Data availability and development of a decision support system for heavy oil production prediction

• An integrated approach was developed to overcome the problem of data availability.
• By combining physical experiments and simulation, sufficient data can be generated.
• A larger dataset can support data analysis studies and a case study was presented.
• The case study addresses selection of an EOR sequence for heavy oil production.
• A decision support system for prediction of heavy oil production was developed.

This paper has the dual objectives of (1) presenting an approach for overcoming the problem of data availability by combining experimental and simulation studies for generating a large data set, and (2) presenting a case study in which this integrated and interdisciplinary approach was adopted for tackling the problem of selection of enhanced oil recovery or EOR technology for heavy oil production. This paper describes development of a decision support system that can predict heavy oil based on different sequential applications of EOR techniques and the reservoir and operation parameters of oil viscosity, reservoir pressure, reservoir size, original oil in place, initial oil saturation, reservoir permeability, water-flooding, CO2 flooding, injection pressure, percentage recovery using the first EOR technique, and a given time frame. Based on the constraints defined by the set of experimental data, a larger set of data was generated using a simulation software. The larger dataset in turn provided the basis for developing the correlation models for the decision support system. An oil production prediction system was developed which can predict oil production over time for a given EOR sequence. However, if a prediction is made beyond the domain of parameter values specified in the experiment, the prediction system performs poorly. The approach illustrated by the case study can potentially be applicable for tackling data modeling studies for which data availability is a problem.