A pseudo-three-dimensional work-balance correction for sidewall friction effects in centrifuge slope models

Abstract Centrifuge model tests are widely used to investigate slope stability; however, frictional interactions between the sliding mass and rigid container sidewalls introduce boundary effects that compromise similitude with ideal plane-strain prototypes. These effects can lead to an apparent increase in the stability of finite-width models relative to the corresponding plane-strain reference. In this paper, a pseudo-three-dimensional work-balance correction based on a classical plane-strain upper-bound mechanism is developed by introducing an additional out-of-plane sidewall-friction work-rate term evaluated from an assumed lateral normal-stress distribution into the global work rate balance. A random-search optimisation technique is employed to identify the critical failure surface geometry and the corresponding stability factor. The proposed formulation is checked against the classical plane-strain upper-bound solution and further compared with published centrifuge test data. Across the reported configurations, the calculated and measured critical acceleration scale factors show the same decreasing trends with increasing model height and slope inclination, although substantial differences remain for individual cases. A parametric study reveals that sidewall friction significantly increases the apparent stability factor at low width-to-height ratios, but this discrepancy diminishes rapidly as the ratio increases. For the representative cases examined at β = 60°, the sidewall friction angle and width-to-height ratio exert larger influences on the sidewall-induced relative deviation than the soil internal friction angle, while the influence of the soil internal friction angle is retained explicitly in the broader reference charts and tables. Reference charts and tables are further developed to quantify the width-dependent deviation for selected parameter combinations and relative-deviation thresholds. These resources provide preliminary, assumption-dependent references for selecting model widths and interpreting sidewall-induced deviations in centrifuge slope tests.