Production Optimiser Pilot for the Large Artificially-Lifted and Mature Samotlor Oil Field

Abstract

This paper describes a production optimiser Pilot, developed by Rosneft/Samotlorneftegaz, with support from bp and deployed in JSC Samotlorneftegaz - a vast, mature, water-flooded, high water-cut and artificially-lifted oil field. Objectives include creating a digital twin for a sub-system of 600 wells and ~180 km of pipeline network, applying discrete, continuous and constrained optimisation techniques to maximise production, developing sustainable deployment workflows, implementing optimiser recommendations in the field and tracking incremental value realisation. This proof-of-concept Pilot and field trial approach was adopted to understand the optimisation technology capability and work-flow sustainability, prior to a field-wide roll-out. The periodic optimisation activity workflows include the creation of a "Digital Twin", a validated surface infrastructure model that is fully calibrated to mimic field performance, followed by performing optimisation that includes all the relevant constraints. Optimisation was trialled using two different classes of algorithms – based on sequential-modular and equation-oriented techniques. This strategy minimises optimisation failure risks and highlights potential performance issues for such large-scale systems. Optimiser recommendations were consolidated, field-implemented and values tracked. The optimiser Pilot development was undertaken during the fourth quarter of 2019. The delivered minimum viable product and workflows were used for field trials during 2019-20 and continuously improved based on the learnings. Specialists from both bp and Rosneft, along with three consulting organisations (1 in Russia and 2 in the UK) collaborated and worked as one-team to deliver the Pilot. Optimiser recommendations for maximising production include continuous and discrete decisions such as ESP frequency changes, high water-cut well shut-ins and prioritised ESP lists for installing variable speed drives. Field production increase of 1% was achieved in 2020 and tracked. Enduring capabilities were built, and sustainable work-flows developed. Field-wide optimisation for Samotlorneftegaz is non-trivial due to the sheer size, with over 9,000 active wells and due to continuously transient operations arising from frequent well-work, well shut-in's, new well delivery, pipeline modifications and cyclic mode of operations in some wells. This Pilot has provided assurance for the optimisation technical feasibility and workflow sustainability. A second Pilot of similar complexity but with different pressure-flow system response is planned. The combined results will help to decide about the full-field roll-out for this vast field, which is anticipated to deliver around 1% of additional production. This Pilot has demonstrated the applicability of discrete and continuous variable constrained optimisation techniques to large-scale production networks, with very high well-count. Furthermore, the developed workflows for configuring and calibrating the digital twin have several unique components including automation of hydraulic network model generation from static data, well model build automation and fit-for-purpose automated well model calibration. Overall, the results of this approach demonstrate a viable and sustainable methodology to optimise large-scale oil production systems.