Density Well Log Prediction in X Field Niger Delta using Ensemble Learning Models and Artificial Neural Network

Abstract

Performing reservoir characterization in exploration with limited data can be very is challenging. Various approaches are used to estimate values away from the well location. In this study, the density log, which is important for porosity analysis, was missed in one of the five available well log datasets. To solve this problem, an artificial neural network (ANN) approach was used to synthesise a density log (RHOB) from available and measured Gamma Ray (GR) log, Sonic (DT), water saturation (SW), and related Depth of 3 wells in the field. The performance of the prediction was evaluated using the fourth well. Five models were constructed with different optimizers from machine learning with a neural network made of an input layer with 5 neurons, a hidden dense layer with 32 neurons and an output dense layer with 1 neuron. The models were constructed based on Nesterov-accelerated Adaptive Moment Estimation (NADAM), Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent (SGD), and Root Mean Square Propagation (RMSP) optimizers, and an Ensemble model which combined the four optimizers. The test on actual data showed very low mean absolute errors of 0.0262, 0.0278, 0.0270, 0.0309, and 0.0248 and high coefficients of determination (R2) of 0.8832, 0.8746, 0.8986, 0.8858, and 0.9051 between the predicted and the actual data obtained for NADAM, ADAM, SGD, RMSP, and the Ensemble model, respectively, after 25 epochs. These indicated high performance of the Ensemble Learning model, suggesting that the constructed model can be used to predict the well that lacks RHOB.