Full text loading...
-
Artificial Intelligence as the Problem Solving Method of Unconventional Reservoir Simulation
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, Saint Petersburg 2018, Apr 2018, Volume 2018, p.1 - 6
Abstract
The main challenge of unconventional reservoirs is to find out their high production zones (so-called “sweet spots”). For “sweet spots” mapping, there is information from wells, which are drilled only in a few points of a reservoir, and from seismic survey covering the entire reservoir area. During upscaling the well data for the whole reservoir, there is a problem how having a limited amount of information from a finite number of points (wells), turn it into an infinite number of polygon points (a reservoir). In this work, the new adaptive modeling approach for seeking “sweet spots” is proposed. The main difference of the adaptive approach from the deterministic one is that it does not connect the well and seismic data by rigid functions. The adaptive approach gives real forecasting of future performance not its calculation, because it is based on artificial intelligence in a form of fuzzy logic algorithms, which are able to form relatively free correlation between the well and seismic data. This work demonstrates the capabilities of the adaptive approach for a real unconventional reservoir in a carbonate formation with heavy oil to predict the location of the reservoir “sweet spots” for new drilling.