Relevance of correlation length in geotechnical engineering
Relevance of correlation length in geotechnical engineering
Soils are heterogeneous in nature and as such, their properties inherently exhibit a distinct spatial variation that reveals past information of their geological history. This spatial variability is an important source of geotechnical uncertainty and hence its proper characterisation is of importance to geotechnical design. Random field theory provides a consistent mathematical framework to account for this soil's heterogeneity and it is a powerful computational tool to perform reliability analyses of geotechnical structures when combined with the finite element method. A key parameter in these stochastic analyses is the correlation length (or sometimes also referred to as scale of fluctuation) because it controls variations of a soil property in a given spatial direction. This spatial statistic can be estimated from in-situ data (e.g., cone penetration test CPT) and the quality of such estimation is the main question tackled in this paper.