Computational Intelligence Research in Geomechanics: A Fresh Perspective for an Optimize Horizontal Stress Calculation Using Seismic-Pressure Information

Horizontal stress, including bath minimum (Shmin) and maximum stress (SHmax), stands as a pivotal geomechanical factor in subsurface assessments. Since measuring horizontal stress directly is costly and time-consuming, and sometimes difficult or impossible, exploring indirect approaches is essential. A wide range of indirect methods has been developed to calculate horizontal stress using varying datasets. This research aims to predict horizontal stress swiftly, accurately, and cost-effectively through indirect means, utilizing intelligence methods and geophysical logs within a carbonate interval. To achieve this objective, the study employs the Artificial Neural Network (ANN) method, integrating input data such as density, pore pressure, overburden pressure (PP-OB), and compressional and shear transition time (DTC-DTSM) through three designed scenarios. In the initial scenario, the input parameters consisted of pore pressure and density. Subsequently, the second scenario introduced compressional and shear transition times (DTC and DTSM) as additional input variables. Finally, in the third scenario, the input dataset was expanded to include overburden pressure, along with compressional and shear transition times (DTC and DTSM). Based on the results, the most accurate estimation for Shmin and SHmax was achieved in the second scenario with an R-value of ۰.۹۸, where the most comprehensive dataset was utilized. There was no significant difference between the results of the first and third scenarios, with a marginal difference of ۰.۰۱; the former and latter had ۰.۹۴ and ۰.۹۵ R-values, respectively. In conclusion, the ANN method demonstrated excellent performance in predicting minimum and maximum horizontal stress using seismic and pressure data in the study area.

کلیدواژه ها:

Horizontal stress ، Seismic data ، Pore pressure ، Overburden pressure ، Artificial Neural Network ، Computational rock mechanics