Optimizing Steam Injection Scheduling Using Analytical Models in a Probabilistic Approach

Abstract

In a heavy oil field producing through cyclic steam stimulation, one of the most complex and time spending tasks, for the reservoir team, is scheduling steam injection across the field. Because there are so many wells, optimizing steam quota and soaking time are usually calculated or simulated for a well, defined as representative for the entire field or at least for an area. That method optimizes the steam stimulation for each cycle of the representative well, but not necessarily optimizes the net present value or reserves for the entire field. Optimizing an entire field through numerical simulation can be frustrating as it consumes lots of time and requires expensive computational infrastructure. This paper shows that analytical methods can be used to optimize net present value (NPV) or reserves in a heavy oil field through massive computing in a probabilistic approach in a much shorter time than it would be done with a numerical simulator, above all in an uncertainty assessment. It can help, also, to analyze the economical convenience of buying new steam generators or moving them between fields or areas.