JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, cilt.25, sa.3, ss.625-633, 2010 (SCI-Expanded)
In this study, the applicability of discriminant analysis, which is a multivariate statistical technique, on the estimation of groutability is investigated. First, a general overview of discriminant analysis is presented. Considering that the groutability is a function of the rheological properties of the grout, the injection pressure, the grain size distribution, the fines content, and the relative density of the soil, a discriminant analysis is applied to a database that consists of 88 laboratory grouting test results. Discriminant functions are developed for two separate models. The developed functions indicate that the most significant variable for the estimation of groutability is the ratio of the grain diameter of soil corresponding to 15% passing (D-15) to the diameter of the injected material particles corresponding to 85% passing (d(85)). Even though it uses only two parameters, the first method proposed herein ranks second in the percentage of successful predictions when compared with other groutability estimation methods.