[shiyougong]

做组分模型有段时间了，但还是有两个基本的问题，没有搞清楚，希望高手指点。我用的是CMG软件中的GEM模块。
（1）组分模型做注水方案可不可以？因为涉及到一个注气和注水方案的比较问题，是应该用黑油模型模拟注水，用组分模型模拟注气，然后比较，还是都在组分模型中做呢？有点困惑。。。
（2）组分模型之前有个地层油的相态拟合，CMG中是Winprop模块。现在的问题是做了两种气体（氮气和天然气）的注入气膨胀实验，那拟合膨胀实验的时候应该怎么做呢？ 是该两种膨胀实验放在一块儿拟合（不过这在Winprop里实现不了，因为只能加一种注入气，听说Eclipse里的PVTi可以实现？）还是两种气体的膨胀实验单独拟合，那样就得到两组状态方程的参数了，导入GEM中，又得到两组模型，分别模拟计算两种气体的注入方案？
拜谢高手指点一二。。。

[yujimo]

组份模型中也可以做水驱的过程，不用分开做的

[jsbreeze]

引用框（shiyougong ＠ 2009-12-10）

做组分模型有段时间了，但还是有两个基本的问题，没有搞清楚，希望高手指点。我用的是CMG软件中的GEM模块。
（1）组分模型做注水方案可不可以？因为涉及到一个注气和注水方案的比较问题，是应该用黑油模型模拟注水，用组分模型模拟注气，然后比较，还是都在组分模型中做呢？有点困惑。。。
（2）组分模型之前有个地层油的相态拟合，CMG中是Winprop模块。现在的问题是做了两种气体（氮气和天然气）的注入气膨胀实验，那拟合膨胀实验的时候应该怎么做呢？ 是该两种膨胀实验放在一块儿拟合（不过这在Winprop里实现不了，因为只能加一种注入气，听说Eclipse里的PVTi可以实现？）还是两种气体的膨胀实验单独拟合，那样就得到两组状态方程的参数了，导入GEM中，又得到两组模型，分别模拟计算两种气体的注入方案？
拜谢高手指点一二。。。

关于问题2， Winprop是可以同时拟合多个实验的。就你的情况而言，你可以先把secondary composition设为氮气，加膨胀实验； 然后再加一个composition输入，将secondary composition设为实验用的天然气组分，加相应的膨胀实验，然后对两个实验同时进行拟合。Winprop的界面从上到下看起来是这样的：
Titles/EOS/Units
Component Selection/Properties
Composition (add your primary composition and Swelling gas composition1 as secondary composition here)
regression parameters
Swelling test
Composition(add your primary composition and Swelling gas composition2 as the secondary composition here)
Swelling test
End Regression

Winprop不仅可以对同一个样本的多个不同实验同时进行拟合， 也可以对不同样本的多个实验同时进行拟合。只要样本属性不相差悬殊，比如说heavry oil 和 wet gas. 属性相似的样本是可以用同一个状态方程来表达的。

[LXH536]

1. I just tested the same model using Black oil and Compositional model. The results shows that there is little difference of FOIP (about 5%) and the calculated behavior,such as Gas production, filed pressure,and water cut is also different from each other. In my opinon, the injection gas may be the main reason that the black model is different from compositinal model. If somebody know much about this problems,welcome to communication with me.
2. Remember that the PVTi experiment just get the EOS of each components. May be you should regression the result using Swelling 1 ,SWelling 2, DL, CCE at the same time.
3.Shiyouyong. I note that you are a work hard young guy . Our company need the guy like you.When you leave school, I suggest you to consider our company.aha

[shiyougong]

引用框（jsbreeze ＠ 2009-12-11）

关于问题2， Winprop是可以同时拟合多个实验的。就你的情况而言，你可以先把secondary composition设为氮气，加膨胀实验； 然后再加一个composition输入，将secondary composition设为实验用的天然气组分，加相应的膨胀实验，然后对两个实验同时进行拟合。Winprop的界面从上到下看起来是这样的：
Titles/EOS/Units
Component Selection/Properties
Composition (add your primary composition and Swelling gas composition1 as secondary composition here)
regression parameters
Swelling test
Composition(add your primary composition and Swelling gas composition2 as the secondary composition here)
Swelling test
End Regression

Winprop不仅可以对同一个样本的多个不同实验同时进行拟合， 也可以对不同样本的多个实验同时进行拟合。只要样本属性不相差悬殊，比如说heavry oil 和 wet gas. 属性相似的样本是可以用同一个状态方程来表达的。

非常感谢你的帮助，你彻底解决了我的第二个疑问，其实我也是觉得可以的，但就是不清楚怎么解决。因为我从winprop的manuals里没有找到解决之道，不知道大侠是通过什么方法解决这个问题的？自己琢磨出来的，还是更高人指点？
对于同一个油样来说，本来就该得出同一组状态方程的参数，这一点跟老师有点分歧，这个现在看来也是没有争议了。
不过我在实际拟合的时候，由于一直不能得出非常好的拟合效果，最后还是得出了几个备选的相态拟合后的方案，导入组分模型中看各个方案的历史拟合的效果，然后最后选定了一个。呵呵。
再次感谢，我又见到了一个组分模型的高手。很高兴。

[shiyougong]

引用框（LXH536 ＠ 2009-12-11）

1. I just tested the same model using Black oil and Compositional model. The results shows that there is little difference of FOIP (about 5%) and the calculated behavior,such as Gas production, filed pressure,and water cut is also different from each other. In my opinon, the injection gas may be the main reason that the black model is different from compositinal model. If somebody know much about this problems,welcome to communication with me.
2. Remember that the PVTi experiment just get the EOS of each components. May be you should regression the result using Swelling 1 ,SWelling 2, DL, CCE at the same time.
3.Shiyouyong. I note that you are a work hard young guy . Our company need the guy like you.When you leave school, I suggest you to consider our company.aha

1. What you do in your "1" suggestion indicate that u are a very executable person. I mean that when u meet one new question, u will quickly think the sloving methord out, just like one of my best friend who also do work of the reservior numerical simulation.
2. "In my opinon, the injection gas may be the main reason that the black model is different from compositinal model." in my opinion, when gas injecting, the best choice is the compositional model, cause the compositional can consider the composition exchange. Now the problem is whether it is neccessary or right to use the compositional oil model to simulate the waterflooding process to compare the effect of the waterflooding and the gas injection.
3. i don't think the pvti or winprop just get the EOS of each componets, they should get the EOS of the mixture(oil-gas system).
4. i am very surprised and very glad to see ur aprreciations, and i must say sorry to u, because i have get an offer from COONC limited. And i don't want to lose faith to that company. Anyway, thank u very much for your appreciation. i will remember u

[wangwanggou]

学习了 谢谢