Decisions and Risk

Engineers are fully immersed in risk. Subsurface uncertainty is a daily challenge, yet decision making under risk is a relatively immature area for the petroleum engineer.

Optimisation without Risk

Optimisation without addressing risk has emerged in the last decade. The founders of EssencePS have carried out complex optimisation studies in the North Sea and Gulf of Mexico, looking at well placement and drilling schedule optimisation.

Intelligent wells

Optimisation can be used for intelligent wells, using all the data available. Optimal completion designs, optimal operational controls, optimal injection plans, can all be investigated together with a full reservoir simulation model, so that both static and dynamic effects are included.

Optimisation under uncertainty

Whilst optimisation with a single geological model provides benefits, the full decision analysis requires investigation of the robustness of the decision against different possible subsurface scenarios. The ‘Flaw of Averages’ demonstrates that a decision based on a most likely scenario may be the wrong decision for the ‘low case’ model such as water breakthrough which will only be known too late.

Optimisation under uncertainty examines the complete range of probabilistic geological models, and provides an optimum decision for the complete uncertainty range.

Decision tools and Risk

EssRisk is the first software in the industry which performs optimisation under uncertainty. Through the use of powerful and performant proxy models, it is now possible to examine the robustness of decisions against uncertainty and optimise those decisions.

Case study

A recent case study has looked at intelligent well design, in the context of controlling water within an uncertain geological structure. Different types of ICD/ICV’s were examined and optimised in a single optimisation configuration, which handled automatically the mixed integer (continuous and discrete variables) nature of the problem.

A single simulation model was constructed which embedded all the different options and variables to be optimised.

This case study optimised NPV, including capex and opex for the different well designs.

Understanding decisions under uncertainty

Instead of a single incremental NPV value, optimisation under uncertainty allows the decision maker to look at the full effect on the forecast NPV, as represented by a cumulative distribution or S curve.