Simulation of Inflow whilst Underbalanced Drilling (UBD) with Automatic Identification of Formation Parameters and Assessment of Uncertainty

Abstract

Abstract Underbalanced drilling (UBD) is defined as a drilling operation where the pressure of the circulating drilling fluid is lower than the pore pressure of the target formation of interest. The most widely recognised benefit of UBD is the reduction of formation damage by minimising the drilling fluid leakoff and fines migration into the formation. It also facilitates the possibility for reservoir characterisation during drilling. The purpose of this paper is to present (i) how to use the inflow data for the evaluation of formation properties and (ii) how to cope with the uncertainty of the results. An in-house multiphase reservoir simulator is used for the simulation of the formation reservoir fluid flow during UBD. The model incorporates discrete consideration of the well with appropriate, time varying UBD boundary conditions. Capillary forces, which facilitate countercurrent imbibition of the drilling mud into the formation, are taken into account. The production rates during UBD depend on the formation properties and the drilling conditions. The inflow information is analogue to transient test data and can be used for the estimation of the reservoir model parameters. With the advancing of the drilling process, the amount of available inflow data increases and allows for a sequential history match procedure. Based on a statistical analysis of the inverse problem we identify the optimal level of parameterisation as quality and quantity of the measured data justifies. Technical contributions of this paper include (i) the usage of a reservoir simulator to derive the inflow of fluids in heterogeneous reservoirs and (ii) its coupling with gradient- based optimisation techniques for determination of certain reservoir parameters. The "simulation while drilling" approach proposed is fully automated and can set-up on a known or totally unknown reservoir model, introducing an iterative and automatic updating process of the reservoir model.