Assessing Waterflood Efficiency with Deconvolution Based Multi-Well Retrospective Test Technique

Abstract

A waterflood study has been performed on a heterogeneous oil deposit with a rising water-cut and production decline after 10 years of commercial production. The objective was to analyze the efficiency of waterflood patterns across the field and suggest injection optimization opportunities. The production is facilitated by ESP with Permanent Downhole Gauges (PDGs) which provides an opportunity to analyze the productivity index and cross-well interference. The PDG analyzes was performed in PolyGon pressure modelling facility and followed Multi-well Retrospective Testing (MRT) workflow which is based on the mathematical procedure of multiwell deconvolution (MDCV). MDCV trains the correlation between bottom-hole pressure (BHP) variations from PDG data records and rates variations from daily production history of a given well and other wells around it. This provides a robust short-term predictor for production response for different rate/BHP scenarios and makes a basis for injection optimization opportunities. MDCV allows reconstructing formation pressure and productivity index back in time, pick up the changes and understand if they were caused locally (by skin) or massively (by transmissibility). The diffusion modelling of deconvolved data allows a robust quantification of some reservoir properties in cross-well intervals, such as the current drained volume around each well, potential drained volume (as if the offset wells are shut-down), apparent cross-well transmissibility, boundary types and compare them against the various geological scenarios and possible well-reservoir contact scenarios. The quantitative analysis allows picking up anomalously high cross-well interference which may be caused by thin-bedding circuiting or induced fracture. It also provides a strong hint for thief-injection and thief-production in cases of poor cross-well interference.