Estimation of DSI log parameters from conventional well log data using a hybrid particle swarm optimization-adaptive neuro-fuzzy inference system

Dipole shear sonic imager (DSI) log offers valuable data for geophysical, petrophysical and geomechanical studies of the hydrocarbon bearing intervals. Compressional (Vp), shear (Vs) and Stoneley (Vst) waves are acquired through processing and interpreting the DSI data. The present study aims to present an improved approach for establishing a quantitative relationship between conventional well log data and sonic wave velocities (Vp, Vs and Vst as DSI log parameters) by coupling an evolutionary computational method based on a neuro-fuzzy inference system. Particle swarm optimization -neuro fuzzy inference system (PSO-ANFIS) model suggested in this study, is based on a combination of fuzzy rule-based system and particle swarm optimization algorithm, which simultaneously adjust both the antecedent and the consequent variables. This approach was conducted in siliciclastic/ carbonate Asmari Formation of Mansuri oilfield in order to estimate the compressional, shear and Stoneley wave velocities. The conventional wireline logs from two wells were employed to build intelligent models while the third well was applied to evaluate the performance and reliability of the created models. The results indicated that the proposed hybrid scheme can satisfactorily improve the computational efficiency and performance of the DSI log parameter prediction, compared to individual intelligent systems and hybrid particle swarm optimization -neural network strategy (PSO-ANN). Finally, the workflow outlined here can be used as an efficient tool for prediction of other petrophysical rock properties, due to an enhanced estimation accuracy afforded by PSO-ANFIS model.