Abstract:

This paper presents a new method of performance prediction of oil and water production in oil-fields with water injection. The method is adaptive to various random factors in the process of oilfield exploitation. By using neural network models of nonlinear time series prediction, we constructed neural network predictors for the oil production and water production of oilfields. Prediction results show that the predictors have high accuracy and can be used for prediction of oil and water production in different production periods. This method improves the theory of the performance prediction of oil and water production. Two examples of performance prediction were given in this paper.

Key words: Oilfield development, Production forecast, Time series, Nerve network