Reservoir Characterization

Reservoir Characterization

Challenge

The classification of reservoir fluids is a matter of considerable practical importance. By following the rule of thumb, fluid types in a reservoir can be identified, but laboratory observations are often required to verify them because of the imprecise and uncertainty that exist in reservoir parameters.

Solution

• We have proposed the application of Artificial Neural Networks to classify the reservoir fluid types based on laboratory observation and field data.
• More than 700 samples of different types of reservoir fluid types were used to develop the ANN model.
• Different types of architecture for different groups of input data were tested using fitness criteria to identify an optimal architecture.
• The optimal neural network was able to classify the reservoir fluid with high accuracy.