Modeling and Control Solutions for Riser Slugging in Offshore Oil Fields
Systematic methods and tools for managing the complexity
Process Control (T4-8P)
Keywords: riser slugging, gas-lift, slugging control, ARX predictor
In offshore oil fields, risers are used to transport a multiphase mixture (composed by oil, gas, water and sand) from the wellheads to separation tanks on producing platforms. If there are no separation and pumping facilities near the wellheads, on the sea bottom, this multiphase mixture must be propelled to the sea surface only at expenses of the reservoir pressure. In cases of high sea depth a common situation is that the reservoir is not pressurized enough to promote the flow of oil at satisfactory rates. In these cases artificial ascension of oil must be conducted via gas-lift techniques where injection of compressed natural gas is maintained at some points in the well column. In consequence, high flow rates of gas have now to be accommodated in the system of risers for further recovery by gas-liquid separators at the platform, before re-compression and re-injection. In this scenario, and depending on the flow conditions, a flow regime called riser slugging can develop in the pipelines. This regime, which is naturally highly transient, is undesirable because it introduces severe oscillations of pressure and flow rates in the system. Abatement of the intensity of riser slugging may be achieved via anticipative control actions on valves positioned at the entrance and at gas/oil exits of the separators.
In this work we present modeling and control strategies for an offshore oil producing system. The final aim is to stabilize production by dampening or neutralizing riser slugging effects.
Phase 1 of this work is concerned with the development of a simplified phenomenological model for the entire producing system, including a set of wells performing gas-lift, their corresponding risers, entrance manifold and a gas-oil separator. The input variables to this integrated model are the percentage of opening of all choke and exit valves on the separator. The corresponding output variables are its internal pressure and level of liquid. The oscillating slugging in the pipelines arises naturally as a consequence of the values chosen for specific model parameters like down-hole pressures, gas injection flow rates, fluid properties and geometry of wells and lines.
Phase 2 of this work is involved with the abatement of slugging consequences via determination of optimum adaptive decentralized control strategies actuating on the choke and exit valves on the gas-liquid separator.
Finally, in Phase 3, we approach the identification of a stochastic predictor for the entire system (risers + tanks) adopting an ARX MIMO structure. The ARX predictor was chosen because it can be easily and periodically tuned to the real process by using recorded time series of inputs and outputs. The utilization of the ARX predictor was demonstrated with simulated data generated in Phase 1. With a tuned ARX predictor, optimum anti-slugging control strategies can be periodically established for the real process.
See the full pdf manuscript of the abstract.
Presented Tuesday 18, 13:30 to 15:00, in session Process Control (T4-8P).
