The choice of an integrated process modeling approach is function of information availability, data uncertainty, and the cross-discipline integration demand. Based on these parameters, the modularity, the multidisciplinary knowledge request, and the decision between feedforward or feedback control is defined. First, this paper discusses drivers and enablers to implement field life cycle management. To this end, right decisions are necessary to ensure that integrated asset optimization reaches the global optimum instead of local ones. The second part of this paper presents a case study describing a compositional production stream modeling to perform multiple flash calculations from the initial reservoir to the stock-tank conditions. The integration objective is to deliver robust fluid characterization to model the multiphase flow expansion process through a surface choke valve at the wellhead.
The surface separation process was defined as two stages: a high pressure stage and the stock-tank. This process represents the well testing route scenario. SRK-EOS previously tuned with PVT data was used for fluid characterization. The initial gas cap composition was generated from the oil flashed at the reservoir saturation pressure and temperature. The original oil composition, at initial reservoir condition, was used for the fluid initialization. A recombination method calculated the wellbore composition at depleted pressure and temperature by an iterative logic. In this procedure, the gas cap, the reservoir oil, and the water mole fractions were blended to match the production gas/oil ratio and water cut from the well. The oil and gas density at stock-tank condition, the oil flow rate at separator condition, and the separator oil formation volume factor were used as simulation validation parameters.
A multi-rate well test was performed in real field scale to provide input data for the investigation. The multi-rate well test consisted to execute a sequential step response disturbance. Each disturbance was followed by a transient and a steady state period "a priori" to a next step. Two types of well performance were analyzed: production with stream composition variation caused by near wellbore reservoir gas/water coning and production with constant stream composition. The simulation was implemented in the software Hysys from Aspentech to provide the fluid properties and the stream composition for modeling the multiphase flow expansion process through a choke valve in wellhead conditions.
The results demonstrated that multiphase flow expansion process through the valve is neither isenthalpic nor isentropic. For the outlet valve (downstream choke) temperature calculation, it was found an absolute relative error of 8% for the isenthalpic flash and 14% for the isentropic flash. These errors corresponded to the higher pressure drop across the valve. The agreement between calculated and measured valve outlet temperature was achieved when equilibrium calculation considered an adiabatic efficiency coefficient (?eff) to correct the flow thermodynamic path assumption through the valve. Finally, a parameter sensitivity analysis is presented to determine how the flash input variables influence the adiabatic efficiency (?eff) behavior.
