24小时热门版块排行榜

返回列表

当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖

lfzlpd

铜虫 (小有名气)

应助: 11 (小学生)
金币: 31.4
帖子: 170
在线: 187.9小时
虫号: 1194711

[交流] 【求助】Casteps几何优化中，遇到这样的提示通常该怎没办？

+------------+-------------+-------------+-----------------+ <-- min BFGS
| Step | lambda | F.delta | enthalpy    | <-- min BFGS
+------------+-------------+-------------+-----------------+ <-- min BFGS
|  previous  | 0.000000 | 0.006740 | -24117.462487 | <-- min BFGS
| trial step | 1.000000 | -0.010705 | -24117.315957 | <-- min BFGS
|  line step | 0.386363 | -0.003071 | -24117.436342 | <-- min BFGS
|  quad step | 0.224039 | -0.001694 | -24117.449946 | <-- min BFGS
+------------+-------------+-------------+-----------------+ <-- min BFGS

BFGS: reverting to earlier configuration
BFGS: finished iteration 39 with enthalpy= -2.41174625E+004 eV

+-----------+-----------------+-----------------+------------+-----+ <-- BFGS
| Parameter |    value    | tolerance | units | OK? | <-- BFGS
+-----------+-----------------+-----------------+------------+-----+ <-- BFGS
|  dE/ion | 1.848121E-005 | 5.000000E-005 |       eV | Yes | <-- BFGS
|  |F|max | 5.391142E-001 | 1.000000E-001 |    eV/A | No  | <-- BFGS
|  |dR|max  | 2.622797E-002 | 5.000000E-003 |       A | No  | <-- BFGS
+-----------+-----------------+-----------------+------------+-----+ <-- BFGS

BFGS: Warning - Repeated consecutive reset of inverse Hessian
BFGS:          without satisfying convergence criteria which
BFGS:          looks like BFGS has run out of search directions.
BFGS: Warning - Lets try allowing some uphill steps and see if
BFGS:          we can get around this barrier.
BFGS: Warning - It is possible that the system may now converge to
BFGS:          a stationary point OTHER than the desired minimum.
BFGS: Hint - this may be an indication that either:
BFGS:          a) you are using a poor guess at geom_frequency_est
BFGS:          and/or geom_modulus_est, or
BFGS:          b) you are using unrealistic convergence criteria.
BFGS:          Suggest therefore that you consider changing them!

================================================================================
Starting BFGS iteration       40 ...
======================================================================小弟金币仅60，分一半出来，答谢各位！

回复此楼

» 猜你喜欢

» 本主题相关价值贴推荐，对您同样有帮助:

有研究随机几何，积分几何或者几何概率的朋友没啊已经有7人回复
哪位知道杜克大学所在的Durham，房租几何？已经有4人回复

» 抢金币啦！回帖就可以得到:

查看全部散金贴

1楼 2011-02-21 11:07:23

已阅回复此楼关注TA 给TA发消息送TA红花 TA的回帖

souledge

专家顾问 (著名写手)

★ ★ ★ ★
lfzlpd(金币+2):谢谢参与
zhang668(金币+3): 多谢详细指导！ 2011-03-23 20:33:10
lfzlpd(金币+5): 谢谢，参数我是改过SCF tolerance我改大了，但我想是结构的问题。另外我正准备用你所说的‘以先简单的用能量判距来找一下合适的模型参数，然后再进行进一步的结构优化’。但你索引的的那两句话，我确实不太明白内涵，望高人指点！ 2011-03-23 21:35:24

引用回帖:

GEOM_MODULUS_EST The estimated bulk modulus which will be used to optimize the cell relaxation where appropriate.
Default Value: 500 GPa
GEOM_FREQUENCY_EST The estimated average phonon frequency at the gamma point which will be used to optimize the ionic relaxation where appropriate.
Default Value: 50 THz

根据软件报错信息，提到了有两个可能，一是上面的两个参数中至少一个设置有问题，如果LZ从来没有设置过，那就是这两个参数的默认值不适合LZ的模型；第二个就是收敛的判距不太靠谱~
另外再说一个，似乎LZ的结构是处于一定的震荡中，这也有可能是初始的结构离平衡位置有点远或者变化比较复杂决定的，可以先简单的用能量判距来找一下合适的模型参数，然后再进行进一步的结构优化。

[ Last edited by souledge on 2011-3-23 at 20:33 ]