3.2.6 Grain size effects

The soil particles in a centrifuge model can not be scaled down to scale the soil particles of prototype, while other model dimensions can be modeled down. This effect is called as 'grain size effect'.

Simple index to show the degree of the effect may be the proportion of the average grain diameter (D₅₀) to the representative length of the model (B).

For circular foundations, Ovesen found deviation in common behavior of foundation when B/D₅₀ is smaller than 15, therefore B/D₅₀ may need to be greater than 15.

For slope stability test, Gooding and Gillette noticed no grain size effect when the slope height (H) is greater than 500 D₅₀

Fugslang and Ovesen suggested that at least 30 particles to be in contact with linear dimension of the model structure for the observed behavior to be representative of the prototype behavior.

However more appropriate way of examining the particle size effect may be by considering the ratio of particle size to shear band width.

Grain size effect on bearing capacity factor is shown in Fig. 3.

Siddiquee, M.S.A., Tanaka, T. and Tatsuoka, F. (1992) : A numerical simulation of bearing capacity of footing on sand, Proc. of 27th. Annual Meet. JSSMFE, Vol. 2, pp. 1413-1416.

Ko, H.Y. (1988) : Summary of the state-of-the-art in centrifuge model testing, In Centrifuges in Soil Mechanics (eds Craig, W.H., James, R.G. and Schofield, A.N.), Balkema, Rotterdam, pp. 11-18.

Taylor, R.N. (1995) : Centrifuges in modelling: principles and scale effects, In Geotechnical Centrifuge Technology (ed Taylor, R.N.), Blackie Academic and Professional, Glasgow, pp. 19-33.

Phillips, R. (1995) : Centrifuges in modelling: practical consideration, In Geotechnical Centrifuge Technology (ed Taylor, R.N.), Blackie Academic and Professional, Glasgow, pp. 34-60.