Case Study on Uncertainty of CPT Input Parameters on Undrained ‘Red Mud’ Strength

January 7, 2026

Uncertainty in interpreting cone penetration test (CPT) data can significantly affect geotechnical design for upstream tailings storage facilities, where stability depends on accurate strength characterisation of the tailings body. This paper presents a case study using data from a bauxite residue (“red mud”) storage facility in Western Australia and quantifies uncertainty in the Nkt method for estimating peak undrained shear strength and its normalised ratio (su/ σ′v0). Three key inputs of this method — empirical cone factor (Nkt), total stress, and equilibrium pore water pressure — were modelled as random variables under two knowledge levels: KL1 (literature-based estimates) and KL2 (site-specific testing with Bayesian updating). A series of Monte Carlo simulations were used to propagate these input uncertainties and assess their influence on the output strength parameters. The results show that for this case, refining Nkt through sitespecific data reduces the coefficient of variation for su from ~30% to ~14.5%, while uncertainty in total stress and pore water pressure had a modest impact on strength uncertainty. However, the dissipation testing provided valuable insight into drainage performance and design assumptions of the facility. The approach is compatible with reliability-based design principles and shows how uncertainty can be quantified through probabilistic methods and Bayesian updating. The method can be used for planning and communicating the intent of geotechnical site investigations.

Medeiros, M., Waters, R.D. 2025 “Case Study on Uncertainty of CPT Input Parameters on Undrained ‘Red Mud’ Strength,” in Proceedings of the NZSOLD/ANCOLD Conference 2025, 19 – 21 November 2025, Ōtautahi Christchurch, New Zealand: NZSOLD/ANCOLD