为什么网格画细了整个频段的响应都出现大的差异
我在ANSYS中用shell63建了一个五块子板的拼接板结构，子板的尺寸1.7*1.5，4个耦合边和拼接板周边边界均用简支边界条件，出现的问题是当我用0.04和0.02两种尺寸划分网格，得到的耦合边的传递功率流曲线差异异常的大，0.02图中的绿线和黄线在整个频段均比0.04小了很多，两种工况我只改变网格的尺寸的大小，其他都未变，求大侠们帮助http://forum.vibunion.com/data/attachment/album/201104/02/1757482zu42owummfco7tw.jpghttp://forum.vibunion.com/data/attachment/album/201104/02/1730551okqkk39u8q5szdq.jpghttp://forum.vibunion.com/data/attachment/album/201104/02/173055kecudci3er307zuv.jpg
模型命令流：
/clear
*set,fr,301
*dim,ff,array,fr,1
*vread,ff(1),F:\ansys_save\APDL\pl,txt,,jik,1,fr
(1f10.2)
*set,bn,5!bn块板
*set,lx,1.7
*set,ly,1.5
*dim,qx,array,fr,1
*dim,qz,array,fr,1
*dim,qq,array,fr,bn-1
/prep7
et,1,shell63,2
real,1
r,1,0.002
mp,ex,1,21.6e10
mp,prxy,1,0.28
mp,dens,1,7800
dmprat,0.01
!mp,dmpr,1,0.01
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
*do,ii,1,bn+1,1
k,ii,0,(ii-1)*ly,0
k,ii+bn+1,lx,(ii-1)*ly,0
*enddo
*do,jj,1,bn,1
a,jj,jj+bn+1,jj+bn+2,jj+1
*enddo
!!!!!!!!!!!!!!!!!!
type,1
real,1
mat,1
esize,0.1
mshkey,1
amesh,all
asel,all
allsel,all
finish
/solu
nsel,s,loc,y,ly
*do,ii,1,bn-1,1
nsel,a,loc,y,ly*ii
*enddo
esln,s,0,all
cm,cmn,node
cm,cme,elem
outres,all,none
outres,nsol,all,cmn
outres,nload,all,cme
allsel,all
antype,harmnic
nsel,s,loc,x,0
nsel,a,loc,x,lx
nsel,a,loc,y,0
*do,ii,1,bn,1
nsel,a,loc,y,ly*ii
*enddo
d,all,uz
nsel,all
allsel,all
f,node(0.6,0.7,0),fz,1
*do,ii,1,fr,1
harfrq,ff(ii)
solve
*enddo
finish
/post26
allsel,all
*set,nc
*set,nmin
nsel,s,loc,y,ly
nsel,u,loc,x,0
nsel,u,loc,x,lx
*get,nc,node,0,count
*dim,nodenum,array,nc,bn-1
!!!!!!!!!!!!!!!将耦合边的节点编号赋给nodenum矩阵
*do,ii,1,bn-1,1
nsel,all
nsel,s,loc,y,ii*ly
nsel,u,loc,x,0
nsel,u,loc,x,lx
*get,nmin,node,0,num,min
*do,jj,1,nc,1
*set,nnext,nmin
*set,nodenum(jj,ii),nmin
*get,nmin,node,nnext,nxth
*enddo
*enddo
nsel,all
allsel,all
!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!计算每条耦合边的传递功率流
*do,ii,1,bn-1,1
*set,qu
*do,jj,1,nc,1
asel,all
asel,s,area,,ii
nsel,s,node,,nodenum(jj,ii)
esla,s
esln,r
*get,e1,elem,0,num,min
*get,e2,elem,e1,nxth
esol,2,e1,nodenum(jj,ii),m,x
esol,3,e2,nodenum(jj,ii),m,x
add,4,2,3
nsol,5,nodenum(jj,ii),rot,x
conjug,6,5
prod,7,4,6,1,,,,1/2,1,3.1416
imagin,8,7
vget,qx,8
*voper,qz,qz,add,qx
*enddo
*do,kk,1,fr,1
*set,qq(kk,ii),abs(qz(kk,1))
*set,qz(kk,1),0
*enddo
*enddo
allsel,all
eplot 频率范围10Hz-10000Hz，具体见附件 按说，网格画密了，四条曲线在低频不可能出现大的差异啊 有试着先执行网格收敛性探讨吗? 我把命令流中的边界条件段由加耦合边节点上d,all,uz,改为边界条件加到对应的线上，即
lsel,s,loc,x,0
lsel,a,loc,x,lx
lsel,a,loc,y,0
*do,ii,1,bn,1
lsel,a,loc,y,ly*ii
*enddo
dl,all,uz
lsel,all
allsel,all
结果0.02和0.04两种工况下，四条曲线在低频段几乎吻合，现在的问题是，简支边界条件加在耦合边的线上和加在对应的节点上有什么不同，这里我不太懂，求大家帮忙解惑
改后两种工况对应曲线：
http://forum.vibunion.com/data/attachment/album/201104/04/2024352ewnw5s9p2pfrwf5.jpghttp://forum.vibunion.com/data/attachment/album/201104/04/2024341kk6zi0kx3zcczdh.jpg
页:
[1]