Miltiadis Konstantinopoulos, "Sensitivity analysis of an artificial lift system", Master Thesis, School of Mineral Resources Engineering, Technical University of Crete, Chania, Greece, 2018
https://doi.org/10.26233/heallink.tuc.72651
The objective of this study is to design a gas lift system for the maximization of the oil production and the verification of this design against various sensitivity analysis scenaria. The injection depth has been set to the maximum possible, i.e. equal to the depth of the casing, to maximize the efficiency of the gas lift operation. Τhe maximum producible flow rate that the well can produce for the given/available reservoir and tubing data using a gas lift system has been computed and that maximum flow rate has been the basis for the design. As variations in the production rate are anticipated by using different tubing ODs, two different sizes of tubing have been considered in this study. All design parameters such as injection depth, gas injection rate, casing pressure and injection pressures have been calculated as well.Apart from the design, verification of the design has also been performed against the variation of several parameters that might change over the life time of the well and those parameters that may include some sort of uncertainty in measurement/calculation. To include the whole range of scenaria the system a sensitivity analysis has been performed by using nodal analysis. This way the system is divided into three subsystems by means of a node at the bottom-hole. By superimposing the suitable IPR and VLP curves the maximum possible production specific to that sensitivity can be read from the intersection point whereas the injection gas rate, casing pressure and injection pressures can be read from the corresponding sensitivity curves for the gas flow in the annulus.Moreover, unloading the annulus and production tubing with the drilling fluid has also been analyzed for casing pressure operated (IPO) and tubing pressure operated (PPO) unloading valves. All analyses have been run by means of the PROSPER and PIPESIM software.