This paper presents a case study where a modern computer science numerical analysis technique has been used to shortcut the screening process for a subsea design, to deliver a demonstrably optimized solution. Subsea design was abstracted into a fitness test, a numerical score based on practicality, performance, and cost criteria. An evolutionary algorithm was then applied to maximize the score of the fitness test, using a matrix of flow assurance modelling data as a basis. The ultimately recommended solution is that with the global maximum fitness test score.
The field is a HP/HT gas/condensate gathering network of two fields and eleven wells, flowing to a processing platform and ultimately to an onshore LNG plant. The optimum design is a " Goldilocks" solution: maximum design temperatures that are not too hot; platform arrival temperatures that are not too cold. Eleven design scenarios were provided by the client. A team of flow assurance engineers and computer science developers, using the method described herein, were able to optimize the design considering all eleven cases.