Potential of Hydrocarbon Generation of Pabdeh and Gurpi Formations on the Basis of ∆LogR and Neural Network Techniques in Wells A and B in an Oil Field in the South-West Iran

Hydrocarbon source rocks, the subject of various analyses and investigations, are well appreciated because of their capacity for oil generation. In the present study, attempts were made to evaluate source rock using a leap cost analytical method. For this purpose, some samples from these formations were analyzed by Rock- Eval. CO2 generated from these samples at 350ºС are calculated as mgHy/gr rock or Kg Hy/Ton rock. As this method requires more time and is relatively expensive, we offered a new method in which by calibration of data, obtained from analyzed samples, the result could be attributed to the whole interval of a formation. By calculation of S2, Tmax and TOC of analyzed samples collected from wells A and B for both Pabdeh and Gurpi formations in accordance with ∆LogR method, rescaling of Resistivity-sonic logs, Resistivity-Density logs and Resistivity-Neutron logs, TOC content of these formations were estimated. Comparison between these data and data obtained from direct sample measurements in lab and extrapolation of an equation that relates these data to S2 and TOC of sample analysis, TOC and S2 content of whole intervals of these formations were calculated through ∆LogR method. Then hydrocarbon generation potential of the Pabdeh and Gurpi formations were finally evaluated. These data were processed in a neural network method with forward back propagation capability designed as try and error structure within Matlab software. Final results are in good agreement with those data obtained from direct measurements.