Predicting Oil Production Rate with Artificial Neural Network

Meet Charles Kporxah, a former student of Mawuli School-Ho and a former petroleum engineering student at University of mines and technology-Tarkwa, in his final year project made profound research on ways of predicting oil production rate with an artificial neural network.

In his research, he brought to light the already known means of estimating the rate of oil production which is through the decline curve analysis methods (PDA, which is a graphical procedure used for analyzing declining production rates and forecasting the future performance of oil and gas) and empirical correlations. His finding significantly reaffirms a different approach for predicting oil flow rate using the artificial neural network technique. His developed model predicts oil production rate as functions of gas rate, production time, flowing bottom-hole pressure and tubing head pressure. The accuracy of the developed artificial neural network model was compared with the decline curve analytical methods. Results from the comparison showed that oil production rates predicted by the artificial neural network model had a mean absolute percentage error of 3.18% and a correlation coefficient of 0.9966, which are in perfect agreement with the actual measured data from his experiments compared to the exponential, harmonic and hyperbolic methods.

The oil and gas production industries is a very crucial economic sector in the world’s businesses today and the world needs to discover and give a chance to such individuals with versatile minds like Charles’ to explore and make this aspect of the world a better place to live in.

If you are interested in more of his research work and will like to connect with him, you can via his email

3 thoughts on “Predicting Oil Production Rate with Artificial Neural Network

Leave a Reply