Application of particle swarm optimization and simple intelligent algorithms in building nonlinear models for predicting formation fracture initiation pressure

Abstract

ABSTRACT The main objective of this study is to estimate the FIP on Sarvak and Ilam formations using neural networks and intelligent algorithms. For this purpose, we first calculate the geomechanical parameters and the formation FIP of three wells, using well log data and experimental relationships. Secondly, we employ neural networks to estimate the FIP in three wells which randomly selected. Thirdly, we obtain a nonlinear mathematical model to calculate the formation FIP using simple genetic algorithms and particle swarm optimization (PSO). Based on our experiments, the PSO performs better than the genetic algorithm. Using the two new nonlinear models we estimated the pore pressure to be between 34.47 and 48.26 MPa, vertical stress between 75.84 and 131 MPa. Maximum and minimum horizontal stress between 62.05 and 87.7 MPa, FIP between 87.7 and 110.3 MPa.