Assessment of the suitability of the fall cone method and the Casagrande cup method for determining the liquid limit of South African soils

Abstract

The plasticity index (PI) is one of the most important soil variables in the design of soil related structures. Most engineers rely on PI values determined by commercial laboratories. Commercial laboratories strive to make their services competitive and affordable to their clients. The liquid limit (LL) is one of the variables used to determine the PI and is determined by laboratory soil tests, therefore it is important to determine the LL accurately.

Two techniques, the Casagrande percussion cup and fall-cone (penetrometer) methods, have been adopted as the standard LL measurement approaches globally. The Casagrande cup method is implemented in South Africa (SANS, 2011) as well as in the USA, whilst the fall-cone method is accepted in the UK (BS 1377-2, 1990).

According to many researchers such as Karlsson (1961), Sherwood and Ryley (1970), Weston (1978), Sampson (1983), Sridharan and Prakash (2000), Feng (2000), the Casagrande cup method is operator dependent, less reliable and soils of low plasticity tend to slide in the Casagrande cup while the LL cannot be measured. On the other hand, the fall-cone test is a simple shear strength test that is much less operator dependent and can measure the liquid limit for a soil of any plasticity accurately.

According to Sampson and Netterberg (1984), the British variant of the Casagrande cup produced LL results that directly correlated with the BS fall-cone LL results, while the SANS variant Casagrande cup correlated indirectly with BS fall-cone LL results with a constant difference of about 4 units. It is important to note that the indirect correlation, does not refer to an inverse correlation, but a correlation with an offset.

This indirect correlation was caused by the international variance of the base hardness of the Casagrande cup devices and lead to poor precision limits (Haigh, 2015; Sampson and Netterberg, 1984).

The aim of this research is to modify the standard British fall-cone method in order to produce a fall-cone procedure that is optimised for South African soils, in order to replace the Casagrande cup method for the determination of the liquid limit for South African soils.

To optimise the sample preparation procedure, the BS fall-cone cup was modified by removing the bottom plate of the cup and sharpening its bottom edges to form a specimen ring. The specimen ring LL results were compared to the results of the BS cup and showed that there was a negligible difference in results due to the removal of bottom plate.

To optimise the amount of sample required to perform the test, four specimen rings were manufactured with diameters 54mm, 35mm, 30mm and 22mm to test whether the smaller diameter LL results would differ from the standard diameter. The results showed that the 35mm specimen ring could be used for testing, while the 30mm and 22mm diameters showed the effects of mould confinement, where the soil heave created by the cone affected the penetration depth of the cone.

The indirect correlation would be corrected by adjusting the standard 20mm corresponding penetration. The 54mm and 35mm diameter rings in the modified fallcone tests showed a direct correlation with the SANS Casagrande cup when the corresponding penetration was adjusted from 20mm to 16mm.

Further tests were conducted where the cone weight was increased in order to investigate the correlation between the fall-cone and the SANS casagrande cup. However, this method was not considered since it was impractical to change the standard apparatus and move further from achieving an international standard.

A draft SANS fall-cone procedure manual was compiled from the research conducted and the results obtained in this dissertation and is attached as Annexure A. This manual will be submitted to SABS for the approval of the modified fall-cone as a South African standard method.