An Accurate Prediction of Well Inflow Performance Using Artificial Neural Networks

Well Inflow Performance Relationship (IPR) is one of the major concepts of production engineering and its accurate prediction is always a matter of importance. These predictions help the engineers to improve the well performance and thus handle the future production plans much easier. IPR models allow us to consider various operating conditions; determine the optimum production scheme, and design production equipment and artificial lift systems. The conventional method for these predictions in solution gas drive reservoirs is using Vogel’s reference curve. This curve is produced from a series of simulation runs from the reservoirs model proposed by Weller. This study is based on the reproduction of Vogel’s work using artificial neural networks. Artificial neural networks have been widely used and are gaining attention in petroleum engineering because of their ability to solve problems that previously were difficult or even impossible to solve. Neural networks have particularly proved their ability to solve complex problems with nonlinear relationships. This paper presents a model to predict the dimensionless oil flow rate (qo/qomax). It uses different parameters as inputs of a Multilayer Perceptron (MLP) Network. This model is able to establish the IPR curve with an overall accuracy better than Vogel’s equation. The results show very good model predictions for both the training and validation data and they are also in a great agreement with the actual values. This study proves the ability of the neural network to predict the dimensionless oil flow rate and to establish the inflow performance relationship with a considerable accuracy

کلیدواژه ها:

IPR ، Vogel’s Curves ، Neural Networks