Water Injection Optimization Using Historical Production and Seismic Data

摘要

This work proposes a method to optimize waterflooding byrnusing production and time-lapse seismic data. Using timelapsernseismic difference as an indication of inter-wellrninfluence, inter-well connectivities or influence functions arerndetermined using a combinatorial algorithm (stochasticrnhillclimbing) by constraining the computation with time-lapsernseismic. The water-cut is estimated using a fitting function. Byrnperturbing the injection rate in the injectors, the oil productionrnis optimized. The field in this study is mature with 15 years ofrnproduction history and with 70% water-cut at the present time.rnTwo legacy repeated seismic surveys, of in 1991 and 2001rnvintages respectively, are reprocessed to enhance thernrepeatability from the original data. The reprocessing is donernto eliminate the survey, source and noise differences. Thernconnectivity or influence function is estimated using a windowrnof balanced flooding. It is found that seismic constraining forrnthe connectivity calculation increases the consistency of therninter-well relationship with the spatial data. In order to have anrnoperational waterflooding scheme, a multiple realizationrnapproach is used. That is, the optimized injection is randomlyrnperturbed under specified upper and lower bounds. Thernproduction rate is computed using the inter-well influencernfunction. It is found that the optimal injection configurationrncan increase oil production. With a pilot implementation, it isrnshown that the new injection scheme can delay the oilrnproduction decline.