A New Model for Forecasting Rate and Genetic Algorithm for the Parameters Estimation

WANG, GUANGLAN, Southwest Petroleum Inst.

The main task of development in an oil field is to get the optimal final recovery with the minimum investment. In the process of enhanced oil recovery, research of development projects and reservoir behavior assessment is based on accurate forecasting rate. On the basis of Poisson distribution, this paper describes a model to predict rate variation with time. In order to use the model, Three unknown parameters must be determined. In the last several years, much progress has been made in this area. This paper provides a new method of intelligent genetic algorithm to determine optimal parameters value in model. This method randomly produces group of initial solutions in the parameters' solution space, it can search for optimal solution in the large by parallel process, and it use iterative process to simulate organism evolution process. Species gene produces new generations group through replication, intersection, variation genetic mode etc. According to competitive principle of retaining excellent solution, eliminating the others. Genetic Algorithm guarantees that parameters’ estimation value convergence to global optimum principle of retaining excellent solution, eliminating the others. Genetic Algorithm guarantees that parameters' estimation value convergence to global optimum.

AAPG Search and Discovery Article #90911©2000 AAPG Pacific Section and Western Region Society of Petroleum Engineers, Long Beach, California