页岩气吸附模型的适用性研究

Study on the applicability of adsorption models to shale Gas

页岩气吸附特征对于页岩气储量预测、渗流机理分析、产能评价有重要影响,而吸附模型是描述页岩气吸附特征的主要方法之一。目前国内外学者在页岩气吸附模型的选用上存在一定分歧,即对页岩气吸附模型的适用性认识不清。针对此问题,通过吸附模型与页岩对甲烷等温吸附实验数据的拟合,就4种常用吸附模型(Langmuir模型、BET模型、D—R模型、半孔宽模型)对于页岩气的适用性展开了深入研究。结果认为:Langmuir模型、BET模型未考虑到页岩孔隙表面各个活性位吸附能级的差异,当压力较低时,模型拟合误差相对较大;D—R模型考虑低压力区间的超高吸附量,只适用于微孔占孔隙比重较大的页岩;半孔宽模型将吸附势理论与页岩的孔径分布进行了综合考虑,与实验数据的拟合结果较好。4种典型吸附模型与鄂尔多斯盆地富县区页岩对甲烷等温吸附实验数据的拟合误差从小到大排序为:半孔宽模型、Langmuir模型、BET模型、D—R模型。

The adsorption characteristics of shale gas have great influence on its reserves prediction, seepage mechanism analysis and productivity evaluation. The adsorption model is one of the principal methods to describe shale gas adsorption characteristics. So far, scholars at home and abroad have different opinions on the selection of adsorption models for shale gas. It is revealed that the applicability of adsorption models to shale gas is not understood clearly. To solve this problem, the applicability of four common adsorption models （i.e., Langmuir model, BET model, D-R model and half pore width model） to shale gas was studied thoroughly by fitting the adsorption models to the experimental data of methane isothermal adsorption on shale. And the following research results were obtained. First, neither Langmuir model nor BET model considers the adsorption energy level difference of active sites on the surface of shale pores, so their fitting errors are larger when the pressure is lower. Second, D-R model considers the ultrahigh adsorption in the interval of low pressure, so it is only applicable to the shale possessing more micro pores. Third, half pore width model takes into consideration the adsorption potential theory and the pore diameter distribution of shale comprehensively, and it is well fitted with the experimental data. And fourth, the fitting errors between four common adsorption models and experiment data of methane isothermal adsorption on shale in Fuxian area of the Ordos basin are ranked from the low to the high, i.e., half pore width model, Langmuir model, BET model and D-R model.