TOC determination of Gadvan Formation in South Pars Gas field, using artificial intelligent systems and geochemical data

Potentially, TOC content is affected by logging data in a source rock (density, sonic, neutron and resistivity logs). Hence, to analyze these logs, which we make a quick and reliable assessment of a source rock. So, it is a quick and economically cheaper method rather than direct geochemical analysis. A source rock interval poses to less density, lower velocity, higher sonic porosity, higher gamma ray values and increase in resistivity. In this research, Gadvan Formation was studied in two boreholes as potential of source rock. The log data of two wells were used to construct of intelligent models in a source rock of the South Pars Gas field in southwest of Iran. A suite of geophysical logs (neutron, density, sonic and resistivity logs) and cutting chip data samples data were applied for determining TOC content of this formation. Rock-Eval pyrolysis data reveal that Gadvan Formation is poor source rock (less than 0.5%). Hence we attempted a correlation between geophysical data and direct TOC content measurements of using ∆ Log R, Rock-Eval, neural network and fuzzy logic techniques.The results showed that intelligent models were successful for prediction of TOC content from conventional well logs data. Meanwhile, similar responses from other different intelligent methods indicated that their validity for solving complex problems.

Research Highlights
► The intelligent models are successful for predication of TOC content from conventional well logs data.
► The intelligent systems were used for TOC prediction, have different concepts to solve problem.
► The measured TOC from samples in Rock−Eval pyrolysis, and calculated data based on ∆ Log R method and predicted data versus depth shows a good agreement for two intelligent systems.