PREDI CTION OF PERMEABILITY REDUCTION IN POROUS MEDIA DUE TO CALCIUM SULFATE SCALING USING NEURAL NETWORK

In this research a new model was presented for prediction of permeability ratio and its reduction due to calcium sulfate scaling during waterflooding using multilayer perceptron and test data. To achieve the maximum efficiency, number of neurons, training function, number of hidden layer and activation function were optimized for multilayer perceptron (MLP) model. The MLP model predicted permeability ratio after waterflooding with total average absolute deviation of 3% whereas the recommended correlation from later studies had about 13% error.

کلیدواژه ها:

Permeability reduction ، Multilayer perceptron ، Calcium sulfate scale ، Porous media