS(R) ICEM CFD(TM) 10.0 continues to take meshing to a new level by providing a complete set of tools to model real world systems, like under-hood and crash analysis, through new capabilities related to hybrid meshing as well as handling of detailed CAD models.
"ANSYS 10.0 represents another significant step forward for ANSYS, within each of the physics domains, and in the way they can be utilized together in the ANSYS Workbench," said Mike Wheeler, vice president and general manager at ANSYS, Inc. "No other company in the world of CAE can match the comprehensiveness of our solution set or provide the same level of integration in a single CAE environment."
About ANSYS, Inc.
ANSYS, Inc., founded in 1970, develops and globally markets engineering simulation software and technologies widely used by engineers and designers across a broad spectrum of industries. The Company focuses on the development of open and flexible solutions that enable users to analyze designs directly on the desktop, providing a common platform for fast, efficient and cost- conscious product development, from design concept to final-stage testing and validation. Headquartered in Canon***urg, Pennsylvania, U.S.A., with more than 25 strategic sales locations throughout the world, ANSYS, Inc. employs approximately 600 people and distributes its products through a network of channel partners in over 40 countries. Visit http://www.ansys.com for more information.
ANSYS, ANSYS Workbench, CFX, AUTODYN, and any and all ANSYS, Inc. product and service names are registered trademarks or trademarks of ANSYS, Inc. or its subsidiaries located in the United States or other countries. ICEM CFD is a trademark licensed by ANSYS, Inc. All other trademarks or registered trademarks are the property of their respective owners.
SOURCE ANSYS, Inc.
Kelly Wall of ANSYS, Inc., +1-724-514-3076, or kelly.wall@ansys.com 最近为了考虑结构在水中的自振频率，考察了一个简单的例子，由于以前对流固耦合没有接触过，所以由很多疑问，请高手指教。
例子是一圆环在水中的模态分析。命令流如下：
finish
/clear
/PREP7
!定义单元类型
ET,1,PLANE42 ! structural element
ET,2,FLUID29 ! acoustic fluid element with ux & uy
ET,3,129 ! acoustic infinite line element
r,3,0.31242,0,0
ET,4,FLUID29,,1,0 ! acoustic fluid element without ux & uy
!材料属性
MP,EX,1,2.068e11
MP,DENS,1,7929
MP,NUXY,1,0
MP,DENS,2,1030
MP,SONC,2,1460
! 创建四分之一模型
CYL4,0,0,0.254,0,0.26035,90
CYL4,0,0,0.26035,0,0.31242,90
! 选择属性,网格划分
ASEL,S,AREA,,1
AATT,1,1,1,0
LESIZE,1,,,16,1
LESIZE,3,,,16,1
LESIZE,2,,,1,1
LESIZE,4,,,1,1
MSHKEY,1
MSHAPE,0,2D ! mapped quad mesh
