Rocks have a wide range of densities because of mineralogy and porosity variations. This bulk density is a key property and playing an essential role in geologic structure interpretation and reservoir characterization, which help in well planning, drilling optimization, and completion and production strategies development. The practical measurements of rock density in the field are either through logging while drilling tools or wireline logging techniques, but with discontinuity and tools limitations. This work aims to develop a model for predicting real-time bulk density of complex lithology during drilling by utilizing the functional networks (FN) technique and using the drilling mechanical parameters as inputs. A vertical well containing sand, shale, and carbonate with 2912 data points was used to build the model. Several methods and types in FN algorithm were tested to obtain the best model. The obtained results showed that the developed model has high accuracy in both training and testing processes with correlation coefficient (R) and average absolute percentage error (AAPE) values of 0.98 and 1.1%, respectively. The model was validated with another well form the same field showing R and AAPE values of 0.98 and 1.2%, respectively. The developed reliable and robust model in this study can predict the real-time bulk density with high accuracy, which assists in formation characterization in a convenient method within the least time, effort, cost, and errors.
Because of mineralogy and porosity variations, rocks have a wide range of bulk density (ρb), which is a key property that has an essential role in many interpretations (Reichel et al., 2013). It is required for identification of porosity (Spross et al., 1993; Ellis, 2003), lithology and fluid content (Alger and Raymer, 1963), overburden and pore pressure (Burrus, 1998; Swarbrick, 2001; Zhang, 2011; Satti et al., 2015; Oloruntobi et al., 2018; Oloruntobi and Butt, 2019a), and geomechanical features such as shear modulus, bulk modulus, Young's modulus and rock matrix compressibility (Coates and Denoo, 1981; Chang et al., 2006; Ameen et al., 2009; Najibi et al., 2015; Adewole et al., 2016; Feng et al., 2019). Hence, prediction of rock density is highly important for interpreting subsurface geologic structure and reservoir characterization to develop well planning, drilling optimization, and completion and production strategies development (Onalo et al., 2018; Oloruntobi and Butt, 2018; Yusuf et al., 2019).
