ABSTRACT
An automatic control procedure for the determination of optimal production schedules is developed. Some production parameters such as injection rates or downhole flowing pressure are considered to be the decision variables, the values of which are to be chosen to maximize oil production. At the same time other production and economic parameters (such as water oil ratio) are to be constrained. The procedures discussed here are a way of using reservoir simulators to optimize production schedules in an automatic manner. The approximation of the non-linear processes in the models in terms of locally linear processes permits such problems to be solved iteratively using linear programming. This in turn results in a procedure which is feasible within reasonable expense. The multistep optimization procedure can be carried out in several different ways; step by step, over all time steps at once or using dynamic programming. Step by step optimization converges more quickly but does not take into account influences from one step to the next. Iterating over all time steps is the most desirable but is computationally the most expensive and converges least quickly. Dynamic programming was found to be the most efficient, requiring little additional computational effort beyond the step by step method while still allowing the solution in one time step to influence subsequent time steps.
