Abstract:
Conventional multiple regression for permeability estimation from well logs requires a functional relationship to be presumed. Due to the inexact nature of the relationship between petrophysical variables, it is not always possible to identify the underlying functional form between dependent and independent variables in advance. An accurate reservoir description is very important in reservoir evaluation, and permeability prediction is the key for a successful characterization. Permeability is one among the most important parameters affecting the productivity of hydrocarbon bearing reservoir. Thus understanding the heterogeneity of reservoir and characterizing it with consistent input of permeability is very crucial. Formation permeability is measured directly from the core sample studies performed in laboratory. However it is very costly and is not feasible for the whole reservoir. Also, permeability cannot be inferred directly from any well log measurements. Earlier, various methods have been used for the permeability prediction using empirical relationship, statistical regression, etc. These methods are not applicable for all reservoirs, since permeability varies largely due to different depositional environments. When large variations in pertrological characters are exhibited, parametric regression often fails or leads to unstable or erroneous results, So a nonparametric approach for estimating optimal transformations of petrophysical data to obtain the maximum correlation between observed variables has been introduced which are Artificial neural network (ANN) and Alternating conditional expectation (ACE).. These methods prove to be more robust as they require no prior assumption regarding functional form.
