[讨论]混凝土分析时材料属性输入疑问
最近一直在做混凝土梁的分析,发现了一个问题,就是定义混凝土的材料时,除了要定义破坏准则(concr)外,要不要定义混凝土的屈服准则(miso等),定义与否究竟有多大的影响?我也是看了很多例子,发现:只是定义了混凝土的破坏准则而不定义屈服准则,收敛比较容易,若是定义了屈服准则(就是多点模拟的应力-应变曲线),收敛就比较困难,可以说是很难收敛,无论模型的单元是多么规则!
由此发出疑问:做混凝土梁分析时,要不要定义屈服准则?影响有多大?欢迎大家讨论
[ 本帖最后由 sysh320 于 2007-11-15 18:19 编辑 ] 在网上广泛流传的两个对比例子,就是简支混凝土梁的有限元分析。先贴出这两个例子
*******************************************************************************
例题1
!----------------------------------------------------
!题目:钢筋混凝土简支梁模拟计算
!方法:分离式;solid65和link8
!材料:混凝土采用concr和钢筋为弹性材料,但不考虑压碎
!---------------------------------------------------
!为方便,假定钢筋置于梁底两侧.
!===================================================
/config,nres,2000
/prep7
!定义单元及其材料特性等
rd0=20.0 !钢筋直径
et,1,solid65
et,2,link8
mp,ex,1,33e3
mp,prxy,1,0.20
r,1
hntra=28
hntrl=2.6
tb,concr,1
tbdata,,0.7,1.0,hntrl,-1
mp,ex,2,2.1e5
mp,prxy,2,0.30
r,2,acos(-1)*0.25*rd0*rd0
!定义梁体即单元划分
blc4, , ,100,200,3000
/view,1,1,1,1
/ang,1
gplot
!定义网分时边长控制
lsel,s,loc,z,1,2999
lsel,r,loc,y,0
latt,2,2,2
lesize,all,,,20 !钢筋网格数目
lmesh,all
lsel,s,loc,z,0
lesize,all,,,4 !截面上的网格数目4x4
vsel,all
vatt,1,1,1
mshape,0,3d
mshkey,1
vmesh,all
allsel,all
finish
/solu
!施加约束
lsel,s,loc,z,0
lsel,r,loc,y,0
dl,all,,uy
dl,all,,uz
lsel,all
lsel,s,loc,z,3000
lsel,r,loc,y,0
dl,all,,uy
lsel,all
ksel,s,loc,x,0
ksel,r,loc,y,0
dk,all,ux
allsel,all
!施加荷载
qmz=0.3
asel,s,loc,y,200
sfa,all,1,pres,qmz
allsel,all
nsubst,40
outres,all,all
time,qmz*10
neqit,40
solve
finish
/post1
pldisp,1
etable,zxyl,ls,1
plls,zxyl,zxyl,1
finish
/post26
nsol,2,33,u,y
prod,3,1,,,,,,1/100
prod,4,2,,,,,,-1
xvar,4
plvar,3
*******************************************************************************
例题2
!----------------------------------------------------
!题目:钢筋混凝土简支梁模拟计算
!方法:分离式;solid65和link8
!材料:混凝土采用concr+Miso和钢筋为弹性材料,但不考虑压碎
!增加网格密度
!---------------------------------------------------
!为方便,假定钢筋置于梁底两侧.
!===================================================
/config,nres,2000
/prep7
!定义单元及其材料特性等
rd0=20.0 !钢筋直径
et,1,solid65
et,2,link8
mp,ex,1,26600
mp,prxy,1,0.20
r,1
hntra=28
hntrl=2.6
tb,concr,1
tbdata,,0.7,1.0,hntrl,-1
tb,miso,1,,14
tbpt,,0.0002,hntra*0.19
tbpt,,0.0004,hntra*0.36
tbpt,,0.0006,hntra*0.51
tbpt,,0.0008,hntra*0.64
tbpt,,0.0010,hntra*0.75
tbpt,,0.0012,hntra*0.84
tbpt,,0.0014,hntra*0.91
tbpt,,0.0016,hntra*0.96
tbpt,,0.0018,hntra*0.99
tbpt,,0.0020,hntra*1.00
tbpt,,0.0025,hntra*0.95
tbpt,,0.0030,hntra*0.90
tbpt,,0.0035,hntra*0.85
tbpt,,0.0040,hntra*0.80
mp,ex,2,2.1e5
mp,prxy,2,0.30
r,2,acos(-1)*0.25*rd0*rd0
!定义梁体即单元划分
blc4, , ,100,200,3000
/view,1,1,1,1
/ang,1
gplot
!定义网分时边长控制
lsel,s,loc,z,1,2999
lsel,r,loc,y,0
latt,2,2,2
lesize,all,,,20
lmesh,all
lsel,s,loc,z,0
lesize,all,,,4
vsel,all
vatt,1,1,1
mshape,0,3d
mshkey,1
vmesh,all
allsel,all
finish
/solu
!施加约束
lsel,s,loc,z,0
lsel,r,loc,y,0
dl,all,,uy
dl,all,,uz
lsel,all
lsel,s,loc,z,3000
lsel,r,loc,y,0
dl,all,,uy
lsel,all
ksel,s,loc,x,0
ksel,r,loc,y,0
dk,all,ux
allsel,all
!施加荷载
qmz=0.3
asel,s,loc,y,200
sfa,all,1,pres,qmz
allsel,all
outres,all,all
time,qmz*10
nsubst,40
neqit,40
solve
finish
/post1
pldisp,1
etable,zxyl,ls,1
plls,zxyl,zxyl,1
finish
/post26
nsol,2,33,u,y
prod,3,1,,,,,,1/100
prod,4,2,,,,,,-1
xvar,4
plvar,3 对例子1,我在9.0上计算,收敛很快,附图如下” 原帖由 sysh320 于 2007-11-14 17:10 发表 http://www.chinavib.com/forum/images/common/back.gif
对例子1,我在9.0上计算,收敛很快,附图如下” 对例子2进行计算时,首先提示一个错误,是很多讨论的,混凝土的弹性模量与应力应变关系第一点不对应,我进行了调整,将原来例子中的33000调整为26600.
*** ERROR *** CP = 227.547 TIME= 16:58:39
For material 1 at temperature 0, the initial slope of the stress-strain
curve based on the TB,MISO table ( 26600 ) is less than EX ( 33000 ). 改过之后,我接着计算,但问题出现了,根本收敛不了,提示如下:
*** ERROR *** CP = 258.906 TIME= 17:00:36
Solution not converged at time 1.42105895 (load step 1 substep 14).
Run terminated.
*** WARNING *** CP = 258.984 TIME= 17:00:47
The unconverged solution (identified as time 3 substep 999999) is
output for analysis debug purposes.Results should not be used for
any other purpose. 其他设置都是一样的,为什么加了应力-应变曲线就不收敛了呢?这个简单例子是出自哪位高手之手估计已很难考究,为什么整出不收敛的例子?这么简单一个小梁尚且这么难收敛,估计模型再复杂一点,更是...... 还有一点要说明的,就是调整那个混凝土弹性模量的问题,我感觉这个影响应该不大,规范中建议的值是经验值,有实验数据最好,没有的话就按应力应变曲线的第一点输入,当然了,要是不考虑应力应变曲线的话,只有输入规范建议值了
写了这么多,还望各位高手踊跃发言
页:
[1]