The low oil price clearly provides the industry with some challenges and the need to optimise oil production or reduce operating costs is now more prevalent than ever. Software tools that can accurately model the field and then perform production optimisation, while minimizing operating costs by informing operations as to what control variables to change on the field, can play an important role.

It is extremely important that the advice provided by the technology to operations can be implemented easily and quickly. This is particularly important for mature fields where the field conditions may be changing continuously and therefore an even greater need for regular (weekly or daily) optimisation of production, to ensure that the field is operating optimally on a day-to-day basis.

The aim of all model-based optimisation technology is to change the available control variables (from their current settings) so that the objective function is either maximised or minimised. In most production systems the objective function would be to maximise oil production, with the control variables being items such as: gas-lift injection gas rate, choke settings, well and pipeline routing, etc. Alternatively the objective could be to minimise costs while keeping production at the target rate. In either case the optimal solution may require that a considerable number of control variables be changed.

For a practical implementation, the large number of required field changes will be a challenge to the production engineer and operations. What is required is a reduced set of control variable changes that yields an increase in production (not necessarily the optimum). However, prior to the optimisation, the production engineer will not know the most important control variables to change, thus the software tool has to undertake this task as part of the calculation.

To date the direct deployment of model-based optimisation to support operations has had some challenges. Some of these challenges involve technical issues that have not helped the wide-spread acceptance, specifically;

  • lack of confidence (validation) of the underlying model's predictions

  • inefficient handling of continuous and discrete control variables by the optimisation algorithm

  • failure to integrate production and facilities in the same model

  • and, perhaps most critically, delivering a solution that cannot be implemented directly in the field due to recommending a large numbers of changes from the current operational point

In this paper, we focus on next generation equation-based optimisation technology (as implemented in the gPROMS Oilfield optimiser) that can address these deficiencies to deliver a solution that can be implemented directly onto the field that includes the simultaneous optimisation of both continuous and discrete control variables.

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