AMESH,1
ASEL,S,AREA,,2
AATT,2,1,2,0
LESIZE,5,,,16,1
LESIZE,7,,,16,1
LESIZE,6,,,5
LESIZE,8,,,5
MSHKEY,0
MSHAPE,0,2D ! mapped quad mesh
AMESH,2
! 关于Y轴镜像
nsym,x,1000,all ! offset node number by 1000
esym,,1000,all
! 关于y轴镜像
nsym,y,2000,all ! offset node number by 2000
esym,,2000,all
NUMMRG,ALL ! merge all quantities
esel,s,type,,1
nsle,s
esln,s,0
nsle,s
esel,inve
nsle,s
emodif,all,type,4
esel,all
nsel,all
! 指定无限吸收边界
csys,1
nsel,s,loc,x,0.31242
type,3
real,3
mat,2
esurf
esel,all
nsel,all
! 标识流固交接面
nsel,s,loc,x,0.26035
esel,s,type,,2
sf,all,fsi,1
nsel,all
esel,all
FINISH
/solu
antype,modal
modopt,damp,10
mxpand,10,,,yes
solve
finish
为了便于对比，也对圆环在空气中做了模态分析
finish
/clear
/PREP7
!定义单元类型
ET,1,PLANE42 ! structural element
!材料属性
MP,EX,1,2.068e11
MP,DENS,1,7929
MP,NUXY,1,0
! 创建四分之一模型
CYL4,0,0,0.254,0,0.26035,90
! 选择属性,网格划分
ASEL,S,AREA,,1
AATT,1,1,1,0
LESIZE,1,,,16,1
LESIZE,3,,,16,1
LESIZE,2,,,1,1
LESIZE,4,,,1,1
MSHKEY,1
MSHAPE,0,2D ! mapped quad mesh
AMESH,1
! 关于Y轴镜像
nsym,x,1000,all ! offset node number by 1000
esym,,1000,all
! 关于y轴镜像
nsym,y,2000,all ! offset node number by 2000
esym,,2000,all
NUMMRG,ALL
/solu
antype,modal
modopt,lanb,10
mxpand,10,,,yes
solve
finish
在水中的自振频率为
SET TIME/FREQ LOAD STEP SUBSTEP CUMULATIVE
1-0.19544E-10 1 1 1
2 0.29640E-03 1 1 1
3-0.21663E-10 1 2 2
4-0.29640E-03 1 2 2
5 0.30870E-03 1 3 3
6 0.0000 1 3 3
7-0.30870E-03 1 4 4
8 0.0000 1 4 4
9-0.53726E-03 1 5 5
10 0.57522E-11 1 5 5
11 0.53726E-03 1 6 6
12-0.89057E-11 1 6 6
13 0.98059E-01 1 7 7
14 35.232 1 7 7
15 0.98059E-01 1 8 8
16 -35.232 1 8 8
17 0.98061E-01 1 9 9
18 35.233 1 9 9
19 0.98061E-01 1 10 10
20 -35.233 1 10 10
在空气中的自振频率为
SET TIME/FREQ LOAD STEP SUBSTEP CUMULATIVE
1 0.0000 1 1 1
2 0.0000 1 2 2
3 0.73609E-03 1 3 3
4 60.805 1 4 4
5 60.805 1 5 5
6 172.97 1 6 6
7 172.97 1 7 7
8 334.40 1 8 8
9 334.40 1 9 9
10 546.59 1 10 10
主要有以下疑问：
1）考虑流固耦合，做模态分析时流体单元是否只能用fluid29（2d）和fluid30（3d),对于fluid129和fluid130在耦合中具体起到什么作用，能不能不设，而用边界约束条件代替？
2）流体范围怎样确定，如本例中（CYL4,0,0,0.26035,0,0.31242,90），外半径为0.31242。如果不是环形的，如一块当水板，该怎样考虑？
3）如果不考虑流体的压缩性，把声速设的很大，MP,SONC,2,1e20，就可以了。
4）从自振频率可以看出，在水中和在空气中，圆环的自振频率差别特别大，且振型也大相径庭，为什么？在水中时，模态提取方法用damp（为什么不能用unsym），特征值的虚部代表角频率，为什么第一阶为正，第二阶为负，而第三阶和第四阶都为0，第六阶、八阶、十阶都为负。应该是从小到大才对？
5）在空气中时，模态提取方法用lanb，为什么第一阶第二阶的频率都为0。
请高手指点迷津，急盼中 对以上问题的解答：
频率为零，一般是发生了刚体位移，估计你是把水抽走，而没有限制圆环。
1。圆环在水中振动必然导致波动(其实就是声波)在水中传播，当声波到达水的另一个界面时就会发生反射(除非水和另一个相邻体的声阻抗是匹配的)。水和金属中的声速相差不大，即可压缩性相差不大。两种可压缩性相差不大的物质的相互作用对两者影响都很大。圆环在水中振动，水对圆环的反作用是由于反射波引起的，流固耦合中采用fluid129和130就是最大程度的减弱反射波。
