[Amazon]

页岩气藏比较特殊，兼有致密气藏和煤层气藏的特点。在数值模拟上有很多挑战。

1. 渗流机理。页岩气藏的孔隙直接接近平均分子自由程，达西定律的适用程度问题。
2. 吸附气的处理。这跟煤层气差不多。在有的页岩气藏中吸附气的比例很高，甚至比游离气还高。
3. 压裂裂缝的处理。关系到产能问题，如何有效的模拟这些裂缝。
4. 气藏内天然裂缝的模拟。同样关系到产能，还有最终采收率的重要问题。

大家有什么见解可以讨论一下。

[ONLY]

引用框（Amazon ＠ 05-12-2010 - 20:42）

页岩气藏比较特殊，兼有致密气藏和煤层气藏的特点。在数值模拟上有很多挑战。

1. 渗流机理。页岩气藏的孔隙直接接近平均分子自由程，达西定律的适用程度问题。
2. 吸附气的处理。这跟煤层气差不多。在有的页岩气藏中吸附气的比例很高，甚至比游离气还高。
3. 压裂裂缝的处理。关系到产能问题，如何有效的模拟这些裂缝。
4. 气藏内天然裂缝的模拟。同样关系到产能，还有最终采收率的重要问题。

大家有什么见解可以讨论一下。

最近几天一直在看页岩气勘探开发方面的文献，由于我自己本身是学数模的，所以很自然的想到 页岩气的数值模拟如何去做，用哪个软件去做，怎么做，相关软件是否有这些功能。。

eclipse是美国产品，而美国自1821年就开始了页岩气的开发探索，到现在年产页岩气当量已超过5000万吨油当量。。
刚才仔细看了下eclipse软件，很遗憾，没找到相关功能模块。。

[Rico_cn]

页岩气对国内来说是新东西。北美的页岩气开发比较有经验，建议你找找这方面的资料。我去加拿大那边时了解到，因为ECL等软件比较贵，且不一定对页岩气适用，加拿大和美国的部分公司用自己的软件来进行页岩气田评价和数值模拟。有兴趣的话我们可以交流。

[Lilac]

CMG GEM
吸附部分很重要

[ONLY]

引用框（Rico_cn ＠ 10-09-2011 - 13:45）

页岩气对国内来说是新东西。北美的页岩气开发比较有经验，建议你找找这方面的资料。我去加拿大那边时了解到，因为ECL等软件比较贵，且不一定对页岩气适用，加拿大和美国的部分公司用自己的软件来进行页岩气田评价和数值模拟。有兴趣的话我们可以交流。

嗯 很想跟前辈请教下，我的QQ是359037864

[LXH536]

Shale in USA have both absorbed gas and free gas. Some Spe paper introduced that the free gas dominate the main productivity and reserve ， nor the absorbed gas. CMB gas is almost absorbed in the coal and tight gas is free gas， however Shale have both absorbed gas and free gas.

Eclipse and CMG etc. can handle shale simulation. Absorption is not the key point in shale gas simulation. Like tight gas， Hydaulic fracture network description is more critical for shale. SRV (stimulated Resrervior Volume）evaluation should be focus on.

[ytlee]

Shale gas simulation is fractures, geomechanical and well design.Geological is not critical but still required.

[akabc47]

引用框（LXH536 ＠ 01-17-2012 - 10:04）

Shale in USA have both absorbed gas and free gas. Some Spe paper introduced that the free gas dominate the main productivity and reserve ， nor the absorbed gas. CMB gas is almost absorbed in the coal and tight gas is free gas， however Shale have both absorbed gas and free gas.

Eclipse and CMG etc. can handle shale simulation. Absorption is not the key point in shale gas simulation. Like tight gas， Hydaulic fracture network description is more critical for shale. SRV (stimulated Resrervior Volume）evaluation should be focus on.

裂缝是如何模拟的？能不能详细讲讲？

[ytlee]

Fracture simulation is more an art than science. But when you boil down to it, it is basically local grid refinement (LGR)and dual/tri porosity modeling. If you have local knowledge of frac orientation and distribution, you can use a correlation function to distribute. For example, austin chalk fractures in texas in general are east-west orientated, for a N-S horizontal well, we can use strip of thin grids along the wellbore. If you have Petrel, you can get an ocean plugin to do LGR modification for you. If I'm not mistaken, the feature is built in the new Petrel. Mike Du & Xu Zhang of SLB wrote a template and article on ECL platform in SIS newsletter, you should check them out.

[LXH536]

LGR is the common technique useing thin grid for descriping hydraulic planar fracture. However， the actual fracture width range from 0.01~0.1 inch，which will probably induce convergence problem and lots of CPU time if the grid width within the range. Usually， engineers use acceptable grid width and the equivalent conductivity to get the similar results.

Pebi grid or hybrid grid is also an option to accurately mimic the actual geometry of fracture. Kappa‘s product and Ecl have this function. Some simulators developed base on finite celement technique are perpect to describe the fracture geometry.

However, accurate geomety models are not always mean thye can get accurate prediction. Remenber this if you are work for petro industry.

[akabc47]

感谢楼上两位的讲解。听了讲解，个人感觉靠画格子的方法来模拟裂缝是非常不靠谱的。不知楼上两位对LGR方法的效果是什么感觉？模拟的比较好还是比较差？

[ytlee]

Each LGR workflow has its pros and cons. When you are into fracture modeling, chances are, you have to write your own simulator or develop special function if you are using eclipse. I don't know CMG well enough to comment. but I believe they are using dual porosity & grid properties too. If my memory doesn't fail me, cmg has a keyword FRACTURE that modifies POR

If you can't afford your own simulator or correlation function, you can use Petrel Tartan grid in make model geometry tab for ECL. Tartan grid is special LGR grid to speed up adding strips or region of thin cells. To set up the Tartan geometry is not straight forward, SLB has an excel spreadsheet to assist you. You can search "Tartan" in support portable for further detail and the spreadsheet. It also has a demo project for setting up the frac system. It is very simplicity but should give you an idea how to use Tartan grid.

For those who knows how to use bitmap to setup properties in Petrel, you can create a fracture system pattern and randomly distribute or by trend to setup a modifier property, (I used a DOE program called FRACGEN http://www.netl.doe....cgenNFFLOW.html to generate the bitmaps) then use the new property to modify porosity, permeability or transmissible.

[yolanda]

I would suggest you use Fekete RTA (perhaps Decline+ is better) or similar software to analyze basic reservoir charactertistics and then use those to guide your simulation. You can assume fracture perms as part of the matrix perms and may use a higher value than a real rock matrix may reveal.

For most shale gas, early production is dominated by free gas, and that is almost always you would like to concern.

引用框（Amazon ＠ 05-12-2010 - 06:42）

页岩气藏比较特殊，兼有致密气藏和煤层气藏的特点。在数值模拟上有很多挑战。

1. 渗流机理。页岩气藏的孔隙直接接近平均分子自由程，达西定律的适用程度问题。
2. 吸附气的处理。这跟煤层气差不多。在有的页岩气藏中吸附气的比例很高，甚至比游离气还高。
3. 压裂裂缝的处理。关系到产能问题，如何有效的模拟这些裂缝。
4. 气藏内天然裂缝的模拟。同样关系到产能，还有最终采收率的重要问题。

大家有什么见解可以讨论一下。