Abstract Permeation models were modified to determine the permeability of C3H8, CH4 and H2 gases through the silica/PDMS nanocomposite membrane accurately. The average absolute relative error (AARE %) of permeation models is about 10–15% for low-sorbing gases of CH4 and H2 and approximately 30% for high-sorbing gas of C3H8. Furthermore, the gas solubility around the nanoparticles in the polymeric matrix is more than in the polymeric bulk due to sorption effects around the nanoparticles. Therefore, a new interface region is assumed to consider the sorption effects as a function of nanoparticle volume fraction and polymeric bulk permeability in the permeation models by genetic algorithm. It was found that the best-modified model for C3H8 and H2 is Lewis–Nielsen and for CH4 is Bruggeman. Finally, the AARE values were obtained less than 5% for C3H8 and less than 2% for CH4 and H2 that were in a good agreement with experimental data.

中文翻译：

---

纳米复合膜渗透模型的新改进：考虑吸附效应

摘要 修改渗透模型以准确测定 C3H8、CH4 和 H2 气体通过二氧化硅/PDMS 纳米复合膜的渗透率。渗透模型的平均绝对相对误差 (AARE %) 对于 CH4 和 H2 的低吸附气体约为 10-15%，对于 C3H8 的高吸附气体约为 30%。此外，由于纳米颗粒周围的吸附效应，聚合物基质中纳米颗粒周围的气体溶解度大于聚合物本体中的气体溶解度。因此，假设一个新的界面区域将吸附效应视为通过遗传算法在渗透模型中作为纳米颗粒体积分数和聚合物体积渗透率的函数。发现 C3H8 和 H2 的最佳修改模型是 Lewis-Nielsen，而 CH4 的最佳修改模型是 Bruggeman。最后，