2。声波从圆环开始传播，随着传播距离的增加，波阵面不断增大，振幅不断减小。同时由于水的衰减，声波也不断减弱。如果水的空间越大，则反射波返回圆环的路径越长，衰减也就越多，影响也就越小。fluid129和130对反射波的衰减(通过很小的反射实现)有限，因此还需要水要有足够的空间。fluid129和130离结构应该大于0.2λ(λ=c/f,c为水中声速)。以上的做法在误差允许的情况下等效于水在无限大水空间中的情况。如果是挡水板，水就是有限空间了，情况也不一样。
3。声速加大情况也不一样，就是不知是不是你所要的情况？
4。空气作为介质，由于其声速比金属小很多，可压缩性大很多，影响可以忽略不计。而水的影响就不同了。这可能就是频率和振型不同的原因吧？我试了你的例子，各种提取方法都可以。
5。空气的影响忽略不计，因此需要对圆环进行约束。你没有约束，那么就会发生静态位移即频率为零。圆环有两个对称轴，因此会发生频率成对出现的情况。也就是说，两个方向上有同样的振型。 接触分析实例---包含初始间隙
fini
/clear,nostart
/prep7
et,1,82
KEYOPT,1,3,3
r,1,0.5
mp,ex,1,1e9
mp,prxy,1,0.3
k,1,0,0
k,2,10,0
k,3,10,5
k,4,6.2,5
k,5,7.5,3.4
k,6,2.5,3.4
k,7,3.8,5
k,8,0,5
a,1,2,3,4,5,6,7,8
LFILLT,6,5,0.18, ,
LFILLT,5,4,0.18, ,
FLST,2,3,4
FITEM,2,9
FITEM,2,11
FITEM,2,10
AL,P51X
FLST,2,3,4
FITEM,2,13
FITEM,2,14
FITEM,2,12
AL,P51X
FLST,2,3,5,ORDE,2
FITEM,2,1
FITEM,2,-3
AADD,P51X
rect,0,10,4.8,5
ASBA, 4, 1
gap=0.02
k,24,6.2-gap,5
k,25,7.5-gap,3.4
k,26,2.5+gap,3.4
k,27,3.8+gap,5
a,24,25,26,27
LFILLT,4,3,0.2, ,
LFILLT,3,2,0.2, ,
FLST,2,3,4
FITEM,2,7
FITEM,2,10
FITEM,2,8
AL,P51X
FLST,2,3,4
FITEM,2,13
FITEM,2,14
FITEM,2,11
AL,P51X
FLST,3,2,5,ORDE,2
FITEM,3,3
FITEM,3,-4
ASBA, 1,P51X
rect,3.8+gap,6.2-gap,5,10
rect,3.8+gap,3.8+gap+8,10,12
FLST,2,3,5,ORDE,3
FITEM,2,1
FITEM,2,3
FITEM,2,5
AADD,P51X
rect,3.8+gap+8,3.8+gap+8+2,10,12
FLST,2,2,5,ORDE,2
FITEM,2,1
FITEM,2,4
AGLUE,P51X
CYL4,2.0,1.8,0.6
CYL4,7.0,1.8,0.6
FLST,2,3,5,ORDE,3
FITEM,2,2
FITEM,2,4
FITEM,2,-5
AOVLAP,P51X
esize,0.2
amesh,all
FLST,5,135,2,ORDE,32
FITEM,5,485
FITEM,5,576
FITEM,5,-577
FITEM,5,621
FITEM,5,-625
FITEM,5,707
FITEM,5,-711
FITEM,5,716
FITEM,5,741
FITEM,5,-745
FITEM,5,750
FITEM,5,-751
FITEM,5,766
FITEM,5,797
FITEM,5,-798
FITEM,5,854
FITEM,5,888
FITEM,5,-938
FITEM,5,1101
FITEM,5,1103
FITEM,5,1420
FITEM,5,1628
FITEM,5,1653
FITEM,5,1696
FITEM,5,1699
FITEM,5,-1702
FITEM,5,1726
FITEM,5,-1728
FITEM,5,1852
FITEM,5,-1874
FITEM,5,2044
FITEM,5,-2066
CM,_Y,ELEM
ESEL, , , ,P51X
CM,_Y1,ELEM
CMSEL,S,_Y
CMDELE,_Y
EREF,_Y1, , ,1,0,1,1
CMDELE,_Y1
ET,2,TARGE169
ET,3,CONTA172
R,3, , ,
R,3,0,0,0.1, 10,0,0
R,4, , ,
R,4,0,0,0.1, 10,-0.02,0
lsel,s,,,9
lsel,a,,,5
lsel,a,,,12
nsll,s,1
type,3
real,3
esurf,all
alls,
lsel,s,,,19
lsel,a,,,20
nsll,s,1
type,3
real,4
esurf,all
alls,
lsel,s,,,7
lsel,a,,,3
lsel,a,,,11
nsll,s,1
type,2
real,3
esurf,all
alls,
lsel,s,,,25
lsel,a,,,26
nsll,s,1
type,2
real,4
esurf,all
alls,
FLST,2,2,5,ORDE,2
FITEM,2,4
FITEM,2,-5
DA,P51X,ALL,
FLST,2,1,4,ORDE,1
FITEM,2,6
SFL,P51X,PRES,500,
/solu
antype,0
nlgeom,on
outres,all,all
nsubst,200,200,2
neqit,1000
solve 我正要进行电机噪声的声固耦合计算，谢谢你的资料，以后再讨论