No abstract

In this work, artificial neural network modeling for CO2 adsorption on various types of Marcellus shale samples is studied. The eight shale geometries are investigated for their CO2 adsorption at 298k and up to 50bar pressure utilizing a gravimetric technique and magnetic suspension balance. ANN modelling was applied to investigate three main objectives which are the impact of various training algorithms, various data initiation points, and altered training/validating ratios and number of neurons required for ANN model. The work can provide insightful knowledge linked to the impact of each of the studied parameters which play an important role in ANN modeling and training algorithms. The outcomes can provide an optimized matrix for unconventional resources, enhanced gas and oil recovery applications intend to apply the artificial intelligence modeling in their assessments.

The CO2 injection in unconventional reservoirs is one of the promising techniques for improved hydrocarbons recovery (IOR) and eliminated greenhouse gas emissions. In this work, the Physico-chemical characteristics of shale rocks, and the influence of those characteristics on CO2 adsorption and desorption are investigated. The mineralogy of four Marcellus shale samples, in addition to the morphology, elemental distribution, functional groups, total organic content, and porosity are indicated and discussed. The adsorption measurements utilizing gravimetric approach show up to 35mg/g and presented obvious correlation of clay minerals existence such as Dolomite, Illite, and Kaolinite at different weight ratios i.e. up to 60wt.% with CO2 adsorption and retention. Open-ended Pore shapes and mesoporous zone are illustrated on the mesoporous shale geometries. The applied numerical equilibrium isotherms modelling, which are Langmuir, Freundlich, Toth, and Tempkin models, are reflected multilayer and heterogeneous gas molecules distribution following the monolayer and homogeneous adsorption. The kinetic measurements showed a rate of adsorption variation, which reflected the total organic content and porosity variations impacts. The temperature program desorption of the pre-adsorbed CO2 gas revealed strong chemisorption on shale matrix, which requested up to 500°C and 2675.538KJ/mol to break down the CO2 bonding from shale geometry. The findings supported the CO2 utilization as water competitor and shale matrix as a storage environment for further utilization or environmental assortments. Keywords: CO2 adsorption/desorption, Marcellus Shales, Kinetics, Hydraulic fracturing, Enhanced Gas Recovery (EGR).

No abstract