Scale modelling of fluid flow in geotechnical centrifuges

Geotechnical centrifuges enable prototype-magnitude effective stresses to be generated in small-scale soil models, a necessary condition for using them to predict prototype behaviour. The paper presents a thorough dimensional analysis of the modelling process. The general relationships between soil-particle size, pore-fluid viscosity and time-scaling processes which are necessary (in addition to effective-stress consistency), to ensure correct Reynolds Number modelling are presented in detail. The conditions for valid modelling of consolidation and accelerating-particle processes are also defined and discussed. Some historical errors are noted together with an analysis of the shortcomings of the use of 'prototype soil' in the model and oversimplified fluid viscosity scaling. A new pseudo-prototype concept is also introduced. This defines the prototype soil which is actually being modelled in a centrifuge test, enabling it to be compared with the one intended.