求助，等着结果出来好回家过年啊，求助大人

Analyst

这是一个非线性计算的问题
主要是弧长法计算材料的非线性问题，采用的MISO准则。计算反复不收敛，由于模型比较大，所以建模的程序没法发出来，模型大体上是圆柱形筒仓，顶部12个注脚受力，加载在顶部节点（组合名为load）下面是计算部分命令流，

FINISH
/prep7
TB,MISO,1,1,3
TBTEMP,0
TBPT,,0,0
TBPT,,0.0011407766863,235e6
TBPT,,0.023262,280.841e6
FINISH
/solu
antype,static
eqslv,SPARSE,1.0e-5
AUTOTS,ON
fcum,repl
CMSEL,S,LOAD
F,ALL,FZ,-(1.3E6)*1
ALLSEL
nlgeom,on
kbc,0
nsubst,100
!ARCLEN,OFF
ARCLEN,ON,50,0
outres,ALL,ALL
solve
fini

请大牛指教

tonnyw

Don't put too many things together. First run a model like the following:
1. Instead of multi-linear isotropic hardening, use bilinear isotropic hardening
2. Don't consider geometric nonlinearity
3. Don't use automatic time stepping
4. Don't use line search and arch-length method.

In sum, set up a model with bilinear isotropic hardening and specify a certain number of load steps. You may also play with convergence criteria.

oceanlhf

你把*.ERR文件贴出来，哥给你看。

oceanlhf

Don't put too many things together. First run a model like the following:
1. Instead of multi-linear isotropic hardening, use bilinear isotropic hardening
2. Don't consider geometric nonlinearity
3 ...
tonnyw 发表于 2010-2-7 13:25

这四个都是加强收敛的呵呵。

oceanlhf

1# Analyst
你把*.ERR文件贴出来，哥帮你看看。
同时，非线性计算很多情况下是不必收敛的，算多少步算多少步。比如结构的倒塌。

Analyst

5# oceanlhf

ERR文件只有这样的几个信息，
*** ERROR ***     SUPPRESSED MESSAGE    CP =    4586.648   TIME= 09:56:19
Solution not converged                                                  
  The arc-length value reached the minimum allowed value                 
  Run terminated.

Analyst

新的求解步骤为

FINISH
/prep7
tb,biso,1,2
tbdata,1,235,2890.04e6
FINISH
/solu
antype,static
fcum,repl
CMSEL,S,LOAD
F,ALL,FZ,-1.4E6
ALLSEL
nlgeom,off
!kbc,0
nsubst,20
ARCLEN,ON,4
outres,,1
solve
fini

tonnyw

7# Analyst

I don't think your bilinear stress-strain curve input is correct. It has to be Tangent modulus and initial yield stress. Turn off your arch-length option too.

si13

8# tonnyw
赞·

Analyst

8# tonnyw

关掉弧长法直接用N-R法的话 是没法追踪后屈曲行为的

Analyst

8# tonnyw

感谢感谢啊，的确有些问题 检查了
几何非线性和材料非线性在N-R法下市可以算的

弧长法应该也不是问题

的确是材料模型选择有些问题