Sammendrag
Multivariate analysis is employed to investigate the structure of variations within highly heterogeneous data. Traditionally, principal component analysis (PCA) is run by analyzing the entire wireline log and using PCA scores to characterize variability within and between lithologies. In this paper, we propose a technique using only specific subsets of all well records to quantify reservoir heterogeneity due to second order lithological variability. These subsets are chosen from uniform lithofacies parts of the wireline log in order to reduce the variability in the correlation matrix that otherwise would cause lithological changes. The purpose is to assess the efficiency of structured PCA in analyzing small-scale heterogeneity that is captured by wireline logs but often masked by traditional PCA approaches. This paper shows that a structured PCA procedure based upon special lithological units is superior to an unstructured PCA, when the focus is within lithology variations. This structured procedure is applied to data from the Heidrun field, offshore mid-Norway. The results demonstrate clear benefits from added insight into the variability of a complex fluviodeltaic heterolithic sequence that poses great challenges to hydrocarbon development.
Vis fullstendig beskrivelse
