Rigorous model for determination of PVT properties of crude oil in operational conditions

ABSTRACT

The different tasks of petroleum engineers such as well testing and simulation are highly function of reservoir fluids properties so the accuracy of these properties can affect the oil recovery. Solution gas oil ratio is one of these major properties so in this investigation, a new and accurate approach based on adaptive neuro-fuzzy interference system (ANFIS) algorithm was evolved to predict the solution gas oil ratio (R_s) as function of other properties of crude oil in operational conditions. In the current work, a comprehensive databank contains 1136 data points and was implemented for preparation and validation of ANFIS. By the evaluation of prediction algorithm, the coefficient of determination values in training and testing processes of R_s prediction were obtained 0.98125 and 0.96477, respectively. Furthermore, the proposed ANFIS was compared with the existing correlations in literature. These comparisons showed that the ANFIS performed in acceptable degree of accuracy for determination of R_s.

KEYWORDS:

• Solution gas oil ratio
• artificial intelligence
• PVT properties
• ANFIS
• predicting model

View correction statement:

Correction