| Summary: | Severe slug flow is one of the most undesired multiphase flow regimes, due to the associated instability, which imposes major challenges to flow assurance in the oil and gas industry. This thesis presents a comprehensive analysis of the systematic approach to achieving stability and maximum production from an unstable riser-pipeline system. The development of a plant-wide model which comprises an improved simplified riser model (ISRM) required for severe slug controller design and control performance analysis is achieved. The ability of the ISRM to predict nonlinear stability of the unstable riser-pipeline is investi¬gated using an industrial riser and a 4 inch laboratory riser system. Its predic¬tion of the nonlinear stability showed close agreement with experimental and simulation results.
Through controllability analysis of the unstable riser-pipeline system, which is focused on achieving the core operational targets of the riser-pipeline produc¬tion system, the maximum stable valve opening achievable with each controlled variable considered is predicted and confirmed through the simulation results. The potential to increase oil production through feedback control is presented by analysing the pressure production relationship using a pressure dependent dimensionless variable known as Production Gain Index (PGI).
The performance analyses of three active slug controllers are presented to show that the ability of a slug controller to achieve closed loop stability at large valve opening can be assessed by the analysis of the H∞ norm of the comple¬mentary sensitivity function of the closed loop system, T(s) ∞. A slug controller which achieves the lowest value of the T(s) ∞, will achieve closed loop stability at a larger valve opening. Finally, the development of a new improved relay auto-tuned slug controller algorithm based on a perturbed first-order-plus dead-time (FOPDT) model of the riser system is achieved. Its performance showed that it has the ability to stabilise the riser system at a valve opening that is larger than that achieved with the original (conventional) algorithm with about 4% increase in production.
|
|---|
