An Automatic Well Planner for Complex Well Trajectories

Abstract A data-driven automatic well planner procedure is implemented to develop complex trajectories by efficiently adapting near-well reservoir properties and geometry. The draws inspiration from geosteering drilling operations, where modern logging-while-drilling tools enable the adjustment of during drilling. Analogously, proposed develops based on a selected geology-based fitness measure using an artificial neural network as decision maker in virtual sequential process within model. While networks have seen extensive use other areas management, best our knowledge, this work first apply trajectory design models. Importantly, both input data generation used train actual operations conducted trained are efficient calculations, since these rely solely geometric initial reservoir, thus do not require additional simulations. Therefore, main advantage over traditional methods highly articulated adapted low-order representation. Well generated realistic qualitatively quantitatively compared differential evolution algorithm. Results show that resulting improve productivity straight line trajectories, for channelized geometrically reservoirs. Moreover, overall with comparable solutions obtained evolution, but at much lower computational cost.