Late Cretaceous Mishrif "Rubble" limestone in Bahrain is a complex challenging heavy oil reservoir. It is a heavily fractured formation with tight permeability matrix where most heavy oil resides.

To unlock Rubble's potential, steam piloting has been going since 2011. Initial pilots tested different conventional thermal processes such as cycle steam stimulation (CSS) and steam-drive using vertical, deviated and horizontal wells. The most encouraging commercial results come from "Forced Imbibition" (FIM). FIM implies simultaneous injection of large slugs of steam into closely spaced horizontal wells in direct communication through the fracture system, then soak and produce all wells together. Five FIM steam cycles have been conducted with a potential recovery of 10% of the oil originally in place (OOIP).

Although results are encouraging, FIM faces challenges that decrease the heating efficiency of the matrix where most oil exists. First, matrix heating is slow because injected steam preferentially flows through the fractures and karst features. This causes matrix to be heated only by slow conduction from the fracture face. Heat available for conduction is diminished due to steam lost out of zone and the huff and puff nature of FIM which produces hot fluids between steam cycles. Second, productivity of the horizontal wells is highly affected by pump location within the wellbore, with deeper pumps producing most of the fluid.

For many thermal projects CSS is followed by steam-flood or thermal gas-oil gravity-drainage (T-GOGD). These accelerate heat conduction to the matrix because steam is continuously pumped into the reservoir. However, significant heat loss and ineffective pump placement must be resolved for the project to be economic. So a steam streamline mapping experiment was conducted to test the potential for steam-flood or T-GOGD. This paper describes the experiment's design and discusses the results.

The steam mapping experiment sought to identify wells with the slowest condensed steam breakthrough to offset wells and the lowest loss of condensed steam out of zone. Because deviated wells are less connected to fractures than horizontal wells, steam was first injected into a single deviated well for five days while monitoring offset producers. After mapping the results, the exercise was repeated with the next deviated well, and so on. When steam was injected into a horizontal well, steam volumes and rates were reduced to see if steam losses dropped.

The experiment helped to identify four important learnings. First, most of the Injected steam and hot fluids flow parallel to the dominant fracture trend and structural dip. Second, the experiment showed that the dominant fracture network has dynamic gas-oil and oil-water contacts which rise during steam injection and fall during fluid production, such that well geometries and pump locations less sensitive to fracture fluid contact flux outperform other wells. Third, with the dynamic fracture fluid contacts, the existing pattern may be conducive to test T-GOGD. Finally, controlling steam injection rates could help in managing the steam losses outside the pattern area.

You can access this article if you purchase or spend a download.