Research and Knowledge Exchange Support Team (EGIS)
+44 (0)131 451 3130
Nothing about oil reservoirs is ever absolutely certain. The critical question from a business viewpoint is whether the lack of knowledge makes a difference to how we develop a reservoir, and if it does whether we can choose a development plan that minimises the impacts of that uncertainty.
The group develops and applies mathematical models to quantify uncertainty in an oil reservoir. It does this using advanced mathematical and statistical techniques to blend different sources of information (including production data) to come up with a statistically consistent estimate of uncertainty. Optimisation techniques are then used to identify reservoir development plans (or interventions such as infill drilling) that maximise expected return and minimise risk in that return.
The research the group undertakes covers a number of areas: we carry out research in stochastic optimisation (algorithms such as particle swarm optimisation), including multi-objective optimisation where we look for trade-offs between two objectives such as maximising oil recovered and minimising water production; we carry out research into Markov Chain Monte Carlo techniques – the 'gold standard' of uncertainty quantification; we use machine learning to help us understand the geological make-up of reservoirs, so when we adjust unknowns in our reservoir models to match history we can do it in a way that is consistent with nature; and we have some research into 'rare events' – quantifying the likelihood of events that are unlikely yet could cause major problems for a project if they occurred.
The research in the group is funded both by industry and the UK's Engineering and Physical Science Research Council (EPSRC). Subject to funding, there are opportunities for both PhD students and Postdoctoral Research Associates. The group has access to both high powered workstations and a Linux cluster for the computational aspects of the research.