How Do We Perform Stochastic Reservoir Optimization?

Friday, February 1, 2002 - 10:10am - 1:00pm
Vincent 570
Benoit Couet (Schlumberger-Doll Research)
New control and monitoring technologies are being introduced to further real-time reservoir management, which is considered key to improving reservoir productivity. Examples of new control technologies include advanced completions, also referred to as smart or intelligent wells. New monitoring technologies include permanently installed sensors for measurements of pressure, flow and voltage. The decision on implementing such technologies is solely based on a cost-benefit analysis. Hence, one must be able to estimate the value of these technologies in monetary terms, such as associated net present value. A crucial underlying factor is the uncertainty in the reservoir model and its properties, and in the financial variables. The former includes uncertainty in properties such as reservoir geometry or permeability in the reservoir, while the latter refers to uncertainty in financial parameters such as discount rates or hydrocarbon price.

Quantifying the value of real-time control and monitoring technologies in the presence of such uncertainties requires a stochastic optimization of the production strategy. Standard process to perform the optimization can be applied and will be described in the context of real case situations.

Can we do better? How do we mathematically describe the uncertainties? What if we do not know the probability density functions? Should we optimize for the mean? Is there a better workflow? All these questions will be raised. We hope they could be addressed in a convenient and practicable way.

* Work in collaboration with Bahvani Raghuraman (SDR), Philip Savundararaj (SDR), and Robert Burridge (M.I.T.)