Abstract: Characterization of interwell reservoir connectivity is the most important and fundamental research work to understand water flooding oilfield.Although the commonly used reservoir engineering,numerical simulation and geochemical methods may improve the reliability of interwell reservoir connectivity characterization,they are heavily dependent on extensive collection of geological data and long period observation of well performance through production cycle,which are either expensive or time consuming.The interaction between injectors,producers,and interwell reservoir is a complex nonlinear dynamic system.The production data of oil and water reflect certain aspect of the system,which may hint into the characters of interwell reservoir connectivity.The multivariable facies distribution reconstruction can be applied to resolve the dynamic characteristics of the system based on the multivariable production data of oil and water wells.Thus the reservoir connectivity can be characterized by calculating correlation dimensions of chaotic attractor of the injection-production dynamic system.The amount of work required by this approach is small,and it works well in the application to WR oilfield,with conclusion being consistent with actual production.The approach is clear and simple both in theory and in algorithm.Without complicated mathematical models or parameters of static geological models,it has initiated a new way to quantitatively characterize interwell reservoir connectivity.

Key words: interwell connectivity, production datum, phase space reconstruction, correlation dimension, quantitative interpretation, reservoir, water flooding oilfield development