Enhanced Oil Recovery (EOR) aims at producing more oil from low productive reservoirs. EOR is a costly process due to the time and expenses required to allocate a suitable EOR method to an oil reservoir. For this purpose, this work presents a systematic approach to incorporate past experiences into a compact system to assign the EOR process to oil reservoirs. Because the number of clusters within the reservoirs data is not known, the method proposes a novel validity index to calculate optimal number of clusters copt. Then, the data are clustered to copt compact clusters according to their similarities, dissimilarities, rock, and fluid properties using fuzzy clustering algorithm. To decide on the EOR process to be applied to a new reservoir, this reservoir is assigned to one of the copt clusters based on similarity, and then the EOR method is chosen based on that cluster. This approach is used for available reservoirs data from different oil fields, and its effectiveness and superiority over previous methods are confirmed.

Keywords