Crude Oil Pressure Bubble Point (Pbp) Prediction and Comparison with Marhoun Equation using Neural Networks
In this project, I have used Neural Networks with different activation functions and settings to find the best neural network setting that would successfully predict the Pressure Bubble Point in Crude Oils and compared it with Marhoun Equation which is widely used to estimate the Pressure Bubble Point.
- Analyzed the data and used Spearman's rank correlation coefficient and Random forest to identify features that contribute the most to the output (Pbp).
- Explored different Neural Network architectures and Activation functions combinations and successfully obtained a Neural Network model with a R2 Score of 0.91.
- Neural Network model output (R2 = 0.91) outperformed the Marhoun Equation output (R2 = 0.866).