Network Parameters to Predict the Well Productivity Index

For development and evaluation of oil and gas fields, having enough information about hydrocarbon reservoirs is necessary and inevitable, in other words, set various parameters in the evaluation of reservoir rock and fluid is really important in hydrocarbon reservoirs. Reservoir productivity index of the wells can be considered as the most important parameter in determining the economic value of a reservoir. Productivity index of the wells accompanying with certain reservoir parameters play an important role in the evaluation of oil and gas reserves. Therefore, an alternative method for determining the properties of artificial intelligence and machine learning has pointed out the use of tools. The main objective of this study is utilizing a non-linear optimization technique called artificial neural networks to predict reservoir parameters. Data for the various stages of learning and assessment networks are divided into three categories of training, validation, and testing after data processing network 70% of them are placed for education, 15% for validation, and 15% for the MLP experiments. The results show that neural network with two hidden layer simulators do best in terms of simulation.

Keywords:

data mining , artificial neural network , well testing , permeability , porosity and well productivity index