Functional neural networks-based model for prediction of the static young's modulus for sandstone formations

This study is aimed to develop a functional neural networks (FNN) model to estimate Estatic for sandstone formations as a function of the conventional well log data of the bulk formation density, compressional time, and shear time. 355 well log datasets from Well-A were used to train the FNN model which was then tested on another 237 datasets from Well-A and validated on 38 data points from Well-B. The developed FNN-based model predicted the Estatic for the training dataset with a very low average absolute percentage error (AAPE) of 0.78%, a very high correlation coefficient (R) of 0.9995, and a coefficient of determination (R²) of 0.999. For the testing dataset, the Estatic was predicted with AAPE, R, and R² of 0.85%, 0.9993, and 0.999, respectively. The optimized FNN model predicted the Estatic for the validation data with AAPE of 2.54%, R of 0.997, and R² of 0.995. The obtained results confirmed the high accuracy of the developed FNN model in estimating the Estatic.