Use of Artificial Intelligence in Determining the Location of Infill Wells in Hydrocarbon Exploration and Production Activities
Among all of the energy resources, oil is the most widely used primary energy source. Oil reservoirs need to be characterized accurately for effective field development purposes. Typically, available field data in the reservoir characterization stage may include seismic surveys, well logs, core analysis data and field production history. One of the challenges for reservoir engineers is to utilize various types of data collected in different scopes to characterize the reservoir and propose optimized development strategies. Numerical reservoir simulation is one of the most broadly implemented approaches to quantitatively evaluate a field development plan. However, establishing a decent reservoir simulation model requires rigorous conversions from the raw field data to structural maps and spatial petrophysical property distributions as input parameters. History matching needs to be carried out to tune the property distributions to match numerical model predictions to field histories. The conventional reservoir characterization and field development workflow could be time and labor intensive. This keynote lecture presents the development of artificial-neural-network based expert systems which effectively correlate seismic survey data, well log data and field production history. When compared against the conventional reservoir characterization and field development optimization protocols, the developed expert system can provide much more rapid predictions than conventional reservoir simulators.