求解模型发现未定义的值，明明是定义了的，怎么回事？...

dong321

Undefined value found in the equation residual vector.
There are 611 degrees of freedom giving NaN/Inf in the vector for the variable mod1.c2
at coordinates: (0.003,0.0198565), (0.003,0.02), (0.003,0.0196868), (0.003,0.0195124), (0.003,0.0193378), ...
There are 611 degrees of freedom giving NaN/Inf in the vector for the variable mod1.c3
at coordinates: (0.003,0.0198565), (0.003,0.02), (0.003,0.0196868), (0.003,0.0195124), (0.003,0.0193378), ...
There are 643 degrees of freedom giving NaN/Inf in the vector for the variable mod1.c4
at coordinates: (0.00457808,0.02), (0.00523563,0.02), (0.006125,0.02), (0.00484491,0.0196179), (0.00440356,0.019638), ...
There are 643 degrees of freedom giving NaN/Inf in the vector for the variable mod1.c5
at coordinates: (0.00457808,0.02), (0.00523563,0.02), (0.006125,0.02), (0.00484491,0.0196179), (0.00440356,0.019638), ...
There are 643 degrees of freedom giving NaN/Inf in the vector for the variable mod1.cH
at coordinates: (0.00457808,0.02), (0.00523563,0.02), (0.006125,0.02), (0.00484491,0.0196179), (0.00440356,0.019638), ...
There are 4717 degrees of freedom giving NaN/Inf in the vector for the variable mod1.p
at coordinates: (0.00457808,0.02), (0.00490685,0.02), (0.00523563,0.02), (0.00568031,0.02), (0.006125,0.02), ...
There are 1254 degrees of freedom giving NaN/Inf in the vector for the variable mod1.phil
at coordinates: (0.00457808,0.02), (0.00523563,0.02), (0.006125,0.02), (0.00484491,0.0196179), (0.00440356,0.019638), ...
There are 1254 degrees of freedom giving NaN/Inf in the vector for the variable mod1.phis
at coordinates: (0.00457808,0.02), (0.00523563,0.02), (0.006125,0.02), (0.00484491,0.0196179), (0.00440356,0.019638), ...

mxio

不合适的初始值，避免用0来初始化
模型上来会更明了，但不能保证能算收敛

dong321

mxio 发表于 2013-7-7 13:53
不合适的初始值，避免用0来初始化
模型上来会更明了，但不能保证能算收敛 ...

谢谢，我改下试试

dong321

模型在这，求指点，谢谢

dong321

mxio 发表于 2013-7-7 13:53
不合适的初始值，避免用0来初始化
模型上来会更明了，但不能保证能算收敛 ...

亲，模型传上来了

mxio

现在你的模型 不是你说的这个问题啊。。。
而是这几个变量值 有可能不能为负值 mod1.siec.Eeq_er1 mod1.siec.iloc_er1
而Darcy流动模块的话，推荐和组分传递一样，对求解域分别进行设置，因为它们并不相连，离散求解事实上是会遇上问题的
从整个模型来看，现在的问题主要出在siec上，请注意变量的大小和符号是否填写正确，是否合乎物理

dong321

mxio 发表于 2013-7-8 10:14
现在你的模型 不是你说的这个问题啊。。。
而是这几个变量值 有可能不能为负值 mod1.siec.Eeq_er1 mod1.sie ...

谢谢之前那个未定义的错误，软件自己解决了，我只是多点了几下计算而已。这个负值问题，上一次也出现了，我不知道点了哪个地方，就没有报错。上次计算结果趋势不对，这次只是重复上次工作，出现这些问题不知道如何下手。敢问一下，COMSOl是不是不靠谱啊？

mxio

是你自己设置有问题。。。 估计是你参数化扫描的参数值 不合适
虽然它经常算不了，但是报错的提示还是比较靠谱的

dong321

mxio 发表于 2013-7-8 16:14
是你自己设置有问题。。。 估计是你参数化扫描的参数值 不合适
虽然它经常算不了，但是报错的提示还是比较 ...

恩，好吧。多多学习。谢谢

蓝色天空0219

同样遇到问题了，新手求指导如何解决

xiaolcheer

把study删掉，全部自己手动定义每个步骤。我就是这个问题，不知道你们有没有用。貌似项目初始化的时候不要选择预设study为好