Coupled matching of the production data and the seismic data

Matching procedures are often used in reservoir production to improve geolo gical models. In reservoir engineering, history matching leads to update pe trophysical parameters in fluid flow simulators to fit the results of the c alculations with observed data. In the same line, seismic parameters are in verted to allow the numerical recovery of seismic acquisitions. However, it is well known that these inverse problems are poorly constrained. The idea of this original work is to match simultaneously both the permeabi lity and acoustic impedance of the reservoir, for an enhancement of the res ulting geological model. To do so, both parameters are linked using either observed relations and/or the classic Wyllie (porosity-impedance) and Carma n-Kozeny (porosity-permeability) relationships. Hence, production data are added to the seismic match, and seismic observations are used for the perme ability recovery. The work consists in developing numerical prototypes of a 1D fluid flow sim ulator and a 1D seismic acquisition simulator. Then, in implementing the co upled inversion loop of the permeability and acoustic impedance of the two models, we can test our theory on a 1D case. Comparison of the coupled matching with the two classical one demonstrates the efficiency of our method. We reduce significantly the number of possibl e solutions, and then the number of scenarios. In addition, the augmentatio n of information leads to a natural improvement of the obtained models, esp ecially in the spatial localization of the permeability contrasts. The impr ovement is significant, at the same time in the distribution of the two inv erted parameters, and in the rapidity of the operation. This work is an imp ortant step in a way of data integration It allows to use the available har d and soft data, as the knowledge of the depositional environment (like geo statistical models), seismic information (like surface seismograms), and/or dynamic information (like well-test and production data), and leads to a b etter reservoir characterization. This original algorithm could also be use ful in reservoir monitoring, history matching and in optimization of produc tion.