A New Approach for Continuous Gas Lift Simulation and Optimization

s Continuous gas lift is one of the most typical forms of artificial lift in oil production. Gas lift is rather inexpensive, easy to implement, very effective in a wide range of operating conditions and requires less maintenance in comparison to other alternatives. The basic principle consists on decreasing the pressure gradient in the liquid via the injected gas. The resulting mixture becomes less heavy than the original oil so that it eventually starts flowing. Current gas lift analytical methods applied to individual wells have applied the state of the art on numerical algorithms such as optimization via genetic or polytope algorithms. However, the models have been oversimplified so that several thermodynamic properties are considered constants though mechanistic basis are often employed. This paper develops a model based on mass, energy and momentum balances to produce a set of differential algebraic equations (DAEs). Cubic equations of state are used to solve the phase equilibrium behavior and to estimate other required thermodynamic properties. The problems solved include a) Determination of the outlet conditions, at the surface, when the amount of gas and oil production are fixed and b) Determination of the appropriate amount of gas required to lift the oil but matching the outlet pressure at surface manifold. The DASSL package is used to solve the DAEs problem and KNITRO is used to solve the overall optimization problem. The model allows determination of phase behavior along the well according to the automatic integration step produced within DASSL. Introduction Artificial lift systems are required when the pressure of a well is not enough as to maintain the oil production with satisfactory economic return. This situation is typical in mature oil fields where increasing watercut or direct decreasing reservoir pressure eventually cause wells to cease natural flow. Selection of the artificial lift system is often based on experience, non-technical preferences and even technical myths. Nowadays, the highest operation cost in the oil fields tends to be due to artificial systems. Gas lift is probably the most widely used artificial lift system. In fact, gas lift is often considered in some reservoirs as the only artificial lift system; see for instance [1]. In these systems, a gas is injected in the production well to modify the mixture density and then provide sufficient energy to produce the petroleum flow. Gas injection has been used in both continuous and discontinuous forms. Clegg et al. [2] compare several artificial lift methods and present a review of design issues that include all gas lift options. Continuous injection is analyzed in this paper since it is the most currently used gas lift system and intermittent injection becomes a better alternative only in those cases where continuous injection is not efficient [3]. Lifting by using the energy of the lift gas from the same reservoir or even from a different reservoir seems to be also a viable alternative [4]. Automation of the gas lift process can be applied to get extra benefits. In particular, fields that are either on steep decline or have been staffed down over years represent an area of opportunity [5]. Natural gas is the most common gas used in gas lift systems. However, different gas may also be feasible in the process. Using nitrogen as the gas lift source seems to be technically and economically feasible and may be, in some cases, more profitable than using natural gas [6]. Several gas lift optimization tools have been developed where the nodal analysis is typically applied [7]. For instance, GOAL is a PC based software where production systems can be simulated and optimal gas lift allocated to determine actions such as shutting in or choking back in wells. Automated real time gas lift measurement and monitoring systems have been implemented in several fields and experience on success or failure is captured and updated in several databases. Optimal allocation of gas lift rates to achieve maximum production rates has been addressed in different ways in existing literature [8]. For several wells in oil fields, Fang and Lo [9] apply linear programming to allocate gas lift and production streams subject to multiple flowrate constraints. Wang et al. [10] expanded this method to include surface facilities. Sequential quadratic programming has also been applied to the rate allocation problem where constraints include facilities and the reservoir [11]. Wang and Litvak [10] include pressure interactions among wells through common lines so that the final model becomes more robust. Gas lift