During a laboratory waterflood study, undertaken at reservoir conditions to provide representative displacement data for predictive reservoir simulation studies, an investigation into acquiring representative wettability showed preserved samples exhibited significantly more oil wet character than restored samples. Further investigations concluded that core retrieval and/or storage resulted in increased oil wet character. Preserved state samples were not therefore representative.
Pressure support into this giant oil reservoir is currently being achieved by water injection. It was originally predicted that water breakthrough would occur around the turn of the century. However breakthrough was observed in a number of production wells significantly earlier than had originally been predicted.
A study was initiated to improve the understanding of the underlying causes of this water breakthrough. It was concluded that one of the key uncertainties governing the ability to predict waterflood performance was the quality of water/oil relative permeability data. Although there was a large number of historic water/oil relative permeability data sets, there was also a large degree of scatter in these data. This scatter appeared to be due, in part, to the variety of experimental procedures used. It was therefore unlikely that historic displacement data would be valid for reservoir performance.
A process study was therefore initiated to generate new data using more rigorous procedures, including ensuring that samples in the laboratory were at a wetting state which was representative of the reservoir. This representative wettability was investigated by comparing displacement characteristics using plug samples in two wetting states at full reservoir conditions using in-situ saturation monitoring techniques.
The approach for ensuring that a reservoir representative wetting state was attained in the laboratory compared displacement characteristics on samples prepared in two different ways, preserved state and restored state.
Preserved state samples are samples which have not been cleaned with any solvents. Removal of any mud filtrate and saturation with brine is achieved by flushing with simulated formation brine, prior to initial water saturations being acquired. The preserved sample is then aged at full reservoir conditions for three weeks.
Restored samples are samples which have been previously cleaned to as water wet state as possible. Samples are saturated with simulated formation bring and then, once initial water saturations has been acquired, are aged at full reservoir conditions with live crude oil.
Prior to any reservoir condition studies it is essential to ensure that initial water saturations are acquired which are distributed uniformly down the length of the plug samples. This is necessary to ensure:
Representative wettability is achieved during restorating/ageing at reservoir conditions with live oil.
Representative Sor is obtained from the waterfloods because of the dependency on Sor with Swi.
Non uniform Swi does not result in an unstable waterflood and nonrepresentative Sor.
Trapped water (in strongly oil wet systems) does not cause an unstable flood and erroneous relative permeability data as well as suppressing initial oil permeabilities.
As reservoir condition studies compare the displacement characteristics in preserved and restored samples, it is necessary to ensure that initial water saturations were successfully acquired in both preserved and cleaned samples.