Application of Experimental Design for Improved Reservoir-Characterisation by Pressure-Transient Analysis

Abstract Pressure-transient Analysis (PTA) is a useful technique for well-performance analysis and reservoir characterisation. Compared to other subsurface-characterisation techniques, PTA generally provides a more quantitative understanding of the subsurface characteristics over a relatively larger area. However, similar to all inverse problems, the results are often non-unique, largely due to uncertainties in the input data and interpretation model. Although the integration of PTA with other sources of information usually narrows the choice of the interpretation model, the use of uncertain input parameters still undermines the quality of the results while putting the associated business decisions at risk. To explore the effects of the full range of input-parameter uncertainties on PTA results, this paper describes the combined application of experimental design and response-surface method to analyse and interpret the pressure-transient data recorded by a permanent downhole-gauge during the shut-in of an oil producer in a West-African deepwater reservoir. Specifically, the paper discusses (i) a procedure for establishing the range of input-parameter uncertainties; and (ii) the effects of these uncertainties on the estimated well, reservoir and boundary properties. Within the parameter-space examined, it is found that the interpreted subsurface properties show strong dependence on the input-parameter range. More importantly, the susceptibility of relying on a single PTA outcome for key subsurface parameters such as permeability and connected volume is clearly demonstrated.