سال انتشار: ۱۳۸۶

محل انتشار: دومین کنگره مهندسی نفت ایران

تعداد صفحات: ۱۰

نویسنده(ها):

Sadegh Fathollahi – Reservoir Engineer, graduated from P.U.T
Babak Dehghani – Reservoir Engineer, graduated from Tehran University

چکیده:

Accurate description of reservoir is necessary for reservoir performance prediction. In fact reservoir description is the basement of reservoir simulation as the main tool for reservoir management. The flow unit model is the most practical approach for reservoir simulation purposes. A flow unit is defined as a group of reservoir rocks with similar properties that affect fluid flow. As hydraulic flow units Effective porosity also has important role in permeability modeling. In this paper after identification of flow units; based on log data the porosity and flow units are modeled in cored wells and these parameters are predicted in uncored wells. The feed forward backpropagation and generalized regression neural networks were used in porosity and hydraulic flow unit modeling. Finally all logging parameters, predicted effective porosity and hydraulic flow units were applied in permeability modeling and prediction.The permeability was predicted by using Amaefule[1] correlation, multivariable regression analysis and artificial neural networks. Both linear and nonlinear multivariable regression analyses were used in permeability prediction. The permeability prediction by artificial neural networks showed that ANNs can model permeability better than conventional methods