In accordance with goal set by the Paris agreement, oil and gas companies are increasingly focusing on reducing the carbon footprint in both existing and new field developments. At the same time we see that oil and gas are still a key part of the energy mix, hence oil and gas companies are looking for solutions that both help maintain and/or increase recoverable volume in new and existing assets, while minimising the carbon emissions. In this research project, Resoptima AS, NORCE, and selected industry partners have joined forces to help develop new solutions and methodologies that both help oil and gas companies quantify and reduce the carbon footprint in their planned and producing assets, while maintaining or increasing their contribution to the energy mix.
The project seeks to establish a mathematical foundation for adding CO2-emission costs into the calculations involved in finding an optimal drainage strategy for a petroleum reservoir. To make such a tool eventually available to the market would make it tractable for stakeholders to deal with CO2-emissions as an integral part of a reservoir production planning exercise, happening prior to any investments on a reservoir. We believe this is what is required to include the ambitions, decisions and implications from the Paris agreement into practical use, notably the ambition to reach 50% emission reduction by 2030, and NZE by 2050. Minimizing the number of wells, reducing amount of water injection, reducing or delaying water influx to production wells and increasing the amount of CO2 injected all can be described by a cost, positive or negative, that will be related to the at any time given price of CO2-emissons. Including that cost in the optimization calculations takes the petroleum industry towards a sustainable future.
Since there are many considerations to deal with simultaneously, including uncertainties on data, the problem to solve is highly multi-dimensional. Moreover, it is optimization under uncertainty. We plan to explore the Stochastic Simplex Approximate Gradient (StoSAG) as well as other optimization methods. Searching for the robust solution we need; it will be an important part of the project to document these various optimisation algorithms after running realistic tests on data from several real reservoirs.
In a multi-dimensional setting several solutions could be within a reasonably close vicinity of the optimum. Due to existing uncertainty of the assumptions, the task is therefore not just to find an optimum solution, but a solution with an estimated range of relevance. E.g. if one of the assumptions (cost of emission) changes 1%, that should not be expected to change the chosen production strategy completely. Robustness is therefore key.