Application of Artificial Neural Network to Evaluate Acidizing Results in an Iranian Oil field

One of the most expensive operations in oil well services is stimulation of oil and gas wells. So, having good estimation of results could help us to have a priority in order to candidate wells for operation. The artificial neural networks have a good ability to simulate and predict behavior of a complex function and it can be used to evaluate stimulated wells in an oil field and results can be extended to other stimulation jobs in that field. It is designed by 71 input nodes and acidizing data of 69 well were available which 70% of data used to train the network and 30% used for validation test and recall data. In order to optimize the artificial neural network several hidden layers, hidden layer nodes, iteration and transfer functions are tested and finally a case with sigmoid transfer function and one hidden layer with 110 node in it with 100000 iterate is specified to simulate the stimulation operation to obtain production rate after stimulation and evaluate acidizing systems, additives and diverters.

کلیدواژه ها:

Acidizing ، Stimulation ، Neural Network ، Transfer Function