Experimental design: importance in petroleum engineering

摘要

The inverse problem associated with data measured to chracterize systems and processes are frequently ill-posed, and, consequently, small errors in the measured quantities may transfer into large errors in the desired estimates. Better accuracy may be obtained by improving the experimental design. This paper deals with experimental design and its importance in the field of petroleum engineering. The experimental design studies described in this paper are based on a linearized covariance analysis. We utilize measures such as confidence intervals, contribution of information, and optimal design as ways of ranking different com-peting measurement strategies (or designs). Among a series of inverse problem applications within petroleum engineering, we consider two cases: The first application deals with estimation of properties defined within a set of partial differential equations, namely those describing flow of fluids through porous media. We show how experimental design can be utilized to select flow experiments (both flow rates, types of data and when and where to measure) to obtain accurate estimates in saturation regions im portant for reservoir exploitation. THe second example deals with identification of the inflo profile of oil, water, and gas, for a production well drilled into an underground reservoir.