Dynamic Estimation for Controlling a Subsea Production System - Virtual Flow Metering using B-spline Surrogate Models
Abstract
Knowledge of the flow rates from individual wells in a subsea production system cangreatly improve decision-making processes. This thesis investigates flow estimation inthe subsea template Tilje in the BP-operated Skarv field, with automatic rate control in mind. Two optimization-based estimation methods, one with static models and onewith dynamic models, are developed and compared against each other by means ofbenchmark data from an OLGA model and historical field data from Tilje. Ideas fromrecent developments in production optimization are applied in the derivation of theproduction network models. This amounts to representing the flow network as individ-ual network components with B-spline approximated models., leading to a transparentmodel which allows for easier estimator tuning, easier constraint handling and fastersolution times.
The results presented in this thesis show that the method with static models is ableto predict flow rates with acceptable accuracy, has good robustness properties and fast solution times, indicating that the method has potential for use in an automatic rate control system. The method with dynamic models is less robust, more complex, anddoes not seem to improve the flow rate estimates. However, it does include estimates of mass and holdup and could potentially be useful if the method was improved. In addition to providing feedback for automatic rate control systems, the methods described in this thesis could potentially be used as cornerstones in advanced decision support tools, such as flow assurance systems and condition monitoring systems.