Perpustakaan Digital ITB

Low salinity water injection (LSWI) is one of emerging enhanced oil recovery (EOR) techniques that utilizes lower salinity of injection water which results to cost-effective and environmentally friendly EOR technique. In sandstone reservoir, one of LSWI mechanisms in increasing oil recovery is through multiple ionic exchange (MIE) which leads to wettability alterations. MIE mechanism is affected by some reactions in crude oil-brine-rock (COBR) system, one of them is cation exchange which mainly controlled by clay properties, namely cation exchange capacity (CEC). However, CEC data is obtained by core data and usually assumed to have homogeneous distribution for each grid. Moreover, CEC is one of several parameters that affects incremental recovery. Hence, heterogeneous CEC distribution should be modeled in order to obtain a reliable incremental recovery. To design and classify the heterogeneous CEC distribution, well log data will be used as an input to machine learning model. Using Python, data pre-processing and feature engineering have to be done before the model will be trained. After predicting CEC as in interval-type data, CEC has to be correlated with static properties for example porosity, shale volume, or other static properties to model the 3D-distribution of CEC. LSWI simulation using CMG GEMTM will be conducted in three CEC cases: (1) Homogeneous CEC of 381 eq/m3, (2) Homogeneous CEC of 1316 eq/m3, and (3) Heterogeneous CEC that previously determined. The salinity concentration in the development scenario is designed at 1980 ppm. The result analyzes macroscopic and microscopic effect which are cumulative oil production, oil saturation, and sodium fraction in ion exchange Using machine learning, CEC can be predicted using well log data and resulted an acceptable RMSE at 0.178. It is found that flow zone indicator (FZI), proposed by Amafuele et al., multiplied by shale volume (Vsh) have a good correlation with CEC, hence CEC can be formulized in 3D-distribution. Simulation results show that low salinity water injection with homogeneous cation exchange capacity distribution and heterogeneous distribution gives a significant difference with natural flow in cumulative oil production by 1.22%. Noticeable effect are identified microscopically especially in oil saturation and sodium ion exchange. However, the difference in results between homogeneous CEC distribution with a value of 318, 1316 and heterogeneous has a very small difference which is 0.03%-0.1% by recovery factor difference.