May-31-2022 - Tuesday Daily Update - Mohammad Elkady
Daily Update - Mohammad Elkady - (May-31-2022)
1. What did I do today?
I have read a paper from Journal of Petroleum Science and Engineering paper named “Neuro-simulation modeling of chemical flooding”. Also, I read SPE 200096-MS and I did meet with Dr. Misra.
2. What did I find interesting/Work on?
1) JPSE paper
· In this paper they developed an ANN model to predict both recovery factor and Net Present Value.
· The optimum architecture is one hidden layer with 8 neurons and training function of Bayesian regularization.
· Artificial Neural Network
o ANN is an information processing system which mimics the biological system.
o It is characterized as parallel, adaptive, non-linear, and distributive model.
o ANN can discover relationship between input and output no matter how complex it is.
o Too many neurons cause over fitting and too few neurons cause poor performance.
· Data used in this paper is from previous M.Sc. Thesis in which it was simulated using UTCHEM.
2) SPE 200096-MS
· The relative contribution of the gravity and capillary forces on the macroscopic scale can be estimated by the inverse Bond Number
· On the microscopic scale, buoyancy effect depends on a local Bond Number, NB, Trapping Number, NT, was introduced to account for both NC and NB
· CMG Only accounts for NC
· This paper simulates the laboratory scale static imbibition, they simulated Amott-cell and tried to match the experimental results. Then they simulated the dynamic imbibition using an injection source.
· This paper was not so useful because it simulates only lab scale experiments
3. What will I do next?
As discussed with the meeting with Dr. Misra, I will be fully focused on simulating new synthetic data as the first essential step. So, in the next few days, I will focus on learning more about CMG simulator through the 6-hour course I have. And I will continue reading papers about CMG Simulation for both CO2 and Chemical EOR.
Comments
Post a Comment