The authors evaluated the efficiency of non-stationary waterflooding on stochastic geological and reservoir simulation models with various reservoir heterogeneity parameters and different injection modes. Prior to geological modelling, Roxar’s Irap RMS software solution was used to analyze the distribution of porosity variogram ranges for the Romashkinskoye oil field. The analysis showed that the variogram has a lognormal distribution and ranges from 100 to 1500 m. The range of variogram of the resultant stochastic geological models (options) was found to be within 100-1000 m with the increment of 50 m. For each option, 10 equally probable model scenarios were prepared. For each scenario, 20 injection modes were simulated. Totally, 3820 reservoir simulation models were built. The resultant models were used to analyze sensitivity of the efficiency of non-stationary waterflooding to each of the following parameters: variogram range, non-stationary waterflooding pattern, and well shutdown period.
The main conclusions from the research can be summarized as follows. Standard deviation (range of cumulative production values) increases with the increase of variogram range. Increase in number of simultaneously shutdown wells improves the efficiency of non-stationary waterflooding that can be expressed as straight line. The longer the period of injection well shutdown, the higher the efficiency of non-stationary waterflooding for the cases when 1-4 wells are shutdown simultaneously, while in the case of six simultaneously shutdown wells the efficiency of non-stationary waterflooding deteriorates. In accordance with experimental results, the best scenario for non-stationary waterflooding entails simultaneous shutdown of six wells for the period of one day.
