In this study, a new algorithm is proposed by employing artificial neural networks in a sequential manner, termed the sequential artificial neural network, to obtain a global solution for optimizing the drilling location of oil or gas reservoirs. The developed sequential artificial neural network is used to successively narrow the search space to efficiently obtain the global solution. When training each artificial neural network, pre-defined amount of data within the new search space are added to the training dataset to improve the estimation performance. When the size of the search space meets a stopping criterion, reservoir simulations are performed for data in the search space, and a global solution is determined among the simulation results. The proposed method was applied to optimise a horizontal well placement in a coalbed methane reservoir. The results show a superior performance in optimisation while significantly reducing the number of simulations compared to the particle-swarm optimisation algorithm.

The relative permeabilities of CO₂ and brine are important parameters that account for two-phase flow behavior, CO₂ saturation distribution, and injectivity. CO₂/brine relative permeability curves from the literature show low endpoint CO₂ permeability values and high residual brine saturation values. These are the most distinguishing aspects of the CO₂/brine relative permeability from oil/water and gas/oil. In this study, this aspect is investigated experimentally by employing a wide range of CO₂ injection flow rates. As a result, all the measurements align with previous studies, having low endpoint relative permeability and high residual brine saturation values. They have obvious relationships with the changes in CO₂ flow rates. As the CO₂ flow rate increases, the endpoint relative permeability increases, the residual brine saturation decreases, and they converge to specific values. These imply that a high CO₂ injection flow rate results in high displacement efficiency, but the improvement in efficiency decreases as the flow rate increases. The reasons are identified with the concept of the viscous and capillary forces, and their significance in the CO₂ injection into a reservoir is analyzed.

This study developed a data-driven surrogate model based on a deep neural network (DNN) to evaluate gas–liquid multiphase flow occurring in horizontal pipes. It estimated the liquid holdup and pressure gradient under a slip condition and different flow patterns, i.e., slug, annular, stratified flow, etc. The inputs of the surrogate modelling were related to the fluid properties and the dynamic data, e.g., superficial velocities at the inlet, while the outputs were the liquid holdup and pressure gradient observed at the outlet. The case study determined the optimal number of hidden neurons by considering the processing time and the validation error. A total of 350 experimental data were used: 279 for supervised training, 31 for validating the training performance, and 40 unknown data, not used in training and validation, were examined to forecast the liquid holdup and pressure gradient. The liquid holdups were estimated within less than 8.08% of the mean absolute percentage error, while the error of the pressure gradient was 23.76%. The R² values confirmed the reliability of the developed model, showing 0.89 for liquid holdups and 0.98 for pressure gradients. The DNN-based surrogate model can be applicable to estimate liquid holdup and pressure gradients in a more realistic manner with a small amount of computating resources.

This article presents an efficient approach to determine the optimal drilling location for maximizing the cumulative production without the need for a reservoir simulation, of which scheme is based on artificial neural network incorporating the productivity potential. A reservoir simulator can provide an accurate result, but is sometimes inefficient due to the enormous computing requirements. The typical artificial neural network scheme used in multiwell placement shows lower predictability as the size of the input data increases. This work introduces the productivity potential that merges various reservoir properties, such as the permeability, porosity, and saturation, and integrates it into an artificial neural network. The cumulative production is compared with the result of the reservoir simulator to determine the accuracy of the developed method. The efficiency of the conventional artificial neural network is improved by the proposed model, as well by using the productivity potential instead of a lot of separate inputs. The predictability is verified by determining the drilling location in the same way as that of the reservoir simulator in the case of a single infill well. The stability is confirmed by its ability to produce a reliable result even as the number of input data increases.