Estimation of geomechanical units using petrophysical logs, core and supervised intelligent committee machine method to optimize exploration drilling operations
Soheyl Pourreza,Farnusch Hajizadeh,Ali Kadkhodaie
DOI: https://doi.org/10.1007/s12517-023-11223-5
2023-02-08
Arabian Journal of Geosciences
Abstract:Determination of geomechanical units is the most important target in oil industry which usually has support on the main properties of rocks. Basic properties of rocks are usually understood by their accurate description in the field and the laboratory. Geomechanical unit determination in most subsurface studies is impossible, due to lack of cores and cuttings. In such situations, where the wire line logs are available, geomechanical units are determined instead. Using cluster analysis, in this study, practiced the geomechanical unit determination in one of the oil fields in southwestern Iran by coding with MATLAB software. The study is carried out on the petrophysical logs of a well containing acoustic, density, gamma-ray, neutron, induction, and caliper logs for dipole shear velocity estimations. Shear interval transit time (DTs)-fast component, shear interval transit time-slow part, and shear interval transit time-mean component are estimated from compressive interval transit time (DTp), gamma, neutron, and density logs using a neural network, fuzzy inference system, and the adaptive neuro-fuzzy inference system (ANFIS). The results are combined using a committee machine (CM) model, first through a simple averaging CM followed by a weighted averaging CM. Committee machine weights were optimized using the genetic algorithm method. The method of the weighted committee machine had the least estimation error and best correlation coefficient compared to other methods. Geomechanical moduli such as shear, bulk, and Young's modulus, uniaxial compressive strength (UCS), and the Poisson ratio were calculated using density log, compressional, and shear wave velocities. Elastic moduli were estimated in cases of three shear velocities, including DTs-fast, DTs-slow, and DTs-mean. The estimated geomechanical moduli were clustered through an agglomerative hierarchical clustering method employing Mahalanobis distance and Ward's linkage functions. I reduced the volume of elastic moduli calculated by creating geomechanical units using these clusters. Well data such as leak-off tests or mini-fracture and image logs are essential for constructing the geomechanical model. Since these data are rarely available, the geomechanical unit (GMU) concept is used in this study to cover such a gap (Gharechelou et al. 2022). Accordingly, three geomechanical units (GMUs) were designed throughout the reservoir interval. The results show a good agreement between the GMUs and rock properties based on other well logs such as gamma-ray logs, neutron logs, and core observations. Validity of the method in geomechanical unit determination is checked, and a suitable cut off level is defined for the cluster analysis. Result from this study is collected to be used in geomechanical unit determination of the other wells with no cores and cuttings.
geosciences, multidisciplinary