A Data-Driven Approach for Modeling the Porosity of Rocks Using Dynamic Elastic Mechanical Properties at Water-Saturated Condition

Porosity plays a crucial role in petroleum engineering as it directly affects the amount of hydrocarbon that can be stored in a reservoir rock. In this study, the relationship between the porosity of rocks and their dynamic elastic mechanical properties was investigated using two machine learning algorithms. A dataset consisting of ۹۶۰ rock samples with various lithologies was utilized, and two regression algorithms, namely Gradient Boosting (GB) and Random Forest (RF), were employed. The coefficient of determination values obtained for the models were ۰.۶۷ and ۰.۶۴, respectively; indicating a potentially meaningful relationship between the elastic mechanical properties and the porosity of rocks. It was found that the porosity of rocks could be modeled using the Young's modulus and the shear modulus as input parameters for the models. The results suggest that the dynamic elastic mechanical properties of rocks are related to their porosity, although the exact nature of this relationship remains unclear and further research is needed to explore this relationship in more detail.

Keywords:

Rock Properties , Porosity , Mechanical Properties , Young's modulus , Shear modulus , Machine Learning