Production Decline Analysis

Abstract

We analyzed data from Wairakei, The Geysers, Larderello and Cerro Prieto using traditional production decline methods borrowed from the petroleum industry. Arp's exponential equation works quite well in many cases. His hyperbolic equation works very poorly and should be avoided. We tried several type curves with little success. Coast and Bodvarsson each developed easy to use methods for predicting either pressure or production given the other and historical data. Coat's influence function model requires only production and pressure data from which an influence function step or impulse response can be calculated. No reservoir model need be assumed. Bodvarsson's linearized free surface assumes that a free liquid surface is present and that enough data are available to define sinking velocity. The sinking velocity equals permeability times gravity divided by porosity time kinematic viscosity. Both methods worked well on the data tested.