Data Science in Carbonate Reservoirs: Feasible to train data with ease?
If carbonate genesis and diagenetic overprint could produce ‘complex multi-scale pore structures’ that could vary from ‘sub-micron scale intra-particle/inter-particle/moldic pores’ to ‘kilometer-scale cave systems’ in a typical carbonate reservoir, and, if inter-connected inter-particle pores govern the reservoir permeability; while, the coupled effect of porosity and permeability together determine the injectibility, storativity and recovery as a complex function of pore-size, pore-size distribution, pore-geometry, packing density, confining stress, deviatoric stress and fracture characteristics; then, in the context of ‘data science’ (machine learning), would it remain ‘practically’ feasible to “train” such “sparse” data set on porosity and permeability towards deducing an improved hydrocarbon recovery?
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs