Estimating reservoir permeability distribution from analysis of pressure/rate transient data: A regional approach

Estimation of permeability variation within the reservoir using classic welltest interpretation methods is hard and not rigorous. To estimate permeability distribution from analysis of pressure/rate transient data, we should resort to automated history-matching methods. In such methods, a numerical reservoir simulator is integrated to an optimization algorithm to estimate reservoir properties such as permeability map through minimizing the well response misfit. During the transient period, the pressure diffusion propagates into the reservoir and the region of investigation enlarges as the time elapses. Therefore, the number of contributing grid blocks (blocks that fall in the region of investigation) in the well response increases with time. This means that the number of history matching parameters are not fixed and changes with time. In addition, every period of pressure derivative curve is the response of a certain region of the reservoir. The common practice is to calibrate the properties of a fixed number of grid blocks by history matching of the entire derivative data. We believe that such approach may not be strict and exact because it does not account for the mentioned aspects. In this paper, we presented new technique that help us calibrate each region of the reservoir with relevant pressure derivative interval. The new approach shows superiority in matching pressure derivative and estimation of permeability map when compared with the conventional approach.