3-D well path design using a multi objective genetic algorithm

Abstract

Optimizing wellbore trajectories to reach an offset subsurface location, involving a complex combination of vertical, deviated and horizontal well components, requires the minimization of both wellbore length and frictional torque on the drill string. This is particularly the case for shallow horizontal wells which are often limited in their extent by torque. By minimizing both wellbore length and torque it is likely that a wellbore designed to reach a specific target can be drilled more quickly and cheaply than other potential trajectories. However, these two objectives are often in conflict with each other and related in a highly non-linear manner. A multi-objective genetic algorithm (MOGA) methodology is developed and applied with two objective functions, viz. wellbore length and torque, to develop a set of Pareto optimal solutions that can aid the selection of less risky/less costly well trajectory designs. The MOGA performance is compared with single-objective function studies of a specific wellbore scenario. The results indicate that the MOGA methodology outperforms single-objective function approaches leading to rapid convergence towards a set of Pareto optimal solutions. Analysis reveals that by adopting an adaptive approach that allows the behavioral parameters of the genetic algorithm (GA) to evolve as iterations progress, the MOGA proposed converges more rapidly toward better ultimate solutions than if the GA behavioral parameters are held constant over all iterations of the algorithm. Algorithm code listings for the MOGA and GA applied in the analysis presented are included as appendices.