No abstract

High-pressure methane adsorption isotherms were measured on five shale core samples obtained during exploratory drilling from three boreholes located in the Colombian Middle Magdalena Valley Basin. The experiments were carried out at 50 and 75 °C and for pressures ranging up to 3.5 MPa under dry conditions through the use of a homemade manometric setup. The effect of the total organic carbon (TOC) content, thermal maturity, clay content, and specific surface area (SSA) on methane adsorption capacity has been discussed. The excess adsorption data were fitted to a three-parameter (n L , p L , and ρ ads ) Langmuir model with the value of the adsorbed phase density, ρ ads , maintained fixed at 421 kg/m 3 , which corresponds to liquid-phase density of methane at a normal boiling point. An excellent fit to the experimental data was achieved. The results show that the temperature has a negative effect on the adsorption capacity, while TOC has a positive effect, even if no linear regression was found between TOC and methane adsorption capacity. No correlation was observed between the clay content and the TOC-normalized adsorption capacity to methane, which indicates that clay minerals do not significantly contribute to methane adsorption in the case of our samples. In addition, there is not a general trend between TOC normalized and thermal maturity. Among the factors investigated in the present study, TOC has the major contribution to the adsorption uptake. A similar contribution is found for the SSA, which is consistent, considering the positive correlation between TOC and SSA. This set of data represents meaningful information for indirect estimations of the gas in place during the future recovery strategies. This study furthers the ongoing projects on the understanding of the adsorption effect on shale gas production and assessment.

*a quien debe dirigirse la correspondencia RESUMEN el comportamiento de diferentes modelos de isotermas de adsorción aplicados a los datos experimentales obtenidos por Gasparik et al. (2012) en una prueba de adsorción de metano, fue modelado y analizado, con el fi n de determinar el que mejor representa los datos experimentales. entre los modelos utilizados se encuentran algunos de dos parámetros como Langmuir, temkin y freundlich, otros de tres parámetros como Langmuir, DubininRadushkevich, Toth y Sips; de cuatro parámetros como Dubinin-Astakhov y Fritz-Schlunder y fi nalmente el modelo de cinco parámetros propuesto por fritz-schlunder. según los resultados obtenidos, los modelos presentaron mejor ajuste a medida que empleaban mayor número de parámetros en sus ecuaciones.Posteriormente se calculó la capacidad de almacenamiento total de gas para una formación hipotética con algunos datos de la formación Posidonea empleando el método propuesto por Ambrose et al (2010), el cual se modifi có reemplazando el término que tiene en cuenta el gas adsorbido con los modelos mencionados anteriormente, a fi n de determinar su infl uencia en las estimaciones de volumen de gas libre y de gas adsorbido. Con los modelos que presentaron un menor porcentaje de error, se obtuvieron datos más conservadores en el cálculo del gas adsorbido y de gas total, lo cual indica la importancia de la adecuada selección del modelo para el cálculo de reservas, dado que conlleva alto impacto técnico y económico.Palabras clave: shale, gas, yacimiento, isotermas, adsorción. Application of adsorption isotherm models of methane on a shale sample and their impact on reserves estimation ABSTRACTSeveral adsorption isotherm models were analyzed for the purpose of fi tting the experimental data published by Gasparik et al. (2012) obtained in a methane adsorption experiment on a shale sample from the Posidonea formation in the netherlands. the following are the ten models used in this study; some of them consist of two parameters such as Langmuir, temkin and freundlich, the three parameter models: Langmuir, DubininRadushkevich, Toth and Sips, four parameters: Dubinin-Astakhov and Fritz-Schlunder and the fi ve-parameter Fritz Schlunder correlation. The results show a better fi t to the experimental data for those models with the highest number of parameters.