Using well test analytical solutions for production prediction is rarely done, as this seems to be abstract and to require assumptions that are not very physical. But it presents a good alternative to numerical simulation, which is often expensive and cumbersome, and to decline-curve analysis, which has little theoretical justification in complicated cases.
This paper presents existing methods, and generalizes them to "n" wells in the reservoir. This allows the prediction of the flowrate produced for a given pressure drop history and a given global drainage area, but without a priori knowledge of the drainage area of each well. This is especially useful in reservoir monitoring.
This work shows how changes in the imposed pressure drop can be treated, which is straightforward, as it does not affect the reservoir response. But il also shows how non-linearities, i.e. changes in the well or reservoir themselves can be treated. To some extent, this technique allows modelling phenomena which are non linear, but still classic during the life of a well: fracturing, damage, or change in skin due to the appearance of multiphase flow.
This technique thus broadens the applicability of analytical production prediction.