During under-balanced drilling (UBD) of a horizontal well, down-hole pressure and temperature, as well as surface rate and pressure data are acquired. Theoretically, estimation of reservoir permeability profile can be achieved by quantitative analysis of these data. However, in practice, not only is a fine-grid reservoir simulator required to simulate the drilling schedule change due to tripping, completions, workovers, increase of well length with drilling and fluid flow in reservoir, but also an effective inversion approach is needed to estimate the reservoir parameters. The problem of reservoir characterization using UBD data has been studied by several previous workers. At current state of the art, neither analytical models nor commercial reservoir simulators are completely suitable for modeling the coupled system of UBD and reservoir fluid flow.
In this paper, we focus on the development of an automatic history matching technique to interpret the measured oil rate and down-hole pressure data while drilling. Our contributions are three fold. Firstly, a two-phase 3D reservoir simulator was developed in which the UBD drilling process was coupled into the reservoir simulator. Next, the reservoir simulator is applied to simulate the entire drilling process from penetrating the pay-zone until the well is complete. Secondly, a least square method is applied to match the down-hole pressure and oil rate data. The inversion for a synthetic UBD data set and realistic UBD data are demonstrated on a multi-zoned reservoir model. The results show that permeability and initial reservoir pressure can be obtained by matching down-hole pressure or oil rate in a short time of exposure to the reservoir section. Finally, we offer the details for implementing the UBD process in the simulator and discuss the practical issues of the application of UBD data in reservoir characterization through this automatic history matching technique.