Ewaluacja i porównanie badań granicy skurczalności i płynności z wykorzystaniem metod statystycznych

Abstract

The paper presents a comparative study on test methods for shrinkage limit and liquid limit of soils. Tested material comprised of the soils that varied in terms of genesis, grain size distribution and mineralogical composition, such as: Miocene-Pliocene clays, glacial tills, alluvial clay soils, monomineral clays (the Sedlec kaolinite, the Wyoming montmorillonite, illite from Bukowa Góra). Shrinkage limit (ws) was tested according to: Polish Standard PN-88/B-04481 [22], British Standard BS1377: Part 2: 1990: 6.3 [3]; „CLOD” test method (Hamberg 1985, after Garbulewski [5, 6]), as well as calculated by the Krabbe equation from 1958 (after Schultze, Muhs [23]) and the equation proposed by Wysokiński at al. [28]. The results obtained revealed that ws values measured according to BS [3] are 0,6 – 7,4% lower than according to PN [22] (Fig. 4). Shrinkage limit values calculated by the equations differ significantly and show no relationship with the experimental results (Fig. 4). Multivariable regression analysis showed low correlation between shrinkage limit and plastic limit, liquid limit, plasticity index and clay content (formulas 4 and 5). Therefore the suitability of the equations based on these parameters for ws determination is to be called in question. The study revealed that the undisturbed samples demonstrate lower ws values than the remoulded samples (Fig. 5). The smallest differences (0 to 1,7%) were obtained for Miocene-Pliocene clay. Liquid limit (wL) was tested by Casagrande method [22], two cone penetrometer methods: according to British Standard [3] and Polish Standard [22], one-point method: using Casagrande apparatus [1] and using cone penetrometer [3]. It was found that wL values obtained from one-point tests were entirely consistent with the results of standard multipoint tests (Fig. 6), which proves suitability of one-point methods in geological engineering practice. Comparison of Casagrande test and cone penetrometer test acc. BS supported the previous findings ([8], [27], [19]) reporting consistency of the results obtained by these two methods (Fig. 7), whereas comparison of Casagrande test and cone penetrometer test acc. PN showed discrepancies between the results obtained (Tab. 1).