Automated Methodologies for Decision Support in Field Development
MetadataVis full innførsel
This thesis deals with automated methodologies for decision support of offshore field development planning. The thesis is in two parts: methodology development and application to a real-world field case. In the first part, three aspects of the methodology are developed. In the first, the early-stage field development problem is formulated into a mathematical optimization model using mixed-integer linear programming. The purpose of this methodology development is to find an effective algorithm that is able to solve the optimization accurately within a reasonable time frame. To capture the nonlinearity of the field production performance, a Logarithm piecewise-linear algorithm was used and its performance was compared with the commonly-used Special Ordered Sets of type 2 algorithm. The optimization decision variables include the drilling program (both number of producer wells and the drilling sequence) and production schedule (or rate allocation). In the second aspect, the focus was on field development planning considering the downside price risk. The method named price stress testing is adapted to assess the flexibility of a field development concept and used for development concept selection. The stress testing methodology provides a way to evaluate the field development plan under extreme conditions of price shocks. In this work, a new indicator was developed to quantify the field development resilience of the concept in the early field development phase. The third aspect of the developed methodology is including abandonment timing into the decision variables, together with drilling and production scheduling. This allows an accurate estimate of optimal field lifetimes considering the uncertainty in the available information. In the second part, all developed methodologies are applied to a real-world case study of field development planning for two oil discoveries located in the Barents Sea. The optimization model was used to select the drilling sequence of the 9 pre-specified candidate wells and the allocation of yearly production rate for a standalone development case. Other development concepts were studied to tieback the production to nearby infrastructures. Using the developed stress testing methodology, three different field development concepts were evaluated considering the resilience of the concept to extreme price conditions and the net present value (NPV). The abandonment timing optimization algorithm has also been used on the standalone concept to demonstrate the application of automatic searching for the drilling sequence and production allocation in a flexible production horizon with the objective function of maximizing NPV. The novelty of this work is threefold: 1) The Logarithm piecewise-linear algorithm was formulated for early-phase field development optimization, and it has significantly reduced the computational time and achieved improved accuracy over SOS2. 2) To the best of my knowledge, it is the first time that the price stress testing methodology is applied to early-phase field development planning to quantitatively evaluate the vulnerability of a field development concept to sudden hydrocarbon price drops. 3) To the best of my knowledge, this is the first published contribution to employ mathematical programming to solve the decommissioning timing optimization during the early stage of field development. In summary, this thesis developed and tested novel decision-support methods for field planners and operators to address common design issues in the early phase field planning such as drilling, production scheduling, abandonment timing and oil price variation. The methods were applied and tested on a real-world field case and were proven successful. The method and computational strategies developed will hopefully contribute to increase the performance of future field developments, add robustness, reduce time and cost and efficiently integrate and process complex information.