Mohammad Reza Mahdiani - Faculty of Petroleum Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran
Ehsan Khamehchi - Faculty of Petroleum Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran

One of the problems that sometimes occur in gas allocation optimization is instability phenomenon.This phenomenon reduces the oil production and damages downhole and surface facilities. Differentworks have studied the stability and suggested some solutions to override it, but most of them (such asmaking the well intelligent) are very expensive and thus they are not applicable to many cases. In thispaper, as a new approach, the stability has been studied in gas allocation optimization problems. Toprevent the instability, instability has been assumed as a constraint for the optimizer and then theoptimizer has been run. For the optimization, first a genetic algorithm and then a hybrid of geneticalgorithm and Newton-Quasi have been used, and their results are compared to ensure the goodperformance of the optimizer; afterwards, the effect of adding the instability constraint to the problemon production reduction have been discussed. The results show that the production loss with addingthis constraint to the system is very small and this method does not need any additional and expensivefacilities for preventing the instability. Therefore, the new method is applicable to different problems.

Stability Phenomenon, Genetic Algorithm, Newton-Quasi, Gas Allocation, Stable Flow